Hypothesis
Q1. First, listen to/read the NPR story at the following link for the background and answer the questions below.
http://www.npr.org/templates/story/story.php?storyId=130732488
a. Consider the hypothesis “As people grow wealthier, they grow taller.” What are the independent and dependent variables in this hypothesis?
b. The story suggests that the causal mechanism behind how wealth produces height (specifically, what Komlos suggests as the reason why the Dutch are so tall nowadays.) This implies an intervening variable rather than a straightforward linkage. Explain what an intervening variable is and describe how wealth would to taller people.
c. The presence of the intervening variable from part b. implies that, under certain conditions, greater wealth would NOT lead to taller people. What in the theorized causal mechanism would lead to this exception? Under what kind of circumstances might such exceptions be found?
d. As the story observes, heights in U.S. has not increased since 1960s. Can you think of confounding variables that might have contributed to this non-change in US (and change in Europe)? Describe what a confounding variable is and why the variables you suggest might indeed be confounding the results.
Q2. Read the paper “The Effects of Canvassing, Telephone Calls, and Direct Mail on Voter Turnout: A Field Experiment.” The paper is available on Gauchospace. Alternatively, you can find the paper at the following link (you need UCSB VPN):
http://links.jstor.org/sici?sici=0003-0554%28200009%2994%3A3%3C653%3ATEOCTC%3E2.0.CO%3B2-9
a. Green and Gerber start the paper by pointing out that examining voter mobilization via personal contact suffers from the presence of a major confounding variable. Specifically, parties tend to “direct their appeal disproportionately to committed partisans.” (on the first page of the paper) What is the problem that this is likely to introduce? Explain why this is likely to be problematic.
b. What features of research design do Green and Gerber incorporate in attempt to address the problem? Be sure to explain your answer.
c. What evidence do Green and Gerber offer as the reason for believing that the results they show in Table 4 are indeed the consequence of personal contacts? Explain this reasoning without math.
Q3. This requires access to the SPSS data from the class GauchoSpace site, although not necessarily using the data itself.
a. There are five data sets online. Of these, you may choose one of the following three.
a. NES2012 data: This is a condensed version of the survey responses from 2054 Americans obtained in course of the 2012 campaign season. It includes information on their background, political opinions, and voting decisions.
b. March madness: This is the performance of how 192 college basketball teams performed in course of 2004-2006 NCAA basketball seasons (there is an error in the codebook). It includes information on how well they did in the regular season and in the Tournament.
c. Eurobarometer69 data: This is a condensed version of the survey of the citizens of 27 EU countries. It includes their opinions, backgrounds, and their participation in EU elections.
b. Examine the codebooks and go through the data. Then, in a single sentence, state a causal hypothesis that you would like to test using any one of these data sets.
c. Briefly describe the causal mechanism behind your hypothesis.
d. List your independent variables and the possible range of values that they may take on in the data set.
e. What is your dependent variable and what possible values can it take on in the data set?