PSC 400: Midterm Exam Exercise
1 Identify the appropriate level of measurement for the following variables: (a) Transparency.org’s corruption index by country (from 0 = most corrupt to 100 = least corrupt) (b) Number of delegates to the Democratic Party’s national convention by state (c) Respondents’ answer to the Gallup Poll question “In general are you satisfied or dissatisfied with the way things are going in the United States at this time?” (Satisfied, Dissatisfied, Unsure) (d) General Social Survey respondents’ confidence in Congress (0 = not applicable; 1 = a great deal; 2 = only some; 3 = hardly any) (e) Answer to National Election Survey (NES) question “For whom did the respondent vote for President” in the 2016 elections (1 = Clinton; 2 = Trump; 3 = Other (Specify)) (f) Presence of death penalty in U.S. states (1 = present; 0 = absent)
Exercise 2 Identify the unit of analysis in each of the following studies and formulate a plausible working hypothesis between the independent and dependent variables in the formal way discussed at length in the videos and in class (a) A study about whether the number of women on a company’s board of directors influences a company’s adoption of maternal leave policies (b) A study investigating whether the presence of a strong primary challenger influences a member of Congress retirement decision (c) An experiment investigating whether participants’ support for a presidential primary candidate is related to candidate performance in a televised debate
Exercise 3 Read the attached transcript from a “Hidden Brain” podcast concerning the relationship between food stamps policy and academic performance. Then answer the following (keep it super-brief; do not write an essay): (a) Given the description of the study in the transcript, what kind of study would you say this is? An experiment? If so, what kind of experiment? An observational study? If so, what kind of observational study? (b) What is the study’s unit of analysis? What are the investigators’ comparing? (Hint: you could say there actually are two unites of analysis here. . . ) (c) According to the article, which variable is the dependent variable and which variable is the independent variable? From the description given in the article, how would you say the investigators operationalize the independent and the dependent variable? (d) Formulate a plausible working hypothesis for their relationship. (e) Does the study meet each of the four “hurdles of causality” discussed in the Canvas video (plausibility; covariation; endogeneity; and control)? Explain briefly which ones are met, which ones are not met, and why. 1 < Examining Links
Between Academic Performance And Food Stamps npr.org /templates/transcript/transcript.php MARY LOUISE KELLY, HOST: Poor families who rely on food stamps often find themselves caught in a familiar cycle. In the days after they receive the benefit each month, there’s plenty of food on the table. But as the weeks tick away, food becomes scarce. Here with some new research on the consequences of this monthly cycle is NPR’s social science correspondent Shankar Vedantam. Good morning. SHANKAR VEDANTAM, BYLINE: Hi, Mary Louise. KELLY: Hi. This new research, I gather, it focuses on the link between food stamps and academic performance. How so? VEDANTAM: That’s right. This work comes from Orgul Ozturk, she’s an economist at the University of South Carolina, along with her colleagues, Chad Cotti, and John Gordanier. They find that children who come from families that are several weeks removed from receiving their food-stamp benefits perform worse on an important math exam. ORGUL OZTURK: We find that when the test date happens to be very far away from the receipt date, students score much lower, significantly lower on the mathematics test. KELLY: It seems like it kind of makes sense, that if you’re hungry you can’t concentrate so you might test lower. How do they know, though, that this is the food-stamp cycle, that that’s what’s responsible? VEDANTAM: Well, it has to do with two quirks in the way the food-stamp program is administered in South Carolina. Recipients receive benefits on the first 10 days of the month. Some families get it on the first, some on the second and so on. Simultaneously, children in grades three through eight also have to take an annual math exam on the second Wednesday of May. Now, the second Wednesday of May falls on different dates each year. So if your family received food stamps on the 10th of April, for example, and the exam falls on the 8th of May, you’re likely to have gone hungry for several days before you took the test. KELLY: And to be clear, people don’t all get their food stamps on the same day. It’s… VEDANTAM: Precisely. 1/2 2 KELLY: …Scattered around the month. VEDANTAM: Exactly. Now, on the other hand, if your family received food stamps on the 2nd of May and the test is on the 8th of May, you’ve probably eaten well in the days before the test. The researchers not only compared kids who’ve eaten well to kids who are hungry, they also compare the math performance of kids who’ve eaten poorly to their own math performance in another year when they’ve eaten well. Ozturk told me that while her study measures the effect of being hungry on one math test, this is a monthly cycle, meaning that the real effects are likely to be much larger. OZTURK: This is happening to these kids every month over the course of nine months. It adds up. What we are measuring is a reduction in performance that month. But they are hungry every single month. The cumulative effect is very significant. KELLY: Shankar, can researchers control whether this is in fact food? It seems like they’re clear on the link to food stamps, but if food stamps are running out maybe the parents are stressed, there’s stress in the household and that’s what kids are reacting to? VEDANTAM: That’s a fair point, Mary Louise. We know that there’s a relationship between food stamps and math performance. We don’t know specifically if it’s about food or just because the family as a whole is stressed and the kids in some ways are reacting to that stress. KELLY: So what’s the implication here? Might it be better to, say, distribute food stamps more frequently? VEDANTAM: So that’s an idea that some people are thinking about, and it’s an intriguing idea. If you can distribute the benefit twice a month, perhaps you can smooth-out the cycle. But it’s the kind of thing that needs to be tested because you could have unintended consequences, and now instead of students being hungry once a month they’re going to end up being hungry twice a month. KELLY: Which is clearly not the direction that you would want to go. VEDANTAM: Precisely. KELLY: That’s NPR’s Shankar Vedantam. Shankar, thank you. VEDANTAM: Thank you, Mary Louise. KELLY: He joins us regularly to talk about social science research. You can follow him on Twitter at @hiddenbrain. You can follow me at @nprkelly, and you can follow this program at @morningedition. Copyright © 2017 NPR. All rights reserved. Visit our website terms of use and permissions pages at www.npr.org for further information. NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record. 2/2 3 Exercise 4 Scenario: You are a consultant for a presidential primary candidate and are working on campaign strategies ahead of the Super Tuesday elections. You are asked to compile a report about whether your candidate should emphasize “bread and butter, kitchen table” issues in her/his message. As part of your research you consider the percentage of workers employed in the manufacturing sector. You use the data to classify states as being of low (below the national average of 8.5), moderate (between the national average and 10 percent), or high (above 10 percent). You call this variable (Rank) and whether the state is located in the South as a Census region (South). State Manufacturing Rank South Alabama 13 high 1 Arkansas 12.8 high 1 California 7.7 low 0 Colorado 5.4 low 0 Maine 8.4 low 0 Massachusetts 6.6 low 0 Minnesota 10.8 high 0 North Carolina 10.4 high 1 Oklahoma 8.0 low 1 Tennessee 11.5 high 0 Texas 7.1 low 1 Utah 8.8 moderate 0 Vermont 9.6 moderate 0 Virginia 6.1 low 1 Using these data to put together your presentation, you do the following: 1. Using pen and paper, build a frequency distribution for the variable Rank, with all of the appropriate frequencies. Notice that this variable has three categories. 2. Using pen and paper, plot the variable Rank with a pie-chart and a bar plot. Add the appropriate labels to each graph. 3. Explain in words understandable to a lay audience what the frequency distribution tables and the graphs suggest about the distribution of the Rank variable. 4. Suppose that, as part of the presentation, you want to address whether regional differences should be taken into account. Voters in Southern states tend to be more value-driven than pocket book driven, and kitchen table issues may be less effective in those states. Build 2×2 table where the dependent variable is whether a state ranks as “high” (assign it a value of 1) or not (assign it a value of 0) on the Rank variable and the independent variable is South. Based on the frequencies that appear in each cell, should you advise your candidate to modify is/her message based on regional differences? 5. Because you may need these data in the future, you also decide to store them in a data set. Using R, enter each variable manually and create a data frame. Display the data frame in the R console. Copy and paste your commands and the output in your Word document. 6. Using R, plot a box plot and a histogram for the variable Manufacturing. Copy and paste your commands and the graphs in your Word document. 7. Again, using R estimate the mean, median, variance, and standard deviation for the variable Manufacturing Again, copy and paste your commands and the output in your Word document. 8. Finally, in case you are asked about this during your presentation to your candidate, provide in writing a substantive interpretation of the median, mean, variance and standard deviation. The exam is due on 3/12/20 at class time (2:00 pm) in hard copy format and, as usual, upload a digital version to Canvas. Please slide a copy of your exam under my office door if I am not in. Because the following week is spring break, we cannot push back the deadline or give any extension. Exams turned in after 2 pm without appropriate documentation will receive a score of zero. 4
