# Two Discussion Responses Needed 100 words each 200 words

Guided Response: Respond to at least two of your classmates by commenting on their posts. In your response, provide your own interpretation of their distribution graph. Note any differences between your classmate’s interpretation and your own.Respond in a substantive manner and provide information or concepts that they may not have considered.

Discussion #1 respond to this post below 100 words #1: PDF of Graph is attached

A lurking variable is data that is left out of a visualization tool even though it has an important effect to the results of the data (Joiner, 1981). The data may have been left out because it is unknown or its influence was thought to be unimportant to the results (Joiner, 1981). These variables are sometimes difficult to find, and careful study of patterns is required to prevent incorrect conclusions (Joiner, 1981).

One can find a lurking variable by plotting the data themselves, placing time order in random areas, viewing the data in a wide variety of plots, and looking for commonalities between the highest and lowest points (Joiner, 1981).

The following chart displays the number of online grocery shoppers by age group. What would you consider to be a lurking variable?

Discussion #2 respond to this post below 100 words: #2 Graph in the discussion post.

Design a study that has a false correlation caused by lurking variable. Watch this week’s video for an example study, and to learn more information about this concept.

False correlations can be found when a third variable is present but not included in the data. A lurking variable is one that can provide an important effect on the relationship between the variables used for the particular study. I also found it interesting that two variables that cannot be distinguished on a response variable is known as a confounder. A lurking variable has the ability to show a strong relationship between the variables but it can also hide the true relationship (Sharpe et al, 2019).

My question for my research is: Is there a correlation between height and educational level attained? I will show that a correlation does exist in the following graph. Please excuse the look of my graph. I had to make it using shapes as I did not have excel available to me at this time.

Height and Level of Education Correlation

I am still having trouble getting my graph to canvas. I will continue to work on this. In the meantime this is the best I could do to show what I am trying to convey.

Ht 6 x

in 5 x

Feet 4 x

3 x

2 x

1 x

0 x

School 0 1 2 3 4 5 6

Level (None) (Pre) (Elem) (H.S). (College)

The observer has to be careful to look for any lurking variables so their final conclusion will be accurate. If a trend exists whether it is linear or non linear it is possible that a lurking variable has not been included but is affecting the variables that were used. When creating an experimental study the researcher should remember to build the experiment so that the risk of having a lurking variable is minimal. For observational studies the researcher should always be aware that lurking variables could cause incorrect relationships between the variables and their relationships (Statology, 2019). It is also good to know the difference between a lurking variable and confounded variables. As we know the lurking variable can change the outcome of the data whereas two variables are considered confounding when the effect of them cannot bedistinguished from one another (Culty, 2012).

I read an article that I found very interesting and funny. It said that Mark Twain made the statement that there were three kinds of lies; lies, damned lies and statistics. Lurking variables can make lies out of statistics. Tyler Vigen was a graduate student that did a search for spurious correlations. Using data and a computer algorithm he came up with results that were pretty funny. He found a 95% correlation between cheese consumption and deaths by getting tangled in your bed sheets. Another example from his experiment was that the divorce rate in Maine has a 99% correlation to per capita margarine consumption (Molnar, 2019). Although these are really off base it shows that computers will give you a correlation only on the data you give it to work with. The correlation between height and level of education shown above may be believed by some because it is put in a graph and looks like a statistical certainty even though it does not make sense. However, we know that there is a lurking variable somewhere that will change the outcome of the study so we will not fall for the absoluteness of the visual given.