Using GSS DATA

Length: Approximately 4 pages
Using the General Social Survey:
1. Choose an attitude or behavior you are interested in. State a two-variable hypothesis (other than the specific ones we discussed in class) based on a short “literature review.” (i.e. cite to some article or newspaper story that might suggest the relationship). That article should not be another survey or poll but a substantive article analyzing WHY there might be a relationship between different social phenomena or demographic characteristics. Your hypothesis should suggest that the independent variable “causes” or “shapes” differences in the dependent variable.
2. Using the SDA website using GSS data, run a bivariate table that shows the relationship between your independent and dependent variables. Describe the relationships between your independent variable (IV) and dependent variable.
3. Select a third control variable that may affect the relationship between the independent and dependent variables and describe a hypothesis about the relationship of the three variables. You can use the full SDA system or use the simpler approach using Quick Tables in SDA (where you can run relationships between your independent and dependent variables, but are limited to control variables of different time periods, gender, race). Your argument will have to make the case that the “control” variable you choose is related in your hypotheses to both your independent variable and dependent variable.
4. Run additional crosstabulations, this time controlling for the third variable from step 3. Study the results to see whether the argument you have developed in step 3 is substantiated or not.
5. Construct table showing the relationship among the three variables. Describe in words what the table says.
Sample bivariate table. Fill in the bold parts of the table with labels for the Independent Variable you choose.

Table 1. Percentage of Americans who hold an attitude, by IV
Attitude/Behavior (DV)
Independent Variable
Total
IV Value 1
IV Value 2
Agree

Disagree

Total (%)
100
100
100

Note: Your attitude/behavior may take on more than two values. If so, add extra rows. If your IV is more than 2 categories, add extra columns. If you have more than 3 values on either, we suggest you collapse them to simplify your discussion. This will involve recoding the variable.
Sample table with control variable. Now add information about your control variable.
Table 2. Percentage controlling for control variable
Attitude (DV)
Independent Variable
IV Value 1
IV Value 2
Control Value 1
Control Value 2
Control Value 1
Control Value 2
Agree

Disagree

Total (%)
100
100
100
100

Directions for Writing Your Paper
Introduce the paper by stating and justifying your focal proposition/hypothesis based upon the article(s) you read in your short “literature review” about how your independent and dependent variables are likely to be related to each other.
Next, discuss how you think this relationship is affected by the presence of the third (“control”) variable. This section should engage the reader in your subject matter and, most importantly, provide your thinking about how these variables relate to each other. In your writeup, please list the one-word codes associated with each of your variables (ie. SEX as determining whether respondents were male or female).
Report the findings from your analysis. Present one or more percentage tables as outlined above (clear, well labeled statistical tables). Do not cut and paste output from the web; you may either use the sample table shells above or create your own. Describe in words what the results mean. The minimum presentation consists of discussing the independent-dependent relationship, both before and after controlling for the third “control” variable you introduce into your argument. You should be able to discuss how the intervening control variable is potentially related to the independent and dependent variables. This can help to establish the plausibility of links in your argument.
Summarize the findings and conclude your paper. How does the third variable help explain or complicate the original relationship? If the results support your original hypothesis, what questions do you have about them or what would you want to look at next? If the results do not support your hypothesis, why do you think that was the result? Is the hypothesis flawed? Did looking at results with the control variable help explain the results? Were the GSS variables you used not up to the task of testing the hypothesis? What other kinds of variables might you want to look at to further test your hypothesis?
Reminder: It is not necessary to have a “correct” hypothesis/ argument in order to write a good paper. Getting negative results on a hypothesis can be more interesting than getting support for it. What is important is that you have a thoughtful hypothesis/argument, one that we have good reason to believe before you look at the data.
Tips for using GSS Data:
Use the GSS Quick Tables program at: http://sda.berkeley.edu/quicktables/quickconfig.do?gss18 or you can use the full tables at http://sda.berkeley.edu/sdaweb/analysis/?dataset=gss18
The GSS asks about many attitudes and behaviors. Try exploring some of the other modules.
To run an analysis under Quick Tables, pick a section such as “Politics and Voting by Background Variables”
Select “variable you want to analyze” (the dependent variable)
Select “breakdown that you want – by” (independent variables)
Click on “Create the Table”
Then rerun the analysis with the same dependent and independent variables
Then “Limit the table to a certain” set of years, race, and gender. Create tables for each of these control variables based on your original dependent and independent variables.
If you want to test variables in the full GSS tables, consult with your professor for help on the best approach.
You can use GSS Explorer to help you find additional variables for that approach
See https://gssdataexplorer.norc.org/variables/vfilter