Demonstrate your ability to apply social science stats to real data and correctl
Demonstrate your ability to apply social science stats to real data and correctly interpret the results in a data memo. To do that:
Pick a cool, interesting topic. Work on something that interests you and that you know a bit about, ideally, a topic from another social science class.
Pick a dataset. Important public datasets include the GSS that we have been using in class and labs, the American National Election Studies (https://electionstudies.org), the Chinese General Social Survey (http://cgss.ruc.edu.cn/English/Home.htm), and the International Social Survey Programme (www.issp.org). They are all good sources for data on attitudes and identities. The US Census Bureau conducts the monthly unemployment survey and other important surveys. See ipums.org for access to those surveys.
Pick an outcome variable. I expect most of you will pick an attitude, for example, racial resentment, support for traditional gender roles, opinions about immigration, satisfaction with standard of living, etc.
Pick predictor variables. Start with a fundamental variable: race, class, or gender. Add other factors. [Chelsea has created a Race4 variable that combines the highly detailed HISPANIC and RACECEN1 variables into a manageable variable. For income use CONINC or CONRINC, but you might also use CLASS (subjective), DEGREE (educational credentials), or SEI10 (an occupational scaling variable scored 0-100)].
Analyze the data.
Start with a two-way tabulation of the outcome by a fundamental variable.
Add other predictors.
Test for interactions.
Conclude if your results imply mediation, interaction, or some other pattern.
Visualize the relationships. Use at least one graph or chart that displays the relationship between two variables. Many of you will choose to show several, including plots of the distribution of outcomes. But one XY chart is required.
Write about relationships in some detail. A relationship between two variables is a pretty abstract thought; narrate a scenario consistent with your results to make them more “real.”
Use the quantitative results. “Getting a masters degree could increase pay by $beta” (plugging your estimate in there where I typed “beta”) or “moving to a step to the political right reduced support for Y by half a point.”
You do NOT have to do a major literature review, but please refer to articles and books you are already familiar with where relevant and cite any articles and books you consulted in preparing your memo.
Include what, where, when, and who for the chart.
Think of explaining the research memo to the audience who are not familiar with this topic. (Why am I into this topic and excited, a lot of explanations)
Use appendix to put graphs, charts, and tables at the end of the essay.
Pick a coold topic and have many variables like 4 to 5.
Length: 7-10 pages (not including figures and tables). Much of that text will come from steps 6 and 7 under “analyze the data” above.
Attached files are sample research memo and SDA guide.