Home Uncategorized critical appraisal

Uncategorized

critical appraisal

There are five articles you can pick one from them.

and here are the requirements:

Students will be required to submit a 1,000-word critical appraisal of a research article from an academic journal. Students will be provided with a choice of articles to appraise, which will be posted on the course Moodle site.

The critical appraisal is designed to provide students with an opportunity to critically review the overall appropriateness of the research design underpinning a piece of social science research. This will involve assessing the strengths and weaknesses of the particular methodological approach adopted in the research and commenting on its suitability to address the research questions. Your appraisal should be specifically methodological in focus.

Here are few guiding questions to help you write and focus on reviewing:

• The research questions the author(s) set out to address. Are they concrete and appropriate?

• The combination of methods being employed. Are they appropriate and fully justified?

• To what extent are the methods chosen suitable to answer the research questions? Are they coherent and well-aligned?

• Did the authors describe their methods fully? What are some gaps you identified?

• What the data and evidence is. How were they collected/recorded/retrieved, analysed and interpreted? • The methodological and epistemological strengths/weaknesses of the research design. What ontological and epistemological assumptions are made by the author[s]?

• Ethical and/or political issues concerning the research. Are they apparent and being addressed by the author[s]?

Your review should be fully referenced drawing on relevant methodological literature. Further guidance on reading and the articles for appraisal will be provided.

Grading and feedback of the assessment will take into consideration the following:

• How well you have demonstrated an awareness of problems and shortcomings of the data and methods of analysis;

• Whether you have considered the extent to which the interpretation of the findings is consistent with the analysis and research design undertaken;

• How well structured are the ideas in the assignment and whether there is a logic to their development and argumentation;

• How reflexive (i.e. reflecting on your own position and assumptions as a student) are your comments on the research methods;

• How well the assignment is presented, including style of writing, grammar, structure, and referencing.

http://uar.sagepub.com/
Urban Affairs Review

http://uar.sagepub.com/content/35/1/

The online version of this article can be found at:

DOI: 10.1177/10780879922184293

1999 35: 72Urban Affairs Review
Richard Sauerzopf and Todd Swanstrom

The Urban Electorate in Presidential Elections, 1920-1996

Published by:

http://www.sagepublications.com

On behalf of:

The Urban Politics Section, American Political Science Association

can be found at:Urban Affairs ReviewAdditional services and information for

http://uar.sagepub.com/cgi/alertsEmail Alerts:

http://uar.sagepub.com/subscriptionsSubscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://uar.sagepub.com/content/35/1/72.refs.htmlCitations:

What is This?

– Sep 1, 1999Version of Record >>

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

http://uar.sagepub.com/content/35/1/72

http://www.sagepublications.com

http://www.apsanet.org/~urban/

http://uar.sagepub.com/cgi/alerts

http://uar.sagepub.com/subscriptions

http://www.sagepub.com/journalsReprints.nav

http://www.sagepub.com/journalsPermissions.nav

http://uar.sagepub.com/content/35/1/72.refs.html

http://uar.sagepub.com/content/35/1/72.full

http://online.sagepub.com/site/sphelp/vorhelp.xhtml

http://uar.sagepub.com/

URBAN AFFAIRS REVIEW / September 1999Sauerzopf,Swanstrom / THE URBAN ELECTORATE

THE URBAN ELECTORATE
IN PRESIDENTIAL

ELECTIONS,
1920-1996

RICHARD SAUERZOPF
Wayne State University

TODD SWANSTROM
State University of New York–Albany

This study of voting in presidential elections in 12 central cities from 1920 to 1996 shows that
cities played a crucial role in the New Deal realignment that dominated presidential elections
from 1932 to the 1960s. Since then, cities have declined as a share of the total electorate, but they
still provide crucial votes for successful Democratic presidential candidates. As cities have
increasingly deviated from national voting trends, however, their turnout rates have increasingly
fallen behind the national rates. A call is issued for researchers to break down the suburban vote
and to examine contextual effects on voting behavior.

Many scholars have documentedthe rise and fall of the urban electorate in
national elections (Degler 1964; Lubell 1965; Sundquist 1973; Andersen
1979; Mollenkopf 1983; Edsall and Edsall 1991). The rise of the political for-
tunes of cities followed the migration from farms to cities that, by the middle
of the twentieth century, made cities the population centers of the country.
The Democratic Party relied heavily on urban votes to build the New Deal
coalition that dominated national elections from the 1930s to the 1960s.
Similarly, the migration outward from cities to suburbs, which accelerated in
the 1950s, fueled the rise to power of the Republican Party in national elec-
tions beginning in 1968 and signaled the progressive marginalization of
urban electorates in national politics.

Accounts of the urban electorate in presidential elections rely heavily on
the theory of electoral realignment.1 According to realignment theory,
approximately once in a generation (1860s, 1890s, 1930s), the existing party

URBAN AFFAIRS REVIEW, Vol. 35, No. 1, September 1999 72-91
© 1999 Sage Publications, Inc.

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

system reaches a crisis when “emergent tensions in society . . . not adequately
controlled by the organization or outputs of party politics as usual, escalate to
a flash point” (Burnham 1970, 10). Political entrepreneurs take advantage of
the crisis with new appeals, ideological differences widen, the coalitions
behind the parties are shuffled, and a new party system is formed with a new
dominant coalition. The realigned party system incorporates political
demands that had been ignored by the old party system.

Franklin Roosevelt’s victory in 1932 is viewed as a quintessential critical
election. Roosevelt was able to swing urban ethnics, blacks, and Jews into the
Democratic Party, forming the nucleus, along with the “solid South,” of the
Democratic coalition that dominated presidential elections for the next gen-
eration.2 The flight to the suburbs, which accelerated after World War II,
weakened the urban base of the Democratic Party. With suburban voters per-
ceiving that liberal Democratic policies siphoned off their tax dollars to
expensive and wasteful programs targeted on central cities and blacks,
Republicans successfully used “wedge issues” to split off elements of the
New Deal coalition (Wilson 1987; Edsall and Edsall 1991). By the 1992 elec-
tion, Schneider (1991, 1992) estimated that suburbanites cast an absolute
majority of the votes, and he proclaimed “the dawn of the suburban era in
American politics.”

The above account of the rise and decline of urban electorates provides the
backdrop for Bill Clinton’s vaunted “suburban strategy” that supposedly won
him the 1992 and 1996 elections. According to Clinton pollster Stanley
Greenberg, the “forgotten middle class” of white suburbanites, as exempli-
fied in Greenberg’s (1995) research on Macomb County, Michigan, had
abandoned the Democratic Party in the face of what they saw as its preoccu-
pation with urban blacks. The key to Democratic presidential victories was to
challenge the Republican Party for suburban votes, Greenberg argued, entic-
ing back into the fold the so-called Reagan Democrats (Greenberg 1995).

To attract suburban votes, Clinton avoided identification with policies tar-
geted to cities or to minorities, emphasizing instead policies that benefit peo-
ple wherever they live. Clinton’s well-publicized attack on Sister Souljah in
the 1992 campaign distanced him from Jesse Jackson and the liberal wing of
the Democratic Party centered in cities. The highly successful Clinton/Gore
bus tours avoided the inner cities and were resplendent with small-town, even
rural, imagery. Other than enterprise zones, Clinton proposed few policies
targeted on cities. Instead, he proposed broad-based policies, such as national
health insurance and investments in education, that would benefit cities and
suburbs, blacks and whites.3 Clinton’s two victories were widely credited to
his ability to contest Republicans for the swing votes in the suburbs. Even

Sauerzopf, Swanstrom / THE URBAN ELECTORATE 73

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

though the conventional account, described earlier, of the rise and decline of
the urban electorate is widely accepted, there is relatively little published
analysis of city voting trends over time to back it up.4 The main reason is that
voting statistics in the United States are reported by county, not by city.5

Unless the city boundaries correspond to the county boundaries, city election
returns are not widely available. Before 1950, county voting returns usually
corresponded closely to city returns, but with massive suburbanization over
time, county returns have increasingly deviated from city returns.

One exception to the dearth of urban voting studies is a 1949 article by
Samuel Eldersveld, who painstakingly collected data on presidential elec-
tions for 12 cities from 1920 to 1948. In this article, we update Eldersveld’s
data for the presidential elections since 1948.6 Eldersveld chose all cities with
populations more than 500,000 according to the 1940 census, eliminating
Washington, D.C., and Buffalo. The 12 cities are the following: New York,
Chicago, Philadelphia, Pittsburgh, Detroit, Cleveland, Baltimore, St. Louis,
Boston, Milwaukee, San Francisco, and Los Angeles. This sample of cities is
by no means representative of the nation as a whole. It is biased toward older
northern cities, although it does include border cities such as Baltimore and
St. Louis, as well as the West Coast cities of Los Angeles and San Francisco.
Our main justification for choosing these cities is that it enables us to piggy-
back on Eldersveld’s work and thus track city voting over a long time period.
Moreover, the 10 states represented by these 12 cities still control 215 elec-
toral votes and thus are intrinsically important. Admittedly, our work is only
exploratory, and further research is needed on urban voting in key sunbelt
states such as Florida and Texas.

We use this aggregate voting data to examine the voting behavior of
central-city and non-central-city electorates over time and to suggest some
revisions in the conventional accounts, both scholarly and in the mass media,
of the rise and decline of urban electorates. Our data are limited, however,
because they only allow us to show how the urban electorate voted, not why it
voted the way it did. The striking voting patterns that we uncover could be the
result of the characteristics of individual voters who live in cities (e.g., their
race or class), or voting behavior could be due to a “contextual effect” of liv-
ing within central-city municipal boundaries. Our data do not allow us to
decide between individual and contextual effects. However, municipalities
are important units within our federal system, with billions of dollars of fed-
eral money targeted to them. Moreover, municipalities may be important
units for shaping political identities. Therefore, in the conclusion, we call for
further research on our hypothesis that contextual effects arising from
central-city residence significantly affect voting behavior.

74 URBAN AFFAIRS REVIEW / September 1999

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

RISE OF THE URBAN ELECTORATE:
NEW DEAL REALIGNMENT

Eldersveld’s (1949) main argument in his original article was that the tra-
ditional way of understanding American presidential elections in terms of
sectional divisions (e.g., the South vs. the North) was no longer adequate.
Rather, this view needed to be supplemented by an appreciation of a powerful
urban-rural division in the national electorate. That is, rather than looking at
presidential elections as being won by a candidate’s appeal to base and key
swing regions of the nation, Eldersveld argued that we should look at elec-
tions as hanging on a candidate’s ability to appeal to cities. This was the case
because the most powerful predictor of how individuals voted had clearly
become whether they lived in urban areas. By 1949, to a great extent regard-
less of region, small-town and rural area voters chose Republican candidates,
and big cities voted Democratic. Eldersveld showed that 12 of the nation’s
largest cities in 10 states consistently turned out victories for Roosevelt and
Truman.

Figure 1 shows that the shift of the urban electorate to the Democratic
Party that drove the New Deal realignment actually began in 1928 with the
nomination of Al Smith, the first non-WASP ever nominated by a major
American party. As Lubell (1965, 49) put it, “Before the Roosevelt Revolution
there was an Al Smith Revolution.” A Catholic of immigrant background
whose career was based on the New York Democratic machine, Tammany
Hall, Smith’s campaign was replete with urban imagery. Especially impor-
tant was Smith’s opposition to the Klu Klux Klan and prohibition, stands that
were popular with urban ethnics. Key (1955) noted that the 1928 election had
all the earmarks of a critical realigning election, with voter turnout rising 26.5%
from the previous election and the coalitions behind the parties undergoing
major change (see also MacRae and Meldrum 1960; Sundquist 1973, 177).7

If Smith began the transition of urban voters to the Democratic Party,
clearly Roosevelt completed it in the 1930s. Having grown up in the country,
Roosevelt was not partial to cities. Furthermore, the Democratic Party prior
to 1932 was more closely identified with Jeffersonian localism than with an
activist federal government. Roosevelt ran for office in 1932 promising to
balance the budget. But once in office, Roosevelt recognized the need for fed-
eral action and enacted a series of urban-targeted programs, “spending nearly
$20 billion in new federal funds to put nearly one-fifth of the unemployed to
work at transforming the urban infrastructure” (Mollenkopf 1983, 70).
Andersen (1979) has shown how Roosevelt’s urban appeal mobilized non-
voters and people voting for the first time into the New Deal coalition.

Sauerzopf, Swanstrom / THE URBAN ELECTORATE 75

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

Eldersveld (1949) argued that the pluralities the Democrats gained in cit-
ies were crucial to their victories in the electoral college. He classified a city,
or cities, as having “won” the state if the urban plurality was 50% or more of
the statewide plurality.8 Clearly, the urban vote in these 12 cities was crucial
to the Democratic victories (see Figure 2). If it were not for the hefty Demo-
cratic pluralities of these major cities, Eldersveld maintained, the Democrats
would have lost the presidency in 1940, 1944, and 1948.

THE DECLINE OF THE URBAN
ELECTORATE: SUBURBANIZATION

As shown in Figure 3, the relative size of the city presidential vote reached
its highest point in 1948. For our 12 cities, the city proportion of the total
national vote reached an impressive 21.3% in 1948, subsequently falling to
only 5.9% by 1996. In 1948, for example, New York City cast more than 50%
of the statewide vote; in 1996, New York City represented only 32.1% of the

76 URBAN AFFAIRS REVIEW / September 1999

Figure 1: Sum of Sample City Pluralities, 1920-1996

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

state vote. Similarly, the city of Chicago fell from 46.5% of the Illinois vote in
1948 to only 20.9% in 1996 (see Table 1).

Notwithstanding the relative decline of the urban electorate, looking back
at Figure 2, one can see that city voters have continued to play important roles
in presidential elections. With the exception of Al Smith, probably no presi-
dential candidate of a major political party was more identified with cities
than John Kennedy. The Democratic Party platform in 1960 attacked the
Republicans for having turned their backs on cities and contained specific
planks to meet the problems of cities as governmental units, calling for the
creation of a cabinet-level department to coordinate urban programs (Rourke
1965, 155). According to Eldersveld’s (1949) criterion, our 12 cities were
responsible for swinging 146 electoral votes to Kennedy in a very close elec-
tion. Kennedy’s massive plurality in Chicago was just enough to overcome
the Republican advantage downstate (see Table 2). Republicans complained
that voting fraud in Chicago swung Illinois to the Democrats.

Sauerzopf, Swanstrom / THE URBAN ELECTORATE 77

Figure 2: State Electoral College Votes Won by Sample City Pluralities, 1920-1996
NOTE: A sample city’s plurality is classified as having “won” its respective state’s electoral col-
lege votes if the city’s plurality represents more than half of the state’s plurality. This standard
was devised by Eldersveld (1949) and is discussed in Note 8.

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

Kennedy’s victory precipitated an agonizing debate within the Republican
Party over what to do to close the “big-city gap.” Senator Barry Goldwater
publicly asserted that the Republican Party should stop trying to attract bloc
support from Negro and other minority groups concentrated in cities (Rourke
1965, 157). With Goldwater’s nomination in 1964, the Republicans pretty
much wrote off the urban vote. Indeed, since 1956, when Eisenhower won
pluralities in 6 out of the 12 cities in our sample, none of our sample cities has
voted Republican in a presidential election.

Notwithstanding the importance of urban votes to the Democrats, Table 2
demonstrates that the power of cities to influence presidential elections has
declined since the Roosevelt-Truman victories. Our sample cities were
important for the victories of Carter in 1976 and Clinton in 1992 and 1996,
but the number of electoral college votes “won” by cities has generally
declined (see Figure 2).9 In every election from 1936 onward, with the excep-
tion of six cities in 1956, positive figures in Table 2 represent instances when
cities made a contribution to the margin of victory of a Democratic candidate
within the state. All of the zeroes that appear in Table 2 since 1928, except in
1956, indicate instances when cities produced Democratic presidential plu-
ralities while their respective states went Republican. In 1972, every one of
Eldersveld’s (1949) sample cities voted Democratic, and all of their

78 URBAN AFFAIRS REVIEW / September 1999

TABLE 1: Sample City Proportion of State Actual Electorates

Election Year

1920 1932 1940 1948 1956 1964 1972 1980 1988 1996

New York 44.1 46.8 51.2 50.6 44.7 41.7 34.6 31 31.2 32.1
Chicago 37.5 40.9 44.9 46.5 37.7 34.2 28 23.5 23.4 20.9
Philadelphia 22.6 21.2 21.8 23.5 19.5 18.9 17 15.8 14.9 11.7
Pittsburgh 5.6 5.6 7.5 7.2 6 5.4 4.3 3.8 3.5 2.9
Detroit 26.8 25.7 26.7 31.8 25.5 21.3 14.8 10.4 8.4 8
Cleveland 8.3 9 11.7 10.6 9.1 7.6 5.4 4.3 3.6 2.9
Baltimore 51.4 48.3 47.9 42.3 34.2 28.4 19.5 17.2 13.6 9.6
St. Louis 21.2 22.2 21.9 21.8 18.1 14.7 10.4 8.1 7.2 5.5
Boston 16.7 16.3 16.9 16.2 13.2 10.9 8.6 7.1 7.1 6.8
Milwaukee 17.8 15.7 19 20.3 18.9 17.7 14.1 12.6 12.2 9.5
San Francisco 15.6 9.8 9.5 8.4 6.2 4.6 3.6 3 2.8 2.6
Los Angeles 27.3 42.8 43.3 42.9 18.5 15.3 12.5 10.7 10.4 7.5

Sample city votes
as percentage of
national electorate 15.8 18.3 20.7 21.3 15.7 13.2 10.2 7.6 7.1 5.9

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

respective states went Republican. In 1984, Ronald Reagan nearly duplicated
Nixon’s feat, losing every city but winning every state in our sample, except
Massachusetts.

REVISING THE CONVENTIONAL WISDOM

When we update Eldersveld’s (1949) data and analyze it, the results at first
seem to confirm the conventional explanations of the decline of the political
significance of cities. The conventional account emphasizes that the main
reason for the declining electoral influence of central cities is the flight to the
suburbs. In 1950, according to the U.S. Bureau of the Census, the population
of all of Eldersveld’s sample cities accounted for 57.8% of the total popula-
tion of their metropolitan areas; by 1990, these central cities accounted for
only 27.7% of that figure. It is well understood that suburbs tend to favor
Republican presidential candidates, but the opposite is true for cities. Increas-
ingly, since 1948, large Democratic pluralities in central cities have not been
able to make up for the rising tide of suburban and rural Republican votes,
although it seems that this has not been for lack of effort. As central-city elec-
torates have shrunk, they have become proportionally more Democratic. As
Schneider (1992) noted, however, suburbs are simply growing faster than cities

Sauerzopf, Swanstrom / THE URBAN ELECTORATE 79

Figure 3: Sample City Electorate as a Percentage of the National Electorate, 1920-1996

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

TABLE 2: Sample City Percentage of State Party Pluralities, 1920-1996

1920 1924 1928 1932 1936 1940 1944 1948 1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996

New York 40.5 15.8 0 145.9 163.3 320.6 244.9 0 0.29 0 206.2 51.8 187 0 248.2 0 0 253.9 88.7 64.4
Chicago 41.5 41.2 4.7 55.5 77.8 326 298.6 926.5 0 R5 5151 75.8 0 0 0 0 0 0 89.7 74.1
Philadelphia 30.5 29.5 14.6 R44.9 31.6 63.3 142.9 0 0 R20.5 285 29.6 160.4 0 207.7 0 0 0 67.8 77.3
Pittsburgh 7.6 2.3 0.7 0 15.4 24.9 54.3 0 0 0 78.6 8.3 41.5 0 33 0 0 0 16.6 14.2
Detroit 31.9 30.3 15.5 65.2 57.7 0 908.7 0 0 0 466.4 38 95.4 0 0 0 0 0 98.5 51.4
Cleveland 12.7 0 0 56.7 26.3 104.8 0 1228 0 0 0 19.3 0 0 901.8 0 0 0 101.1 28.5
Baltimore 70.9 71.4 11.7 62.3 71.1 75.5 231.8 0 0 R20.1 115.5 47.7 483.9 0 111.4 294.2 0 0 53.7 38.8
St. Louis 37.5 58.9 0 22.4 32 74.7 152.2 38.2 0 1811 1011 29 0 0 85 0 0 0 31.4 50.3
Boston 7.9 4.3 582.4 161.9 65.5 65.7 71.9 58.2 0 0 28.3 15 18.5 0 12.1 465.2 79.1 29.1 14.1 10.8
Milwaukee 7 M20.2 0 27 32.7 280.8 0 96.4 0 R4.3 0 28.9 0 0 201.5 0 0 120.2 77.3 38.7
San Francisco 15.9 1.6 0 15.5 14.1 12 15.8 55.6 3 R2 0 10.7 0 0 0 0 0 0 11.9 11.5
Los Angeles 30.8 58.8 55.5 37.8 43 47.7 46.2 44.4 5.6 R2.7 0 24.2 0 0 0 0 0 0 28 25.9
Mean city % 27.8 31.4 17.2 64.1 52.5 126.9 216.7 347.7 3 266.6 917.9 31.5 164.6 0 225.1 379.7 79.1 134.4 56.6 40.5
Number of

cities 12 10 6 10 12 11 10 7 3 7 8 12 6 0 8 2 1 3 12 12

NOTE: This table shows sample city presidential party pluralities as percentages of sample state pluralities. Zeros indicate instances when state pluralities and
city pluralities were for different parties. For 1920 to 1928, all figures indicate sample city support for Republican pluralities, except where indicated by an M
for LaFollette Progressive and D for Democratic pluralities. For 1932 to 1996, all nonzeros indicate sample city support for statewide Democratic pluralities,
except where indicated by an R for Republican pluralities. All zeroes for this period indicate instances when states produced Republican pluralities despite
sample city Democratic pluralities.

at G
lasgow

U
niversity Library on O

ctober 15, 2012
uar.sagepub.com

D
ow

nloaded from

http://uar.sagepub.com/

are growing more Democratic. The vaunted “suburban strategy” that Clinton
followed in his recent victories is rooted in basic demographics.

A closer examination of the data, however, suggests that the conventional
wisdom is too simplistic. Although the population of our 12 cities has
declined, the total Democratic presidential votes coming out of them since
1968 has remained remarkably stable (see Figure 4). There are two main rea-
sons for this. First, 4 of our cities located on the two coasts (New York, Bos-
ton, Los Angeles, and San Francisco) reversed the postwar downward trend
and gained population in the 1980s, largely because of immigration. In 1990,
for example, 28% of New York City’s population was foreign born. The rate
at which immigrants naturalize, of course, varies, and studies show that
immigrants who become citizens tend to vote at low rates (Desipio 1996). In
the long run, however, just as immigration fueled the political rise of cities at
the beginning of the twentieth century, it is doing the same thing at the end of
the century (New York City’s proportion of state actual electorate actually
increased between 1980 and 1996; see Table 1).

A second reason why Democratic votes have remained stable in cities is
because the urban electorate has voted progressively more Democratic even

Sauerzopf, Swanstrom / THE URBAN ELECTORATE 81

Figure 4: Total Democratic Presidential Votes for Sample Cities and States, 1956-1996
NOTE: Vote tallies are in millions of votes cast for president.

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

as the rest of the nation has tended to vote more Republican in presidential
elections. A striking trend that emerges from our data is the increasing diver-
gence between the voting behavior of cities and the rest of the nation. We can
see this clearly here when we examine the index of difference (see Figure 5).
The index of difference measures the magnitude of the difference between
the partisan division of the vote in cities and in the rest of the state. The higher
the final score in a given year, the greater the difference in partisan voting
behavior between the city and noncity vote. For instance, if there were no dif-
ference between the partisan presidential voting patterns of city and noncity
voters, the index would be zero. If, on the other hand, all city voters voted
Democratic and voters in the remainder of the state voted Republican, then
the index of difference would be 200.

Figure 5 shows that since 1952, the index of difference has increased
markedly. In comparatively good years for Republicans, the index of differ-
ence is quite high. This illustrates dramatically that Republican presidential
victories have occurred, since 1968, in utter opposition to the preferences of
the urban electorate.

One would expect the index of difference to fall in years when the Demo-
crats do comparatively well, winning more votes outside the cities. However,
even in elections when the Democrats win most or all of the states covered by
our data, the index generally remains high, reaching its highest point in the
1996 election. Urban voters are increasingly “eccentric,” responding differ-
ently to political cues than voters in the rest of the nation. “Solid cities” have
largely replaced the “solid South” as the bedrock of the Democratic coalition.
Contrary to conventional wisdom, which emphasizes that elections are won
or lost in the suburbs, central-city votes are still crucial to Democratic victo-
ries in presidential elections (see Figure 2).

According to our analysis, our sample of 12 cities was controlling (con-
tributing more than 50% of the statewide plurality) in states representing 138
electoral votes in 1992 and 107 votes in 1996. In 1992, for example, New
York City provided 89% of Clinton’s nearly 1 million vote statewide margin.
Indeed, Clinton lost (narrowly) to Bush in suburban Long Island (Nassau and
Suffolk counties). Obviously, if we went beyond our limited sample to look
at, say, all central cities with populations more than 100,000, cities would
have been decisive in delivering many more electoral votes into the Demo-
cratic column. Although it would be exaggerating our point, it would be pos-
sible to turn the conventional wisdom on its head (at least for the 1992 and
1996 elections): How could the Republicans have expected to win the presi-
dency when they were incapable of seriously contesting for votes in the big
cities?

82 URBAN AFFAIRS REVIEW / September 1999

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

Given the fact that central-city voters increasingly vote differently from
the rest of the nation and increasingly find themselves on the losing side, per-
haps it is not surprising that the propensity for people in cities to vote has
declined relative to votes in the rest of the state (see Figure 6). The declining
electoral influence of cities is not just due to suburbanization but to demobili-
zation of the urban electorate. The effect is significant. Our data show that in
1996, although the total population of voting-age persons in our sample cities
represented a healthy 17.4% of the sample state figure, our city voters
accounted for only 13.5% of the sample state total.

Sauerzopf, Swanstrom / THE URBAN ELECTORATE 83

Figure 5: Index of Difference, 1956-1996
NOTE: The index of difference measures the gap between the partisan division of the sample cit-
ies and the party split in the rest of the sample states in presidential elections. The index is calcu-
lated by pooling Democratic and Republican votes for president for all of the sample cities and
for all of the sample states minus sample city returns. The percentage of the pooled sample city
presidential vote that went Democratic is subtracted from that which went Republican. The same
is done for the pooled presidential votes for sample states minus the sample city returns. The two
figures are then either added together if the sample city and state presidential votes were won by
different parties or subtracted one from the other if the same party won both the city and the state.
Index of difference = {(Extracity state % R – Extracity state % D) + or – (City % R – City % D)}.
The index of difference was developed by David Olson in 1992, when he was a doctoral student
in the Political Science Department at the State University of New York at Albany.

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

For the most part, Democrats have been able to take the urban vote for
granted because there is nowhere else for it to go. Although our data do not
enable us to determine the causes of central-city turnout decline, they do sug-
gest that the Democrats should not take their urban base completely for
granted. Clinton’s “suburban strategy” may have political costs. As Demo-
cratic candidates have distanced themselves from urban policies and sym-
bols, the turnout gap between cities and suburbs has grown. Even as central-

84 URBAN AFFAIRS REVIEW / September 1999

Figure 6: Sample City and State Relative Propensity to Vote, 1952-1996
NOTE: The relative propensity to vote compares sample city and state voter turnout rates ad-
justed for election-to-election changes in national turnouts. Here, in any given election, the na-
tional turnout is set at 1, and sample turnouts are recalculated proportionally. Sample state and
city turnouts compare pooled votes for president to pooled voting-age populations for all sample
cities and states. Adjusted sample state turnout rates do not include sample city population and
voting figures. Nardulli, Dalager, and Greco (1996) used the relative propensity to vote in their
comparative study of turnouts. Voting-age population (VAP) figures are calculated by straight-
line extrapolation between VAP figures from decennial Census of Population and Housing re-
ports. VAP figures through 1968 are based on a 21-year-old voting age.

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

city pluralities contributed significantly to both of President Clinton’s victo-
ries (see Figure 2), city turnout rates declined sharply in 1992 and 1996 rela-
tive to their respective states and the nation as a whole (see Figure 6). These
trends, taken together, suggest diminishing returns for an electoral strategy
that does not engage central-city voters.

In a more catastrophic scenario for the Democrats, the alienation of the
urban electorate could be mobilized by a charismatic third-party candidate.
According to aNew York Times/CBS voter survey, in 1992 Clinton actually
lost the white vote 39% to 41%. Nationwide, Clinton won 82% of the black,
largely urban, vote. If Jesse Jackson had run as an independent in 1992,
George Bush probably would have been elected to a second term.10

In short, our data suggest that cities are both more important to the Demo-
crats and more volatile than the conventional wisdom suggests. There may be
costs to the Democrats’suburban strategy: declining turnout among potential
Democratic voters in the cities.

A CALL TO EXAMINE THE SUBURBAN VOTE
AND CONTEXTUAL EFFECTS

Although central cities no longer have the power to determine most presi-
dential elections, they are still essential to Democratic presidential victories.
If anything, our research on voting in central cities underestimates their
importance. Our dichotomy between the central-city electorate and the rest
of the state, however, fails when examining the most rapidly growing seg-
ment of the statewide vote: the metropolitan voters outside central cities. At
the same time that cities have declined in relative importance, metropolitan
areas have grown. Nardulli, Dalager, and Greco (1996) found that in 1990,
the nation’s 32 largest metropolitan areas constituted more than 40% of the
national electorate. Therefore, investigating ways in which metropolitan areas
structure elections is critical to understanding national electoral politics.

The first issue that should be examined is where to draw the line between
the “urban” and “suburban” voters. The conventional wisdom naturalizes this
division, assuming that all suburban voters share something essential in com-
mon and are fundamentally distinct in political orientation from urban voters.
Reed and Bond (1991, 735) referred to this as the “mystified notion of the
suburbs as a new, coherent political constituency.” In fact, “cities” and “sub-
urbs” are problematic social constructions. The current dividing line between
cities and suburbs is arbitrary—being based, as it is, on the jurisdictional
boundaries between the original central city and everything else in the metro-
politan area. Although some cities are elastic, expanding their boundaries as

Sauerzopf, Swanstrom / THE URBAN ELECTORATE 85

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

their metropolitan area grows, others, including most of the cities in our sam-
ple, are inelastic, hemmed in by independent suburban municipalities (Rusk
1993). In these cases especially, as metropolitan areas continue to become
more urbanized, characteristics associated with central cities spill over into
neighboring suburbs.

Although cities increasingly resemble the conventional view, with minor-
ity populations, social problems, and physical decline concentrated within
their borders, suburbs are becoming more varied. When estimating the role
that metropolitan place plays in electoral politics, one too often fails to con-
sider the fact that many suburbs do not resemble the conventional image of
leafy, white, middle-class, residential refuges. As the dichotomous assump-
tions of the conventional view are challenged, so are the consequences of this
perspective for cities, metropolitan areas, and the Democratic Party.

Weiher (1991) explored the fantastic proliferation of incorporated
municipalities surrounding most of the nation’s large central cities. As he
showed, these suburbs over time have developed increasingly distinct eco-
nomic and social characteristics. These characteristics structure the cognitive
map, Weiher argued, that prospective residents use to inform their moving
decisions. Individual residential choices accumulate to further cement the
identities of specific suburbs. The result is a sociopolitical fracturing of met-
ropolitan areas. This fragmentation, as opposed to stratification, is character-
ized by substantial variation among suburbs. However, as variation among
suburbs is increasing, many suburban municipalities are becoming more
homogeneous internally. The result of this fragmentation may have serious
political consequences as identification with municipalities supersedes
understandings of metropolitan areas as interdependent polities.

The suburbs that represent the most critical problem for the conventional
view are those that more closely resemble central cities, typically inner-ring
suburbs. Many inner-ring suburbs—such as Harvey, Illinois; Lackawanna,
New York; Compton, California; and East Cleveland, Ohio—have more in
common with their central cities than they do with the stereotypical middle-
class suburbs. Orfield (1997) has hypothesized that the “swing” votes in
recent presidential elections have not been in the suburbs generally but in the
inner-ring, distressed (largely white, working-class) suburbs.11

Our analysis provides support for those scholars who argue that the cur-
rent period is not characterized so much by a Republican realignment as by
dealignment, the inability of either party to assemble a permanent winning
coalition.12Under dealignment, the ability of parties to mobilize voters weak-
ens, and voters withdraw from politics as such. Our results, which show
declining voter turnout both by urban and suburban electorates, support those
who argue that the current period is characterized more by dealignment than

86 URBAN AFFAIRS REVIEW / September 1999

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

realignment—at least in metropolitan areas outside the South. Future voting
research needs to disaggregate the suburban vote. Inner-ring voters may be
cross-pressured between economic concerns that pull them toward the
Democratic Party and social concerns, especially about race and crime, that
draw them toward the Republicans. By breaking out the inner-ring subur-
ban vote, researchers could test the hypothesis that this is where voter vola-
tility (switching parties) and withdrawal (declining turnout) are centered—
differentiating them from central cities and outer-ring suburbs.13

The second major thrust of electoral research on urban areas should be to
determine whether there is an independent “contextual effect” of place on
voting behavior. The aggregate data used here only allow us to describe vot-
ing patterns by place but not to determine why people vote or do not vote the
way they do. The results we have reported could be an artifact of the
individual-level characteristics of central-city and suburban electorates (e.g.,
their class or racial background). On the other hand, the results could also be a
result of the social networks, sources of information, and reference groups
rooted in places. Scholars have hypothesized, for example, that the move
from rural areas to cities caused people to become more conscious of the
interdependencies of society and the need for government to protect people
from the destabilizing and damaging effects of industrial capitalism (Lubell
1965, 47). Similarly, others have argued that the suburban context has a con-
servativizing effect, pulling erstwhile Democratic voters toward the Republi-
cans: As white ethnics migrated to the suburbs, they became home owners
and lived more privatized lifestyles, making them more susceptible to Repub-
lican appeals that favor the market over government (Harris 1954; Greenstein
and Wolfinger 1958; Schneider 1991, 1992).

More research needs to be done on the contextual effects of different met-
ropolitan residential environments. Aggregate data can be used to test for
contextual effects by controlling for income, education, race, and other
individual-level variables and seeing if place still has independent effects on
voting behavior.14 The best data, however, for examining contextual effects
are individual-level data (e.g., from sample surveys in which each case has
been geocoded so that researchers can move back and forth between individ-
ual and contextual levels of analysis) (Huckfeldt and Sprague 1993).

In one of the only studies of contextual effects of central cities, Wolman
and Marckini (1997) found that after controlling for a wide range of other
characteristics, central-city congressional representatives voted differently
than non-central-city representatives. Apparently, the central-city context
makes representatives vote more liberal. In one of the few studies of the effect
of suburban contexts on voting behavior, Oliver (1994, 1996) has used geo-
coded individual-level data to show that political participation declines,

Sauerzopf, Swanstrom / THE URBAN ELECTORATE 87

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

controlling for individual-level variables, in homogeneous suburban con-
texts. Oliver’s research suggests, beyond selection and rational interest struc-
turing, that residence within a specific kind of place has a transformative
effect. For Oliver, not only do some middle-class residents select suburbs as a
form of escape from dynamic urban political environments (as Schneider
1991, 1992 argued) but residents are also depoliticized by the socioeconomic
homogeneity and lack of significant political conflicts associated with these
places and their reform-style municipal governments.

If one takes seriously the potential of metropolitan places to exert politi-
cally transformative effects on their residents and extend it to other metro-
politan municipal types, then understanding how metropolitan areas struc-
ture political attitudes and behaviors becomes a critical project, central to
understanding national political divisions.

NOTES

1. The seminal statements on realignment theory come from Key (1955) and Burnham
(1970).

2. Realignment theory applies to congressional as well as presidential elections. In this
article, we deal only with presidential elections. For recent analyses of the decline of urban rep-
resentation in the U.S. Congress, see Caraley (1992, 19-23), Wolman and Marckini (1997), and
Cook (1997). It is important to note, however, that big cities were consistently underrepresented
in Congress and state legislatures until the Supreme Court reapportionment cases of 1962-1964.
Therefore, until this time, cities viewed presidential elections as the best way to make their con-
cerns heard in national politics (Barone 1990, 106).

3. Clinton knows William Julius Wilson and is on record endorsing Wilson’s views on the
problem of concentrated urban poverty (“A Visit with Bill Clinton” 1992). Likewise, Wilson, in a
1992New York TimesOp Ed piece, praised Clinton’s broad-based policy approach (Wilson
1992). Wilson, however, later criticized Clinton’s signing of the welfare reform act (Personal
Responsibility and Work Opportunity Act of 1996).

4. Two notable exceptions are Lubell (1965), who reported separate city returns for 12 cen-
tral cities from 1920 to 1964, and Nardulli, Dalager, and Greco (1996), who reported results for
32 center cities from 1828 to 1992. The weakness of the latter data set is that for 10 cities, county
data are used to represent the city vote.

5. The main published source of presidential election returns is the volumes ofCongres-
sional QuarterlyElection Research Center’sAmerica Votes(Scammon, McGillivray, and Cook
1956-1994). These report results for states and counties but not cities separate from counties.
The collection and storage of election data in the United States are remarkably backward. Lack-
ing a national repository of historical election results, the data are controlled by technologically
backward and often patronage-ridden county boards of election.

6. The data used for all tables and graphs for the period from 1920 to 1948 are from Elders-
veld (1949). Returns from 1952 to 1992 are from a variety of sources. All state and national elec-
tions returns are from Scammon, McGillivray, and Cook (1994). City election data for Milwau-
kee, Cleveland, Los Angeles, and Pittsburgh for this period are from county legislative manuals

88 URBAN AFFAIRS REVIEW / September 1999

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

and other boards of election records. Presidential returns for all other cities for this period are
from America Votes(Scammon, McGillivray, and Cook 1956-1994). Calculations that require
total sample city presidential votes are based on total votes cast for president. In a few cases,
however, total votes cast for major party presidential candidates had to be used as an approxima-
tion because returns for minor candidates were not available. All state and national votes cast for
president in 1996 are from the Federal Election Commission. Presidential returns for all sample
cities for 1996 were originally obtained from newspaper accounts. We want to thank the follow-
ing individuals who helped us to collect these data: Larry Bennett, Barbara Ferman, Robin Jones,
Robin Boyle, Dennis Keating, Patrica Atkins, Meredith Ramsay, Marc Levine, Rich DeLeon,
Ali Modarres, and Kenny Thomas. We were later able to confirm and edit these 1996 returns
based on returns as provided by local and state election offices for all of the cities except Phila-
delphia, Baltimore, St. Louis, and San Francisco. Presidential returns for all sample cities are
from the central cities themselves, with one exception: Eldersveld (1949) could not obtain elec-
tion returns for the city of Los Angeles and so substituted Los Angeles County returns. However,
Los Angeles election returns from 1952 to 1996 are for the city of Los Angeles.

7. Alvarez and True (1973) argued that urban ethnics were aligned toward the Democrats
throughout the period from 1896 to 1924.

8. Eldersveld’s (1949) criterion was based on the idea that if the city vote had swung in the
opposite direction, then the state would have been won by the other party. The idea that city votes
could swing from one party to the other is questionable, but we think Eldersveld’s criterion is
nevertheless a good general indicator of the electoral significance of cities.

9. Regional population shifts away from the older industrial states are a small part of the
reason for the declining influence of our sample cities. The 10 states representing our 12 cities
reached a peak of 233 electoral votes in 1960, declining to 215 electoral votes following the reap-
portionment based on the 1990 census. The only state in our sample to gain electoral votes during
this period was California.

10. Recently released tapes of Richard Nixon record him in 1972 discussing a scheme to
finance an independent black candidate. Nixon aide H. R. Haldeman discussed underwriting
Shirley Chisolm in the Democratic primaries but then recommended another course. “The argu-
ment is that if we can launch a Jesse Jackson or somebody outside the party we’re better off, we
can keep them bought.” Nixon replied, “Yeah.” Haldeman then proposed paying Jackson
$10,000 for every percentage of the vote he gets (quoted in Feinsilber 1996).

11. There is a growing appreciation of the distinctiveness of these places. See McCormick
and McKillop (1989) and Minerbrook (1992).

12. To sample the debate on realignment versus dealignement, see Niemi and Weisberg
(1993).

13. Once again, the difficulty in disaggregating the suburban vote is the lack of published
voting statistics by municipality. One notable exception is the Record on American Democracy
(ROAD) data set put together by King et al. (1997) at Harvard University. For the period from
1984 to 1990, they collected election data for all federal and statewide elections broken down to
the precinct level (approximately 170,000 nationwide). The data also were aggregated by “major
civil division,” which for most states allows analysis of voting results by suburban municipality.
The data set includes 1990 census data. The data also include geographic boundary files so that
maps can be drawn. The entire data set is available free on the World Wide Web at
http://data.fas.harvard.edu/ROAD/.

14. Because of the small size of ourN(12 cities), using regression analysis to test for contex-
tual effects in our case would have been highly problematic.

Sauerzopf, Swanstrom / THE URBAN ELECTORATE 89

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

REFERENCES

Alvarez, D. J., and E. J. True. 1973. Critical elections and partisan realignment: An urban test
case.Polity 5:563-76.

Andersen, K. 1979.The creation of a democratic majority 1928-1936. Chicago: Univ. of Chi-
cago Press.

Barone, M. 1990.Our country: The shaping of America from Roosevelt to Reagan. New York:
Free Press.

Burnham, W. D. 1970.Critical elections and the mainstreams of American politics. New York:
Norton.

Caraley, D. 1992. Washington abandons the cities.Political Science Quarterly107 (1): 1-30.
Cook, R. 1997. Rare combination of forces may make history of ‘94.Congressional Quarterly

Weekly Report53:1076-81.
Degler, C. N. 1964. American political parties and the rise of the city: An interpretation.Journal

of American History51(1): 41-59.
Desipio, L. 1996. Making citizens or good citizens? Naturalization as a predictor of organiza-

tional and electoral behavior among Latino immigrants.Hispanic Journal of Behavioral Sci-
ences18:194-213.

Edsall, T., and M. D. Edsall. 1991.Chain reaction: The impact of race, rights, and taxes on
American politics. New York: Norton.

Eldersveld, S. J. 1949. The influence of metropolitan party pluralities in presidential elections
since 1921: A study of twelve key cities.American Political Science Review43 (6):
1189-1206.

Feinsilber, M. 1996. Nixon tapes reveal talk of financing a black candidate in ‘72.Albany Times
Union, 28 November.

Greenberg, S. B. 1995.Middle class dreams: The politics and power of the new American major-
ity. New York: Times Books.

Greenstein, F., and R. Wolfinger. 1958. The suburbs and shifting party loyalties.Public Opinion
Quarterly22:473-82.

Harris, L. 1954.Is there a Republican majority?New York: Harper & Row.
Huckfeldt, R., and J. Sprague. 1993. Citizens, contexts, and politics. InState of the discipline II,

edited by Ada Finifter, 281-303. Washington, DC: American Political Science Association.
Key, V. O., Jr. 1955. A theory of critical elections.Journal of Politics17:3-18.
King, G., B. Palmquist, G. Adams, M. Altman, K. Benoit, C. Gay, J. B. Lewis, R. Mayer, and

E. Reinhardt. 1997.The record of American democracy, 1984-1990. Produced by Harvard
University. Ann Arbor, MI: Interuniversity Consortium for Political and Social Research
[distributor].

Lubell, S. 1965.The future of American politics. New York: Harper & Row.
MacRae, D., and J. Meldrum. 1960. Critical elections in Illinois: 1888-1958.American Political

Science Review54:669-83.
McCormick, J., and P. McKillop. 1989. The other suburbia: An ugly secret in America’s suburbs:

Poverty.Newsweek, 26 June, 22-24.
Minerbrook, S. 1992. A tale of two suburbias: The decline of blue-collar suburbs and growth of

“edge cities” create a new kind of isolation.U.S. News and World Report, 9 November,
32-40.

Mollenkopf, J. H. 1983.The contested city. Princeton, NJ: Princeton Univ. Press.
Nardulli, P. F., J. K. Dalager, and D. E. Greco. 1996. Voter turnout in U.S. presidential elections:

An historical view and some speculation.PS: Political Science and Politics29:480-90.

90 URBAN AFFAIRS REVIEW / September 1999

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

Niemi, R. G., and H. F. Weisberg. 1993. Dealignment and realignment in the current period. In
Controversies in voting behavior, 3d ed., edited by R. G. Neimi and H. F. Weisberg, 321-32.
Washington, DC: Congressional Quarterly.

Oliver, J. E. 1994. City size and political participation in the fragmented American metropolis.
Paper presented at the annual meeting of the American Political Science Association, New
York, September.

. 1996. The influence of social context on patterns of political mobilization. Paper pre-
sented at the annual meeting of the American Political Science Association, San Francisco,
September.

Orfield, M. 1997.Metropolitics: A regional agenda for community and stability. Washington,
DC: Brookings Institution.

Reed, A., Jr. and J. Bond. 1991. Equality: Why we can’t wait.Nation, 9 December, 733-37.
Rourke, F. E. 1965. Urbanism and the national party organizations.Western Political Quarterly

18:149-63.
Rusk, D. 1993.Cities without suburbs. Washington, DC: Woodrow Wilson Center Press.
Scammon, R. M., A. McGillivray, and R. Cook. 1956-1994.America Votes. Washington, DC:

Elections Research Center, Congressional Quarterly.
Schneider, W. 1991. Rule suburbia.National Journal, 28 September.
. 1992. The suburban century begins.Atlantic Monthly, July, 33-44.
Sundquist, J. L. 1973.Dynamics of the party system: Alignment and realignment of political par-

ties in the United States.Washington, DC: Brookings Institution.
A visit with Bill Clinton. 1992.Atlantic Monthly, October.
Wilson, W. J. 1987.The truly disadvantaged: The inner city, the underclass, and public policy.

Chicago: Univ. of Chicago Press.
. 1992. The right message.New York Times, 17 March.
Weiher, G. 1991.The fractured metropolis: Political fragmentation and metropolitan segrega-

tion. Albany: SUNY.
Wolman, H., and L. Marckini. 1997. Changes in central city representation and influence in Con-

gress. Paper presented at the annual meeting of the Urban Affairs Association, Toronto, Can-
ada, April.

Richard Sauerzopf is a graduate student at the State University of New York at Albany.
He is writing his Ph.D. dissertation on voting behavior in inner-ring (white, working-
class) suburbs. A case study of the Detroit metropolitan area, the research includes an
analysis of aggregate voting, as well as survey data. He is a visiting instructor in the
Geography and Urban Planning Department, College of Urban Labor and Metropolitan
Affairs at Wayne State University.

Todd Swanstrom is a professor of political science at State University of New York–Al-
bany. The coauthor of a critical theme text on American politics,The Democratic Debate
(2d ed., Houghton Mifflin, 1998), he has written extensively on urban politics and politi-
cal economy. Currently, he is researching the political causes and consequences of class
segregation across space in U.S. metropolitan areas and the “nonprofitization” of fed-
eral housing policy.

Sauerzopf, Swanstrom / THE URBAN ELECTORATE 91

at Glasgow University Library on October 15, 2012uar.sagepub.comDownloaded from

http://uar.sagepub.com/

Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=cesw20

Download by: [University of Glasgow] Date: 12 September 2016, At: 08:05

European Journal of Social Work

ISSN: 1369-1457 (Print) 1468-2664 (Online) Journal homepage: http://www.tandfonline.com/loi/cesw20

Placement, protective and risk factors in the
educational success of young people in care: cross-
sectional and longitudinal analyses

Robert J. Flynn , Nicholas G. Tessier & Daniel Coulombe

To cite this article: Robert J. Flynn , Nicholas G. Tessier & Daniel Coulombe (2013) Placement,
protective and risk factors in the educational success of young people in care: cross-
sectional and longitudinal analyses, European Journal of Social Work, 16:1, 70-87, DOI:
10.1080/13691457.2012.722985

To link to this article: http://dx.doi.org/10.1080/13691457.2012.722985

Published online: 10 Sep 2012.

Submit your article to this journal

Article views: 521

View related articles

Citing articles: 6 View citing articles

http://www.tandfonline.com/action/journalInformation?journalCode=cesw20

http://www.tandfonline.com/loi/cesw20

http://www.tandfonline.com/action/showCitFormats?doi=10.1080/13691457.2012.722985

http://dx.doi.org/10.1080/13691457.2012.722985

http://www.tandfonline.com/action/authorSubmission?journalCode=cesw20&show=instructions

http://www.tandfonline.com/doi/mlt/10.1080/13691457.2012.722985

http://www.tandfonline.com/doi/citedby/10.1080/13691457.2012.722985#tabModule

Placement, protective and risk factors
in the educational success of young
people in care: cross-sectional and
longitudinal analyses

Des facteurs de placement, de
protection, et de risque dans le succès
scolaire des jeunes placés: Analyses
transversales et longitudinales
Robert J. Flynn, Nicholas G. Tessier & Daniel Coulombe

In the present study, we formulated and tested a basic model of the educational success of

young people in out-of-home care. We used data from 2007 to 2008 and 2008 to 2009 on

a sample of 1106 young people in care in Ontario, Canada. The youths were 12�17 years
of age; 56.24% were male and 43.76% female. The indicators of educational success in

both years were the youth’s average marks and the youth’s school performance in reading,

math, science and overall, as rated by his or her caregiver. Based on resilience theory and

on a model of the influence of maltreatment on educational achievement, our model

included four categories of predictors: control variables (youth gender and age and, in the

longitudinal analyses, the year 7 value of the year 8 dependent variable), three placement

types (foster, kinship care or group homes), three risk factors (previous repetition of a

grade in school, a health-related cognitive impairment index and a measure of

behavioural difficulties) and three protective factors (caregiver involvement in the

youth’s school, caregiver educational aspirations for the young person and the youth’s

total number of internal developmental assets). Cross-sectional and longitudinal

Correspondence to: Robert J. Flynn, School of Psychology and Centre for Research on Educational and

Community Services, University of Ottawa, 34 Stewart Street, Ottawa, ON K1N 6N5, Canada. Email:

rflynn@uottawa.ca

# 2013 Taylor & Francis

European Journal of Social Work, 2013

Vol. 16, No. 1, 70�87, http://dx.doi.org/10.1080/13691457.2012.722985

http://dx.doi.org/10.1080/13691457.2012.722985

hierarchical regression analyses provided mixed support for the proposed model. The

youth’s gender, level of behavioural difficulties and number of developmental assets, and

the caregiver’s educational aspirations for the young person, emerged as the most

consistent predictors of educational success. The implications and limitations of the

findings were discussed.

Keywords: Educational Success; Youth In Care; Risk And Protective Factors; Ontario

Looking After Children project

Dans cette étude, nous avons formulé et évalué un modèle de base du succès scolaire des

jeunes qui ont été placés en dehors de leur famille d’origine. Nous avons utilisé des

données dont la saisie s’est faite en 2007�2008 et 2008�2009 auprès d’un échantillon
composé de 1106 jeunes placés dans la province d’Ontario au Canada. L’âge des jeunes

étaient entre 12 et 17 ans; 56.24% étaient de sexe masculin et 43.76% de sexe féminin.

Les indicateurs du succès éducatif chaque année étaient la moyenne des notes scolaires du

jeune ainsi que sa performance en lecture, mathématiques, science, et dans l’ensemble des

matières. Notre modèle, basé sur la théorie de la résilience ainsi que sur un modèle de

l’influence de la maltraitance sur le rendement scolaire, incluait quatre catégories de

prédicteurs : des variables de contrôle statistique (le sexe et l’âge et, dans les analyses

longitudinales, la valeur dans l’an 7 de la variable formant la variable dépendante dans

l’an 8), trois types de placement (des familles d’accueil, des familles de parenté, et des

foyers de groupe), trois facteurs de risque (le fait d’avoir redoublé à l’école, un indice

composé de plusieurs difficultés cognitives reliées à la santé, et un indice des difficultés du

jeune sur le plan du comportement), et trois facteurs protecteurs (l’implication du parent

d’accueil dans la vie de l’école du jeune, les aspirations du parent d’accueil envers le

jeune, et le nombre total des acquis internes de développement du jeune). Des analyses

de régression transversales et longitudinales ont fourni un soutien partiel au modèle

proposé. Le sexe, le niveau de difficultés de comportement, le nombre d’acquis de

développement, et les aspirations du parent d’accueil envers le jeune se sont révélés les

meilleurs prédicteurs du succès éducatif. Les implications ainsi que les limites des

résultats ont été explorés.

Mots clés: succès scolaire; jeunes placés; facteurs de risque et de protection; le projet

S’Occuper des enfants en Ontario

Introduction

Many young people in out-of-home care (hereafter, ‘in care’) in Europe, North

America and other regions experience educational difficulties, including cognitive

deficits, poor problem-solving and reasoning skills, inconsistent school attendance,

below-average academic performance and low scores on standardised tests of

academic achievement in reading, writing and mathematics (Flynn et al., 2004;

European Journal of Social Work 71

Jackson, 2007; Slade & Wissow, 2007; Trout et al., 2008). In the USA, Trout et al.

(2008) conducted a comprehensive review of the American educational research

conducted on young people in care and published in journals during the period of

1940�2006. The 29 studies reviewed described the academic status of a total of 13,401
young people in care; they had a mean age of 12.9 years, and 52% were male. On the

whole, the young people in care had a much higher level of risks in school

functioning than the general youth population, with frequent changes in schools and

high levels of grade retention, suspension and school dropout. The young people in

care were nearly three times as likely to be involved in special education as their age

peers in the general population, tended to score in the low to low-average range on

measures of academic achievement, and were often rated by their teachers as

academically at risk. Trout et al. (2008) suggested that many youths in care presented

academic deficits similar to those identified in reviews of the academic status of other

at-risk populations, including children with emotional and behavioural disorders and

maltreated children who had been reported to child welfare agencies.

In the UK, Jackson (2007) reviewed the progress made over the last 20 years in

improving the educational outcomes of young people in care. She found that

considerably better data on their academic status now existed and that coordination

had improved between local educational and child welfare services. In addition, the

issue of education for children in care had risen to the top of the policy agenda, with

local authorities now required to promote their educational attainment. Berridge

(2007) noted that there was recent evidence of a slight improvement in outcomes in

the UK. The proportion of young people in care obtaining one General Certificate of

Secondary Education (GCSE) or equivalent had increased from 53% to 60% between

2002 and 2005, but this improved level of achievement was still much lower than the

level of 96% on the same criterion in the general child population. Moreover, the

proportion of young people in care who had obtained five or more GCSEs had

increased only from 8% to 11% during 2002�2005, whereas in the general population
the level had increased from 52% to 56%.

Jackson (2007) agreed with Berridge (2007) that there had been little attempt in

the UK to understand the basic reasons for the achievement gap, while disagreeing

with his view that the answer lay in the characteristics of the families of origin of

children in care rather than in weaknesses of the care system itself. Jackson (2007)

suggested that discussions of foster parent recruitment and selection had largely

ignored the large body of research that identified a strong link between children’s

academic performance and the educational level and expectations of their parents or

caregivers. She added that the problem in the UK of ensuring an adequate education

for young people in the care system also characterised the child welfare systems of

other English-speaking countries. This appears to be true for Australia (Cashmore

et al., 2007) and the USA (Trout et al., 2008), and also, as we shall see, of Canada. The

educational attainment gap also appears to characterise other countries, however,

such as Sweden (Vinnerljung et al., 2005) and Norway (Iversen et al., 2010). Weyts

(2004), cited in Iversen et al. (2010), for example, found that the reading skills of

72 R. J. Flynn et al.

Norwegian children in care were no better than those of children in care in England,

the Netherlands and Spain.

In Canada, the relatively few studies conducted to date on the educational

achievement of young people in care provide a sketch similar to the portrait in the

countries previously mentioned. In the province of Ontario, Flynn and Biro (1998)

found that children in care had higher rates of grade retention and school suspension

than their age peers in the general population. Flynn et al. (2004) compared the

ratings made by caregivers in Ontario of the educational performance of children and

adolescents in their care with the ratings made by parents in the general Canadian

population of their own children’s educational progress. Eighty per cent of the young

people in care aged 10�15 years and 78% of the children in care aged 5�9 years were
rated by their foster parents as performing in the same range as the lowest third of the

children in the general population, who had been rated by their parents on the same

composite measure of reading, spelling, math and overall educational

performance.

In more recent research in Ontario, Miller et al. (2008) pointed to some possible

reasons for the achievement gap. Sixty-eight per cent of the young people in care aged

10�15 years in their sample had changed schools three or more times for reasons
unrelated to normal progression through the school system, and the percentage

repeating a grade was 16% among 5�9 year olds in care, 27% among 10�15 year olds
and 32% among 16�20 year olds.

In Ontario, as elsewhere, academic difficulties seem especially prevalent among

boys in care. Miller et al. (2009) found that girls in care were less likely than boys to

undergo assessments for learning-related problems (58% vs. 79%) or to receive

special academic help at school (49% vs. 69%). Caregivers also rated the girls’ school

performance more highly: 24% of the girls were rated as performing ‘Very Well’ or

‘Well’ in written work, compared with 13% of the boys; 41% of the girls (vs. 28% of

the boys) were seen as doing ‘Very Well’ or ‘Well’ in reading; and 29% of the girls

(vs. 20%) were rated as doing ‘Very Well’ or ‘Well’ overall. Only in math were equal

proportions of boys (23%) and girls (24%) rated as performing ‘Very Well’ or ‘Well’.

The girls also tended to be more positive about education-related matters than the

boys: 37% (vs. 28%) said, for example, that they read ‘for fun’ every day, and 40%

(vs. 26%) aspired to attain one or more university degrees.

Recently, the Ontario Association of Children’s Aid Societies (OACAS, 2010), in

collaboration with 43 of its local member agencies, carried out a review of the files of

4694 Crown Wards or former Crown Wards (i.e., young people in relatively long-

term out-of-home care). The youths were 16�20 years of age and had been in school
during 2008�2009. OACAS compared the results from 2008 to 2009 with those from
a similar study conducted in 2006�2007. The results showed that the youths in care
had results that fell well short of those of their age peers in the general population,

although some progress had been made in the two-year interval since the initial

study. The percentage of 16 and 17 year olds not attending any educational

programme (Ontario requires school attendance up to age 18) had declined from

14% to 7%. Graduation from secondary school had increased by 2%, from 42% to

European Journal of Social Work 73

44%, compared, however, with a larger increase in the general youth population,

from 75% to 79%. The number of former Crown Wards aged 18�20 who were
enrolled in post-secondary education (PSE) had increased from 21% to 23%,

compared to 39% in the general population. Of those in PSE, 81% were now in

community colleges, including apprenticeship programmes (vs. 84% two years

earlier), compared with 19% in university (vs. 16% two years earlier).

The purpose of the present study was to formulate and test a basic model of

educational success among young people in care that included the standard control

variables of gender and age and selected placement, protective and risk factors. In

formulating the model, we were guided by two theoretical frameworks. First, we drew

upon Masten’s (2006) conceptualisation of resilience theory: ‘Resilience refers to

positive patterns of functioning or development during or following exposure to

adversity, or, more simply, to good adaptation in a context of risk’ (p. 4). Masten

(2006) noted that ‘Direct predictors of better outcomes often are described as assets

or resources’ (p. 6). As predictors of academic achievement, good examples of assets

would be high-quality parenting or higher IQ scores. In the risk-related child-welfare

context of providing care for formerly abused or neglected young people, assets may

be called protective factors because they appear to play an especially important role in

positive adaptation. Risk factors, on the other hand, are predictors of undesired

outcomes. In the context of educational performance, abusive or neglectful parenting

or severe poverty would be good examples. Masten (2006) suggested that a typical

‘short list’ of factors associated with resilience in children and youth includes

relationships and parenting (e.g., strong links with one or more effective parental

figures; high-quality parenting that provides affection, monitoring and expectations);

individual differences (e.g., learning and problem-solving skills; self-control of

attention, emotional arousal and impulses); and community context (e.g., effective

schools; positive organisations).

Second, we used the framework proposed by Slade and Wissow (2007), in which

maltreatment is hypothesised as influencing educational outcomes through two main

pathways. The first pathway consists of mental health problems stemming from abuse

or neglect, including disruptive classroom behaviours, suspensions or difficulties of

concentration and motivation. The second pathway comprises inadequate cognitive

stimulation at home, lower-quality informal and formal education, and poorly

developed academic skills in word knowledge, literacy and numerical reasoning. Slade

and Wissow (2007) further hypothesise that the maltreated youth’s mental health

difficulties and low academic skills raise his or her risk of not adhering to behavioural

norms at school, obtaining less support from teachers and classmates, doing poorly

on homework assignments and tests and ultimately performing inadequately in

school.

In the present study, we defined educational success in terms of two outcomes: the

youth’s average marks during the last year in school, and his or her school performance

as rated by the caregiver on a composite measure of reading, math, science and

overall performance. Based on the literature reviewed, we included four categories of

74 R. J. Flynn et al.

predictors in our basic model of educational success. The first category consisted of

the standard control variables of gender and age, although we also saw female gender

as a protective factor because of girls’ greater educational success than boys (Miller

et al., 2009). Regarding age, we had no expectation that older youths would perform

any better or worse than younger youths. The second category of predictors,

corresponding to Masten’s (2006) ‘community context’ factor, comprised the type of

placement in which the young person had been living, whether a foster home, kinship

care home or group home. In line with McClung and Gayle’s (2010) findings

regarding the role of placement type, we anticipated that youths living in smaller

settings (i.e., foster or kinship care homes) would succeed better in school than those

residing in larger settings (i.e., group homes). The third category of predictors

consisted of three risk factors that we believed would be negatively associated with

educational success: having previously repeated a grade in school (Flynn & Biro,

1998), a lower level of cognitive functioning (Masten, 2006) and a higher level of

behavioural difficulties (Slade and Wissow, 2007). The fourth category of predictors

comprised three protective factors suggested by Masten’s emphasis on the key role of

assets or resources in promoting better adaptation. Two parenting-related assets were,

respectively, a greater degree of involvement by the parental figure (caregiver) with

the young person’s school, and a higher level of aspirations on the part of the

caregiver regarding the young person’s eventual level of educational attainment. The

third resource was the young person’s level of internal developmental assets, chosen

because of prior evidence that a greater number of developmental assets is associated

with greater educational success both in the general population (Scales et al., 2006)

and in young people in care (Flynn & Tessier, 2011).

Method

Participants and Service Context

The sample consisted of 1106 young people in care, aged 12�17; 56.24% were male
and 43.76% female. The young people had participated in both year 7 (2007�2008)
and year 8 (2008�2009) of the Ontario Looking after Children (OnLAC) project
(Flynn et al., 2006; Flynn et al., 2009), which annually monitors the service needs

and developmental outcomes of children and youth who have been in care for a

year or more in the province. The OnLAC project is mandated by the provincial

government in local Children’s Aid Societies (CASs) across Ontario to encourage

more data-based decision-making about children’s needs, improve the quality of the

substitute parenting they receive and enhance their short-term and long-term

outcomes.

At the time the data were gathered, child welfare services in Ontario were

provided or supervised by a network of 53 government-funded CASs, the number of

which was beginning to be reduced through amalgamations in a search for greater

efficiency and sustainability. There were approximately 18,500 children and youths in

out-of-home care, over half of whom were teenagers (Commission to Promote

European Journal of Social Work 75

Sustainable Child Welfare, 2010). Excluding older youths in supported transitional or

independent living, 80% of the days of care provided in Ontario in 2009�2010 were
spent in family-based care (i.e., family foster care or kinship care), 15% in group care

and 5% in other settings (e.g., hospitals, youth justice settings or children’s mental

health settings). Approximately 40% of expenditures in child welfare in Ontario were

allocated to in-care services (Commission to Promote Sustainable Child Welfare,

2010).

Instrument

The child welfare worker responsible for a given young person in care administered

the data collection instrument from which all the measures in the present study were

taken, namely, the second Canadian adaptation of the Assessment and Action Record

from Looking after Children (AAR-C2-2006; Flynn et al., 2009). The AAR-C2-2006

consists of eight age-appropriate formats, each of which comprises a family of

instruments. Administration of the tool is done annually, in the form of a structured

conversational interview in which the young person in care (if aged 10 or over), his or

her caregiver and his or her child welfare worker take part. The information gathered

in the AAR-C2-2006 interview is used to carry out a major revision, each year, of the

young person’s plan of care for the ensuing 12 months.

The AAR-C2-2006 consists of questions that cover nine areas: a background

section, completed mainly by the child welfare worker, that provides basic descriptive

information on the young person, caregiver and child welfare worker; seven sections,

rated mainly by the young person in care and his or her caregiver, that assess the

youth’s service needs and developmental outcomes in each of the Looking After

Children domains, namely, health, education, identity, social and family relation-

ships, social presentation, emotional and behavioural development and self-care

skills; and a final section, adapted from the work of the Search Institute (Scales et al.,

2000), in which the child welfare worker rates the young person’s acquisition of 40

different developmental assets (Flynn et al., 2009).

Criterion (Outcome) Measures

Two criterion measures of educational success were selected from the year 7 AAR-C2-

2006 data for the eventual cross-sectional analyses, and the same two measures were

taken from the year 8 data for the longitudinal analyses. The first measure was the

average marks that the young person in care had attained during the previous year in

school or during the last year he or she had been enrolled in school. The possible

values were 4 (B50%), 5 (51�60%), 6 (61�70%), 7 (71�80%), 8 (81�90%) or 9
(90�100%). The second measure was the young person’s school performance, as
assessed by the caregiver on a four-item composite scale consisting of ratings of how

well the youth had done in school in years 7 and 8 on language and reading,

mathematics, science and overall. Each item was rated on a 3-point scale: 0 �Very

76 R. J. Flynn et al.

Poor or Poor; 1 �Average; and 2 �Very Well or Well. The total score on school
performance could range from zero to eight.

Predictor Measures
Control

variables

Female gender was assigned the value of one and male gender the value of zero. The

young person’s age was his or her age in years as of the date that the AAR-C2-2006

interview had begun in OnLAC year 7.

Placement type

Three dichotomous variables were used to represent the type of placement setting in

which the young person in care had resided in OnLAC year 7: Foster Home (1 �Yes,
0 �Other), Kinship Care Home (1 �Yes, 0 �Other) or Group Home (1 �Yes,
0 �Other). In the regression analyses, the group home dichotomy served as the
reference category and was thus omitted.

Risk-factor measures

The measures of the three risk factors were taken from the AAR-C2-2006. The first

was a dichotomy that indicated whether the young person in care had ever repeated a

grade in school (1 �Yes, 0 �No). The child welfare worker provided this
information, with assistance, as needed, from the caregiver and young person. The

second risk-factor measure was a health-related Cognitive Impairments Index that we

created. The index consisted of the youth’s total number of cognitively related long-

term health conditions (out of a maximum of four), as rated by the youth’s child

welfare worker. These health conditions had lasted or been expected to last for 6

months or more, had been diagnosed by a health professional, and, by their very

nature, were likely to pose a challenge to the youth’s cognitive functioning. The child

welfare worker indicated which of the following health conditions the youth had:

Learning Disability (1 �Yes, 0 �No), Developmental Disability (1 �Yes, 0 �No),
Attention-Deficit Disorder (1 �Yes, 0 �No) and Fetal Alcohol Syndrome (1 �Yes,
0 �No). The score on the index could range from zero to four.

The third risk-factor measure was the youth’s score on the Total Difficulties Scale of

the Strengths and Difficulties Questionnaire (SDQ; Goodman, 1997), which is

embedded in the AAR-C2-2006. The SDQ Total Difficulties Scale, composed of 20

behavioural items rated by the caregiver (0 �Not True, 1 �Somewhat True and
2 �True), covers the domains of emotional symptoms, conduct problems,
hyperactivity/inattention and peer problems. The total score could range between 0

and 40.

Protective-factor measures

The three protective-factor measures were also taken from the AAR-C2-2006. The

first two were educationally relevant aspects of high-quality parenting on the part of

the caregiver. The Caregiver School Involvement Index consisted of the number of

European Journal of Social Work 77

school activities (out of a maximum of eight) in which the caregiver reported having

been involved during the current or last school year, such as volunteering in the

young person’s class or attending a school event in which the young person had

participated. The second protective-factor measure, Caregiver Aspirations, consisted

of the caregiver’s expressed hope that the young person in care would achieve a

certain level of education (1 �Primary or elementary school; 2 �Secondary or high
school; 3 �Trade, technical, vocational school or business college; 4 �Community
college or nursing school; 5 �University).

The third protective-factor measure consisted of the number of Internal

Developmental Assets (out of a maximum of 20) possessed by the young person

in care. The 20 asset items were rated by the youth’s child welfare worker

(1 �Yes, 0 �Uncertain or No). The internal assets covered four areas: the
youth’s commitment to learning (e.g., ‘The young person is motivated to do well

in school’); the young person’s positive values (e.g., ‘The young person accepts and

takes personal responsibility’); the youth’s social competencies (e.g., ‘The young

person knows how to plan ahead and make choices’); and the young person’s positive

identity (e.g., ‘The young person feels that he/she has control over ‘‘things that

happen to me’’’).

Data Analysis
Preliminary analyses

We began by assessing whether we would need to conduct multi-level analyses of our

data, in which the young people in care would be nested within their respective local

CASs. We found that this would not be necessary, as the amount of overall variance

accounted for by CASs in our two measures of educational success was very small

and statistically non-significant. We also evaluated whether we would need to

control for the particular geographic region (out of a total of six) in Ontario within

which the youth’s CASs was located. This, too, turned out to be unnecessary, as

geographic region was not significantly related to either measure of educational

success.

Hierarchical regression analyses

We related our two criteria of educational success, average marks and school

performance, to the four categories of predictors (controls, placement type, risk

factors and protective factors), in a series of hierarchical regression analyses. Two

cross-sectional analyses were conducted, in which the year 7 outcomes served as the

dependent variables. Similarly, two longitudinal analyses were carried out, in which

the year 8 outcomes were the dependent variables and in which the year 7 values of

the criterion variables were entered as control variables in step 1 (along with gender

and age). These longitudinal analyses tested the ability of our predictive models to

account for change from year 7 to year 8 in the two educational outcomes.

78 R. J. Flynn et al.

Results

Descriptive Results

Paired t-tests (not shown) revealed that, on the outcome of young people’s average

marks, there was no significant mean change from year 7 (M�6.54, SD �1.01) to
year 8 (M�6.56, SD�1.04; t(784) �0.29, p�0.77). There was also no change on
the other outcome, school performance, between year 7 (M�4.21, SD�2.37) and
year 8 (M�4.32, SD�2.26; t(857) �1.42, p�0.16).

Table 1 presents basic descriptive information on the study variables. On some, the

effective sample size was B1106 because, for example, some young people were in

ungraded classrooms and thus had no data on the outcome of average marks. Other

youths were in placement settings that were neither foster, kinship, nor group homes,

such as mental health or juvenile justice residential settings and were eliminated from

the analyses. On other variables, the caregiver or child welfare worker were uncertain

about the young person’s previous scholastic history (e.g., regarding whether the

young person had previously repeated a grade).

Over half of the young people (52.1%) had no health-related cognitive

impairments, whereas 29.1% had only one, 13.7% had two, 4.3% had three and

0.8% had the maximum of four. The internal consistency coefficients (Cronbach’s

alphas) for four of the multi-item measures were excellent, in the 0.80s. On our two

constructed indexes, internal consistency was lower. It was acceptable (0.62) in the

case of the eight-item Caregiver Involvement in School Index but marginal on the

Table 1 Means (or percentages), standard deviations and Cronbach’s alphas for study

variables

Variable N Mean (or %) SD Cronbach’s Alpha

Outcomes
Average marks*Year 7 894 6.51 1.03 �
Average marks*Year 8 878 6.54 1.03 �
School performance in Year 7 975 4.11 2.39 0.89
School performance in Year 8 942 4.28 2.27 0.88

Control variables
Gender (1 �Female, 0 � Male) 1106 43.76% � �
Age (in years) 1106 13.99 1.34 �

Placement type
Foster home (1 �Yes, 0 � Other) 1058 72.40% � �
Kinship care home (1 � Yes, 0 �Other) 1058 7.09% � �
Group home (1 �Yes, 0 � Other) 1058 20.51% � �

Risk factors
Previously repeated a grade (1 � Yes; 0 �No) 825 20.73% � �
Cognitive impairment index 1106 0.73 0.91 0.45
SDQ total difficulties 1045 12.64 7.44 0.87

Protective factors
Caregiver involvement in school 902 3.10 1.67 0.62
Caregiver aspirations 855 3.79 1.01 �
Internal developmental assets 1106 12.75 5.10 0.88

European Journal of Social Work 79

brief Cognitive Impairments Index (Cronbach’s alpha �0.45). Despite this, the latter
correlated significantly and in the expected direction with virtually all of the other

variables (see Table 2).

Predictors of Educational Success
Inter-correlations

Table 2 showed, as anticipated, that the four measures of educational success were

positively and significantly inter-correlated. Other findings were also as expected: the

girls had better outcomes than the boys in both years; all 12 of the correlations

between the risk factors and outcomes were negative and statistically significant, and

10 of the 12 correlations between the protective factors and outcomes were positive

and significant. On the other hand, the correlations between the three types of homes

with the educational outcomes were weak, although in the expected direction, and

only half were statistically significant.

Hierarchical Regressions
Average marks

Table 3 displays the results for the cross-sectional (left-hand panel) and longitudinal

(right-hand panel) regression models for the outcome of average marks. In the cross-

sectional model, the control (step 1), risk (step 3) and protective factors (step 4) all

accounted for statistically significant increments in the amount of variance accounted

for in year 7 average marks, and there was a trend in the same direction in the case of

the placement types (step 2).

The cross-sectional model as a whole accounted for 20.5% of the variance in year 7

average marks, with the risk and protective factors together explaining 17.2%. As

predicted, youths in the foster and kinship care homes had higher average marks than

those in group homes (step 2), although the relationship was modest. Once the three

risk factors had been entered into the model, however (at step 3), the beta coefficients

for the placement types were reduced to nearly zero, indicating that their association

with educational success was probably mediated by the risk factors. In the final model

(at step 4), the youth’s total number of internal developmental assets was the

strongest predictor of average marks. Caregiver aspirations and youth behavioural

difficulties were also statistically significant predictors, and there was a trend in this

direction in the instance of female gender and previous repetition of a grade in

school.

As previously noted, there was no significant mean change in the youths’ average

marks between year 7 and year 8. Thus, it was not surprising that the longitudinal

model explained little additional variance in year 8 average marks, once the role of

year 7 average marks and female gender (which was associated with improved marks

at all four steps) had been taken into account. Only two additional predictors*the
caregiver’s level of involvement in school activities and the youth’s level of internal

developmental assets*were significantly associated with year 8 average marks.

80 R. J. Flynn et al.

Table 2 Inter-correlation matrix

Variables 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

1. Average marks*Year 7 �
2. Average marks*Year 8 43*** �
3. School performance*Year 7 64*** 35*** �
4. School performance * Year 8 37*** 60*** 48*** �
5. Gender (1 � F, 0 � M) 13*** 18*** 13*** 13*** �
6. Age (in years) �01 �03 00 �07* 02 �
7. Foster home (1 � Y, 0 � N) 01 04 07* 04 10*** �09** �
8. Kinship care home

(1 � Y; 0 � N)
04 08* 07* 06 02 �02 �45*** �

9. Group home (1 � Y, 0 � N) �04 �11** �13*** �09** �13*** 11*** �82*** �14*** �
10. Previously repeated a grade

(1 � Y; 0 � N)
�10** �10** �13*** �13*** �04 �00 04 �06 �00 �

11. Cognitive impairment index �11*** �12*** �23*** �19*** �20*** �10*** �02 �12*** 10** 13*** �
12. SDQ total difficulties �27*** �19*** �37*** �24*** �10*** �01 �18*** �12*** 27*** 08* 31*** �
13. Caregiver involvement 10** 16*** 03 05 �01 �22*** �03 01 03 08* 11*** 04 �
14. Caregiver aspirations 26*** 16*** 35*** 28*** 15*** �06 07* 06 �13*** �13*** �34*** �29*** 03 �
15. Internal developmental assets 34*** 28*** 43*** 33*** 18*** �06* 21*** 11*** �30*** �06 �24*** �54*** 07* 29*** �

Note: Decimals omitted in correlations. Correlations are pair-wise; the number of cases on which the correlations were based varied between 652 (for the correlation between

Previously Repeated a Grade and Caregiver Aspirations) and 1106.

*p B 0.05 (2-tailed); **p B 0.01 (2-tailed); ***p B 0.001 (2-tailed).

E
u

ro
p

ea
n

Jo
u

rn
a

l
o
f

S
o
cia

l
W

o
rk

8
1

Table 3 Beta coefficients in hierarchical regressions of average marks on control, placement, risk and protective variables

Outcome variable: Average marks in Year 7 Outcome variable: Average marks in Year 8
Cross-sectional regression (b) (N�531) Longitudinal regression (b) (N�488)

Predictors Step 1 Step 2 Step 3 Step 4 Predictors Step 1 Step 2 Step 3 Step 4

Average marks (Year 7) 0.42*** 0.42*** 0.40*** 0.37***
Female gender 0.15*** 0.14* 0.11** 0.07

Female gender 0.14*** 0.14*** 0.12** 0.11**

Age �0.03 �0.02 �0.05 �0.02 Age �0.02 �0.02 �0.03 �0.00
Foster home 0.10

$
0.01 �0.04 Foster home �0.00 �0.02 �0.03

Kinship care home 0.11* 0.03 0.02 Kinship care home 0.05 0.03 0.03
Previously repeated a grade �0.11* �0.08$ Previously repeated a grade �0.02 �0.03
Cognitive impairments index �0.04 0.00 Cognitive impairments index �0.07 �0.07
SDQ total difficulties �0.28*** �0.13* SDQ total difficulties �0.02 0.02
Caregiver involvement in

school
0.03 Caregiver involvement in

school
0.10*

Caregiver aspirations 0.18*** Caregiver aspirations 0.02
Internal developmental assets 0.25*** Internal developmental assets 0.10*

DR
2

0.024* 0.009
$

0.095** 0.077** DR
2

0.214*** 0.003 0.006 0.018**

Note: *p B 0.05 (2-tailed); **p B 0.01 (2-tailed); ***p B 0.001 (2-tailed);
$
p B 0.10 (2-tailed).

8
2

R
.

J.
F

ly
n

n
et

a
l.

School performance

Table 4 shows the results for the two models for school performance. In the cross-

sectional model, the results were similar to those for average marks. The control, risk

and protective factors all accounted for statistically significant increments in the

variance, and the findings for step 2 were at the level of a trend. The cross-sectional

model as a whole explained 28.2% of the variance in year 7 school performance, with

the risk factor of behavioural difficulties especially important as a predictor. The

foster and kinship care homes were again modestly associated with better school

performance, but their beta coefficients were reduced to near zero once the three risk

factors (all of which were significantly and negatively associated with school

performance) had entered the model. In the final model, at step 4, the risk factor

of behavioural difficulties and the protective factors of caregiver aspirations and

youth internal developmental assets were all significantly related to school

performance.

In the longitudinal analyses, the lack of significant change in year-to-year school

performance meant that the model had relatively little power to predict change in

year 8 school performance, once year 7 school performance had been taken into

account. Interestingly, the girls had relatively consistent improved school perfor-

mance, at all four steps, whereas the older youths had fairly consistent worse

performance, at all four steps. In the final model (at step 4), caregiver aspirations for

the youth’s educational attainment was the only protective factor that was

significantly associated with better school performance.

Discussion

The findings provide some, albeit mixed, support for our basic model of educational

success and have implications for improving the educational success of young people

in care. First, in the cross-sectional regression model for average marks (i.e., in year

7), all four steps in the regression model were associated with increments in the

amount of variance accounted for, either at statistically significant levels (steps 1, 3

and 4) or at the level of a trend (step 2). The risk (9.5%) and protective factors

(7.7%) explained important proportions of the variance in average marks. Similarly,

with respect to school performance, statistically significant increments in the variance

accounted for were also found at steps 1, 2 and 4, and a trend in this direction was

seen at step 2, with the risk (17.4%) and protective factors (8.3%) accounting for

important increments in the amount of variance explained in school performance.

The fact that in the longitudinal analyses steps 2�4 accounted for much less
additional variance in the two outcomes was no doubt due to the fact that, on

average, there was little or no change to explain.

Second, the girls in our sample, as predicted, experienced greater educational

success than the boys, on both outcomes. The sequential reductions in the size of the

beta coefficient for female gender at each step of the mode, especially pronounced in

the cross-sectional models, suggested that the girls’ educational advantage was partly

European Journal of Social Work 83

Table 4 Beta coefficients in hierarchical regressions of school performance on control, placement, risk and protective variables

Outcome variable: school performance in Year 7 Outcome variable: school performance in Year 8
Cross-sectional regression (b) (N�565) Longitudinal regression (b) (N�518)

Predictors Step 1 Step 2 Step 3 Step 4 Predictors Step 1 Step 2 Step 3 Step 4

School Performance (Year 7) 0.42*** 0.42*** 0.37*** 0.33***
Female gender 0.12** 0.12** 0.06 0.02 Female gender 0.10* 0.10* 0.09* 0.08

Age 0.02 �0.02 �0.00 0.03 Age �0.08* �0.08* �0.09* �0.07$

Foster home 0.09
$ �0.02 �0.06 Foster home �0.02 �0.05 �0.06

Kinship care home 0.11* �0.00 �0.02 Kinship care home 0.04 0.01 0.01
Previously repeated grade �0.11** �0.08* Previously repeated a grade �0.04 �0.03
Cognitive impairments index �0.11* �0.04 Cognitive impairments index �0.05 �0.03
SDQ total difficulties �0.37*** �0.22*** SDQ total difficulties �0.09$ �0.05
Caregiver involvement in

school
0.03 Caregiver involvement in

school
0.04

Caregiver aspirations 0.23*** Caregiver aspirations 0.10*
Internal developmental assets 0.21*** Internal developmental assets 0.06

DR
2

0.016* 0.009
$

0.174*** 0.083*** DR
2

0.200*** 0.003 0.012
$

0.012*

Note: *pB0.05 (2-tailed); **pB0.01 (2-tailed); ***pB0.001; (2-tailed);
$
pB0.10 (2-tailed).

8
4

R
.
J.
F
ly
n
n
et
a
l.

mediated by lower levels of the risk factors and higher levels of the protective factors.

Despite this, the girls’ educational advantage tended to persist even after the risk and

protective factors had been introduced into the model.

Third, age had no relationship to educational success, in three of the four

regression models. In the fourth, however, age was consistently and negatively

associated with improved school performance, indicating perhaps that older youths

tend to fall further behind as the academic curriculum becomes more demanding.

Fourth, we found, as expected, that young people in foster and kinship care homes

had better educational outcomes than those in group homes, at least in the cross-

sectional models. This advantage was modest, however, and disappeared when the

risk factors had been taken into account. This was probably due to mediation and

may reflect a selection rather than a programme effect, with more turbulent youths

being selected out of foster or kinship care and into group care.

Fifth, it is clear that young people in care*of both genders*would benefit from
lower levels of behavioural difficulties. The cross-sectional results for this latter

variable were congruent with Slade and Wissow’s (2007) hypothesis that poor

behavioural skills have a serious negative impact on the educational success of

maltreated youth. Previous repetition of a grade in school was also predictive of lower

educational success in the cross-sectional models, even when the youth’s level of

health-related cognitive functioning had been taken into account. This indicated that

effective action to prevent young people from repeating a grade is likely to pay

dividends.

Sixth, in all four models, the protective factors were associated with a statistically

significant increment in the amount of variance explained in the two indicators of

educational success. The young person’s level of internal developmental assets was

particularly important for educational success (except in the case of improved school

performance), which is congruent with the findings of Scales et al. (2006).

Seventh, it was noteworthy that caregiver attitudes and behaviour were related to

both indicators of educational success. Higher educational aspirations on the part of

caregivers were associated with better outcomes in three of the four regression

models, and caregiver involvement in a greater number of school activities predicted

significant improvement in the youth’s average marks. These results are consistent

with Jackson and Ajayi’s (2007) position that caregivers are an important resource for

improving educational outcomes and should be recognised as such. The recruitment

and training of carers should thus take explicit account of their influential role in the

educational achievement of young people in care (Dill et al., 2012; Ferguson &

Wolkow, 2012; Jackson & Cameron, 2012).

The present study had a number of limitations. The data were correlational in

nature, and the longitudinal analyses covered only a 12 month period. Also, the

effective sample size was considerably reduced due to incomplete data on some

variables. In addition, our index of health-related cognitive impairment was a rather

rudimentary measure of current cognitive functioning, and our measure of caregiver

aspirations for the young person in care was but a single item. Despite these

European Journal of Social Work 85

limitations, our analyses were based on relatively large samples and employed

comprehensive measures of two important predictors, the risk factor of behavioural

difficulties and the protective factor of internal developmental assets. In future, with

the accumulation of large samples and additional longitudinal data, we plan to carry

out multi-year analyses of the educational trajectories of subgroups of young people

in care, in the hope of identifying those in particular need of intervention. As Trout

et al. (2008) commented, many young people in care appear to need intensive and

effective assistance, if their educational careers are to be as successful as possible.

Acknowledgements

We gratefully acknowledge the financial support of this study by the Ontario

Association of Children’s Aid Societies and the Ontario Ministry of Children and

Youth Services. We also thank the many young people in care, caregivers and child

welfare workers in local Children’s Aid Societies who participated in the Ontario

Looking After Children project, from which the study data were drawn.

References

Berridge, D. (2007) ‘Theory and explanation in child welfare: Education and looked after children’,

Child and Family Social Work, vol. 12, no. 1, pp. 1�10.
Cashmore, J., Paxman, M. & Townsend, M. (2007) ‘The educational outcomes of young people 4�5

years after leaving care: An Australian perspective’, Adoption and Fostering, vol. 31, no. 1, pp.

50�61.
Commission to Promote Sustainable Child Welfare. (2010, December) Future directions for in-care

services in a sustainable child welfare system, Working paper no. 3, Toronto, ON, Commission

to Promote Sustainable Child Welfare.

Dill, K., Flynn, R. J., Hollingshead, M. & Fernandes, A. (2012) ‘Improving the educational

achievement of young people in out-of-home care’, Children and Youth Services Review,

vol. 34, pp. 1081�1083. doi:10.1016/j.childyouth.2012.01.031
Ferguson, H. B. & Wolkow, K. (2012) ‘Educating children and youth in care: A review of barriers to

school progress and strategies for change’, Children and Youth Services Review, vol. 34,

pp. 1143�1149. doi:10.1016/j.childyouth.2012.01.034
Flynn, R. J. & Biro, C. (1998) Comparing developmental outcomes for children in care with those of

other children in Canada. Children and Society, vol. 12, pp. 228�233.
Flynn, R. J. & Tessier, N. G. (2011) ‘Promotive and risk factors as predictors of educational

outcomes in supported transitional living: Extended care and maintenance in Ontario,

Canada’, Children and Youth Services Review, vol. 33, pp. 2498�2503.
Flynn, R. J., Dudding, P. M. & Barber, J. G. (eds) (2006) Promoting Resilience in Child Welfare,

Ottawa, ON, University of Ottawa Press.

Flynn, R. J., Ghazal, H., Legault, L., Vandermeulen, G. & Petrick, S. (2004) ‘Use of population

measures and norms to identify resilient outcomes in young people in care: An exploratory

study’, Child and Family Social Work, vol. 9, pp. 65�79.
Flynn, R. J., Vincent, C. & Legault, L. (2009) User’s Manual for the AAR-C2-2006, Ottawa, ON,

Centre for Research on Educational and Community Services,

University of Ottawa.

Goodman, R. (1997) ‘The strengths and difficulties questionnaire: A research note’, Journal of Child

Psychology and Psychiatry, vol. 38, pp. 581�586.

86 R. J. Flynn et al.

Iversen, A. C., Hetland, H., Havik, T. & Stormark, K. M. (2010) ‘Learning difficulties and academic

competence among children in contact with the child welfare sytem’, Child and Family Social

Work, vol. 15, pp. 307�314.
Jackson, S. (2007) ‘Progress at last?’, Adoption and Fostering, vol. 31, no. 1, pp. 3�5.
Jackson, S. & Ajayi, S. (2007) ‘Foster care and higher education’, Adoption and Fostering, vol. 31, no. 1,

pp. 62�72.
Jackson, S. & Cameron, C. (2012) ‘Leaving care: Looking ahead and aiming higher’, Children and

Youth Services Review, vol. 34, pp. 1107�1114.
Masten, A. S. (2006) ‘Promoting resilience in development: A general framework for systems of

care’, in Promoting Resilience in Child Welfare, eds R. J. Flynn, P. M. Dudding & J. G. Barber,

Ottawa, ON, University of Ottawa Press, pp. 3�17.
McClung, M. & Gayle, V. (2010) ‘Exploring the care effects of multiple factors on the educational

achievement of children looked after at home and away from home: An investigation of two

Scottish local authorities’, Child and Family Social Work, vol. 15, pp. 409�431.
Miller, M., Flynn, R. & Vandermeulen, G. (2008) Looking After Children in Ontario: Good Parenting,

Good Outcomes: Ontario Provincial Report (Year Six), Reports for 0�4, 5�9, 10�15, and 16�
20 year olds, Ottawa, ON, Centre for Research on Educational and Community Services,

University of Ottawa.

Miller, M., Vincent, C. & Flynn, R. (2009) Looking After Children in Ontario: Good Parenting, Good

Outcomes. Ontario Provincial Report (Year Seven), Report for 10�15 year olds, Ottawa, ON,
Centre for Research on Educational and Community Services, University of Ottawa.

Ontario Association of Children’s Aid Societies (OACAS). (2010) Gateway To Success: Cycle Two,

OACAS survey of the educational attainment of Crown Wards and former Crown Wards,

ages 16 through 20, during the 2008�2009 academic year, Toronto, ON, Ontario Association
of Children’s Aid Societies.

Scales, P. C., Benson, P. L., Leffert, N. & Blyth, D. A. (2000) ‘Contribution of developmental assets

to the prediction of thriving among adolescents’, Applied Developmental Science, vol. 4, pp.

27�46.
Scales, P. C., Benson, P. L., Roehlkparain, E. C., Sesma, A. & van Dulmen, M. (2006) ‘The role of

developmental assets in predicting academic achievement: A longitudinal study’, Journal of

Adolescence, vol. 29, pp. 691�708.
Slade, E. P. & Wissow, L. S. (2007) ‘The influence of childhood maltreatment on adolescents’

academic performance’, Economics of Education Review, vol. 26, pp. 604�614.
Trout, A. L., Hagaman, J., Casey, K., Reid, R. & Epstein, M. H. (2008) ‘The academic status of

children and youth in out-of-home care: A review of the literature’, Children and Youth

Services Review, vol. 30, pp. 979�994.
Vinnerljung, B., Öman, M. & Gunnarson, T. (2005) ‘Educational attainment of former child welfare

clients: A Swedish national cohort study’, International Journal of Social Welfare, vol. 14, pp.

265�276.
Weyts, A. (2004) ‘The educational achievements of looked after children: Do welfare systems make a

difference?’, Adoption and Fostering, vol. 28, no. 3, pp. 7�19.

European Journal of Social Work 87

O R I G I N A L P A P E R

‘‘Because You’re Mine, I Walk the Line’’? Marriage,
Spousal Criminality, and Criminal Offending Over
the Life Course

Marieke van Schellen • Robert Apel • Paul Nieuwbeerta

� Springer Science+Business Media, LLC 2012

Abstract
Objectives This study is an analysis of the relationship between marriage and crime in a
high-risk sample of Dutch men and women. Marriages are classified as to whether the spouse

had been convicted of a crime prior to the marriage, in order to ascertain if one’s criminal

career after marriage unfolds differently depending on the criminal history of one’s spouse.

Methods Data are from the Criminal Career and Life-Course Study, a random sample of
all individuals convicted of a criminal offense in the Netherlands in 1977 (N = 4,615).
Lifetime criminal histories for all subjects are constructed from age 12 to calendar year

2003. Official marriage records are also consulted, and the criminal history of all spouses

are similarly constructed. Fixed-effects Poisson models are estimated to quantify the

relationship between marriage, spousal criminality, and conviction frequency, controlling

for age, parenthood, prior conviction, and prior incarceration.

Results Among men, marriage reduces the frequency of criminal conviction, but only if
the marriage is to a non-convicted spouse. Marriage to a convicted spouse, on the other

hand, is indistinguishable from singlehood—it neither discourages nor promotes criminal

behavior. Among women, marriage has a crime-reducing effect, regardless of the criminal

history of the spouse. A set of preliminary follow-up analyses suggests further that men

with more extensive criminal histories, and with more stable marriages, benefit in a more

pronounced way from marriage to a non-convicted spouse. However, even unstable

marriages to non-convicted spouses appear to reduce conviction frequency while they last.

Conclusions Marriage is indeed a salient transition in the criminal career, but there are
important differences depending on the characteristics of the offender (gender, criminal

history), the characteristics of the spouse (criminal history), and the characteristics of the

M. van Schellen (&)
Department of Sociology, Utrecht University, Heidelberglaan 2, 3584 CS Utrecht, The Netherlands
e-mail: m.vanschellen@uu.nl

R. Apel
School of Criminal Justice, Rutgers University,

123

Washington Street, Newark, NJ 07102, USA

P. Nieuwbeerta
Department of Criminology, Leiden University, Steenschuur 25, 2311 ES Leiden, The Netherlands

123

J Quant Criminol

DOI 10.1007/s10940-012-9174-x

marriage (duration). The authors conclude that while marriage matters, it does not nec-

essarily mean the end of a criminal career, and that processes of both partner selection and

partner influence deserve close attention by marriage-crime researchers. Qualifications of

the study’s findings include the use of conviction data from official sources, the use of a

sample of men and women who were all convicted of a crime at some point in their lives,

the study of legal marriage in the Netherlands, and the inability to measure potential

mechanisms for the observed marriage effects.

Keywords Marriage � Spousal criminality � Criminal convictions �
Life-course criminology � Panel models

Introduction

Marriage has long been correlated with a variety of beneficial outcomes for the involved

individuals. Married individuals appear to have a larger network of help and support, show

less risky and unhealthy behavior, earn higher income, and possess more assets and wealth.

Accordingly, married persons tend to be happier, healthier, and better off financially

(Waite 1995; Waite and Gallagher 2000).

The idea of marriage as a protective institution takes a prominent place in criminology

as well. Marriage is considered to be an important transitional event that can reduce

criminal activity and even lead to desistance from crime (Sampson and Laub 1993). The

decline in criminal behavior is often attributed to the social bond that forms and

strengthens as a result of marriage: Spouses monitor each other’s behavior and do not want

to endanger their marital relationship by committing crime (Laub and Sampson 2003).

Recent empirical studies show that marriage is indeed associated with lower offending

levels (Bersani et al. 2009; Blokland and Nieuwbeerta 2005; Farrington and West 1995;

Horney et al. 1995; King et al. 2007; Laub and Sampson 2003; Piquero et al. 2002;

Sampson et al. 2006; Theobald and Farrington 2011).

Despite its prominence, the idea that marriage reduces crime is less straightforward than

assumed. Although on average the effect might be protective, the benefits of marriage

might not be homogenous and are likely to depend, among other things, on the criminal

history of the spouse. Marriage to a criminal partner could sustain or even stimulate an

offender’s engagement in criminal activities over time (Rhule-Louie and McMahon 2007).

For example, offenders probably have similar views on the appropriateness of criminal

offending, learn from each other, and pass on their criminal skills (Giordano et al. 2007;

Leverentz 2006; Simons et al. 2002). At the very least, marriage to a criminal spouse could

result in persistence in criminal offending, and at worst, escalation.

While empirical studies of marriage effects on criminal behavior have been accumu-

lating, the impact of spouses’ criminal careers has received far less attention (Rhule-Louie

and McMahon 2007). This is surprising, simply because attachments to unconventional

persons are considered to be among the most important predictors of delinquent behavior

during adolescence: Adolescents who have delinquent friends are more likely to become

delinquent, and commit more crimes than adolescents without deviant connections (Haynie

et al. 2005; Simons et al. 2002). Although it has been argued that adolescents are more

sensitive to social influence processes (Warr 2002), this finding does clearly demonstrate

that social ties are not necessarily protective, but can stimulate criminal behavior as well.

To gain more knowledge about the development of individual criminal careers, insight

is needed into the criminal behavior of spouses. One of the reasons for the lack of empirical

J Quant Criminol
123

research on the effects of partners’ criminal history is that the requirements for the design

of these studies are substantial. First, longitudinal information on the development of

criminal behavior is necessary. Second, this information is needed for both marriage

partners. Third, very long periods of observation are required in order to examine research

subjects beyond adolescence into adulthood. Thus far, virtually no study meets these

requirements.

In this study, we estimate quasi-experimental models of the relationship between

marriage and criminal conviction, and investigate whether this relationship varies by the

spouse’s criminal history at the time of marriage. We employ data from a unique long-term

study of a Dutch conviction cohort and their marriage partners: The Criminal Career and

Life-Course Study (CCLS; Nieuwbeerta and Blokland 2003). The CCLS contains data on

the officially registered criminal careers of 4,615 Dutch offenders and their marriage

partners, covering ages 12–72. The aim of this study is to contribute to the current literature

in several ways. First, this study will be the first to investigate the life-long criminal careers

of a large number of offenders and their marriage partners. Second, the CCLS contains data

on the exact timing of marriages, convictions, and periods of incarceration over the entire

life span. Therefore, we can clearly distinguish partner influences from partner selection

processes that took place before marriage. Third, the data allow us to investigate the

criminal careers of male as well as female offenders.

It should be noted that the CCLS does not contain data on never-convicted individuals

and their spouses. This research thus concerns the effects of marriage for a sample of

individuals who have been convicted at some point during their lives. Within this group of

convicts, individuals differ in the timing, number, and seriousness of crimes committed.

Because the CCLS does not provide information about other relationship types than

marriage (e.g., cohabitation), the focus of this study is on the effects of marriage and not on

partner relationships in general.

Marriage, Spousal Criminality, and Crime: Desistance or Persistence?

With the rise of developmental and life-course criminology (Piquero et al. 2003), a

growing number of studies have focused on the unfolding of individual criminal careers

over the life span. These studies show that offending rates tend to increase gradually during

childhood, rise more sharply during adolescence and then begin to decline steadily as

individuals enter adulthood. Although there appears to be a group of persistent offenders

that continues committing crimes far into adulthood, most persons tend to stop their

criminal careers after adolescence (Blokland et al. 2005; Laub and Sampson 2003).

Desistance from crime is often explained by the fact that persons experience crime-

inhibiting life-course transitions as they navigate the bridge from adolescence to adulthood.

The formation of marital relationships has been argued to play a key role in this desistance

process (Laub and Sampson 2003).

Several theoretical mechanisms have been proposed to explain why the transition to

marriage would reduce criminal offending (see also Sampson et al. 2006: 467–468). These

mechanisms can be grouped in four different categories: Social bond, routine activities,

social learning, and cognitive transformation. Originally, all explanations have centered on

the development of individual offending trajectories, and neglected the criminal careers of

partners. This is unfortunate, since several studies have shown that offenders have a higher

chance to form relationships with partners who are criminally active as well, i.e., they mate

assortatively (Moffitt et al. 2001; Simons et al. 2002). Although the institution of marriage

J Quant Criminol
123

might be protective in itself, it could very well be the case that the effects of marriage

depend on the criminal history of the spouse (Giordano et al. 2002). Marriage could

explain desistance as well as persistence in crime—depending on the spouse’s involvement

in crime.

Below we discuss the main theoretical mechanisms, and derive hypotheses on the

effects of marrying a non-criminal spouse versus criminal spouse. In doing so, we pay

explicit attention to gender differences in the effects of marriage and spousal criminality.

Although the different theoretical perspectives are not fundamentally incompatible, they

differ in their central focus. Some explain the crime-reducing effect of marriage by pro-

cesses external to the individual (e.g., changes in opportunities to commit crime), while

others focus on internal factors (e.g., changes in preferences). Although the data do not

allow us to test the various underlying mechanisms, they give us more insight in why a

relationship between marriage, spousal criminality, and crime can be expected.

Social Bond

Marriage may change criminal offending because it strengthens social bonds to conven-

tional society. This idea has a prominent place in Sampson and Laub’s (1993) age-graded

theory of informal social control. Spouses monitor and attempt to control each other’s

behavior, and tend to discourage activities that do not pay off in the long run, like hanging

out with deviant friends. Also, especially if ties are strong and stable, marital relationships

create obligations and restraints that increase the costs of offending. Over time, as com-

mitment and investment in the relationship grows, there are fewer incentives to commit

crime, because more is at stake (Laub et al. 1998).

Whether the effects of marriage are protective may, however, strongly depend on the

criminal history of the spouse to whom one is attached. We nuance Sampson and Laub’s

theoretical ideas in two ways. First, marriages are not necessarily strong if both spouses are

involved in crime (Simons et al. 2002). The idea that marital ties are of higher quality and

less likely to dissolve if spouses resemble each other, is prominent in family sociology

(Brines and Joyner 1999; Kalmijn 1998). It is, however, questionable whether this also

applies when it concerns similarity in criminal behavior. For example, if both partners are

involved in crime, they are both likely to have personal traits and to be involved in situ-

ations that undermine the quality and stability of the marriage (Western 2006: 5). Second,

conventional behavior is not necessarily encouraged if both partners are criminally

involved. If their spouses are not involved in crime, offenders might indeed risk their

relationship by violating the law. If spouses have a criminal history as well, offending is

likely to be a conventional way of behaving, which is less likely to be discouraged and

does not threaten the continuation of the relationship.

Routine Activities

A different interpretation of marriage’s role has been given by Warr (1998) (see also Laub

and Sampson 2003). Warr (1998) emphasizes the role of peers in criminal offending. The

decline in crime following marriage is caused by a decrease in time spent with (delinquent)

friends and an accompanying reduction in opportunities and reinforcement for criminal

behavior. Married people spend more time in each other’s company, and stay home

together more often. Although non-criminal partners can indeed be expected to promote a

conventional lifestyle (Simons et al. 2002: 404), this is less likely to apply to criminal

partners. Criminal partners are likely to be enmeshed in a criminal network themselves,

J Quant Criminol
123

and may therefore stimulate contact with other antisocial individuals and bring one to risky

places at risky moments.

Socialization

Another mechanism that may underlie the relationship between marriage and criminal

offending can be derived from differential association and social learning theories. These

theories state that behavior is learned through social interaction within a cohesive and

intimate group, where criminal norms, values, and knowledge are passed on through

ongoing socialization processes (Akers 1973; Warr 2002). Although these theories have

traditionally been used to explain the influence of delinquent peers, the same ideas can be

applied to marital relationships (see also Haynie et al. 2005; Simons et al. 2002). In

contrast to the earlier discussed mechanisms, socialization theories make explicit that

intimate associations can have a positive or negative influence on offending depending on
the normative orientation of others. Although marrying a non-criminal partner is likely to

lead to socialization in a conventional law-abiding environment, marrying a criminal

partner is likely to sustain or stimulate an individual’s criminal activities over time (Rhule-

Louie and McMahon 2007). Offenders probably have similar views on the appropriateness

of criminal behavior, learn from each other, pass on their criminal skills, or may even start

co-offending and become real ‘‘partners in crime.’’ It has been suggested that this form of

socialization may be more powerful and important for females. As they would be more

oriented toward relationships, their behavior would be more frequently determined by the

behavior of their partners (Moffitt et al. 2001; Steffensmeier and Allan 1996).

Cognitive Transformations

Finally, marriage may also lead to changes in criminal offending, because it changes one’s

sense of self through cognitive transformations (Giordano et al. 2002; Sampson et al. 2006:

468). In contrast to the earlier mentioned mechanisms, processes internal to the offender

are emphasized. Marriage—if accompanied by an openness to and readiness for mean-

ingful change—can lead to desistance, because it fosters pro-social role modeling. The

accompanying cognitive transformations result in a change in the meaning and salience of

criminal behavior: Criminal behavior is no longer seen as positive, viable, or personally

relevant (Giordano et al. 2002). Although marrying a non-criminal spouse may indeed lead

to a pro-social and responsible lifestyle, criminal partners are less likely to function as

positive role models and therefore may undermine conventional identity change.

Hypotheses

Although we are not able to directly test the underlying mechanisms, all aforementioned

mechanisms lead to the expectation that marriage in general reduces criminal offending. It

not only reduces the preference to commit crimes, but also leads to fewer opportunities to

offend. We nuance this hypothesis in two ways. First, we assume that the effect of marriage

depends on the criminal behavior of the spouse. Being married to a non-criminal spouse is

expected to lead to a decrease in the number of offenses, and being married to a criminal

spouse to persistence in crime. On the one hand, we may find no change in the level of

offending compared to singlehood. On the other hand, partners may influence each other in

such a way that they stimulate criminal activity, resulting in an increase in the number of

J Quant Criminol
123

offenses. Second, we expect gender differences in the effect of marriage and spousal

criminality. It has been argued that men would benefit more from marriage in general,

because they are more likely to marry non-criminal spouses than women are. In other

words, ‘‘men marry up’’ (Laub and Sampson 2003). However, according to the sociali-

zation perspective, women might be more influenced by the behavior of their partners.

Therefore, we expect the crime-reducing effects of marrying a non-criminal spouse and the

crime-stimulating effect of marrying a criminal spouse to be even larger for women.

Earlier Studies and Their Limitations

Although a growing number of studies have investigated the relationship between marriage

and crime, only a few of them considered the impact of partners’ criminal behavior. Using

data from Glueck and Glueck’s classic study of criminal careers, Sampson et al. (2006)

investigated 226 delinquent men followed prospectively from adolescence to age 32.

Although men with criminal or deviant wives displayed higher criminal offending rates,

within-individual estimates of the effects of marriage showed that it nevertheless signifi-

cantly reduced criminal involvement, controlling for duration of the marriage, marital

attachment, and spousal criminal record.

Moffitt et al. (2001) investigated the effects of partner relationships on antisocial

behavior among a birth cohort of 360 individuals followed from age 13 to 21. The analyses

revealed that women were more likely to persist in crime (measured as self-reported

antisocial behavior at age 21) when they formed unions with antisocial men. However,

antisocial men continued to be antisocial, regardless of whether their female partner was

antisocial at the time of the relationship.

Simons et al. (2002) simultaneously tested the impact of delinquent friends and partners

on delinquent behavior among 236 young adults. Results showed that having an antisocial

romantic partner was related to higher levels of criminal behavior both directly as well as

indirectly, through its effect on the quality of the romantic relationship and involvement

with deviant friends. These relationships were significant for both men and women,

although having an antisocial partner was associated with criminality more strongly for

women. Using data from a nationally representative sample of school-going adolescents,

Haynie et al. (2005) reached similar conclusions. Romantic partners’ deviance was more

strongly related to females’ involvement in minor deviance, although gender did not

condition the strength of the relationship between the romantic partners’ serious delin-

quency and the respondents’ serious delinquency.

The study by Woodward et al. (2002) was unique in that it included single as well as

romantically involved individuals. Individuals involved with a non-deviant partner had

lower rates of offending at age 21 than those with no partner, while those without a partner

had lower rates of offending at age 21 than those involved with a deviant partner. Similar

results were found for men and women.

The most recent empirical study has been conducted by Capaldi et al. (2008). Using a

sample of at-risk men (N = 191), the results showed that a partner’s antisocial behavior
was related to both onset and persistence of arrests, even when controlling for deviant peer

associations. In contrast to earlier studies, respondents’ and partners’ offending behavior

were not measured at the same time, enabling stronger conclusions about the direction of

influence. Unfortunately, data on partners’ criminal history before marriage was lacking.

Another drawback is that the analysis only investigated respondents’ arrests in the first year

after the relationship was formed.

J Quant Criminol
123

In sum, most studies have found that a partner’s offending is associated with an increase

in the referent subject’s own offending. In some cases, this relationship is stronger for

women (Capaldi et al. 2008; Haynie et al. 2005; Moffitt et al. 2001; Simons et al. 2002).

Other studies have found that single individuals are even better off than those involved

with a delinquent or criminal partner (Woodward et al. 2002). This finding is inconclusive,

however, because there is other evidence that marriage has protective effects irrespective

of the criminal history of the spouse (Sampson et al. 2006).

Although earlier studies have made important contributions to the marriage-crime lit-

erature, they have several limitations. First, they lack longitudinal information on partners’

criminal histories. As relationship status and partner criminality trend to be measured at the

same time, there are limits to causal inferences: Any association between partners’

criminal behavior could result from selection processes that take place before relationship

formation. Second, most studies limit their focus to adolescence and early adulthood. Yet

partnerships are particularly salient during adulthood, not to mention that long-term effects

of marriage are impossible to study. Third, earlier studies investigated relationships of

varying durations, and at various stages of attachment (e.g., married, cohabiting, unmarried

but committed relationship). In addition, the (marital) relationships under study might be

the first relationship, but also the second or even the third. Although these different types of

relationships might very well have different effects, they are not analyzed separately

(partly because of small sample sizes).

Data

In this study, we use data from the CCLS (Nieuwbeerta and Blokland 2003). The CCLS

subjects were selected by taking a four-percent sample of all cases of criminal offenses

tried in the Netherlands in 1977, with an oversampling of less common—mostly serious—

offenses (e.g., murder, rape, drug offenses) and an undersampling of common offenses

(e.g., drunk driving). This resulted in a total sample of 4,615 offenders (4,191 men and 424

women).

Extracts (‘‘rap sheets’’) from the General Documentation Files (GDF) of the Criminal

Records Office were used to reconstruct the entire criminal careers of all 4,615 research

subjects from the age of 12—the minimum age of criminal responsibility in the Nether-

lands—until calendar year 2003. In the Netherlands, individuals are not given a ‘‘clean

slate’’ upon becoming an adult. The extracts thus contain information on both juvenile and

adult offenses. Although the GDF contain information on all offenses that led to any type

of judicial action, here we use only information on those offenses that were either followed

by a conviction or a prosecutorial disposition because of policy reasons.
1

We therefore

exclude cases that resulted in an acquittal or a prosecutorial disposition because of

1
In the Dutch criminal justice system, the public prosecutor has the discretionary power not to prosecute all

cases forwarded by the police. First, the public prosecutor may decide to drop the case if prosecution would
probably not lead to conviction due to lack of evidence, or for technical considerations (procedural or
technical waiver). Second, the public prosecutor is authorized to waive prosecution ‘‘for reasons of public
interest’’ (waiver for policy considerations). The Board of Prosecutors-General has issued national prose-
cution guidelines under which a public prosecutor may decide to waive a case for policy reasons. In some
cases measures other than penal sanctions are preferable or more effective, or prosecution would be dis-
proportionately unjust or ineffective in relation to the nature of the offense or the offender, or prosecution
would be contrary to the interest of the state or the victim (Tak 2003).

J Quant Criminol
123

insufficient evidence. This means that the criminal offenses which are studied have most

likely been committed by the offenders.

To measure the unfolding of life circumstances, the judicial data were supplemented

with data from the population registration records. These records contain information on

the exact timing of marriage, divorce, fertility, and mortality. All data are derived from

official sources, which means that they are of high quality and have very few missing

values. Because the mean age of the research subjects was 27.8 years in 1977 (med-

ian = 25 years), we have data on convictions and life circumstances that reach far into

adulthood for a large portion of the sample. For example, while the sample was followed

until a mean age of 43.6 years (median = 44 years), the age of follow-up for almost six

percent of the sample is 70 or older.
2

In preparation for this study, the CCLS was supplemented with data on the complete

criminal careers of all of the marriage partners of the research subjects from age 12 to

calendar year 2007. The population registration records revealed that 74.5 % (N = 3,437)
of the original 4,615 research subjects married on at least one occasion, to a total of 4,409

partners. This supplement to the CCLS allows us to determine the exact timing of marriage

and, for all research subjects and their married partners, the exact timing of criminal

offenses, the type of offenses committed, and periods of prison confinement.

To estimate the empirical models of conviction frequency, we eliminate the ‘‘criterion

conviction,’’ referring to the conviction that brought each subject into the CCLS sampling

frame. Because of the way that the CCLS data were collected, all subjects are convicted at

least once during their lives. For most offenders (96.8 %), this conviction was in either

1976 or 1977. If subjects who marry have no convictions prior to their first marriage then

they must, by construction, have at least one conviction during or after their first marriage.

This could result in the estimation of a criminogenic effect of marriage that is artifactual.

To avoid this, we exclude the criterion conviction altogether.

We also focus the analysis on the first marriage.
3

By focusing on first marriages, we

avoid having to account for feedback effects between marriage and crime whereby (a

criminal) marriage affects the likelihood of crime, which in turn affects the likelihood of (a

criminal) marriage. Although the focus on first-time marriage limits the generalizability of

our findings, we eliminate this kind of endogeneity bias as a source of confounding of the

empirical estimates (see also Nieuwbeerta et al. 2009: 232).

Analytic Strategy

We are interested in what effect marriage to a non-convicted or convicted spouse has on an

individual’s post-marriage conviction frequency. One empirical challenge is that marriage

and partner selection are not randomly determined. Individuals who marry are likely to

have different characteristics than persons who do not marry, and individuals who marry

2
The mean age in 1977 is older than the peak of the well-known age-crime curve. This is because the

CCLS contains data on criminal convictions rather than arrests, which will lead to a slightly older sample.
On average, the CCLS offenders have been followed for 32 years (min = 1, max = 60). The sample size
varies across ages. For example, at age 12, the data contain information on all 4,615 individuals. The sample
drops at age 22 (4,605), age 32 (4,547), age 42 (4,255), age 62 (788), and age 72 (245).
3

Most of the subjects in our sample—76 %—marry only once.

J Quant Criminol
123

convicted spouses are likely to have different characteristics than offenders who marry

non-convicted spouses. This is known as the selection problem, and it can cloud causal

interpretations of correlations between marriage and criminality, since differences in crime

risk probably exist even in the absence of marriage. One solution to the selection problem

is to attempt to adjust away these differences by including as many control variables as

possible in a regression or propensity score model. This represents a ‘‘selection on

observables’’ approach to causal effect estimation (see Heckman and Hotz 1989). Although

the CCLS data are unusually rich with regard to offenders’ marital and criminal histories,

as with most official sources of data, information on other variables known to be correlated

with marriage and crime (e.g., personality characteristics, educational attainment, socio-

economic status) is unfortunately unavailable. Yet even with an exhaustive set of such

control variables, the selection problem would persist because differences between indi-

viduals are always partly unobserved. A unique strength of the CCLS data, with lifetime
conviction histories on all offenders, is the ability to estimate the effect of marriage and

spousal criminality on conviction frequency in the presence of ‘‘selection on unobserva-

bles’’ (Heckman and Hotz 1989). The most rigorous way to do so is through the use of a

fixed-effects model.

Fixed-effects models adjust for so-called ‘‘unobserved heterogeneity’’ by restricting

attention to within-individual change in marriage and crime over time. The model thus

eliminates biases that are attributable to any source of variation in criminality that

remains constant over time, for example, biological or genetic differences (Halaby

2004; Johnson 1995). In other words, any estimate of the ‘‘marriage effect’’ on crime is

purged of enduring differences between individuals. Yet fixed-effects models still

produce inconsistent estimates in the presence of ‘‘dynamic selection,’’ or omitted time-

varying regressors that are correlated with joint changes in marriage and crime. We

explicitly measure two such characteristics (age, parenthood), and control as rigorously

as possible for prior offending (convictions, incarceration). By using fixed-effects

models in this way, we aspire to take advantage of the strengths of the CCLS data (i.e.,

the unique longitudinal data on time-varying variables) and compensate as much as

possible for the weaknesses (i.e., the lack of relevant time-stable confounding vari-

ables).We return later to a more thorough discussion about causal identification in

marriage-crime studies.

To estimate the fixed-effects models, a person-year file is constructed with records

containing information on each individual in each calendar year. For every person the

records begin at age 12 and end in the year 2003 (the end of data collection), in the last

year of the first marriage (divorced subjects are excluded in all years after their first

marriage), or in the year prior to death (in this way we account for ‘‘false desistance’’

caused by mortality). The fully constructed data file contains information on 150,315

person-years for the 4,615 CCLS subjects.

In our empirical models the dependent variable, Yit, is a discrete random variable
representing a count of the number of convictions received by subject i (i = 1,…,N) in
calendar year t (t = 1,…,Ti). It is distributed Poisson with density:

fðYit Xit; Sitj Þ¼
expð�SitkitÞðSitkitÞYit

Yit !

where Sit represents the inverse of a subject’s exposure, or the ‘‘street time’’ in a given
calendar year, measured as the proportion of the year not confined in a correctional

J Quant Criminol
123

institution.
4

By controlling for the opportunity to commit crimes, we eliminate the pos-

sibility of false desistance attributable to incarceration (see Piquero et al. 2001).

The analysis begins with the baseline model that controls for characteristics that have

been demonstrated to influence the development of criminal behavior:

ln kit ¼ a1f Ageitð Þþ a2Childit þ a3Coni;t�1 þ a4AccumConi;t�2 þ a5Inci;t�1
þ a6AccumInci;t�2 þ ui

Ageit is modeled as a cubic to capture age-related changes in the rate of conviction for the
entire sample. Childit is a time-varying dummy variable for whether the subject has one or
more children under the age of 18. We also include several time-varying measures of

criminal history in the models. Two measures of prior convictions are added, including a

dummy indicator for having been convicted in the previous calendar year (Coni,t-1) as well
as the total number of convictions accumulated as of 2 years ago (AccumConi,t-2). Two
measures of imprisonment are also added, including a dummy indicator for having been

incarcerated in the previous calendar year (Inci,t-1) and the total accumulated time spent in
prison as of 2 years ago (AccumInci,t-2).

5
The individual effect, ui, captures unobserved

heterogeneity in conviction risk, or that portion of the total variation in conviction that is

unobserved (and unmeasured) but is stable over time. The individual effect or error

component, ui, is modeled as fixed in this analysis.
To the variables that comprise the baseline specification, henceforth denoted for the

purpose of economy as akXitk, the first model of substantive interest adds a time-varying
indicator for marriage. The model is thus specified as follows:

ln kit ¼
XK

k¼1
ak Xitk þ bMarriedit þ ui ð1Þ

Marriedit is coded ‘‘1’’ in each year that subjects are married and ‘‘0’’ in all earlier
years. If marriage promotes desistance from crime, we expect b to be negative. Recall that,
since all person-years after the first marriage ends are excluded, this indicator quantifies the

effect of one’s first marriage on conviction frequency.

In our second model we take into account the criminal history of the spouse by adding a

second marriage indicator in the following manner:

ln kit ¼
XK

k¼1
ak Xitk þ bMarriedit þ cMarriedConvictit þ ui ð2Þ

The indicator Marriedit is coded as before. The new indicator MarriedConvictit is coded
‘‘1’’ in the years that subjects are married to a spouse who had a criminal conviction

preceding the marriage. The paired marital states—Marriedit and MarriedConvictit—are
not mutually exclusive. Therefore, b represents the effect of being married relative to

4
The fixed-effects Poisson model proceeds by maximizing the conditional likelihood, where conditioning is

achieved by summing across each individual’s Ti observations on the dependent variable. This technically
makes it a conditional fixed-effects model. Therefore it necessarily excludes individuals whose observations
(here, total number of convictions) sum to zero during the period of observation, resulting in the loss of
degrees of freedom. In our model, 3,356 of the 4,191 men (80 %) are retained, whereas 173 of the 424
women (41 %) are retained.
5

The first order-lags, Coni,t-1 and Inci,t-1, capture recency in criminal offending and are binary while the
second-order lags, AccumConi,t-2 and AccumInci,t-2, capture the accumulated criminal history and are non-
binary. By including both types of variables we can distinguish the short-term, state-dependent effects of
criminal conviction and incarceration from the long-term effects.

J Quant Criminol
123

singlehood and c represents the additional effect of being married to a convicted spouse
relative to being married to a non-convicted spouse. In other words, c is a contrast. In order
to recover the impact of marriage to a convicted spouse relative to singlehood, these two

coefficients must be summed together and tested against zero.

If marriage to a non-convicted spouse leads to desistance from crime, then b will be
negative and significant. If, however, marriage to a convicted spouse leads to persistence in

crime, then we expect c to be positive and significant. Additionally, if marriage to a
convicted spouse exacerbates crime relative to remaining single, then b ? c will be
positive and significant. If, on the other hand, marriage to a convicted spouse does not

differ from singlehood, then b ? c will not be significantly different from zero.
In a third and final set of models, we evaluate whether the effects of marriage and

spousal criminality depend on the offender’s own criminal history as well as on the length

of marriage. These analyses are limited to the male CCLS offenders. We provide a more

extensive description of these models in later sections.

Results

The aim of this paper is to investigate the extent to which the effects of first marriage

depend on the criminal history of the spouse. Before we turn to our panel models, we begin

with descriptive statistics, provided in Table 1. It appears that three in four male offenders

(73.2 %) and six in seven female offenders (86.5 %) marry before calendar year 2003. Of

these married offenders, only one in 20 males (5.5 %) but one in four females (26.7 %)

marry a convicted spouse. The fact that women are far overrepresented among individuals

who marry convicted spouses is consistent with the observation that ‘‘women marry down’’

while ‘‘men marry up’’ when it comes to crime (Laub and Sampson 2003). This may

indicate that there are simply more men with criminal records in the marriage market, and

therefore the chances are higher for a woman to marry an offender. Also, there might be

stronger selection processes at work for female offenders compared to male offenders.

Because of the relative rarity of female criminal behavior, a convicted female is likely to

be more deviant or ‘‘pathological’’ with respect to the unobserved characteristics correlated

with crime (e.g., lower self-control, lower socioeconomic status, personality disorders).

Finally, it might be the case that some of the female partners actually do have a criminal

history, but are convicted less often (e.g., because they commit less serious crimes).

Examining criminal backgrounds, male subjects who marry a convicted spouse in their

first marriage accumulate twice as many convictions over their lifetime, on average, than

subjects who marry a non-convicted spouse (21.1 vs. 10.2 convictions). Correspondingly,

they are more likely to have received an early first conviction (35.5 vs. 24.8 %) and to have

ever been incarcerated (65.7 vs. 43.9 %). The same pattern holds true for female subjects.

Females who marry a convicted partner in their first marriage have more lifetime con-

victions (4.0 vs. 3.4), a higher risk of early first conviction (12.2 vs. 5.9 %), and a higher

incarceration risk (18.4 vs. 13.4 %).

Although the mean age of first marriage differs by no more than a couple of years (and

only for male subjects), the length of the first marriage is substantially shorter for subjects

who marry a convicted spouse compared to a non-convicted spouse, indicative of greater

marital instability (males: 13.4 vs. 17.1 years; females: 14.7 vs. 20.2). But note that

marriages are quite durable, irrespective of the criminal behavior of the spouse. Most of the

CCLS offenders grew up in an era in which divorce was less common than today. As a

result, individuals were more likely to stay married. Subjects who marry a convicted

J Quant Criminol
123

T
a

b
le

1
D

e
sc

ri
p
ti

v
e

st
a
ti

st
ic

s
b
y

e
n

e
r

V
a
ri

a
b

le
M

a
le

su
b

je
c
ts

(N
=

4
,1

9
1

)
F

e
m

a
le
su
b
je
c
ts
(N
=

4
2

4
)

F
ir

t
m

a
rr

ia
g

e
to

a
…

F
ir

st
m

a
rr
ia
g
e
to
a
…

N
o

n
-c

n
v

ic
te

d
sp

u
se

o
n

v
ic

e
d

sp
o

u
se

e
v

e
r

m
a
rr

ie
d

N
o
n
-c

o
n
v

ic
te
d
sp

o
u
se

C
o

v
ic

te
d

sp
o
u
se

N
e
v
e
r

m
a
rr
ie
d

N
2

,9
0
1

1
6

9
1

2
1

2
6

9
9

8
5

%
o

f
sa

m
p

le
6

9
.2

4
.0

2
6

.7
6

3
.4

2
3

3
.4

e
a
rs

in
st

u
d

y
4

0
.9

(9
.4

)
[3

]

4
0

.2
(9

)

[3
7

]
3

5
.6

(8
.6

)
[3

5
]

4
6

.2
(1

)
[4

6
]

4
1

.9
(9

.3
)

[4
2

]
3

8
.2

)
[3
6
]
Y
e
a
rs

in
a
n

a
ly

si
s

sa
m

p
le

a
3

1
.6

(

1
2

.5
)

[

3
1

]
3

0
.5

(1
2

.2
)

[3
1
]
3
5
.6
(8
.6
)
[3
5
]

3
2

.2
(1

5
.0

)
[3

1
]

2
6

.4
(1

2
.7

)
[2

4
]

3
8

.2
(9

.8
)

[3
6

]

P
e
rs

o
n

a
l

b
a
c
kg

u
n
d

Y
e
a
r

o
f

b
ir

th
1

9
4

8
.4

(

1
0

.9
)

[

1
9

5
1
]

1
9

4
9

.6
(1

0
.2

)
[1

9
5

3
]

1
9

5
3

.0
(8

.0
)

[1
9

5
5

]
1

9
4

2
.3

(1
1

.9
)
[1
9

4
4
]

1
9

4
7

.9
(9
.3
)
[1
9

4
8
]

1
9

5
0

(

9
.5

)
[1
9
5
4
]
N
o

n
-D

u
tc

h
(%

)
1

1
.1

9
.5

1
7

.0
1

7
.6

8
.2
2
1
.1

D
e
a
th

p
ri

o
r

to
c
a
le

n
d
a
r

y
e
a
r

2
0

0
3

(%
)

1
5

.0
1

4
.8

1
9

.2
1

6
.5

8
.2
1
7
.5

C
ri

m
in

a
l
b
a
c
kg
ro
u
n
d

A
g

e
o

f
fi

rs
t

c
o

n
v

ic
ti

o
n
2
1

.5
(8

.8
)
[1
9
]
1
9
.2

(7
.6

)
[1

7
]

1
8

.9
(6

.3
)

[1
7

]
3
0
.5
(1
1
.3
)

[2
9

]
2

4
.3

(8
.0

)
[2
3
]
2
3

.1
(7

.9
)

[2
0

]

A
c
c
u
m

u
la

te
d
c
o
n
v
ic
ti
o
n

s
1

0
.2

(

1
4

.9
)

[5
]

2
1

.1
(2

1
.0

)
[1
5
]
1
7

.5
(2

0
.8

)
[1

0
]

3
.4

(6
.5

)
[1

]
4

.0
(5

.2
)

[

2
]

6
.4

(1
6

.1
)

[2
]

E
a
rl

y
fi

rs
t

c
o
n

v
ic

ti
o

n
(\

1
7

y
e
a
rs

)
(%

)
2

4
.8

3
5

.5
3

3
.6

5
.9

1
2

.2
5

E
v

e
r

in
c
a
rc

e
ra

te
d
(%
)

4
3

.9
6

5
.7

6
1

.6
1

3
.4
1
8

.4
2

1
.1
A
c
c
u
m
u
la
te
d
p
ri

so
n

ti
m

e
(m

n
th

s)
b

1
4

.8
(2

3
.5

)
[5

.6
]

2
1

.2
(2

6
.6

)
[1
2
.3
]
2
1
.6

(2
7

.3
)

[1
0

.0
]

8
.4

(1
4

.0
)

[4
.0

]
6

.0
(1

1
.5

)
[1

.9
]

1
9

.0
(4

5
.9
)
[1

.2
]

E
v
e
r
c
o
n
v
ic
te

d
o

f
v

io
le

n
t

c
ri

m
e

(%
)
5
0

.5
6

6
.9

5
8

.8
1

4
.9

1
9

.4
1

5
.8

E
v
e
r
c
o
n
v
ic
te
d
o

f
p

ro
p

e
rt

y
c
ri

m
e
(%
)

6
4

.0
7

9
.9

7
5

.2
6

3
.6

6
6

.3
5

4
.4

E
v
e
r
c
o
n
v
ic
te
d
o

f
d

ru
g

c
ri
m
e
(%
)
1
2

.9
3

0
.2
3
1
.5
6

.7
1

0
.2

2
2

M
a
rr

ia
g
e

b
a
c
k
g
ro

u
n
d
A
g
e
o
f
fi
rs
t
m
a
rr
ia
g

e
2

6
.0

(5
.9

)
[2
5
]

2
8

.9
(8

.4
)

[2
7

]
2

3
.3

(5
.6

)
[2
2
]
2
3

.1
(6

.4
)

[2
1

]

L
e
n

g
th

o
f

fi
rs

t
m
a
rr
ia
g

e
(y

e
a
rs
)
1

7
.1

(1
2
.3
)

[1
5

]
1
3
.4
(1
1
.9
)

[9
]

2
0
.2
(1

5
.2

)
[1
7
]
1
4

.7
(1

2
.1

)
[9

]
J Quant Criminol
123

T
a
b
le

1
c
o

n
ti

n
u
e
d

V
a
ri
a
b
le
M
a
le
su
b
je
c
ts
(N
=
4
,1
9
1
)
F
e
m
a
le
su
b
je
c
ts
(N
=
4
2
4
)
F
ir
st
m
a
rr
ia
g
e
to
a
…
F
ir
st
m
a
rr
ia
g
e
to
a
…
N
o
n
-c
o
n
v
ic
te
d
sp
o
u
se
C
o
n
v
ic
te
d
sp
o
u
se
N
e
v
e
r
m
a
rr
ie
d
N
o
n
-c
o
n
v
ic
te
d
sp
o
u
se

C
o
n
v
ic

te
d
sp
o
u
se
N
e
v
e
r
m
a
rr
ie
d
C
o
n
v
ic
te

d
5

y
e
a
rs

b
e
fo

re
m

a
rr
ie
d
(%
)
4
9

.7
7

0
.4

1
3

.1
2

1
.4

T
o

ta
l

#
c
o

n
v
ic
ti
o
n

s
a
t

m
a
rr
ia
g

e
4

.1
(6
.9
)
[2
]
1
0

.1
(1

2
.2

)
[6

]
0

.6
(2

.1
)

[0
]

1
.1

(2
.4

)
[0

]

e
a
n

s
o

f
b

in
a
ry

v
a
ri

a
b

le
s

a
re

sh
o

w
n

a
s

p
e
rc

e
n

ta
g

e
s.

F
o

r
ra

ti
o
-l

e
v

e
l

v
a
ri
a
b

le
s,

st
a
n

d
a
rd

d
e
v

ia
ti

o
n

s
a
re

p
ro

v
id

e
d

in
p

a
re
n
th

e
se

s
a
n

d
m

e
d
ia

n
s

a
re
sh
o
w
n

in
b

ra
c
k

e
ts

a
M

a
rr
ie
d

in
d

iv
id

u
a
ls

h
a
v

e
a
n

a
rt

ifi
c
ia

ll
y

sh
o

rt
e
r

le
n

g
th
o
f

fo
ll

o
w

-u
p

in
th

e
a
n
a
ly

si
s
sa
m
p
le

b
e
c
a
u

se
th

e
y

a
re

e
x

c
lu

d
e
d

a
ll

y
e
a
rs

a
ft

e
r

th
e

fi
rs
t
m
a
rr
ia
g
e
d

is
so

lv
e
s.

H
o

w
e
v

e
r,

th
e

d
e
sc

ri
p
ti
v
e
st
a
ti
st
ic

s
p
ro

v
id
e
d
in
th

is
ta

b
le
u
se
a
ll

a
v
a
il

a
b
le

y
e
a
rs

b
O

n
ly

in
d
iv
id
u
a
ls

w
h

o
w

e
re

e
v
e
r
in
c
a
rc
e
ra
te
d
a
re

in
c
lu

d
e
d
J Quant Criminol
123

spouse are also far more likely to have been convicted themselves in the 5 years prior to

marriage (males: 70.4 vs. 49.7 %; females: 21.4 vs. 13.1 %) and to have accumulated more

convictions at the time of their marriage (males: 10.1 vs. 4.1; females: 1.1 vs. 0.6),

implying a substantial degree of assortative mating with respect to criminal behavior.

Interestingly, female subjects who never marry tend to be more crime prone than their

counterparts who marry, and in many instances, more crime prone than those who marry a

convicted spouse. For example, perpetually single females exhibit a younger age of first

conviction, a higher volume of lifetime convictions, and a higher lifetime incarceration

risk. The same is not necessarily true for male subjects, however. Male offenders who

marry a convicted spouse are uniformly more crime prone relative to their married and

never-married counterparts. Thus, first marriage to a convicted spouse tends to be the

deviant marital state for men, whereas singlehood is the deviant marital state for women.

Marriage, Spousal Criminality, and Conviction Frequency

Results from the fixed-effects Poisson models are provided in Table 2. Recall that these

models estimate the effect of marriage and spousal criminality on the number of convic-

tions per year of street time. Model 1 estimates the impact of marriage on conviction

frequency irrespective of the criminality of the spouse, and is equivalent to Eq. 1. Model 2

estimates the impact of spousal criminality on conviction frequency, and is equivalent to

Eq. 2. All models are estimated separately for male and female subjects. Note that in

follow-up models that are not shown, we lagged the marriage indicators by 1 year to ensure

temporal priority of marriage vis-à-vis criminal conviction. The results from these models

were virtually identical to those reported in Table 2.

In Model 1, as expected from a variety of theoretical perspectives, being married is

associated with a significant decrease in conviction frequency relative to being single. This

is true for male as well as female offenders. Exponentiating each of the coefficients (eb)

Table 2 Fixed-effects Poisson models of the impact of first marriage on conviction frequency, by gender

Variable Male subjects (N = 3,356) Female subjects (N = 173)

Model 1 Model 2 Model 1 Model 2

Age .487 (.010)*** .489 (.010)*** .574 (.083)*** .577 (.083)***

Age squared -.013 (.000)*** -.013 (.000)*** -.012 (.002)*** -.012 (.002)***

Age cubed .000 (.000)*** .000 (.000)*** .000 (.000)*** .000 (.000)***

Have a child -.157 (.021)*** -.154 (.021)*** -.026 (.140) -.006 (.141)

Convicted last year .470 (.013)*** .469 (.013)*** .267 (.105)* .257 (.105)*

Accumulated convictions -.013 (.001)*** -.013 (.001)*** -.130 (.015)*** -.129 (.015)***

Imprisoned last year .283 (.015)*** .282 (.015)*** .296 (.161) .316 (.162)

Accumulated prison time .063 (.009)*** .064 (.009)*** .958 (.115)*** .947 (.114)***

Currently married -.314 (.023)*** -.351 (.024)*** -.801 (.169)*** -.617 (.194)***

Currently married convict .296 (.054)*** -.498 (.267)

Coefficients and standard errors are provided. Models adjust for exposure time. Exponentiating the coef-
ficient and subtracting one (eb – 1) yields the proportional increase/decrease in the number of convictions
associated with a unit increase in the regressor. Italicized coefficients represent those that are significantly
different from zero when summed together, yielding the main effect of being married to a convicted spouse
relative to being single (p \ .05)
* p \ .05; ** p \ .01; *** p \ .001 (two-tailed tests)

J Quant Criminol
123

provides an incident rate ratio (IRR), and subtracting one (eb – 1) yields the proportional
increase/decrease in the number of convictions given a state of marriage as opposed to a

state of singlehood. The IRRs for males and females are 0.73 (e-.314) and 0.45 (e-.801),
respectively. This can be taken to mean that being married significantly lowers conviction

frequency by 27 % among males and 55 % among females, on average and all else equal.

Interestingly, a comparison of the impact of marriage for men and women indicates that

women benefit significantly more from marriage than do men. Since the models are

estimated independently, we can conduct a test of the difference in marriage coefficients, a

test that yields a z-statistic of 2.86 (p \ .01) (for details on this test, see Brame et al. 1998).
Before proceeding to Model 2, we first draw attention to the remaining regressors. First,

the expected age-crime relationship is observed, as the coefficients imply an inverted-U

shape to the mean number of convictions per year of street time. Second, fertility status

appears to have an inverse relationship with conviction frequency among male subjects,

but no relationship with conviction frequency among female subjects. Third, past con-

viction strongly influences current conviction frequency. For both males and females,

having been convicted in the previous calendar year increases one’s conviction rate in the

current year, although the accumulated number of convictions is inversely associated with

conviction. Because multicollinearity is not a problem in this model, the effect can be

interpreted as a ‘‘slowing down’’ of the effect of accumulated convictions over time, net of

age. Fourth and finally, past imprisonment influences current conviction risk. Having been

confined in the last calendar year is associated with significantly more convictions only for

male offenders, while the accumulated time served in prison is also positively associated

with conviction frequency among both males and females.

Model 2 examines the differential effects of marriage to non-convicted and convicted

spouses on the number of convictions. In these models, the coefficient for marriage actually

represents the effect of being married to a non-convicted spouse, while the coefficient for

marriage to a convicted spouse represents a contrast with marriage to a non-convicted spouse.

Recall that, to recover the effect of marriage to a convicted spouse relative to singlehood,

these coefficients must be summed together. Italicized coefficients in Table 2 indicate those

for whom the summed coefficients are significantly different from zero.

For men, the coefficient for marriage is negative and significant, which means that

marriage to a non-convicted spouse is associated with a significant decline in conviction

frequency relative to singlehood (IRR = .70). On the other hand, the contrast for marriage

to a convicted spouse is positive and significant, meaning that the decline in conviction

frequency is not as pronounced for these individuals. To test the effect of marriage to a

convicted spouse compared to remaining unmarried, the two marriage coefficients summed

together yields a coefficient of -0.055 (s.e. = .052) and an IRR of 0.95, an effect that is

not statistically significant. In sum, being married to a non-convicted spouse reduces

conviction frequency by 30 percent relative to being unmarried, while being married to a

convicted spouse is statistically indistinguishable from singlehood.

The findings are somewhat different for female subjects. As expected, the coefficient for

marriage is negative and statistically significant (IRR = .54), implying that being married

to a non-convicted spouse leads to significantly fewer convictions relative to remaining

single. While the contrasting coefficient for marriage to a convicted spouse is negative, it is

not statistically significant. It should thus not be interpreted from this that marriage to a

convicted spouse is more protective than marriage to a non-convicted spouse, and close

inspection reveals that this contrast is driven by a comparatively small number of females

(N = 44), which introduces instability. On the other hand, the sum of the two marriage
coefficients does yield a significant contrast of marriage to a convicted spouse relative to

J Quant Criminol
123

singlehood of -1.115 (s.e. = .238), with an IRR of 0.33. This can be taken to mean that,

for female offenders, marriage per se is the most salient transition, with no predictive

influence of the criminal history of the husband. Thus, for both males and females, mar-

riage slows the pace of criminal conviction. Yet for male offenders, only marriage to a non-

convicted spouse is protective. On the other hand, for female offenders, marriage to a

convicted or non-convicted spouse exerts a similar protective influence.
6,7

The Moderating Effect of Criminal History

The influence of a (non-) convicted spouse may depend on the offender’s own criminal

history at the time of marriage. We expect the effects of marriage to become smaller if

individuals committed more offenses before marriage. Chronic offenders would be more

present-oriented and self-centered, and would not have developed the capacity and desire

to invest in social relationships. Therefore, they would be less likely to be affected by

social ties like marriage (Moffitt 1993; Nagin and Paternoster 1994; Rhule-Louie and

McMahon 2007). To investigate the degree to which this is the case, we modeled the

interaction between the subject’s current marital status, spousal criminality, and the sub-

ject’s conviction history at the time of marriage. Conviction history is modeled as a series

of mutually exclusive dummy indicators (0 convictions, 1–3, 4–6, 7–9, 10–12, 13–15,

16?). Among female subjects, the effect of marriage did not differ by the criminal history

of the spouse, nor did it differ by the number of convictions at the time of marriage. As

described above, marriage per se is the relevant transition for female offenders. But we

should note that small cell sizes limit this part of the analysis.

Among male subjects, the impact of marriage does indeed differ by the offender’s

criminal history. In order to facilitate interpretation of this model, we summarize the results

in Fig. 1, which plots the IRRs and 95 % confidence intervals for the interaction between

marital status, spousal criminality, and the number of prior convictions. Because the

6
Note that the estimates yield average reductions over the entire span of marriage. However, some subjects

(i.e., those who were older in 1977) are in the analysis for more years than others. To test the sensitivity of
the results, we estimated the models by limiting attention to discrete post-marriage intervals: the first 1, 5,
and 10 years of the first marriage. To be able to compare married and unmarried persons, singles were
followed until the mean age of first marriage plus respectively 1, 5, or 10 years. Importantly, for males and
females alike, the findings are replicated when a limited number of post-marriage years are considered.
7

At the request of an anonymous reviewer, we also investigated cohort effects. Cohabitation has become
much more widespread over the last decades. Therefore, the effects of marriage may have changed over
time. We limited this investigation to male subjects, as the results for females did not exhibit sensitivity to
birth cohort. We began by stratifying the men into one of three cohorts based on their birth year (1907–1945,
1946–1955, 1956–1965), and then constructed separate marriage indicators for each cohort to include into
the fixed-effects Poisson model. Interestingly, for the earliest cohort, the coefficient for marriage was
positive and statistically significant, while it was negative and significant for the last two cohorts. This
suggests that, relative to singlehood, marriage to a non-convicted spouse increases in salience and desistance
potential over time. In the earliest cohort, in fact, these marriages appear to be criminogenic. However, the
contrasting coefficient for marriage to a convicted spouse was positive and significant for all three cohorts,
indicating that marriage to a non-convicted spouse is more beneficial compared to marriage to a convicted
spouse, irrespective of cohort. An additionally interesting result was that, for the latest cohort only, marriage
to a convicted spouse was associated with a significant reduction in convictions compared to being single.
This suggests that even marriage to a convicted spouse possesses desistance potential in later cohorts (but
not as much potential as marriage to a non-convicted spouse). The finding that the crime-reducing impact of
marriage becomes stronger over time is in line with the study of Bersani et al. (2009). They argue that the
quality and stability of recent marriages may be higher, because these marriages are often preceded by
cohabitation. Cohabitation is considered to be a testing phase, and marriage a further investment in the
relationship.

J Quant Criminol
123

distinction between marriage to a non-convicted spouse and marriage to a convicted spouse

remained relevant, two fitted IRR curves are displayed. Notice first that the effect of

marriage to a non-convicted spouse was estimated very precisely (the 95 % error bars are

comparatively tight around the fitted IRR). Additionally, marriage to a non-convicted

spouse appears to grow significantly stronger in the number of pre-marriage convictions. In

other words, men with lengthy conviction histories benefited in a quite pronounced way

from these marriages. Although the results are not conclusive, to some degree, the same

can also be said of men who marry convicted spouses, but only if they have accumulated at

least ten prior convictions. Men with fewer than ten convictions at the time of marriage are

generally indistinguishable from singlehood when they marry a convicted spouse. How-

ever, it is worth noting that the fitted IRR is unstable and the confidence intervals are quite

wide, suggesting that the analysis has limited power to detect any moderating influences of

criminal history for men who marry convicted spouses.

The Moderating Effect of Marital Stability

The impact of marriage and spousal criminality on criminal conviction might also be

influenced by characteristics of the marriage itself. Sampson and Laub (1993), for example,

suggest that higher marital quality and stability increase the protective effect of marriage

on criminal behavior. This is of extra relevance since individuals marrying a convicted

spouse are more likely to have unstable marriages, as we saw in Table 1. We test this

expectation in a set of additional analyses. We created interactions between the marriage

indicators and a set of mutually exclusive indicators for the total length of the first marriage

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

1-3 4-6 7-9 10-12 13-15 16+

Number of Convictions at the Time of Marriage

E
ff

e
c
t

o
f

M
a
rr
ia
g

e
o

n
C

o
n
v
ic
ti
o

n
F

re
q

u
e
n

c
y

Married to Convicted Spouse

Married to Non-Convicted Spouse

Fig. 1 Effect of marriage and spousal criminality on the number of convictions per year of street time
(male subjects only). The estimates shown are exponentiated coefficients (i.e., incident rate ratios) from a
fixed-effects Poisson model in which the effect of marriage is interacted with the total number of convictions
at marriage. Error bars correspond to 95 % confidence intervals. An exponentiated coefficient of 1.0 implies
no relationship between marriage and conviction

J Quant Criminol
123

in years (1–5, 6–10, 11–15, 16?) until divorce, death, or the end of the observation period

(calendar year 2003). As before, we limit this analysis to the male subjects, because of

power problems in the model for female subjects. These results are reported in Table 3.

For men, the pattern suggests that the beneficial impact of marriage to a non-convicted

spouse generally increases in the total length of marriage, whereas marriage to a convicted

spouse is indistinguishable from singlehood no matter the length of marriage (the summed

marriage coefficients are never significant). A notable finding, however, is that even

marriages to non-convicted spouses which dissolve after a few years have desistance

potential. In other words, even unstable marriages (to conventional women) reduce crime

while the marriages last.

Discussion

The aim of this study has been to investigate the relationship between marriage, spousal

criminality, and the subsequent development of criminal offending. Although numerous

studies have shown that marriage is a transformative life-course event that reduces criminal

offending, the criminal history of the spouse has largely been neglected. We thus used data

from the CCLS, a long-term study of a conviction cohort of Dutch offenders. The data have

a number of advantages for studying the effects of marriage on crime. In contrast to the few

existing studies in this area, the CCLS contains data on the criminal careers of both

offenders and their spouses that reach far into adulthood. Moreover, we had information on

the exact timing of convictions and marriages. Therefore, we were able to clearly distin-

guish partner selection from partner influences during marriage. Furthermore, the data

enabled us to investigate the criminal careers of male as well as female offenders.

The results show that marriage is indeed a salient transition in the criminal career, but

there are important qualifications to this conclusion that have to do with the characteristics

of the offenders (gender, criminal history), the characteristics of the spouses (criminal

history), and the characteristics of the marriages (duration). Among men, being married to

a non-convicted spouse uniformly reduces criminal involvement. On the other hand, being

married to a convicted spouse is indistinguishable from singlehood, and thus sustains

criminal involvement. Although ‘‘criminal marriages’’ are thus indeed not protective, we

do not observe an increase in criminal behavior as has been found in earlier studies

(Woodward et al. 2002). One explanation might be that these earlier studies had no

Table 3 Fixed-effects Poisson models of the impact of first marriage on conviction frequency, by gender
and total length of marriage (male subjects only)

Variable Total length of marriage

1–5 years 6–10 years 11–15 years 16? years

Currently married -.107 (.040)** -.357 (.042)*** -.528 (.049)*** -.445 (.034)***

Currently married convict .149 (.097) .184 (.149) .463 (.134)*** .324 (.086)***

Coefficients and standard errors are provided. Fully specified models are estimated as in Table 2. Expon-
entiating the coefficient and subtracting one (eb – 1) yields the proportional increase/decrease in the number
of convictions associated with a unit increase in the regressor. Italicized coefficients represent those that are
significantly different from zero when summed together, yielding the main effect of being married to a
convicted spouse relative to being single (p \ .05)
* p \ .05; ** p \ .01; *** p \ .001 (two-tailed tests)

J Quant Criminol
123

longitudinal information on spouses’ criminal behavior, and have not been able to clearly

distinguish partner selection from partner influences during the relationship. Similarity in

criminal behavior could also result from the fact that partners already resemble each other

before relationship formation.

In contrast to our expectation, the effect of being married to a non-convicted spouse is

especially pronounced for men with extensive criminal involvement prior to marriage. In

line with this finding, it has been suggested that the crime-reducing effects of relationships

would be stronger for individuals with a higher propensity to commit crimes, simply

because they have more potential criminal behavior in need of deterrence (Wright et al.

2001). Additionally, we find that the impact of a non-convicted spouse increases when

males’ marriages are more stable (i.e., of longer duration). However, even the marriages

that dissolve after a few years appear to lower crime risk while they last.

Women who marry also benefit from their union, but interestingly, this relationship

holds up irrespective of the conviction history of the spouse. Thus, the institution of

marriage per se tends to promote desistance among high-risk female subjects. Remarkably,

we do not find support for the idea that women are more strongly influenced by the

criminal behavior of their partners than men. The fact that we do not find any effect of a

convicted spouse might be partly attributed to the birth of children during marriage. The

birth of a child might have a more pronounced impact on females’ lives (both practically

and emotionally) and reduce the preferences and opportunities to commit crimes even

more than for men (Giordano et al. 2002; Uggen and Kruttschnitt 1998). The crime-

reducing effect of childbearing might thus outweigh the crime-stimulating effect of a

convicted husband. And in fact, in our data we do observe a modest tendency for the

presence of children to contribute to the marriage effect (results of which are not shown).

Untangling the complex interactions between marriage and parenthood is an important task

for future research. In addition, less contact with peers might explain the finding that

marriage reduces women’s criminal behavior irrespective of the criminal background of

the spouse: Married women might prioritize family responsibilities over friends.

Although the data used in this paper are unique—they stem from a large-scale, pro-

spective, longitudinal study with a very long observation period—a number of limitations

have to be taken into account when interpreting the results. These limitations offer several

guidelines for future research. First, our sample consists of persons who were in contact

with the criminal justice system in 1977—all individuals in the sample have committed an

offense at least once during their lives. The sampling frame influences the generalizability

of our results in two ways. First, our results speak to the effect of marriage on criminal

convictions among convicted and to-be-convicted individuals, and not necessarily among

the population at large. Second, the results pertain to a particular time and place. The

Netherlands was characterized by a lenient penal climate until the 1990s. This means that

offenders were less easily convicted in 1977 than today. The fact that the CCLS offenders

were convicted during this era therefore means that they were relatively serious offenders.

Studies in different contexts will therefore be needed to test the generality of the findings

from this sample.

In addition, due to the use of official data, we cannot rule out the possibility that we

underestimate the total number of criminal acts. Not all offenses are recorded by the police

or are prosecuted. The underestimation may be selective, meaning that the probability of

being convicted is not equal for all persons. For example, some offenders may be more

intensely monitored by the police, while others (e.g., those with a higher intelligence) may

adopt more effective strategies to keep out of the arms of the law. However, it should be

noted that the use of official data has important advantages as well. It enables us to

J Quant Criminol
123

examine a great variety of criminal acts that differ in severity, e.g., violent offenses,

property offenses, drug offenses, weapon offenses, and offenses against the public order.

Moreover, it enables us to investigate the development of criminal behavior over the entire

life course.

Second, our measure of marriage includes only legal marriage. From the 1980s onward,

it has become more and more common to cohabit, and cohabitation has even become a

substitute for marriage in the Netherlands (Liefbroer and Dykstra 2000). This development

is less of a problem for our analyses, because the CCLS contains data on a cohort of

individuals convicted in 1977. The largest share of these persons already reached mar-

riageable age before this time. Although nowadays the Netherlands is known for its high

cohabitation rate, marriage patterns were comparable to other countries (e.g., the US)

during most of the period under study. The current increase in cohabitation rates is not

unique to the Netherlands and has taken place in other countries as well (Kalmijn 2002).

Therefore, future studies should also focus on the impact of other relationship types on

criminal offending (see also Bersani et al. 2009).

Third, we are unable to gain insight into the intermediate mechanisms underlying the

relationship between marriage, spousal criminality, and criminal behavior. The question

why marriage to a non-convicted spouse reduces criminal conviction, while marriage to a

convicted spouse does not change conviction risk (at least for men), ultimately remains

unanswered. It is unclear whether the changes in the former case are caused by social

bonds, restructured routine activities, social learning processes, or cognitive transforma-

tions. Untangling these mechanisms and determining their relative importance is an

important task for future research.

Fourth, given that analyses of the marriage-crime relationship must be limited to quasi-

experimental designs, we must maintain a healthy skepticism about the degree to which

our models provide estimates of the causal impact of marriage and spousal criminality on
criminal convictions. To date, no single study can claim to have reliably estimated the true

causal impact of marriage on crime, our study included. Because marriage cannot be

randomly assigned, even in principle, researchers must resort to quasi-experimental

designs that can at least narrow the boundaries of plausible causal estimates. The fixed-

effects models employed here produce consistent estimates of the ‘‘marriage effect’’ in the

presence of confounding by time-stable unobservables, but inconsistent estimates if there

are time-varying unobservables which are correlated with marriage and crime (see Bjerk

2009). The models are thus not a panacea to the selection problem, although they do

restrict the potential sources of confounding to time-varying rather than time-stable ones.

As observed by an anonymous reviewer, the ‘‘marriage effects’’ that we estimate in this

study reflect a sum of the ‘‘true effect’’ of marriage and spousal criminality on conviction

and the ‘‘dynamic selection effect’’ that persists due to lack of exhaustive controls for

relevant time-varying confounders.

Finally, as pointed out by an anonymous reviewer, it will be important moving forward

to further untangle, and theorize about, the selection mechanisms that are at work in the

processes of marriage and partner selection. Most analyses (the present one included) treat

the selection process as one sided, and only from the perspective of the sampled individual.

Yet marriage is clearly a two-sided affair, as a sample subject chooses his (her) partner,

while the spouse-to-be must likewise choose the sample subject as her (his) partner. To the

extent that the partner’s unobservables, which jointly influence his/her marriage and crime

decisions, are highly correlated with the focal individual’s unobservables, and both are

fairly time stable, the results from the analysis will remain robust. However, the veracity of

this assumption is ultimately unknown (and unknowable), because these processes are

J Quant Criminol
123

poorly understood (and understudied). We would add further that this will be true of the

study of any ‘‘market’’ behavior where outcomes depend on decisions made by two or

more parties. Most notably, studies of the employment-crime relationship rarely

acknowledge that a job applicant’s decisions are determined, to an unknown degree, by the

tastes and preferences of potential employers.

Conclusion

Notwithstanding the foregoing limitations and knowledge gaps, this study extends our

knowledge about the role of marriage in the criminal career in important ways. Marriage

matters for the development of criminal behavior, but its impact depends in systematic

ways on gender, criminal history, length of marriage, and spousal criminality. With regard

to the latter observation, getting married does not necessarily mean the end of a criminal

career. For men, the ‘‘good marriage effect’’ clearly depends on the criminal history of the

spouse whom one marries. Our conclusions have significant implications for criminolog-

ical theories emphasizing the protective effects of marriage and give important guidance

for future research. On the basis of the current findings, we believe it is important to

provide more nuance to the prominent idea that marriage uniformly reduces criminal

behavior (e.g., Laub and Sampson 2003). Future studies of the marriage-crime relationship

would be well advised to devote attention to partner selection processes and the way in

which partners influence each other during marriage.

References

Akers RJ (1973) Deviant behavior. A social learning approach. Wadsworth Publishing Co, Belmont
Bersani BE, Laub JH, Nieuwbeerta P (2009) Marriage and desistance from crime in the Netherlands: do

gender and socio-historical context matter? J Quant Criminol 25:3–24
Bjerk D (2009) How much can we trust causal interpretations of fixed effects estimators in the context of

criminality? J Quant Criminol 25:391–417
Blokland AAJ, Nieuwbeerta P (2005) The effects of life circumstances on longitudinal trajectories of

offending. Criminology 43:1203–1240
Blokland AAJ, Nagin DS, Nieuwbeerta P (2005) Life span offending trajectories of a Dutch conviction

cohort. Criminology 43:919–954
Brame R, Paternoster R, Mazerolle P, Piquero A (1998) Testing for the equality of maximum-likelihood

regression coefficients between two independent samples. J Quant Criminol 14:245–261
Brines J, Joyner K (1999) The ties that bind: principles of cohesion in cohabitation and marriage. Am Sociol

Rev 64:333–355
Capaldi DM, Kim HK, Owen LD (2008) Romantic partners’ influence on men’s likelihood of arrest in early

adulthood. Criminology 46:267–299
Farrington DP, West DJ (1995) Effects of marriage, separation, and children on offending by adult males.

In: Blau ZS, Hagan J (eds) Current perspectives on aging and the life cycle, vol 4. JAI Press,
Greenwich, pp 249–281

Giordano PC, Cernkovich SA, Rudolph JL (2002) Gender, crime, and desistance: toward a theory of
cognitive transformation. Am J Sociol 107:990–1064

Giordano PC, Schroeder RD, Cernkovich SA (2007) Emotions and crime over the life course: a neo-
Meadian perspective on criminal continuity and change. Am J Sociol 112:1603–1661

Halaby CN (2004) Panel models in sociological research: theory into practice. Annu Rev Sociol 30:507–544
Haynie DL, Giordano PC, Manning WD, Longmore MA (2005) Adolescent romantic relationships and

delinquency involvement. Criminology 43:177–210
Heckman JJ, Hotz VJ (1989) Choosing among alternative nonexperimental methods for estimating the

impact of social programs: the case of manpower training. J Am Stat Assoc 84:862–874

J Quant Criminol
123

Horney JD, Osgood W, Haen Marshall I (1995) Criminal careers in the short-term: intra-individual vari-
ability in crime and its relation to local life circumstances. Am Sociol Rev 60:655–673

Johnson DR (1995) Alternative methods for the quantitative analysis of panel data in family research:
pooled time-series models. J Marriage Fam 57:1065–1077

Kalmijn M (1998) Intermarriage and homogamy: causes, patterns and trends. Annu Rev Sociol 24:395–421
Kalmijn M (2002) Sociologische analyses van levensloopeffecten: een overzicht van economische, sociale

en culturele gevolgen. [Sociological analyses of life course effects: an overview of economic, social,
and cultural consequences]. Bevolking en Gezin 31:3–46

King RD, Massoglia M, MacMillan R (2007) The context of marriage and crime: gender, the propensity to
marry, and offending in early adulthood. Criminology 45:33–65

Laub JH, Sampson RJ (2003) Shared beginnings, divergent lives. Delinquent boys to age 70. Harvard
University Press, Cambridge

Laub JH, Nagin DS, Sampson RJ (1998) Trajectories of change in criminal offending: good marriages and
the desistance process. Am Sociol Rev 63:225–238

Leverentz AM (2006) The love of a good man? Romantic relationships as a source of support or hindrance
for female ex-offenders. J Res Crime Delinq 43:459–488

Liefbroer AC, Dykstra PA (2000) Levenslopen in verandering: een studie naar ontwikkelingen in de
levenslopen van Nederlanders geboren tussen 1900 en 1970. [Life courses in change: a study of the
developments in life courses of the Dutch born between 1900 and 1970]. Sdu Uitgevers, Den Haag

Moffitt TE (1993) Life-course-persistent and adolescence-limited anti-social behavior: a developmental
taxonomy. Psychol Rev 100:674–701

Moffitt TE, Caspi A, Rutter M, Silva PA (2001) Sex differences in antisocial behaviour. Cambridge
University Press, Cambridge

Nagin DS, Paternoster R (1994) Personal capital and social control: the deterrence implications of a theory
of individual differences in criminal offending. Criminology 32:581–606

Nieuwbeerta P, Blokland AAJ (2003) Criminal careers of adult Dutch offenders (codebook and docu-
mentation). NSCR, Leiden

Nieuwbeerta P, Nagin DS, Blokland AAJ (2009) Assessing the impact of first-time imprisonment on
offenders’ subsequent criminal career development: a matched samples comparison. J Quant Criminol
25:227–339

Piquero AR, Blumstein A, Brame R, Haapanen R, Mulvey EP, Nagin DS (2001) Assessing the impact of
exposure time and incapacitation on longitudinal trajectories of criminal offending. J Adolesc Res
16:54–74

Piquero AlexR, Brame Robert, Mazerolle Paul, Haapanen Rudy (2002) Crime in emerging adulthood.
Criminology 40:137–169

Piquero AR, Farrington DP, Blumstein A (2003) The criminal career paradigm. Crime Justice Rev Res
30:359–506

Rhule-Louie DM, McMahon RJ (2007) Problem behavior and romantic relationships: assortative mating,
behavior contagion, and desistance. Clin Child Fam Psychol Rev 10:53–100

Sampson RJ, Laub JH (1993) Crime in the making: pathways and turning points through life. Harvard
University Press, Cambridge

Sampson RJ, Laub JH, Wimer C (2006) Does marriage reduce crime? A counterfactual approach to within-
individual causal effects. Criminology 44:465–508

Simons RL, Stewart EA, Gordon LC, Conger RD, Elder GH Jr (2002) A test of life-course explanations for
stability and change in antisocial behavior from adolescence to young adulthood. Criminology
40:401–434

Steffensmeier D, Allan E (1996) Gender and crime: toward a gendered theory of female offending. Annu
Rev Sociol 22:459–487

Tak PJP (2003) The Dutch criminal justice system: organization and operation. Boom Legal Publishers,
Meppel

Theobald D, Farrington DP (2011) Why do the crime-reducing effects of marriage vary with age? Br J
Criminol 51:136–158

Uggen C, Kruttschnitt C (1998) Crime in the breaking: gender differences in desistance. Law Soc Rev
32:339–366

Waite LJ (1995) Does marriage matter? Demography 32:483–507
Waite LJ, Gallagher M (2000) The case for marriage: why married people are happier, healthier, and better

off financially. Doubleday, New York
Warr M (1998) Life-course transitions and desistance from crime. Criminology 36:183–215
Warr M (2002) Companions in crime. Cambridge University Press, Cambridge
Western B (2006) Punishment and inequality in America. Russell Sage Foundation, New York

J Quant Criminol
123

Woodward LJ, Fergusson DM, Horwood JL (2002) Deviant partner involvement and offending risk in early
adulthood. J Child Psychol Psychiatry 43:177–190

Wright BRE, Caspi A, Moffitt TE, Silva PA (2001) The effects of social ties on crime vary by criminal
propensity: a life course model of interdependence. Criminology 39:321–350

J Quant Criminol
123

‘‘Because You’re Mine, I Walk the Line’’? Marriage, Spousal Criminality, and Criminal Offending Over the Life Course

Abstract
Objectives
Methods
Results
Conclusions
Introduction
Marriage, Spousal Criminality, and Crime: Desistance or Persistence?
Social Bond
Routine Activities
Socialization
Cognitive Transformations
Hypotheses
Earlier Studies and Their Limitations
Data
Analytic Strategy
Results
Marriage, Spousal Criminality, and Conviction Frequency
The Moderating Effect of Criminal History
The Moderating Effect of Marital Stability
Discussion
Conclusion
References

Sociology
2016, Vol. 50(4) 673 –694

sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0038038515570144

soc.sagepub.com

The Effects of Financial
Vulnerability and Mothers’
Emotional Distress on
Child Social, Emotional and
Behavioural Well-Being: A
Structural Equation Model

Morag Treanor
University of Edinburgh, UK

Abstract
This article aims to understand the pathways through which financial vulnerability affects
children’s social, emotional and behavioural (SEB) well-being and whether that impact is directly
experienced or, as hypothesised, indirectly through their mothers’ emotional well-being. It uses
data from Growing Up in Scotland – a longitudinal birth cohort study of 5217 children born
in 2004–2005. The results show that maternal emotional distress is strongly associated with
financial vulnerability, more so than with income, and that child SEB well-being is negatively
associated with financial vulnerability and maternal emotional distress, with two-thirds of the
effect of financial vulnerability being experienced indirectly through maternal emotional distress.
While the qualitative evidence shows that financial vulnerability adversely affects older children
directly, through the comparisons they make to their reference group, the quantitative finding
is that young children are also negatively affected but predominantly via the effect of financial
vulnerability on their mothers’ emotional distres

Keyword

children, financial vulnerability, income, maternal emotional distress, poverty, quantitative
methods, Scotland, social, emotional and behavioural (SEB) well-being, structural equation
modelling (SEM)

Background

Runciman (1966: 9), in the theory of relative deprivation, posited that ‘people’s attitudes,
aspirations and grievances largely depend on the frame of reference within which they

Corresponding author:
Morag Treanor, Social Policy/Q-Step, School of Social and Political Science, University of Edinburgh, 2.25
Chrystal Macmillan Building, 15a George Square, Edinburgh EH8 9LD, UK.
Email: morag.treanor@ed.ac.uk

570144 SOC0010.1177/0038038515570144SociologyTreanor
research-article2015

Article

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

mailto:morag.treanor@ed.ac.uk

http://soc.sagepub.com/

674 Sociology 50(4)

are conceived’. The frame of reference, or reference group, to which an individual compares
themselves ‘can be a class, a country, a single person or even an abstract idea’ (Runciman,
1966: 11). The theory of relative deprivation suggests that when an individual makes
negative comparisons between themselves and their reference group it can lead to
‘subjective deprivation’ experienced as ‘subjective injustice’ and ‘emotional distress’
(Ragnarsdóttir et al., 2013: 756). Who an individual chooses as their reference group
will invariably affect how the individual feels about their situation, position and status.
Merton (1968: 14) states that ‘some similarity in status attributes between the individual
and the reference group must be perceived or imagined for the comparison to occur at
all’. For those living in poverty and financial vulnerability, the reference group can be
those in the same socioeconomic position or others who are similar in other characteristics
but who have a different socioeconomic position. For mothers living in poverty and
financial vulnerability the reference group may be other mothers in their community
with whom they perceive an attributional similarity, for example, the status of motherhood,
but who may have dissimilar levels of financial vulnerability. Comparing themselves
to mothers with different socioeconomic realities may result in feelings of subjective
disadvantage, which ‘can appear as emotional distress manifested through anger and
depression’ (Ragnarsdóttir et al., 2013: 758; Smith et al., 2012).

It is widely theorised that economic crises and financial vulnerability lead to distress
directly, through the economic problems they bring (Lister, 2004), and also indirectly
through the subjective injustice and emotional distress they trigger in the individual
through comparisons of self to the reference group (Ragnarsdóttir et al., 2013). While the
theory of relative deprivation has commonly been applied to sudden economic crises,
such as recessions (Ragnarsdóttir et al., 2013), it has an apposite application to the study
of chronic economic conditions such as poverty and financial vulnerability. Merton
(1968: 201) noted that ‘“poverty” is not an isolated variable which operates in precisely
the same fashion wherever found; it is only one in a complex of identifiably interdependent
social and cultural variables’. Thus, poverty per se, and not just the experience of
poverty, is relational to its social context (Townsend, 1979: 132).

Experiences of poverty can be transient, that is, mild and alleviable by existing or
acquired resources; or acute, severe, chronic and persistent. This dynamic aspect of
poverty is an increasing focus in the study of poverty with the recognition that poverty,
particularly transient poverty, is more common than cross-sectional studies would
suggest, indicating greater financial vulnerability than previously realised (Berthoud
and Bryan, 2011; Fouarge and Layte, 2005; Jenkins et al., 2001). Yet, financial vulner-
ability is a term that is often used erroneously and synonymously with poverty. Chambers
(1989: 1) emphasises that vulnerability is not the same as poverty. He explains that
where poverty indicates lack or want, vulnerability is defined by ‘insecurity, and expo-
sure to risk, shocks and stress’ (1989: 6). That the financially vulnerable increasingly
includes those who are in-work with insecure livelihoods implies consequences that are
presently under-recognised (Shildrick et al., 2013).

Within a family milieu, how financial vulnerability directly, and its lived experience
indirectly, creates subjective distress is not widely studied. How mothers, who have various
frames of reference, and children, who have their peers as their reference group, are
affected has not been widely studied in relation to the mothers and children together. The

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

Treanor 675

theory of relative deprivation suggests, and the empirical evidence supports the theory,
that social comparisons reduce emotional well-being for those with financial vulnerability.
What is less understood are the pathways through which poverty and financial vulnera-
bility affect the subjective distress of children and mothers. Older children show that
they experience the distressing effects of poverty and financial vulnerability directly,
through their own social comparisons; however, there is no research that studies the
impact of their mothers’ distress on the young people’s experience. Moreover, studies on
the pathways through which financial vulnerability has an impact on young children,
who would not be expected to have either a reference group or the ability to make
comparisons, are few and are not situated within a sociological framework. The gap that
this research aims to fill is to understand these pathways through which financial vulner-
ability has an impact on young children and whether that impact is directly experienced
or, as hypothesised, indirectly through their mothers’ emotional well-being.

Reference Groups and Relative Deprivation

Relative deprivation theory states that subjective comparisons are made between the
individual and their reference group. Such comparisons influence how people feel about
their circumstances: negative social comparisons can lead to ‘invidious self-depreciation’
and ‘personal inadequacy’ (Merton, 1968: 294). Furthermore, the salience of the subjective
comparisons can be greater than the objective reality of a given situation (Ragnarsdóttir
et al., 2013). Within the study of poverty, this resonates in the relatively low level of
overlap between subjective feelings of poverty and objective measures of poverty
(Bradshaw and Finch, 2003). There are several reasons why the relationship between
subjective and objective measures of poverty is imperfect: subjective deprivation; false
consciousness; intra-familial transfer of resources; low aspirations or expectations;
measurement error; and the lagged effect of income poverty on living standards
(Bradshaw and Finch, 2003). Financial vulnerability, in contrast to measures of objective
poverty, has an inherently subjective element.

Financial Vulnerability

Chambers (1989) argues that definitions of poverty conceived by professionals overlook
vulnerability despite it being a primary concern to poor people themselves. He asserts
that poverty, as measured by low income, can be reduced by borrowing, but that the
resulting debt makes households more vulnerable (Chambers, 1989). People living in
poverty have a fear of debt and are more aware than poverty professionals of the trade-offs
between poverty and vulnerability. Chambers (1989: 6) posits that ‘poor people all over
the world are reluctant to take debts which increase their vulnerability’.

Whelan and Maitre (2005, 2008) used the European Community Household Panel
(ECHP) data to create a concept translated directly from Chambers’ work that they call
‘economic vulnerability’. They conceptualise vulnerability as insecurity and exposure
to risk and shock rather than directly measured economic deprivation. Their measure
of economic vulnerability includes objective risk of deprivation and subjective sense of
insecurity (Whelan and Maitre, 2008: 640). They compared the groups identified as being

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

676 Sociology 50(4)

economically vulnerable with social class and found that financial vulnerabilities operate
along traditional social class lines; those from a higher social class ‘had very high
levels of protection from economic vulnerability’ whereas those from the traditionally
lower social classes experienced persistent economic vulnerability (Whelan and Maitre,
2008: 655).

Financial Vulnerability and Subjective Distress

Financial vulnerability, as measured by debt, money worries and managing on current
income is qualitatively shown to have a negative impact on emotional well-being,
stress, anxiety and depression (Green, 2007; Magadi and Middleton, 2007). One study
conducted longitudinal, qualitative research on families living in poverty over the
course of a year in 2008. Its overarching finding does not relate to poverty but to the
depth and extent of financial vulnerability: in particular of debt and the inability to
cope with unexpected bills and expenses (Harris et al., 2009; Stewart, 2009). It high-
lights the negative impact of financial vulnerability on adult emotional health with
families reporting this as their main cause of anxiety, depression and relationship
conflict. A further study emphasises that debt ‘compounds vulnerability and negatively
affects emotional wellbeing’ (Whitham, 2012: 5). The quantitative evidence is less
abundant. One study explores the financial vulnerability on emotional distress, using
data from c.6000 adults aged 16 to 64 years in Sweden. It shows that women are twice
as likely, and men three times as likely, to experience anxiety, depression and reduced
psychological well-being if they are experiencing financial vulnerability (Starrin
et al., 2009).

As regards children and young people, qualitative studies clearly show that older
children (usually aged eight and above) feel ashamed, excluded and stigmatised by their
family’s economic disadvantage (Holscher, 2008). This subjective distress children and
young people experience is said to occur because they are unable to participate in the
social, leisure and celebratory activities of their peer group, which can adversely affect
their friendships and self-esteem (Ridge, 2002, 2009). Additionally, children and young
people are reported as being aware of, and worried about, the financial pressure their
family is under, which has further detrimental effects on their subjective distress
(Whitham, 2012). This suggests that, at least in part, poverty and financial vulnerability
have direct negative impacts on children and young people’s emotional well-being due
to negative social comparisons.

There are a variety of causal pathways proposed for the impacts on financial vulner-
ability on child social, emotional and behavioural (SEB) well-being. The stress induced
by financial vulnerability, be it the result of directly experienced resource deprivation, or
the subjective deprivation induced by subjective comparisons to the reference group, is
postulated to have adverse impacts on mothers’ emotional well-being which in turn have
adverse effects on child well-being, creating an indirect path from financial vulnerability
to child well-being via maternal well-being (Conger et al., 2010; Yeung et al., 2002). This
article uses structural equation modelling to decompose the hypothesised relationship
between financial vulnerability, child SEB well-being and maternal emotional distress.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

Treanor 677

Data and

Methods

The analysis is conducted using data from the Growing Up in Scotland (GUS) study,
a birth and child cohort study, with an annual frequency of data collection, initiated
to record the characteristics, circumstances, health and behaviours of children in
the early years in Scotland (Anderson et al., 2007). A stratified, clustered sampling
strategy was used: a named sample was selected on the basis of the children’s dates of
birth using UK Child Benefit records, chosen because 97 per cent of all eligible fami-
lies were registered for this, then universal, benefit (2005). The response rate across
the birth and child cohorts was 80 per cent of all in-scope children producing an
achieved sample for the birth cohort of 5217 babies at sweep 1 (Anderson et al., 2007:
196) reducing to 3833 children at sweep 5; see Table 1 for more details on survey
response rates. The main carer, usually the mother (approximately 98 per cent), is the
respondent in GUS. The analysis is undertaken using Stata 13 and the ‘surveyset’
procedure is tested to take account of the complex sampling, that is, strata, clustering,
sample selection weights and longitudinal attrition weights. For more information see
the annual user guides on GUS (Bradshaw et al., 2009, 2010; Corbett et al., 2005,
2006, 2007).

The Dependent Variable – Child SEB Well-Being

The child SEB outcome is measured by the age-appropriate Strengths and Difficulties
Questionnaire (SDQ) available for children aged four to 17 years old. For the children in
this study (aged four/five) the questionnaire was completed by mothers. There are five
dimensions to the SDQ: conduct problems; emotional symptoms; hyperactivity; peer
relationships; and pro-social behaviour (Goodman, 1997: 581). The first four of these are
summed to provide a total score. The fifth dimension, pro-social behaviour, cannot be
incorporated into the total score ‘since the absence of pro-social behaviours is conceptually
different from the presence of psychological difficulties’ (Goodman, 1997: 582). The
resulting score from the summed first four dimensions is then reversed and standardised,
giving a mean of zero and a standard deviation of one. Positive scores denote higher than
average SEB well-being and negative scores denote lower SEB well-being.

Table 1. Sweep information for the birth cohort.

Sweep Year Achieved sample Response rate
(all eligible cases) (%)

Response rate
(as % of sweep 1)

1 2005–2006 5217 80 100
2 2006–2007 4512 88 87
3 2007–2008 4193 91 80
4 2008–2009 3994 91 77
5 2009–2010 3833 92 73

Source: GUS sweeps 1–5.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

678 Sociology 50(4)

Financial Vulnerability

GUS collects data in sweep 5 on financial vulnerability. The exact wording of the
questions and their possible responses are set out below (Bradshaw et al., 2010):

•• Money worries: how often would you say you have been worried about money
during the last few weeks?

○ 1 almost all the time,
○ 2 quite often,
○ 3 only sometimes,
○ 4 never

•• Household debt: thinking back over the past 12 months, how often would you say
you have had trouble with debts that you found hard to repay?

○ 1 almost all the time,
○ 2 quite often,
○ 3 only sometimes,
○ 4 never

•• Manage financially: taking everything together, which of the phrases on this card
best describes how you and your family are managing financially these days?

○ 1 Manage very well
○ 2 Manage quite well
○ 3 Get by all right
○ 4 Don’t manage very well
○ 5 Have some financial difficulties
○ 6 Are in deep financial trouble

These three variables form the latent construct ‘financial vulnerability’ estimated in the
first measurement part of the structural equation model discussed in the methods section.

Maternal Emotional Distress (SF-12)

In sweep 5 data are collected using the SF-12 health survey form, which is the instrument
of choice for large scale and longitudinal survey studies (Jenkinson et al., 1997;
Ware et al., 1996). For this paper seven questions from the SF-12 pertaining to maternal
emotional distress are used (EMD1–EMD7). These are:

•• EMD1 – In general, would you say your health is excellent, very good, good, fair
or poor?

○ 1 Excellent
○ 2 Very good
○ 3 Good
○ 4 Fair
○ 5 Poor
○ 6 Can’t say

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

Treanor 679

•• EMD2 – During the past four weeks, have you accomplished less than you would
like as a result of any emotional problems, such as feeling depressed or anxious?

○ 1 Yes
○ 2 No

•• EMD3 – During the past four weeks, did you not do work or other regular activi-
ties as carefully as usual as a result of any emotional problems, such as feeling
depressed or anxious?

○ 1 Yes
○ 2 No

•• EMD4 – How much time during the past four weeks have you felt calm and peaceful?

○ 1 All of the time
○ 2 Most of the time
○ 3 A good bit of the time
○ 4 Some of the time
○ 5 A little of the time
○ 6 None of the time

•• EMD5 – How much of the time during the past four weeks did you have a lot of
energy?

○ 1 All of the time
○ 2 Most of the time
○ 3 A good bit of the time
○ 4 Some of the time
○ 5 A little of the time
○ 6 None of the time

•• EMD6 – How much of the time during the past four weeks have you felt down?

○ 1 All of the time
○ 2 Most of the time
○ 3 A good bit of the time
○ 4 Some of the time
○ 5 A little of the time
○ 6 None of the time

•• EMD7 – During the past four weeks, how much of the time has your physical
health or emotional problems interfered with your social activities like visiting
with friends, relatives, etc?

○ 1 All of the time
○ 2 Most of the time
○ 3 A good bit of the time
○ 4 Some of the time
○ 5 A little of the time
○ 6 None of the time
at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

680 Sociology 50(4)

These seven variables form the latent construct ‘maternal emotional distress’ estimated
in the second measurement part of the structural equation model discussed in the
methods section.

Equivalised Income

Economies of scale, where people pool their resources, that is, share their wealth or
their poverty with other family or household members (Alcock, 2006), are taken account
of in the measurement of income through equivalence scales, which assign a ‘weight’ to
each member. The equivalence scale used here is the modified Organisation for
Economic Co-operation and Development (OECD) equivalence scale, which gives a
weight of 1.0 for the first adult in a household, 0.5 for an additional person aged 15
years or over and 0.3 for any children aged 0–14 years (Chanfreau and Burchardt,
2008), and is a continuous measure taken at sweep 5.

Control Variables

The existing research identifies factors that should be controlled for in analysis that
includes socioeconomic disadvantage and child SEB well-being (Kiernan and Huerta,
2008; Kiernan and Mensah, 2009; Schoon et al., 2010, 2012). These factors are: child’s
gender; family composition; maternal education; maternal employment; birth order of
the child; and the age of the mother at the birth of her first child.

Child’s Gender

The gender of a child is found to be associated with his/her SEB well-being: being a boy
is associated with lower scores on this developmental outcome (Blair et al., 2004).
Research shows that boys mature more slowly than girls (Cohn, 1991), that girls are
more content than boys to sit still and listen in school and that boys are more physical
and active (King and Gurian, 2006); all of which may affect perceptions of boys’ SEB
well-being. The gender variable is a straightforward binary girl/boy variable.

Family Composition

The existing evidence on the impact of family composition on child well-being is often
contradictory. Furthermore, family composition often focuses on the differences
between married and unmarried parents and not on family transitions, that is, moving
from a couple to a lone parent family or vice versa. As the data here are longitudinal,
using all five sweeps, the family composition variable can focus on transitions. The
derived variable is categorical and the categories are:

•• ‘stable couple family’, where a couple has been together since the start of the
study (reference category);

•• ‘stable lone parent family’, where the respondent is the sole adult in the household
in each of the five years of the study;

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

Treanor 68

•• ‘lone parents who have re-partnered’ – there is no distinction in the measure on the
point at which the respondent re-partners;

•• ‘couple families who have separated’ – the same caveat applies as before; and
•• ‘separation(s) and re-partnering(s)’ – this category does not differentiate between

those who may be separating and re-partnering with the same or with different
partners.

Maternal Education

Maternal education is highly significantly associated with both the socioeconomic posi-
tion of a family and with the developmental outcomes of a child (Hansen et al., 2010;
Melhuish et al., 2008). Maternal education is relatively stable in GUS and so the measure
taken at sweep 5 is used in the analysis. The categories are:

•• Degree or equivalent (reference category)
•• Vocational qualification below degree
•• Higher grade/A-level
•• Standard grade/GCSE
•• No qualifications

Maternal Employment

Maternal employment is significantly associated with the socioeconomic position of a
family, with maternal mental well-being and with child outcomes (Thomas et al., 2005;
Zick et al., 2001). The measure for maternal employment taken at sweep 5 is used in the
analysis. The categories are:

•• Full-time employed (reference category)
•• Part-time employed
•• Not in paid work

Study Child’s Birth Order

Birth order is posited to be associated with higher levels of child development (Bradshaw,
2011). However, it is not clear what affect birth order has on SEB well-being as it is not
well documented. The birth order variable is a simple binary first born/not first born
measure.

Age of Mother at First Child’s Birth

Having a younger mother is associated in the literature with lower SEB well-being
(Bradshaw and Tipping, 2010; Bromley, 2009). Younger mothers are also at increased
risk of living in poverty and for living in poverty for longer periods of time (Barnes et al.,
2010). The age of the mother at the birth of her first child is a continuous variable in the
model. The summary statistics for all variables can be found in Table 2.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

682 Sociology 50(4)

Methods

Structural equation modelling (SEM) is used to understand the pathways through which
financial vulnerability has an impact on child SEB well-being, and to ascertain whether
these impacts are experienced directly or indirectly via maternal emotional distress,
while controlling for confounding variables. SEM comprises two components in one
model: the measurement part and the structural part (Acock, 2013). The measurement

Table 2. Summary of all variables.

Variables Count Mean SD Min Max

Child SEB 3537 0.00 1.00 −4.93 1.65
Financial vulnerability 3537 0.00 1.00 −1.02 2.38
Financial vulnerability variables:
How often money worries (FV1) 3537 1.81 0.79 1.00 3.00
How manage financially (FV2) 3537 1.57 0.65 1.00 3.00
How often debts hard to pay (FV3) 3537 1.40 0.66 1.00 3.00
Maternal emotional distress variables:
General health (EMD1) 3537 2.32 0.99 1.00 6.00
Time felt calm (EMD2) 3537 2.99 1.10 1.00 6.00
Time felt energetic (EMD3) 3537 3.03 1.14 1.00 6.00
Time felt down (EMD4) 3537 4.85 1.06 1.00 6.00
Time health interfered socially (EMD5) 3537 5.40 1.12 1.00 6.00
Limited accomplishments (EMD6) 3537 1.85 0.36 1.00 2.00
Limited work/activities (EMD7) 3537 1.88 0.33 1.00 2.00
Income (equivalised) 3537 24676.96 12442.59 1286.77 68965.52
Birth order (ref: first born child) 3537 0.51 0.50 0.00 1.00
Male 3537 0.51 0.50 0.00 1.00
Family transition dummy variables:
Couple 3537 0.78 0.41 0.00 1.00
Lone parent 3537 0.08 0.27 0.00 1.00
Re-partnered lone parent 3537 0.05 0.21 0.00 1.00
Separated couple 3537 0.06 0.24 0.00 1.00
Separations/Re-partnerings 3537 0.03 0.18 0.00 1.00
Age at birth of first child 3537 27.02 5.92 11.00 46.00
Maternal employment dummy variables:
Full-time 3537 0.77 0.42 0.00 1.00
Part-time 3537 0.12 0.33 0.00 1.00
Not in paid work 3537 0.10 0.31 0.00 1.00
Maternal education dummy variables:
No qualifications 3537 0.06 0.24 0.00 1.00
Standard grade (GCSE) 3537 0.14 0.34 0.00 1.00
Higher grade (A level) 3537 0.07 0.26 0.00 1.00
Vocational 3537 0.40 0.49 0.00 1.00
Degree 3537 0.33 0.47 0.00 1.00

Source: GUS sweeps 1–5.
at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

Treanor 683

part comprises confirmatory factor analysis which allows a latent variable to be esti-
mated from its manifest indicator variables while accurately isolating any measurement
error (Acock, 2013). The predictive power of the model is stronger when measurement
error is removed as it is assumed to be a random error with no explanatory power (Acock,
2013). For this article the latent construct financial vulnerability is estimated using three
ordinal variables previously described: ‘money worries’, ‘household debt’ and ‘manage
financially’; and maternal emotional distress is estimated using a measurement model
and the seven variables, EMD1–EMD7, described previously.

The structural part of the model is a path analysis that can be decomposed into direct
and indirect effects pathways. The structural part of the model can show theoretically
causal linkages which can provide evidence to support a causal theory (Acock, 2013).
Figure 1 presents the full structural equation model which shows the measurement parts
of the model with direct pathways leading from financial vulnerability to maternal
emotional distress and child SEB well-being; from maternal emotional distress to child
SEB well-being; and an indirect path from financial vulnerability to SEB well-being via
maternal emotional distress. The control variables described earlier have pathways to
maternal emotional distress, to financial vulnerability and to child SEB well-being.

The use of structural equation modelling in this study brings manifold advantages
over the traditional regression modelling techniques. The first is that the latent constructs
financial vulnerability and maternal emotional distress can be estimated as integral parts
of the model allowing the measurement error to be estimated. The second is that SEM

.73*** .75*** .73***

.43***

-.08***

-.32***

.58***

.52***

.55***

-.78***

-.74***

-.71***

-.68***

Financial
vulnerability

Household
debt

Money
worries

Manage
financially

ε 2 ε 3 ε 4

Child
well-being ε 5

Control variables
(see table for

details)

RSMEA= 0.020
CFI = 0.970
CD = 0.351

ε 1

Maternal
emotional
distress

ε 6

EmD2 ε 8

EmD3

EmD5

EmD6

EmD4

EmD7

EmD1 ε 7

ε 10

ε 9

ε 11

ε 12

ε 13

Figure 1. SEM model.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

684 Sociology 50(4)

allows for direct and indirect effects to see what effects, if any, are direct to the child and
which operate through the pathway of maternal emotional distress. The third is that the
model allows pathways from all of the control variables to the two measurement models
(financial vulnerability and maternal emotional distress) and the outcome variable (child
SEB) to be estimated separately, which decomposes the influence of these confounding
factors, while providing additional evidence to further the substantive knowledge in this
area. SEM relies on correlations between the variables in the model for its estimation
procedures. A correlation of all the single variables (including dummy variables) used in
this analysis can be found in Table 3.

The model was estimated using Stata 13. The svyset weighting facility was tested in
the modelling framework but was not used in the final model of this article as it made
little difference to the coefficients or standard errors but did reduce the goodness of fit
statistics available post-estimation.

Results

The three financial vulnerability indicator variables load highly onto the latent construct
‘financial vulnerability’ with respective loadings of ‘money worries’ 0.75, ‘household
debt’ 0.73 and ‘managing financially’ 0.73, all of which are significant at the 0.1 per cent
level. The seven SF-12 indicator variables load highly and moderately highly onto the
latent construct ‘maternal emotional distress’ with respective loadings of ‘EMD1’ 0.52,
‘EMD2’ 0.58, ‘EMD3’ 0.55, ‘EMD4’ –0.78, ‘EMD5’ –0.74, ‘EMD6’ –0.71, ‘EMD7’
–0.68, all of which are significant at the 0.1 per cent level.

The goodness-of-fit measures, shown in Table 4, are as follows: the root mean squared
error of approximation (RMSEA) is 0.020, much lower than the recommended 0.5 and
closer to the ideal of zero (Acock, 2013; Hu and Bentler, 1999; Statacorp, 2013); com-
parative fit index (CIF) is 0.970, higher than the recommended 0.95; the Tucker-Lewis
index (TLI) at 0.981 is higher than the recommended 0.95; and the coefficient of deter-
mination (r-squared, ranging between 0 and 1) at 0.351 is moderate (Statacorp, 2013).

Table 5 shows the direct and indirect effects in the model. The first section of the
model shows that maternal emotional distress has the largest negative effect on child
SEB well-being, accounting for almost a third of a standard deviation decrease
(β = −0.32). Financial vulnerability, as hypothesised, has a direct (β = −0.08) and an
indirect (β = −0.14) negative effect on child SEB well-being through the pathway of
maternal emotional distress (total effect size = −0.22). Other variables that have an
impact on child SEB well-being are: (1) income, direct β = 0.05 and indirect β = 0.06:
higher income is associated with more positive levels of child SEB well-being; (2) not
being the first born, that is, having siblings, is associated with higher SEB well-being,
(direct β = 0.11); (3) boys, as expected, have lower SEB well-being (direct β = −0.13);
(4) family transitions – stable lone parent (indirect β = −0.02), re-partnered lone parent
(indirect β = −0.02), separated couple (indirect β = −0.03), through their effect on maternal
emotional distress, and separations and re-partnerings (direct β = −0.07), all compared to
stable couple parents. This means that having a mother with repeated separations and
re-partnerings is the only family formation that has a direct, negative association with
SEB well-being; (5) age of the mother at first birth (direct β = 0.06); (6) not being in
paid work, direct β = −0.05 and indirect β = −0.07; and (7) all education levels

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

T
ab

le
3

or
re

la
tio

n
m

at
ri

x
of

a
ll

va
ri

ab
le

s
(in

cl
ud

in
g

du
m

m
ie

s)
.

SE
B

FV
1

FV
2

FV
3

M
M

H
1

M
M

H
2

M
M

H
3

M
M

H
4

M
M

H
5

M
M

H
6

M
M

H
7

In
co

m
e

Fi
rs

t
bo

rn
M

al
e

C
ou

pl
e

Lo
ne

re
nt

Re
-p

ar
t

ne
r

Se
p

co
up

le
Se

Re
-p

ar
t

A
ge

Fu
ll-

tim
e

Pa
rt

–
tim

e
U

ne
m

–
pl

oy
ed

N
o

qu
al

s
St

an
d

H
ig

he
r

Vo
ca

–
tio

na
l

D
eg

re
e

SE
B

FV
1

−0
.2

6
1

2
−0

.2
09

0.
56

6
1

3
−0

.2
43

0.
53

8
0.

52
3

M
M
H
1
−0
.2

9
0.

22
7

0.
22

9
0.

4
1

M
M
H
2
−0
.2

99
0.

23
7

0.
20

8
0.

17
9

0.
34

7
1

M
H

3
−0

.2
62

0.
20

7
0.

21
3

0.
15

6
0.

46
0

0.
59

0
1

M
M
H

4
0.

32
1

−0
.2

92
−0

.2
52

−0
.2

4
−0

.3
59

−0
.4

92
−0

.4
30

M
M
H
5

0.
28

4
−0

.2
41

−0
.2

04
−0

.2
56

−0
.4

03
−0

.3
65

−0
.3

86
0.

58
9

M
M
H
6

0.
25

1
−0

.2
65

−0
.2

6
−0

.2
62
−0
.3

9
−0

.4
16

−0
.3

0
0.

55
0

0.
55

3
1

M
M
H
7
0.

22
6

−0
.2

39
−0

.2
03

−0
.2

56
−0

.3
29

−0
.3

79
−0

.3
53

0.
51

4
0.

54
8

0.
75

4
1

co
m

e
0.

24
1

−0
.3

31
−0

.3
81

−0
.3

7
−0

.2
15

−0
.0

73
−0

.0
70

0.
18

1
0.

21
4

0.
17

2
0.

16
9

Fi
rs
t
bo

rn
0.

06
0

0.
03

3
0.

04
0

0.
04

4
0.

01
0

0.
05

2
0.

01
6

−0
.0

32
−0

.0
49

−0
.0

44
−0

.0
45

−0
.1

43
1

M
al
e
−0
.1

5
0.

01
3

0.
02

4
−0

.0
07

0.
01

1
0.

03
3

0.
01
3
−0

.0
26

−0
.0

29
−0

.0
24

−0
.0

26
−0

.0
10

0.
01
4
1

ou
pl

e
0.

19
7

−0
.2
39
−0

.1
87

−0
.3

06
−0

.1
64

−0
.0

43
−0

.0
48

0.
17
6
0.

18
6

0.
16

5
0.

15
6

0.
40

1
0.

05
2

−0
.0

25
1

ne
p

ar
en

t
−0

.1
15

0.
12

9
0.

13
5

0.
19

5
0.

11
0

0.
00

6
0.

02
2

−0
.1
04
−0

.1
47

−0
.1
31
−0

.1
28

−0
.3

02
−0

.0
70
0.
03
2
−0

.5
50

R
e-

pa
rt

Lo
ne
−0
.0

74
0.

08
9

0.
04
7
0.

12
7

0.
06

4
0.

01
8

0.
01
4
−0

.0
44

−0
.0

54
−0

.0
25

−0
.0

8
−0

.1
22

−0
.0
38
−0

.0
22

−0
.4

25
−0

.0
65

Se
p
co
up

e
−0

.0
50

0.
13

9
0.

10
3

0.
13
0
0.

07
7

0.
04
4
0.

04
2

−0
.1
26
−0

.0
91

−0
.1

05
−0

.0
95

−0
.1

60
0.

05
4

0.
01
3
−0

.4
73

−0
.0
73
−0

.0
56

Se
ps

/R
e-

pa
rt

s
−0

.1
26

0.
06
8
0.

03
7

0.
09

2
0.

03
8

0.
01
2
0.

00
7

−0
.0

34
−0

.0
27

−0
.0

19
0.

00
2

−0
.1

17
−0

.0
41

0.
01
9
−0
.3
53
−0
.0
54
−0

.0
42

−0
.0

47
1

ge
0.

19
7
−0
.2

19
−0

.1
98

−0
.3
19
−0

.1
57

−0
.0

0
−0

.0
27
0.
13
6
0.

16
1

0.
11

5
0.

12
1

0.
51
0
−0

.2
17

−0
.0

27
0.

36
5

−0
.2

11
−0

.2
08

−0
.1
06
−0

.1
38

Fu
ll-
tim
e
0.
20
1
−0

.2
55

−0
.2

68
−0

.3
21

−0
.1

82
−0

.0
69

−0
.0
60
0.

19
5

0.
20
6
0.

19
0

0.
19
6
0.
52
3
−0
.0

12
−0

.0
15

0.
56
4
−0

.4
66

−0
.0
25
−0

.3
26

−0
.1

44
0.

31
3

Pa
rt

-t
im

e
−0

.0
46

0.
12
6
0.

13
4

0.
13
6
0.

04
5

0.
01
7
0.

00
0

−0
.0

78
−0

.0
45
−0
.0

51
−0

.0
64

−0
.2

55
−0

.0
15
−0
.0

15
−0

.3
01

0.
20
0
0.

01
1

0.
24

4
0.

06
2

−0
.0

94
−0

.6
92

U
ne

m
pl

oy
ed
−0
.2
27
0.

21
5

0.
22
4
0.

29
5

0.
20
2
0.

07
6

0.
08

2
−0

.1
83

−0
.2

35
−0

.2
07

−0
.2

01
−0

.4
43

0.
03
3
0.
03
7
−0
.4

49
0.

42
4

0.
02
2
0.

18
5

0.
13
1
−0
.3
29

−0
.6

27
−0

.1
29

N
o
qu
al
s
−0
.1
47
0.
09
1
0.

09
8

0.
12
6
0.

10
6

0.
03
3
0.
03
8
−0
.1

13
−0

.1
21

−0
.0
94
−0
.0
91
−0
.2

47
0.

08
4

0.
00
1
−0

.1
65

0.
11
7
0.

10
4

0.
01
2
0.

06
5

−0
.2

00
−0

.2
14

0.
04
4
0.

24
6

St
an

da
rd

−0
.0

80
0.

05
2
0.
06
1
0.

10
5

0.
06
4
−0

.0
18

−0
.0
04
−0

.0
38

−0
.0

59
−0

.0
30

−0
.0
19
−0

.1
89

0.
04
0
0.

01
9

−0
.1

29
0.

11
2

0.
06
5
0.

02
0

0.
02
6
−0

.1
91

−0
.1

25
0.

02
9

0.
14

1
−0

.1
01

H
ig
he
r
0.
02
3
−0

.0
19

−0
.0
02
−0

.0
05

−0
.0

07
−0

.0
12

0.
01
2
0.

01
4

0.
00
8
0.

02
1

0.
01
7
−0

.0
34

0.
01
6
0.

00
5

0.
01
2
0.

00
9

−0
.0
01
−0

.0
09

−0
.0

28
−0

.0
38
0.
00
7
0.
01
0
−0
.0

21
−0

.0
71

−0
.1

11
1

oc
at

io
na

l
−0

.0
20

0.
06
0
0.

07
4

0.
08
9
0.

03
6

−0
.0

33
−0

.0
24
0.
01
3
−0

.0
03

−0
.0
02
−0
.0
15
−0
.1

50
0.

00
2
−0
.0

30
−0

.0
57

0.
02
1
0.

03
2

0.
03
4
0.

01
7

−0
.1

22
−0

.0
15
0.
03
3
−0

.0
16

−0
.2
07
−0

.3
25

−0
.2

28
1

eg
re

e
0.

14
2

−0
.1

36
−0

.1
70

−0
.2
31
−0

.1
34

0.
03
8
0.
00
2
0.
06
4
0.

10
2

0.
06
0
0.

06
6

0.
43

8
−0

.0
82

0.
01
5
0.

23
0

−0
.1
68
−0

.1
33

−0
.0
51
−0

.0
55

0.
38

9
0.

21
1

−0
.0

84
−0

.1
99

−0
.1
78
−0

.2
80

−0
.1

96
−0

.57
4

So
ur

ce
: G

U
S

sw
ee

ps
1

–5
.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

686 Sociology 50(4)

(except Higher/A-level grade) compared to degree educated mothers with direct betas
of −0.07 (no qualifications), –0.05 (standard grade/GCSE) and vocational level (–0.04).
As the coefficients for this model are additive, this means that the combined direct and
indirect coefficients for income, financial vulnerability and employment sum to 0.45 of
a standard deviation decrease in child SEB well-being, which, when combined with the
direct effect of maternal emotional distress (0.32) amounts to a −0.77 of a standard
deviation decrease in SEB well-being.

In the second section, the model shows the variables that are directly associated
with financial vulnerability as there are no indirect effects on financial vulnerability.
The largest effect comes from income (β = −0.34) which accounts for 34 per cent of
a standard deviation increase in financial vulnerability. Other variables statistically
associated with increased levels of financial vulnerability are: (1) being a re-partnered
lone parent (β = 0.06) or being a separated couple (β = 0.07), compared to being
a stable couple; (2) the age of the mother at the birth of first child (β = 0.08) with
financial vulnerability decreasing as maternal age increases; and (3) being in either
part-time work (β = 0.08) or not in paid work (β = 0.14) compared to being employed
full-time.

In the third section, the model shows that financial vulnerability has the largest
effect, greater than income, on maternal emotional distress (β = 0.43), accounting for
43 per cent of a standard deviation increase. Income has an insignificant direct effect
on maternal emotional distress; however, it has a strong indirect effect mediated
through its association with financial vulnerability (β = −0.15). This suggests that
financial vulnerability is a more important factor in increased emotional distress than
income alone and that it is important irrespective of the level of income a family has.
Other variables statistically significantly associated with increased maternal emotional
distress are: having an employment status of ‘not in paid work’ (β = 0.18), and having
been part of a couple that has since separated (β = 0.08): each of these has a direct
pathway and an indirect pathway through its association with financial vulnerability.
The remaining significant variables have an indirect pathway through financial vulner-
ability: (1) being a re-partnered lone parent (β = 0.03); (2) having part-time compared
to full-time employment (β = 0.03); and (3) the age of the mother at the birth of first
child (β = −0.03).

Table 4. Goodness-of-fit statistics.

Goodness-of-fit statistics Value

Root mean squared error of approximation (RMSEA) 0.020
Comparative fit index (CIF) 0.970
Tucker-Lewis index (TLI) 0.981
Coefficient of determination (CD) 0.351
N 3583
d.f. 355

Source: GUS sweeps 1–5.
at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

Treanor 687

T
ab

le
5

.
D

ir
ec

t
an

d
in

di
re

ct
e

fe
ct

O
ut

co
m

e
D

ir
ec

t
ef

fe
ct

s
In

di
re
ct
e

ffe
ct

C
oe

d.
E

rr
.

P-
va

e
⩽

St
d.

C
oe

f.
C

oe
f.

St
d.

E
rr

.
P-

va
lu

e
⩽
St
d.
C
oe
f.

C
hi

ld
S

EB
←

M
at

er
na

l e
m

ot
io

na
l d

is
tr

es
s

−
0.

62
0.

04
2
0.
00

0
−

0.
32

–
–

na
nc

ia
l v

ul
ne

ra
bi

lit
y

−
0.

13
0.

04
0
0.
00

1
−

0.
08
−
0.

24
0.

01
5

0.
00
0
−
0.
14

In
co
m
e
0.

00
0.

00
0
0.
01
4
0.

05
0.

00
0.
00
0
0.
00
0
0.

Fi
rs
t
bo
rn
0.

22
0.

03
1

0.
00
0
0.

11
−

0.
02
0.
01
3
0.
21
4
−
0.

M
al

e
−

0.
25
0.
03
0
0.
00
0
−
0.

3
−

0.
02
0.
01
2
0.

20
1

−
0.
01

Fa
m

ily
t

ra
ns

iti
on

s
(r

ef
: s

ta
bl

e
co

up
le

St
ab

le
lo

ne
p
ar
en

t
0.

09
0.

06
9

0.
19
3
0.

2
−

0.
06
0.
02
9
0.

05
1

−
0.

e-
pa

rt
ne

re
d

lo
ne

p
ar

en
t

−
0.

06
0.

07
2

0.
42

7
−

0.
01
−
0.
09
0.
03
1
0.
00
3
−
0.
02

Se
pa

ra
te

d
co

up
le
0.
09

0.
07

0
0.

19
2

0.
02
−
0.
13
0.

03
0

0.
00
0
−
0.
03

Se
pa

ra
tio

ns
a

nd
r

e-
pa
rt
ne

ri
ng

s
−

0.
39

0.
08
6
0.
00
0
−
0.

07
0.

00
0.
03
6

0.
97

1
0.

A
ge

o
f m

ot
he

r
at

fi
rs

t
bi

rt
h

0.
01
0.
00
3
0.

00
1

0.
06
0.
00
0.
00
1
0.

09
7

0.
01

pl
oy

m
en

t
st

at
us

(
re

f:
fu

ll-
tim

e
w

or
k)

rt
-t

im
e

w
or

k
0.

02
0.

05
4

0.
70

1
0.

01
−

0.
06
0.
02
3
0.

01
2

−
0.
02

ot
in

p
ai

d
w

or
k

−
0.

17
0.

06
7

0.
01
3
−
0.
05
−
0.
22
0.
03
0
0.
00
0
−
0.
07

uc
at

io
n

st
at

us
(

re
f:

de
gr

ee
o

r
ab

ov
e)

o
qu

al
ifi

ca
tio

ns
−

0.
30

0.
07
3
0.
00
0
−
0.

07
−

0.
05
0.
03
0
0.

10
1

−
0.
01

an
da

rd
/G

C
SE

−
0.

15
0.

05
3

0.
00

5
−

0.
05
0.
02
0.
02
2
0.

39
9

0.
01

H
ig

he
r/

A
-le

ve
l

−
0.
02
0.

06
3

0.
80

0
0.
00
0.

04
0.

02
6

0.
15
8
0.
01

V
oc
at
io
na

l b
el

ow
d

eg
re
e
−
0.
07
0.
03
8
0.
05
4
−
0.
04
0.

01
0.

01
6

0.
35

3
0.

01
Fi

na
nc
ia
l v
ul
ne
ra
bi

lit
y
←

In
co
m
e
0.
00
0.
00
0
0.
00
0
−
0.

34
–

–
–

–

Fi
rs
t
bo

rn
−

0.
01
0.
02
1
0.

55
3

−
0.

01
–

–
–
–

M
al
e
0.
01
0.
02
0

0.
78

8
0.

00
–

–
–

–

(C
on

tin
ue

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

688 Sociology 50(4)

O
ut
co
m
e
D
ir
ec
t
ef
fe
ct
s
In
di
re
ct
e
ffe
ct
s
C
oe

f.
St

d.
E
rr
.
P-
va

lu
e
⩽

St
d.
C
oe
f.
C
oe
f.
St
d.
E
rr
.
P-
va
lu
e
⩽
St
d.
C
oe
f.

m
ily

t
ra

ns
iti

on
s

(r
ef

: s
ta

bl
e

co
up

le
):

St
ab
le

lo
ne

p
ar
en
t
0.
05
0.
04
6
0.

31
1

0.
02
–
–
–
–

R
e-
pa
rt

ne
re

d
lo

ne
p
ar
en
t
0.
17
0.

04
8

0.
00
1
0.

06
–

–
–
–

pa
ra

te
d

co
up

le
0.

17
0.

04
7

0.
00
0
0.

07
–

–
–
–

Se
pa
ra

tio
ns

a
nd

r
e-

pa
rt

ne
ri

ng
s

0.
08
0.
05
7
0.

14
8

0.
03
–
–
–
–

ge
o

m
ot

he
r

at
fi

rs
t

bi
rt

h
−

0.
01
0.
00
2
0.
00
0
−
0.

08
–

–
–
–

Em
pl

oy
m

en
t

st
at
us
(
re
f:

fu
ll-

tim
e

w
or

Pa
rt
-t
im
e
w
or
k
0.
14
0.
03
5
0.
00
0
0.
08
–
–
–
–

N
ot

in
p

ai
d

w
or
k
0.

26
0.

04
4

0.
00
0
0.

14
–

–
–
–

Ed
uc

at
io

n
st

at
us
(
re

f:
de

gr
ee

o
r

ab
ov

N
o
qu
al

ifi
ca

tio
ns
0.
00
0.
04
9
0.

96
0

0.
00
–
–
–
–

St
an
da
rd

/G
C

SE
−

0.
01
0.
03
5
0.

79
1

−
0.
01
–
–
–
–

ig
he

r/
A

-le
ve

l
−

0.
04
0.
04
2
0.
39
9
−
0.

02
–

–
–
–

V
oc
at
io
na
l b
el
ow
d
eg
re
e
0.
04
0.

02
5

0.
16
3
0.

03
–

–
–

–
M

at
er

na
l E

m
ot
io
na

l d
is

tr
es

s
(E

M
D

)
←

Fi
na

nc
ia

l v
ul

ne
ra

bi
lit

y
0.

38
0.

02
5
0.
00
0
0.

In
co
m
e
0.
00
0.
00
0
0.
14
1
0.
04
0.
00
0.
00
0
0.
00
0
−
0.
15

Fi
rs
t

bo
rn

0.
03
0.
01
8
0.

07
0

0.
03
0.
00
0.
00
8
0.
55
3
0.
00

M
al
e
0.
02
0.
01
7
0.

20
3

0.
02
0.
00
0.
00
8
0.

78
8

0.
00
(C
on
tin
ue
d)
T
ab
le
5

.
(C

on
tin

ue
d)

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

Treanor 689

O
ut
co
m
e
D
ir
ec
t
ef
fe
ct
s
In
di
re
ct
e
ffe
ct
s
C
oe
f.
St
d.
E
rr
.
P-
va
lu
e
⩽
St
d.
C
oe
f.
C
oe
f.
St
d.
E
rr
.
P-
va
lu
e
⩽
St
d.
C
oe
f.

Fa
m
ily
t
ra
ns
iti
on
s
(r
ef
: s
ta
bl
e
co
up
le
):

St
ab
le
lo
ne
p
ar
en
t
0.
06
0.
04
0
0.

12
4

0.
03
0.
02
0.
01
8
0.

31
2

0.
01

R
e-
pa
rt
ne
re
d
lo
ne
p
ar
en
t
0.
05
0.

04
3

0.
27

0
0.
02
0.
06
0.
01
9
0.
00
1
0.

Se
pa
ra
te
d
co
up
le
0.

11
0.

04
1

0.
01
0
0.
05
0.
07
0.
01
8
0.
00
0
0.
03

Se
pa
ra
tio
ns
a
nd
r
e-
pa
rt
ne
ri
ng
s
−
0.
05
0.
05
1

0.
31

2
−
0.
02
0.
03
0.
02
2
0.

15
0

0.
01

A
ge
o

f m
ot

he
r
at
fi
rs
t
bi
rt

h
0.

00
0.
00
2

0.
63

1
0.
01
0.
00
0.
00
1
0.
00
0
−
0.
03

Em
pl
oy
m
en
t
st
at
us
(
re
f:
fu
ll-
tim
e
w
or
k)

Pa
rt
-t
im
e
w
or
k
0.
01
0.
03
1
0.
80
1
0.
01
0.
05
0.
01
4
0.
00
0
0.
03

N
ot
in
p
ai
d
w
or
k
0.
19
0.
04
0
0.
00
0
0.
12

0.
10

0.
01
8
0.
00
0
0.
06

Ed
uc
at
io
n
st
at
us
(
re
f:
de
gr
ee
o
r
ab
ov
e)

N
o
qu
al
ifi
ca
tio

ns
0.

08
0.

04
3
0.
05
7
0.
04
0.
00
0.
01
9

0.
96

0
0.
00

St
an
da
rd
/G
C
SE
−
0.
02
0.
03
1
0.
43
5
−
0.
02
0.
00
0.
01
3
0.
79
1
0.
00

H
ig
he
r/
A
-le
ve
l
−
0.
04
0.
03
7

0.
29

8
−

0.
02
−
0.
01
0.
01
6
0.
40
0
−
0.
01

V
oc

at
io

na
l b

el
ow

d
eg

re
e
−
0.
05
0.

02
3

0.
04

6
−

0.
04
0.
01
0.
01
0
0.

16
5

0.
01
So
ur
ce
: G
U
S
sw
ee
ps
1
–5
.

–
no

in
di

re
ct

p
at

T
ab
le
5
.
(C
on
tin
ue
d)
at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

690 Sociology 50(4)

Discussion

The results strongly demonstrate that experiencing financial vulnerability is signifi-
cantly associated with higher levels of maternal emotional distress; an effect even
greater than that of income. This corresponds to the literature where the presence of
financial vulnerability is associated with high levels of stress, anxiety and depression in
Swedish adults (Starrin et al., 2009). They show too that the broader, more subjective,
concept of financial vulnerability is more salient than the objective measure of income,
a finding premised in the theory of relative deprivation (Ragnarsdóttir et al., 2013).
Financial vulnerability encapsulates the objective deprivation resulting from low
income and the subjective deprivation associated with comparative aspects such as feel-
ings towards coping on income. The results also suggest, as per Chambers’ (1989)
assertion, that measures to alleviate low income, such as increased borrowing, increase
vulnerability and that this vulnerability is keenly felt. This has relevance to academia,
policy and practice and suggests that consideration should be given to financial vulner-
ability when working with families experiencing poverty. For practice situations, financial
vulnerability is an easily measured concept that could be employed to establish mothers’
heightened vulnerability and raised risk of emotional distress.

The results also strongly demonstrate that experiencing financial vulnerability is sig-
nificantly associated with lower child SEB well-being which supports the findings in
the qualitative literature (Harris et al., 2009; Whitham, 2012). The literature on older
children indicates that the impact of financial vulnerability is directly experienced
through the comparisons they make with their peer (reference) group (Ridge, 2002). It
was hypothesised in this article that young children would be unable to make their own
social comparisons and that the effect of financial vulnerability experienced by young
children would be indirect, through the pathway of maternal emotional distress. This is
indeed the case for almost two-thirds of the effect; however, the other third of the effect
is experienced directly by the child despite their young ages. It is possible that the
remaining third of the effect on the child may have another, unmeasured, pathway
through a maternal characteristic or it may be that young children are directly affected
by financial vulnerability. What is of note here is the young age of the children for
whom this association is statistically visible.

These findings show that child SEB well-being is responsive to financial vulnerability
and maternal emotional distress which suggests that were these conditions to change
then child well-being might change too. That children of such a young age display lower
SEB well-being when maternal vulnerabilities are high is a central finding of this article
raising two thoughts: (1) children’s SEB well-being is highly sensitive to their mothers’
socioeconomic status and emotional distress; and (2) this implies that SEB is a malleable
rather than a fixed trait that may respond to intervention.

There are two other results of incidental note; the effects of paid work on child and
maternal well-being and the effects of family formation. A mother not being in paid work
greatly increases her emotional distress and financial vulnerability. This in turn results in
decreased child well-being. In a time of austerity measures that have disproportionately
affected female employment the effect of not being in paid work on maternal emotional
distress is substantively and statistically important. What is not modelled in this analysis
is the nature of the effects of unpaid work on maternal emotional distress, that is, whether

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

Treanor 691

unpaid care work leads to greater emotional distress or whether paid employment leads
to reduced emotional distress, or a combination thereof. The literature in this area is
large, varied and provides evidence to support both of these theses. This has not been
modelled in this article but would make an interesting area of future study using the GUS
data. The results in relation to family formation show that family transitions per se are
not key to child SEB well-being but that they are key to financial vulnerability and, to a
lesser extent, maternal emotional distress. The ‘couple who separated’ family transition
was directly linked to poorer maternal emotional distress, which suggests that women
experiencing separation would benefit from targeted emotional and financial support.
The categories ‘stable lone parent’ and ‘re-partnered lone parent’ have no significant
association with child SEB well-being directly, which suggests that the effect of a couple
separating on maternal emotional distress is likely to be time-limited. In the end, only
‘separations and re-partnerings’ was directly linked to poorer child SEB well-being,
which may indicate flux and uncertainty in a family’s life. As the literature using cross-
sectional data often finds negative effects associated with lone-parenthood the longitudinal
approach taken in this article benefits the analysis of the effects of family formation on
child well-being.

Conclusions

In conclusion, this article shows that young children experience the effects of financial
vulnerability indirectly via their mothers’ emotional distress and not through their own
social comparisons as is the case with older children. It shows too that financial vulner-
ability has a greater negative effect on maternal emotional distress than income per se
suggesting that the subjective element of financial vulnerability is salient. It also shows
that child SEB well-being is malleable and that children are emotional barometers
responding to and corresponding with their mothers’ well-being.

A second conclusion is that actions taken to ameliorate the effects of income depriva-
tion can paradoxically increase financial vulnerability, and that this is an added stressor
for those living on low incomes, a factor which ought to be considered when working
with, or legislating for, families living in socioeconomic disadvantage. Furthermore, not
being in paid work affects maternal emotional distress and children’s SEB well-being
independently, even after income and financial vulnerability are controlled for, which is
important and may have implications for policy, for example, on access to affordable,
high-quality childcare. Important too is the evidence elicited due to the longitudinal
nature of the data that repeated separations and re-partnerings is the only category of
family formation directly associated with reduced child SEB well-being. This provides
evidence to counter the culture of blame towards ‘family breakdown’ common in current
political discourse (DfE, 2012). It also has implications for practice: this evidence sug-
gests that families who experience flux in their relationships are particularly vulnerable.

The implications of these findings are relevant to academia, policy and practice. It
adds to the knowledge base in this field and makes the recommendation that financial
vulnerability ought to be considered in tandem with income poverty and other measures
of socioeconomic disadvantage.

The Growing Up in Scotland data are publicly available from the UK Data Service:
http://ukdataservice.ac.uk/.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

Homepage

http://soc.sagepub.com/

692 Sociology 50(4)

Acknowledgements

I would like to thank my supervisors, Professor Kay Tisdall and Dr Paul Norris for their support.

Funding

I would like to thank the Economic and Social Research Council (ESRC) for the funding for my
PhD (award number ES/HO13008/1) at the University of Edinburgh.

References

Acock AC (2013) Discovering Structural Equation Modeling Using Stata, Revised Edition.
College Station, TX: Stata Press.

Alcock P (2006) Understanding Poverty. Basingstoke: Macmillan.
Anderson S, Bradshaw P, Cunningham-Burley C, et al. (2007) Growing Up in Scotland: Sweep

One Overview Report. Edinburgh: Scottish Government.
Barnes M, Chanfreau J and Tomaszewski W (2010) Growing Up in Scotland: The Circumstances

of Persistently Poor Children. Edinburgh: Scottish Government.
Berthoud R and Bryan M (2011) Income, deprivation and poverty: A longitudinal analysis. Journal

of Social Policy 40: 135–156.
Blair KA, Denham SA, Kochanoff A, et al. (2004) Playing it cool: Temperament, emotion regula-

tion, and social behavior in preschoolers. Journal of School Psychology 42: 419–443.
Bradshaw J and Finch N (2003) Overlaps in dimensions of poverty. Journal of Social Policy 32:

513–525.
Bradshaw P (2011) Growing Up in Scotland: Changes in Child Cognitive Ability in the Pre-School

Years. Edinburgh: Scottish Government.
Bradshaw P and Tipping S (2010) Growing Up in Scotland: Children’s Social, Emotional and

Behavioural Characteristics at Entry to Primary School. Edinburgh: Scottish Government.
Bradshaw P, Marryat L, Corbett J, et al. (2009) Growing Up in Scotland Sweep 4: 2008–2009 –

User Guide. Edinburgh: Scottish Government.
Bradshaw P, Marryat L, Mabelis J, et al. (2010) Growing Up in Scotland Sweep 5: 2009–2010

– User Guide. Edinburgh: Scottish Government.
Bromley C (2009) Growing Up in Scotland: The Impact of Children’s Early Activities on Cognitive

Development. Edinburgh: Scottish Government.
Chambers R (1989) Editorial introduction: Vulnerability, coping and policy. IDS Bulletin 20: 1–7.
Chanfreau J and Burchardt T (2008) Equivalence Scales: Rationales, Uses and Assumptions.

Edinburgh: Scottish Government.
Cohn LD (1991) Sex differences in the course of personality development: A meta-analysis.

Psychological Bulletin 109: 252–266.
Conger RD, Conger KJ and Martin MJ (2010) Socioeconomic status, family processes, and indi-

vidual development. Journal of Marriage and the Family 72(3): 685–704.
Corbett J, Marryat L and Bradshaw P (2005) Growing Up in Scotland Sweep 1 User Guide. UK

data archive study number 5760. Edinburgh: Scottish Centre for Social Research.
Corbett J, Marryat L and Bradshaw P (2006) Growing Up in Scotland Sweep 2 User Guide. UK

data archive study number 5760. Edinburgh: Scottish Centre for Social Research.
Corbett J, Marryat L, Bradshaw P, et al. (2007) Growing Up in Scotland Sweep 3 User Guide. UK

data archive study number 5760. Edinburgh: Scottish Centre for Social Research.
DfE (2012) Measuring Child Poverty: A Consultation on Better Measures of Child Poverty.

London: HM Stationery Office.
Fouarge D and Layte R (2005) Welfare regimes and poverty dynamics: The duration and recur-

rence of poverty spells in Europe. Journal of Social Policy 34: 407–426.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

Treanor 693

Goodman R (1997) The Strengths and Difficulties Questionnaire: A research note. Journal of
Child Psychology and Psychiatry 38: 581–586.

Green M (2007) Voices of People Experiencing Poverty in Scotland. York: Joseph Rowntree
Foundation and the Poverty Alliance.

Hansen K, Joshi H and Dex S (2010) Children of the 21st Century. Bristol: Policy Press.
Harris J, Treanor MC and Sharma N (2009) Below the Breadline. London: Barnardo’s.
Holscher P (2008) Children and young people’s experiences of poverty and social exclusion. In:

Ridge T and Wright S (eds) Understanding Inequality, Poverty and Wealth. Bristol: Policy
Press, 181–203.

Hu L and Bentler PM (1999) Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling: A Multi-
disciplinary Journal 6: 1–55.

Jenkins SP, Rigg JA, Devicienti F, et al. (2001) The Dynamics of Poverty in Britain. Leeds:
Corporate Document Services for the Department for Work and Pensions.

Jenkinson C, Layte R, Jenkinson D, et al. (1997) A shorter form health survey: Can the SF-12 rep-
licate results from the SF-36 in longitudinal studies? Journal of Public Health 19: 179–186.

Kiernan KE and Huerta MC (2008) Economic deprivation, maternal depression, parenting and
children’s cognitive and emotional development in early childhood. British Journal of
Sociology 59: 783–806.

Kiernan KE and Mensah FK (2009) Poverty, maternal depression, family status and children’s
cognitive and behavioural development in early childhood: A longitudinal study. Journal of
Social Policy 38: 569–588.

King K and Gurian M (2006) Teaching to the minds of boys. Educational Leadership 64: 56–58.
Lister R (2004) Poverty. Cambridge: Polity.
Magadi M and Middleton S (2007) Severe Child Poverty in the UK. London: Save the Children.
Melhuish EC, Phan MB, Sylva K, et al. (2008) Effects of the home learning environment and pre-

school center experience upon literacy and numeracy development in early primary school.
Journal of Social Issues 64: 95–114.

Merton RK (1968) Social Theory and Social Structure. New York: Free Press/London: Collier-
Macmillan.

Ragnarsdóttir BH, Bernburg JG and Ólafsdóttir S (2013) The global financial crisis and individual
distress: The role of subjective comparisons after the collapse of the Icelandic economy.
Sociology 47: 755–775.

Ridge T (2002) Childhood Poverty and Social Exclusion: From a Child’s Perspective. Bristol:
Policy Press.

Ridge T (2009) Living with Poverty: A Review of the Literature on Children’s and Families’
Experiences of Poverty. London: Department for Work and Pensions.

Runciman WG (1966) Relative Deprivation and Social Justice: A Study of Attitudes to Social
Inequality in Twentieth-Century England. London: Routledge & Kegan Paul.

Schoon I, Hope S, Ross A, et al. (2010) Family hardship and children’s development: The early
years. Longitudinal and Life Course Studies 1: 209–222.

Schoon I, Jones E, Cheng H, et al. (2012) Family hardship, family instability, and cognitive devel-
opment. Journal of Epidemiology & Community Health 66: 716–722.

Shildrick T, MacDonald R, Webster C, et al. (2013) Poverty and Insecurity [Electronic Resource]:
Life in Low-Pay, No-Pay Britain. Bristol: Policy Press.

Smith HJ, Pettigrew TF, Pippin GM, et al. (2012) Relative deprivation: A theoretical and meta-
analytic review. Personality & Social Psychology Review 16: 203–232.

Starrin B, Aslund C and Nilsson KW (2009) Financial stress, shaming experiences and psycho-
social ill-health: Studies into the Finances–Shame Model. Social Indicators Research 91:
283–298.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

694 Sociology 50(4)

Statacorp (2013) Stata 13 Structural Equation Modelling Reference Manual. College Station, TX:
Stata Press.

Stewart T (2009) Counting on Credit. London: Barnardos.
Thomas C, Benzeval M and Stansfeld SA (2005) Employment transitions and mental health: An

analysis from the British Household Panel Survey. Journal of Epidemiology and Community
Health 59: 243–249.

Townsend P (1979) Poverty in the United Kingdom: A Survey of Household Resources and
Standards of Living. Harmondsworth: Penguin.

Ware JE Jr, Kosinski M and Keller SD (1996) A 12-item short-form health survey: Construction of
scales and preliminary tests of reliability and validity. Medical Care 34: 220–233.

Whelan CT and Maitre B (2005) Vulnerability and multiple deprivation perspectives on economic
exclusion in Europe: A latent class analysis. European Societies 7: 423–450.

Whelan CT and Maitre B (2008) Social class variation in risk: A comparative analysis of the
dynamics of economic vulnerability. British Journal of Sociology 59: 637–659.

Whitham G (2012) Child Poverty in 2012: It Shouldn’t Happen Here. Manchester: Save the
Children.

Yeung WJ, Linver MR and Brooks-Gunn J (2002) How money matters for young children’s devel-
opment: Parental investment and family processes. Child Development 73(6): 1861–1879.

Zick CD, Bryant WK and Osterbacka E (2001) Mothers’ employment, parental involvement, and
the implications for intermediate child outcomes. Social Science Research 30: 25–49.

Morag Treanor is a Nuffield-funded Q-Step lecturer in quantitative social policy at the University
of Edinburgh. Her current research uses birth cohort data, mainly the Growing Up in Scotland
study, to explore the impacts of longitudinal poverty and persistently low/high incomes on chil-
dren’s developmental outcomes. Her research is concerned with the impact of family, peer and
social relations on child well-being. She is an experienced quantitative and qualitative researcher
in relation to socioeconomic inequalities and vulnerable populations. She is presently conducting
longitudinal qualitative research with families affected by welfare reform with the Child Poverty
Action Group in Scotland and is a member of their Expert Advisory Group.

Date submitted January 2014
Date accepted January 2015

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

http://soc.sagepub.com/

University of Utah

U. S. Economic Aid and Political Repression: An Empirical Evaluation of U. S. Foreign Policy
Author(s): Patrick M. Regan
Source: Political Research Quarterly, Vol. 48, No. 3 (Sep., 1995), pp. 613-628
Published by: Sage Publications, Inc. on behalf of the University of Utah
Stable URL: http://www.jstor.org/stable/449005

Accessed: 13/05/2013 16:35

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Sage Publications, Inc. and University of Utah are collaborating with JSTOR to digitize, preserve and extend
access to Political Research Quarterly.

http://www.jstor.org

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=sage

http://www.jstor.org/action/showPublisher?publisherCode=utah

http://www.jstor.org/stable/449005?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp

U.S. Economic Aid and Political
Repression: An Empirical
Evaluation of U.S. Foreign Policy

PATRICK M. REGAN, UNIVERSITY OF CANTERBURY, CHRISTCHURCH, NEW ZEALAND

The U.S. Congress has mandated that foreign aid be used in a manner that
distances the U.S. from regimes which consistently violate the human
rights of their populations, and promotes more acceptable human rights
records in recipient countries. There has been considerable scholarly
attention devoted to the first of these congressional mandates, though as
yet little effort has been made to evaluate the effectiveness of U.S. foreign
aid programs in actually changing human rights behavior This essay is a
first attempt at evaluating the impact of changes in economic assistance on
changes in the amount of political abuse perpetrated by those on the
receiving end of the assistance programs. Although others have shown that
Carter and Reagan distributed their respective aid programs differently, the
findings presented below demonstrate that economic aid has no
discernable effect on the human rights records of the recipients; this result
holds across both the Carter and Reagan administrations.

A somewhat recent flurry of scholarship has attempted to identify a
relationship between U.S. foreign assistance and the level of human rights
abuse in the recipient countries. Although there have been some moderately
strong findings, the direction of any causal relationship remains in question,
and in fact the empirical evidence can appear to be quite contradictory (e.g.,
Cingranelli and Pasquarello 1985; Poe 1992; McCormick and Mitchell 1988,
1989; Stohl, Carleton, and Johnson 1984). In an attempt to move this debate
forward-and possibly clarify some of the empirical contradictions-this
article will contribute to the debate by both utilizing new data on human rights
abuses, and by turning the putative causal equation around to examine the
effect of aid disbursements on human rights violations. As such this is the first
attempt to test the hypothesis that changes in foreign economic aid can affect
changes in the level of political abuse in recipient countries. This question is
central to the empirical evaluation of U.S. foreign aid policy, the effectiveness
of which has heretofore been largely evaluated by impressionistic means.
Ultimately, this analysis will conclude that although there is evidence that the

613

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

Political Research Quarterly

differing human rights policies of the Carter and Reagan administrations did
result in differences in aid flows between administrations, the aid itself had
no discernable impact on changes in the human rights practices of the
recipient countries. When viewed in conjunction with earlier studies this
finding suggests that aid might best be conceived of as an indicator of a
broader underlying relationship, in essence, an indicator of a political message
being sent by the administration. It is this message that has potentially causal
linkages to the use of political repression.

After briefly presenting the theoretical foundations and policy
prescriptions that have linked aid to levels of political repression, I will test
a hypothesis relating foreign economic aid disbursements to the level of
political abuse in recipient countries. As such this research portends to move
us one step forward in the search for an understanding of the interaction
between these two foreign policy issues, albeit a short one, The different
emphases in the policies toward human rights articulated by the Carter and
Reagan administrations provide a useful “experiment” in the test of
hypotheses.

POLICY PRESCRIPTIONS AND THE AID-REPRESSION LINK

That the U.S. Congress believes it has some influence over those who carry
out repressive policies can be inferred by the restrictive legislation that has
been enacted over the past fifteen years. Not only has Congress required that
aid be denied to consistent violators of human rights, but also that the
administration certify progress toward the alleviation of abuses in specified
countries before those countries can receive additional funding. Much has
been made of the Carter administration’s emphasis on human rights as a
beacon for foreign policy decisions, as well as President Reagan’s apparent
disdain for human rights issues. In both cases, with the urging of Congress,
foreign aid was used as a carrot to advance the objectives of the particular
administration. Congress and the two administrations under scrutiny have
made it clear that they believe there is a connection-either empirical or
normative-between U.S. foreign aid and the human rights practices of the
recipients of such assistance.

Even though the ability of Congress to alter dramatically annual aid
decisions is in question, from the congressional perspective the objective in
using human rights criteria for the distribution of foreign assistance has a
twofold purpose: (a) to promote practices that are consistent with
international standards of human rights, and (b) to distance the United States
from those regimes that are consistent violators of the rights of their citizens.
Each of these criteria suggests different underlying assumptions. In the first
instance, that changes in aid disbursements will cause changes in the behavior

614

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

U.S. Economic Aid and Political Repression

of the recipient governments; in the second, that aid should be denied to the
most persistent abusers. The intentions of Congress were spelled out in three
separate pieces of legislation. In each case Congress stipulated that no aid shall
be granted to those countries deemed consistent violators of human rights.
The language used in this legislation is specific in terms of both the course
of action to be taken, and the offenses deemed particularly inappropriate. Sec-
tion 116 of the Foreign Assistance Act of 1961 reads as follows:

No assistance may be provided under this part to the government of any
country which engages in a consistent pattern of gross violations of inter-
nationally recognized human rights, including torture or cruel, inhuman,
or degrading treatment or punishment, prolonged detention without
charges, causing the disappearance of persons by the abduction and
clandestine detention of those persons, or other flagrant denial of the
right to life, liberty, and the security of person …. (House Committee
on Foreign Affairs 1979: 1)
The Carter and the Reagan administrations brought different perspectives

to bear on the issue of human rights. President Carter was widely understood
to have viewed the human rights record of a government as a central element
in the conduct of bilateral relations. Not only did his administration publicly
profess to be guided by human rights concerns, but in a number of instances
he denied foreign aid to countries exhibiting particularly brutal human rights
practices. That the administration thought that foreign assistance could be
used to temper the human rights abuses of the potential recipients is evident
in statements made by administration officials. Warren Christopher, Deputy
Secretary of State under President Carter, testified before the House Committee
on Foreign Affairs (ibid.: 8) that “our foreign assistance programs are an essen-
tial tool in promoting a broad category of internationally recognized human
rights’

This practice of manipulating aid disbursements in an effort to directly
mediate the political abuses by recipient countries was viewed by the Reagan
administration as a “negative or reactive” approach to addressing issues of human
rights (House Committee on Foreign Affairs 1983: 2). While acknowledging that
the “reactive track” has merit as a foreign policy tool, the Reagan administration
argued that a second approach was necessary. This second track involved the use
of foreign aid to promote democracy within the recipient countries. It was assert-
ed that “democracies have the best human rights records, for an obvious reason:
When people can choose their government and dismiss it, that government is
less likely to abuse their human rights ..

”
(ibid.: 2). The Reagan administration,

therefore, embarked on a policy of using foreign aid as a tool to foster the de-
velopment of democratic movements, and through these fledgling democracies,
to decrease the level of political abuses.

615

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

Political Research Quarterly

Within both the congressional and the administrative sectors of the U.S.
government there exists the perception of a causal connection between the
amount of aid disbursed to a country and that country’s behavior with respect
to human rights abuse. Whether foreign aid has an impact on the level of
political repression in those countries on the receiving end of U.S. largesse is
an empirical question. But before attempting to answer it, let us briefly review
the literature that attempts to evaluate this relationship between U.S. foreign
aid and political repression.

THEORETICAL ARGUMENTS LINKING AID TO REPRESSION

There are two main themes of research into the relationship between foreign
assistance and human rights abuse. One argues that the human rights practices
of countries affects the amount of foreign aid they receive; the other that foreign
aid can affect the level of abuse carried out by the elite in the recipient countries.
Given that there is considerably more research focused on the former of these
two questions (eg., Cingranelli and Pasquarello 1985; Stohl, Carleton, and
Johnson 1984; McCormick and Mitchell 1988, 1989; Poe 1992), and that
sufficiently evaluating US. foreign policy requires some understanding of the
second of these two questions, I will direct my efforts toward the issue of whether
foreign aid influences the repressive practices of the recipients. But before taking
that step it is important to review what we know about the relationship between
societal attributes and political repression.

Our knowledge in this area is quite broad and suggests trends across a
wide spectrum. At the outset of this discussion, however, we should be clear
that much of what we know identifies characteristics associated with levels
of repression, not changes in the levels observed in any one year For instance,
a strong and consistent finding is that democratic regimes tend to have lower
levels of repression (e.g., Henderson 1991; Poe and Tate 1994), but what we
do not know is that a change to or from a democracy is associated with
changes in political abuse. If there is a causal relationship it could go either
way. Likewise, much of the same argument can be forwarded regarding the role
of ideology (Pion-Berlin 1988; McCormick and Mitchell 1988), though some
evidence is more contradictory (Poe and Tate 1994). Besides the role of
ideology and regime type, population pressures have been associated with
increased levels of repression (Henderson 1993), as has the amount of
resources devoted to the military sector (Davenport 1992). But simply because
this research has addressed issues of levels of repression does not mean that
this knowledge is wasted when searching for the causes of changes in
repression. We can productively use what we know about factors associated
with levels of repression to control for the relationship between changes in
economic aid and changes in repression.

616

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

U.S. Economic Aid and Political Repression

Notwithstanding previous efforts to identify patterns in the relationship
between aid allocations and the human rights records of the recipients, we still
know very little about the effectiveness of foreign aid as a tool to manipulate
human rights practices. Lars Schoultz (1981a), in his analysis of human rights
and U.S. policy in Latin America, demonstrates that economic aid can, and
has been used to support repressive governments, and to assist in the over-
throw of nonrepressive regimes in favor of regimes more disposed to repress
their citizens. Arguing that it is the tendency of the U.S. government to support
those regimes which view Latin American political crises as a threat to
hemispheric stability, and therefore U.S. security, Schoultz suggests that the
United States uses aid to suppress attacks on the status quo. As evidence he
shows that ” . . . in the 1970s, aid was forthcoming for the Uruguayan police
and not the Tupamoaros, for the Nicaraguan National Guard and not the San-
dinistas, for the Argentine military and not the Montoneros” (1981a: 168).

Foreign aid can be used either directly or indirectly to support a repres-
sive government. Training and equipment for police and paramilitary units are
the most direct forms of support. The U.S. had trained and equipped the
South Vietnamese National Police, the Brazilian federal police, the Nicaraguan
National Guard, and more recently a number of police and military units from
Central American countries. In each case, whether the aid program fell under
International Military Education and Training (IMET) or the disbanded Office
of Public Safety (OPS), the countries in question suffered at the hands of their
repressive leadership. The argument is that aid granted by the United States
was used to facilitate the repressive policies of the recipient governments.

Indirect support of repression can be the result of seemingly benevolent
aid programs. Economic assistance and food aid do have unmistakable politi-
cal motivations behind humanitarian-based decision making. The subsequent
cut-off of economic aid to a government deemed antagonistic to the United
States can force that government to divert its limited resources from other so-
cial needs in order to fill the gap created by the shortfall. This could be as true
for food aid as it would be for health and medical assistance, or direct cash
grants to a government. The social unrest that can result from the aid-related
deprivations, could precipitate the downfall of a non-repressive regime, with
its replacement by a regime more compelled to restore order through the vio-
lent suppression of dissent. The resumption of economic assistance to the
successor government could serve to reinforce its policies of suppression. It
is in this manner that U.S. assistance has been used to help bring about the
creation of repressive regimes. Citing examples of Brazil, 1964; Chile, 1973;
Peru, 1960s; and Bolivia, 1970-71, Schoultz argues that well-timed cut-offs, or
disbursements, of aid have had the effect of helping to usher out “threatening”

617

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

Political Research Quarterly

governments, and easing in their successors. In each case the new government
was much more repressive than its predecessor According to Schoultz, “Eco-
nomic aid can be used to create a repressive government only in a negative
sense: by withdrawing aid from a relatively nonrepressive government in
order to undermine its control of political power” (ibid: 169). Once control
has changed hands, and power resides in a government more acceptable to
the United States, the aid spigot is turned back on.

The argument here, of course, is not that foreign aid is the sole cause of
either increases in repression or the overthrow of a non-repressive regime and
its replacement with a brutal leadership, but rather that the use of economic
assistance can be a contributing factor in either of these two outcomes. In
some instances this may be an unintended consequence of U.S. foreign policy;
in others a calculated risk. At various times economic or geopolitical factors
may dominate human rights issues. Whether intended or not-and that may
be an equally debatable issue-U.S. economic aid has been positively associat-
ed with increases in the levels of repression in the recipient countries, at least
in the Latin American region.

Another perspective from which to address this question of the role of
U.S. foreign aid in promoting the adherence to acceptable standards of human
rights is through an analysis of what the Reagan administration termed the
“positive approach” to human rights policy: using foreign aid to encourage
movements toward democratic reform. Muller (1985), however, finds that U.S.
economic and military aid is associated with the destabilization of democratic
governments, not their cultivation. His argument stems from the inherently
political nature of U.S. aid. There were three basic principles that guided U.S.
foreign aid policy: (a) Cold War concerns and the furtherance of U.S. national
security interests; (b) promoting economic development; and (c) fostering
democratic movements. However, these objectives often conflict with each
other and when they do, national security concerns take precedence. The
result, according to Muller, was a destabilization of democratically constituted
governments in exchange for more U.S. friendly authoritarian regimes. His
findings indicate that military aid is most closely associated with the destabili-
zation of democratic regimes, while economic aid has only a minimal impact
on the breakdown of democratic movements.

From the preceding summary it should be evident that there is both a the-
oretical and a political reason to advance the hypothesis that foreign aid can
be used to influence the behavior of the recipient governments with regard to
their human rights practices. The U.S. Congress in general seems to hold to
the belief that if this relationship is not empirically correct, then it is at least
a normatively justifiable position. The following analysis subjects this hypo-
thesis to empirical scrutiny in an effort to evaluate the effectiveness of U.S.

618

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

U.S. Economic Aid and Political Repression

foreign policy-at least as it has been articulated in congressional legislation.
After a brief discussion of data and measurement issues, both bivariate and
multivariate models will be tested.

MEASURING POLITICAL REPRESSION

The methodological problems inherent in any attempt to understand the
causes or correlates of political repression have been well documented (see
Goldstein 1986; Banks 1986; Stohl et al. 1984; McNitt 1986). The problems
generally fall into one of two categories: (a) weak sources of data, and (b) too
few cases to permit time-series analysis. Until there is a change in the manner
in which human rights abuses are reported and recorded, there will continue
to be problems associated with the validity and reliability of any attempts at
measurement. A few researchers, however, have attempted systematic meas-
urement (e.g., Stohl, Carleton, and Johnston 1984; Carleton and Stohl 1985;
Cingranelli and Pasquarello 1985; McCormick and Mitchell 1989; Schoultz
1981b; Poe and Tate 1994). This analysis contributes to the efforts to develop
a reliable and valid measure of the levels of political repression by constructing
an index composed of five indicators; these indicators are based on an opera-
tional definition framed in terms of issues of “integrity of person” (see Poe and
Tate 1994). The cases included in this study involve a stratified sample of 32
countries from Latin America and Asia derived from a population of 54 coun-
tries. Those excluded countries generally consist of small island nations of the
Caribbean and the Pacific, countries for which no human rights data was avail-
able (e.g., North Korea, Vietnam, Cambodia), or countries occupied by a for-
eign power (Afghanistan).’ The temporal dimension covers the period
1977-88, and although this time frame is still insufficient to permit reliable
longitudinal analysis, it does permit us to test adequately for the effects of
changes in economic aid on changes in political repression by pooling the
data.2

The 32 developing countries used in this analysis technically constitute
neither a random selection process nor the total population of countries. They
do, however, represent a mix of those countries in Latin America and Asia that
receive most of the foreign aid coming from the United States, along with those
that received little or no aid during the period under study. For example, the
countries included in this stratified sample account for 98 percent of the for-
eign aid given in the Asian region in 1984, and nearly three of every four

1 See appendix for a complete list of the countries included.
2 There are, of course, other data sets available that are in use for this type of analysis.
For a discussion of the benefits and liabilities of various other options see Henderson
(1991) and Poe and Tate (1994).

619

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

Political Research Quarterly

dollars of aid allocated to Latin America. On the other hand, this sample also
includes a number of countries that received no economic aid from the United
States, or have had their aid allotments terminated during the temporal period
of this study. For instance, Nicaragua, Cuba, South Korea, and Chile fall into
this latter category.

In an effort to develop a reliable measure of the level of political repression
for each country and year the annual reports of Amnesty International were
subjected to content analysis. In many ways this was an elaboration of the
procedures used by Cingranelli and Pasquarello (1985). Five categories of
repression-disappearances, torture, arbitrary arrests, political prisoners, and
political killings-were coded on a four-point scale (0 through 3). This proce-
dure yields an ordinal scale on the extent of the reported violation on each
of five forms of political repression, with each level assigned a numerical value
of 0, 1, 2, or 3, respectively. The scores across all five indicators were summed
to form a political repression score (POLREP) for each country/year.3 The
maximum possible score for each country/year was 15. Inter-coder tests were
performed to check the reliability of this procedure, using both different
coders and a recoding by the principle investigator after the original coding
was performed. These inter-coder tests achieved a reliability of .95 against the
summed political repression score, and a mean inter-coder score of .90 against
the individual indicators. Descriptively, there are 328 cases, with a mean of
4.77 and a standard deviation of 2.79; the maximum recorded value for any
one country/year was 14 and the minimum was 0. The amount of variation
in the political repression index (compared to other measures that employ a
five-point scale-e.g., Carleton and Stohl 1987) permits much more dis-
criminating analyses; this is particularly important when trying to identify the
process behind changes in political abuse. For comparison’s sake, the bivariate
correlation of the data used in this analysis with the Stohl et al. data is approx-
imately .60; the correlation with U.S. Department of State data generated in
conjunction with my Amnesty data is .75.

TESTING PROCEDURES AND DATA SOURCES

As a first step to test the hypothesis that changes in foreign aid disbursements
affect changes in political abuse, categorical analysis was employed to look for

3 Two separate data sets exist, each based on the content analysis of the reporting from
either Amnesty or the Department of State. The correlation between the two political
repression scores is .75. The results presented below will reflect only the use of repres-
sion scores derived from the Amnesty International Reports. This is due largely to the
confusing nature of presenting two complete sets of findings, but also reflects the general
assumption that the U.S. Department of State reports are strongly biased in favor of ad-
ministration policy Coding sheets and coding rules can be obtained from the author.

620

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

U.S. Economic Aid and Political Repression

an association between first differences on both variables. The categorical
analysis allows us to examine in a fairly straightforward and uncluttered man-
ner the distribution of aid relative to the levels of abuse in the recipient coun-
tries. This is, in effect, testing against the null hypothesis that there is no
empirical relationship between changes in the distribution of aid and subse-
quent changes in the behavior of the recipient countries. I then take the ana-
lytic process one step further, embedding the hypothesized aid-repression
relationship into a broader, pool time-series, model incorporating various fac-
tors generally associated with the adherence to acceptable human rights stan-
dards. The roots of this broader model can be found in research by Poe
(1992), McCormick and Mitchell (1988), Davenport (1992), and Henderson
(1982, 1991) to name but a few, and consists of controls for the size of the
population, the degree of democracy, the resources devoted to the military,
and the level of economic development.

Data on economic assistance was taken from The Annual Report of the
Chairman of the Development Coordination Committee, Statistical Annex 2: US.
Foreign Assistance. The lagged difference between the current allocation and the
previous years’ allocation is the indicator. Lagging the change in foreign aid by
one year should capture the causal effect on human rights policies; lagging
more than one year would effectively move the “carrot or the stick” too far
away from the expected behavior.4 Data on military personnel, and popula-
tion were taken from the Correlates of War Material Capabilities data set; data
on democracy was taken from the Polity II data. As an indicator of economic
development the per capita energy consumption was used (Henderson 1991);
data was taken from the Correlates of War Material Capabilities data set.

FINDINGS AND DISCUSSION

When testing for the ability of economic aid to manipulate human rights poli-
cies in the recipient countries an interesting result seems to emerge. If we were
to take congressional legislation as a guide, we would expect previous years

A reasonable research strategy might suggest that the length of the lag should be an
empirical question. Two factors mitigated against this strategy: (1) that moving much fur-
ther than one year from the observed effect makes the causal inferences more difficult,
and (2) each lag “cost,’ one year of data, with all of the data coming from the Carter years.
Removing Carter’s presidency from the analysis eliminates any comparative inferences
that might be drawn regarding the impact of differing policy emphases. In spite of these
“costs)’ lags of two and three years were performed with the results showing a loss in
explanatory power of the model. For instance, the coefficient associated with a two-year
lag in economic aid was a mere .0002, with a standard error of .001. Given this model
specification there appears to be little evidence to suggest that in search of the causal
affects we must move back further in time.

621

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

Political Research Quarterly

changes in economic aid to be associated with current year changes in politi-
cal repression in a manner somewhat consistent with the carrot-and-stick
analogy outlined by administration spokespersons. Presumably a decrease in
economic aid last year would compel the recipient to bring human rights prac-
tices more in line with international standards. An increase in economic aid,
likewise, would reward behavior and encourage greater compliance.

By recoding the repression data into three categories of increases,
decreases, or no change.5 and economic aid data into categories reflecting in-
creases, decreases, and no change,6 the frequency with which increases/
decreases in economic aid are associated with increases/decreases in political
repression can be assessed. As can be seen in Table 1, there appears to be no
clear relationship between changes in the distribution of economic aid and
the subsequent change in the human rights behavior of the recipient coun-
tries. What we would expect to find under conditions where changes in eco-
nomic aid has a strong effect on repression is that the cells along the diagonal
would be full, while the off-diagonal cells would be empty. This is clearly not
what we find. When disaggregating the data into the Carter and Reagan eras
there does not appear to be any indication that economic aid has a differing
impact across administrations, possibly confirming that aid itself does not
have much of an impact on the human rights practices of the recipients.7
The lack of any discernible difference between administrations is evident in
spite of the fact that there are differences in the patterns of aid distributed by
Reagan and Carter For example, in 11 percent of the cases the Carter adminis-
tration gave no economic aid, while Reagan’s administration gave no economic
aid in nearly 25 percent of the cases. At the upper bound, however, Reagan
distributed economic aid in sums of over $100 million in 21 percent of the
cases, though Carter did so in only 10 percent. The maximum amount of aid
that Carter gave in any one year was $351 million; for Reagan that figure was
$383 million. The mean distributions were $42 million for Carter and $57
million during Reagan’s years; standard deviations were $65 and $81 million,
respectively.

5 The “no change” category spanned the range -1 to 1 in an attempt to reflect both the
potential uncertainty in the coding procedures and the potential inability of the U.S.
government to discriminate small changes in the human rights practices of countries.

6 In this case the “no change” category reflects a narrow band for which changes in aid
could be the result of normal distributional fluctuations, and therefore conveys neither
an implicit or explicit message regarding acceptable human rights behavior.

7 These findings are not reported in the text, but the same bivariate tests, disaggregated
by administrations, reveal no systematic relationship between aid disbursements and
repression.

622

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

U.S. Economic Aid and Political Repression

Table 1
LAGGED CHANGES IN ECONOMIC AID AND CHANGES IN POLITICAL REPRESSION,
1979-88

Changes in
Repression Changes in Economic Aid (Millions U.S.$)*

Substantial Substantial
Cell Totals Decline Little or No Increase
Column Pct. in Aid Change in Aid

Decreases in
Repression 9 25 19 53

17 16.6 26 19.1
No Change 28 93 37 158

52.8 61.1 50.7 57.0
Increases in
Repression 16 33 17 66

30.2 21.9 23.3 23.8

53 151 73 277
19.1 54.5 26.3 100.0

Chi-Square – 4.64 p – .30

* Aid categories were broken down into decreases in aid of $5 million or more; increases
of $5 million or more; and a narrow band between reductions and increases of $5 million.

The broader model attempting to account for changes in levels of repres-
sion shows little substantive change in the relationship between fluctuations
in economic aid and the subsequent changes in the level of political repres-
sion in those target countries. Although the degree of statistical confidence we
have in the relationship is within the boundaries accepted by convention, the
substantive effect of changes in aid on subsequent changes in repression is
negligible (Table 2). For example, it would take a $250 million change in eco-
nomic aid to have a one-unit change in the fifteen-point repression score. The
most extreme change in economic aid in any one year under study was $162
million.

One must keep in mind when interpreting Table 2 that although the addi-
tional explanatory variables should be linked to levels of repression, there is
generally not sufficient variation over the time period covered in this analysis
to draw inferences about their respective effects on changes in repression. In
essence, these findings should be interpreted as the effect of changes in eco-
nomic aid on changing levels of repression when controlled for the degree of
democracy, the size of the population, the level of development, and the num-
ber of military personnel. Overall, as a model of factors that contribute to
changes in the level of political repression, this has not been a strong showing,

623

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

Political Research Quarterly

Table 2
MULTIPLE REGRESSION MODEL OF CHANGES IN REPRESSION

Variable Coefficient Standard Error T-Ratio

Population -.14 *E-05 -.10*E-05 -1.39
Mil. Per .67*E-03 .61*E-03 1.10
Democracy .015 .017 .91
Development -.35 .16 -2.19
Lag Econ Aid -.004 .002 -1.94
Standard Error
of Estimate .49
Buse R-Squared .05
DW 1.74
N-234

legend: Population (000s)
Military Personnel (000s)
Democracy (I thru 10; 10 democratic)
Development operationalized as per capita energy consumption (tons of coal

equivalent/capita)
Economic Aid (Millions U.S.$)

with only levels of economic development appearing to bear any systematic
relationship to changes in political repression.

While direct causal inferences cannot be drawn from the above analyses,
it does suggest that the effectiveness of U.S. economic aid as a tool to shape
the human rights policies of the recipient countries has been virtually nil. So
while others have demonstrated that economic aid can be distributed in a
manner that distances the United States from repressive regimes (e.g., Poe
1992; Stohl et al. 1984 ), economic aid apparently has little direct impact on
the behavior of those regimes. This result seems particularly relevant when
viewed in conjunction with many of those earlier findings (Poe 1992; Cingranelli
and Pasquarello 1985; Stohl et al. 1984; McCormick and Mitchell 1988). The
general conclusion might be that the aid itself is only part of the signal that is
sent to the recipients of U.S. assistance and that those searching for mechanisms
with which to manipulate human rights practices should focus on the entire
range of bilateral interactions. Foreign economic aid clearly has been, and can be,
used as a tool to promote U.S. policy objectives, but it appears to be those objec-
tives that determine the relationship between economic aid and levels of human
rights abuse. For example, foreign aid can be used to foster economic develop-
ment, alleviate the suffering from natural disasters, promote bilateral cooperation,
and reward allies for previous compliant behavior But the evidence outlined
above suggests that aid is not very effective at altering the repressive behavior of
recipient states, in spite of the fact that that is set out as one of its objectives.

624

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

U.S. Economic Aid and Political Repression

CONCLUSION AND SUGGESTIONS FOR FURTHER RESEARCH

The overriding conclusion that can be drawn from this analysis is that U.S.
economic aid has had little or no impact on the human rights practices of the
recipient governments, at least as a manipulable variable controlled by the
United States. A plausible inference is that although economic aid itself is not
an effective mechanism for altering the human rights records in recipient
countries, it might serve as an indicator of a diplomatic message that conveys
some sense of American approval or disapproval of current repressive poli-
cies. It is this level of international disrepute that increases the cost of the vio-
lent repression of political dissent in countries that are somewhat dependent
on foreign assistance. That message, presumably, could be conveyed in many
ways, as is evident in the two different messages emanated from the White
House during the Carter and Reagan administrations.

Given the plethora of literature consistently arriving at contradictory con-
clusions regarding the relationship between foreign aid and political repres-
sion, this analysis helps to bring some clarity to this foreign policy question.
This is in part made possible by the development of a new, more discriminat-
ing data set that permits an examination of the effects of aid on changes in
the amount of political repression. As an exercise in foreign policy evaluation
these findings contribute to our understanding of the connection between for-
eign aid and political repression and offer insights for those who make foreign
policy To the policy practitioner it should be clear that it is one thing to legis-
late policy designed to achieve some desired foreign policy goal, and another
thing altogether to implement the policy in a manner that actually achieves
that goal. Changes in aid flows do little to affect the level of political repression
in recipient countries. Enacting legislation is clearly not sufficient to achieve
these goals, implementation is the key.

For the researcher this analysis helps point the way for future exploration.
First, we know far too little about those factors that contribute to changes in
levels of political repression. We, for instance, seem to have a fairly clear pic-
ture of the relationship between democracy and repression (e.g., Henderson
1991), or the putative relationships between various social/political factors
and levels of repression, but these factors do not contribute much to a mul-
tivariate model predicting changes in repression. Some of the work by Gurr
(1968) and Gurr and Lichbach (1986) should be particularly relevant at this
juncture. Presumably changes in the demands placed on the ruling coalition
affect the amount of political violence perpetrated by the state (Gartner and
Regan 1993), but most of the other factors which are statistically associated
with levels of repression vary far too slowly, if at all, to be predictors of short-
term fluctuations in repression. From the perspectives of both the researcher

625

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

Political Research Quarterly

and the practitioner alike, foreign aid has been thought of as one manipulable
variable in the calculus that determines changes in human rights practices.
That does not appear to be the case. Secondly, it may be that the relationship
between foreign aid and political repression is not linear, and that there is an
upper and lower bound at which changes in aid flows can have little impact,
yet in between these limiting factors changing aid levels can influence human
rights policies (Gartner and Regan 1993).

In sum, now that we have data with which to examine the factors that
contribute to changing levels of repression, there is a pressing need to articu-
late theoretical models that might guide future empirical analysis. The broad
stroked approach advanced here is sufficient only to point to overarching
trends in the relationship between economic aid and political repression. In
that regard it suggests that the effect of U.S. foreign aid programs are not con-
sistent with one of the statutory objectives they seek to achieve. But this is far
from the complete story

APPENDIX

The stratified sample used in this analysis consisted of 32 developing coun-
tries in Asia and Latin America. Countries were generally excluded for four
main reasons: (1) because the country was occupied by foreign troops; (2)
lack of data; (3) size; and (4) major power status.

For a few countries, the presence of an occupation force confounded the
inferences that could be drawn: examples are Afghanistan and Cambodia. In
some instances there was no theoretical reason to exclude a country, though
the lack of data prevented inclusion (e.g., North Korea, Laos, Vietnam). Follow-
ing on a well-entrenched, though possibly counter-productive tradition, small
island states were excluded, as were generally those countries with a popula-
tion under one million (for examples, St. Lucia, Micronesia, etc). China and
Japan were also excluded because of their major power status, as well as Japan
being a grantor of foreign aid.

The countries included are: Argentina, Cuba, El Salvador, Mexico, Philip-
pines, Bangladesh, Guatemala, Nepal, South Korea, Bolivia, Guyana,
Nicaragua, Sri Lanka, Brazil, Haiti, Pakistan, Thailand, Burma, Taiwan, Chile,
Honduras, Panama, Uruguay, Colombia, India, Paraguay, Venezuela, Costa Rica,
Indonesia, Peru, Malaysia, Ecuador.

REFERENCES

Banks, David L. 1986. “The Analysis of Human Rights Data over Time'” Human
Rights Quarterly 8: 654-75.

Carelton, David, and Michael Stohl, 1985. “The Role of Human Rights in Foreign
Assistance Policy” American Journal of Political Science 31: 1002-18.

626

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

U.S. Economic Aid and Political Repression

Cingranelli, David L, and Thomas E. Pasquarello. 1985. “Human Rights Practices and
the Distribution of U.S. Foreign Aid to Latin American Countries” American Jour-
nal of Political Science 29: 539-63.

Davenport, Christian A. 1992. “The Dynamics of State Repression: A Cross-National
Time Series Examination of Domestic Political Processes’ Ph.D. dissertation, State
University of New York, Binghamton.

Gartnei Scott, and Patrick M, Regan. 1993. “Brutal Elites: Modelling the Decision to
Violently Repress Opposition” Paper presented at the annual meeting of the
Peace Science Society, Syracuse, New York.

Goldstein, Robert Justin. 1986. “The Limitations of Using Quantitative Data in Study-
ing Human Rights Abuse” Human Rights Quarterly 8 (4): 607-27.

Gurn Ted Robert. 1968. “A Causal Model of Civil Strife: A Comparative Analysis Using
New Indices.’ American Political Science Review 4: 1104-24.

Gurn Ted Robert, and Mark I. Lichbach. 1986. “Forecasting Internal Conflict: Competi-
tive Evaluation of Empirical Theories:’ Comparitive Political Studies 19 (1): 3-38.

Gurn Ted Robert, Keith Jaggers, and Will H. Moore 1989. Polity II Codebook. Bouldei
CO: Center for Comparative Politics, Department of Political Science, University
of Colorado.

Henderson, Conway. 1982. “Military Regimes and Rights in Developing Countries: A
Comparative Perspective” Human Rights Quarterly 4: 110-123.

. 1991. “Conditions Affecting the Use of Political Repression’ Joumal of Conflict
Resolution 35 (1): 120-42.

. 1993. “Population Pressures and Political Repression’ Social Science Quarterly
74: 322-33.

McCormick, James M., and Neil Mitchell. 1988. “Is U.S. Aid Really Linked to Human
Rights in Latin America?” American Journal of Political Science 32: 231-39.

. 1989. “Human Rights and Foreign Assistance An Update” Social Science
Quarterly 70 (4): 969-79.

McNitt, Andrew. 1986. “Measuring Human Rights: Problems and Possibilities. Policy
Studies Joumal 15 (1): 71-83.

Mullei Edward M. 1985. “Dependent Economic Development. Aid Dependence on
the United States, and Democratic Breakdown in the Third World.’ Intemational
Studies Quarterly 29 (4): 445-70.

Pion-Berlin, David. 1988. “The National Security Doctrine, Military Threat Perception,
and the ‘Dirty War’ in Argentina” Comparative Political Studies 21 (3): 382-407.

Poe, Steven C. 1992. “Human Rights and Economic Aid Allocation under Ronald Rea-
gan and Jimmy Cartef American Journal of Political Science 36 (1): 147-67.

Poe, Steven C., and C. Neal Tate 1994. “Repression of Human Rights to Personal In-
tegrity in the 1980s: A Global Analysis” American Political Science Review 88 (4):
853-72.

Schoultz, Lars. 1981a. Human Rights and United States Policy Toward Latin America.
Princeton, NJ: Princeton University Press.

. 1981b. “U.S. Policy Toward Human Rights in Latin America: A Comparative
Analysis of Two Administrations’ In Ved P Nanda, James R. Scarritt, and George
W. Shepherd, eds., Global Human Rights: Public Policies, Comparative Measures,
and NGO Strategies. Bouldei CO: Westview Press.

627

This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

Political Research Quarterly

Stohl, Michael, David Carleton. and Steven E. Johnson. 1984, “Human Rights and
U.S. Foreign Assistance from Nixon to Cartel’ Journal of Peace Research 21:
215-26.

United States Department of State, “The Annual Report of the Chairman of the De-
velopment Coordination Committee, Statistical Annex 2: US. Foreign As-
sistance,’ Washington, DC: U.S. Government Printing Office.

United States House of Representatives Committee on Foreign Affairs, 98th Con-
gress. 1983. “Review of Human Rights Policy” Hearing before the Subcommit-
tee on Human Rights and International Organizations March 3, June 28, and
September 21. Washington, DC: U.S. Government Printing Office.

United States House of Representatives Committee on Foreign Affairs, 96th Con-
gress. 1979. “Human Rights and U.S. Foreign Policy” Hearing before the Sub-
committee on International Organizations, May 2 and 10, June 21, and August
2. Washington, DC: U.S. Government Printing Office.

Received: May 19, 1994
Accepted for Publication: January 17, 1995
Political Research Quarterly, Vol. 48, No. 3 (September 1995), pp. 613-

628

628
This content downloaded from 77.98.244.156 on Mon, 13 May 2013 16:35:17 PM
All use subject to JSTOR Terms and Conditions

http://www.jstor.org/page/info/about/policies/terms.jsp

Article Contents

p. 613
p. 614
p. 615
p. 616
p. 617
p. 618
p. 619
p. 620
p. 621
p. 622
p. 623
p. 624
p. 625
p. 626
p. 627
p. 628

Issue Table of Contents

Political Research Quarterly, Vol. 48, No. 3 (Sep., 1995), pp. 461-696
Front Matter
Public Choice in Education: Markets and the Demand for Quality Education [pp. 461-478]
Directing Electoral Appeals away from the Center: Issue Position and Issue Salience [pp. 479-505]
Moral Judgment, Organizational Incentives and Collective Action: Participation in Abortion Politics [pp. 507-534]
Subjective vs. Objective Discrimination in Government: Adding to the Picture of Barriers to the Advancement of Women [pp. 535-557]
Presidential Vetoes: An Event Count Model [pp. 559-572]
Calling It Quits: Strategic Retirement on the Federal Courts of Appeals, 1893-1991 [pp. 573-597]
The Role of Gender in Descriptive Representation [pp. 599-611]
U. S. Economic Aid and Political Repression: An Empirical Evaluation of U. S. Foreign Policy [pp. 613-628]
Explaining Changes in Tax Incidence in the States [pp. 629-641]
Field Essay
The State of State Politics Research [pp. 643-681]
Back Matter [pp. 683-696]

Turn in your highest-quality paper
Get a qualified writer to help you with

“ critical appraisal ”

Get high-quality paper

NEW! AI matching with writer

Hire a Writer

Client Reviews

4.9

Sitejabber

4.6

Trustpilot

4.8

Our Guarantees

100% Confidentiality

Information about customers is confidential and never disclosed to third parties.

Original Writing

We complete all papers from scratch. You can get a plagiarism report.

Timely Delivery

No missed deadlines – 97% of assignments are completed in time.

Money Back

If you're confident that a writer didn't follow your order details, ask for a refund.

New to Your Trusted Assignment Help Service? Sign up & Save

Calculate the price of your order

Type of paper needed:

Pages:

You will get a personal manager and a discount.

Academic level:

We'll send you the first draft for approval by at

Total price:

$0.00

Power up Your Academic Success with the
Team of Professionals. We’ve Got Your Back.

Power up Your Study Success with Experts We’ve Got Your Back.

Order Now Order Now