Module 2 Assignment 1 Should You Believe a Statistical Study

Module 2 Assignment 1: Should You Believe a Statistical Study?

We are bombarded by statistical information from a wide variety of sources. Much of the statistical research performed in the world is conducted with a great deal of integrity, validity, and reliability. However, there are many statistical studies that are conducted that are—whether intentionally or unintentionally—plagued by bias.

As a member of society, and as a consumer, it is prudent to develop the skills necessary to critically examine reported statistical claims. Determining whether the report contains bias is a great way to determine the reliability of study results and make a better informed decision.

In this assignment, you will apply the guidelines found in the textbook Unit 5B “Should You Believe a Statistical Study?” in order to critically analyze the content, design, and reported results of a statistical study.

Conduct an Internet search to find a study whose statistical results have been published in the news or any other public forum. Applying the following guidelines, critically analyze the study’s reported content and results.

1.(a) Identify the goal of the statistical study. The goal of Literary Digest’s statistical study was to determine who would win the 1936 Presidential Election.

(b) Identify the population. The population of the study was all those who were registered to vote in the 1936 Presidential election.

(c) Identify the type of study. The type of study that was conducted was a survey.

2.(a) Who conducted the study? Literary Digest conducted the study.

(b) Do you think that there is any intentional bias in the study?

I believe that Literary Digest had no intention of producing incorrect results. I would assume that the magazine would have wanted to publish the correct results of the public’s opinion because it would make the magazine look more competent.

3.(a) Is there selection bias in the sample used in the study?

Yes, there was selection bias in the sample study. Literary Digest chose to poll subscribers of their magazine, automobile owners and people who were found in the phone book. During the great depression of the 1930’s, individuals who were able to afford magazines, telephones and automobiles belonged to the upper middle class, whose members favored the Republican candidate Alf Landon. Because of the lack of randomization in the polling population, the results of the statistical study were biased, and overall produced results incongruent with the results of the Presidential election.

(b) Does the sample contain participation bias?

Yes, the sample does contain participation bias. In any voluntary response study there will be a participation bias, because one hundred percent of those polled will not take the time to participate. Those that do take the time to participate are people who care deeply about the topic at hand. In the case of a presidential election, those that care the most are those that are seeking change. In the 1936 Presidential election, Franklin Roosevelt was the incumbent and Alf Landon was the challenger. Thus, the majority of those who responded to the survey would have wanted Alf Landon, the new comer, to win the election.

4.(a) Are there any problems in defining the variables of interest in the study? No, the variables are clearly defined, because the only two options were the two candidates.

(b) Are there any problems in measuring the variables? No, there was no problem in measuring the variables. They simply needed to count the votes.

5.Are there any confounding variables present in the study?

A confounding variable present in the study was the demographic of those polled. As previously mentioned, Literary Digest did not attempt to poll an entirely random population and instead surveyed those individuals of a certain class. This is a confounding variable that effected the results of the study.

6.Are the results presented fairly?

Yes, the results are presented fairly. They were listed in a straightforward, factual manner.

7.Is the study’s conclusion reasonable? Does it make sense?

Yes the study’s conclusion is reasonable. It would make sense that 57% of those polled would chose to vote for the challenger as opposed to the incumbent because people like change. Considering this was during the great depression, people may have thought that a new president might boost the economy, and provide new job opportunities.

8.Do the results have practical significance?

The results do have practical significance. The results of the survey were 43% to 57%, that is a pretty big gap. If individuals who did not participate in the poll saw the results, it could have influenced their voting—either encouraged them to vote for Roosevelt because the poll showed that he would lose, or assume that because their candidate was going to win already, taking the time to vote for Landon wouldn’t make a difference.

When you reply to two of your classmates, please address at least the following:

1.Do you agree with your classmate’s response to item 2(b)? Please explain.

2.Do you agree with your classmate’s response to item 8? If your classmate wrote about the

1936 Presidential election, what percentage (of projected votes) do you think that a candidate

should have (in a pre-election poll) in order to possess a “significant” advantage over his/her

opponent? For example, do you consider a 51% to 49% poll advantage to carry practical

significance?