Data access to the GSS

Share on FacebookShare on Google+Email this to someoneTweet about this on Twitter

Half Sigma made a series of interesting posts about the General Social Survey. I recently noticed that GSS data are available thru a very useful web application. Click that link to a go at it. Below the fold is a brief primer on what I’ve learned from playing around with it.

If you want to jump right in, get started with Row=WORDSUM, Column=GOD.
Also fun: Row=POLVIEWS, Column=CHILDS


Getting started help page

useful “Analysis” options:
* Frequencies or crosstabulation – produces heat-map tables like Half Sigma’s posts
* Comparison of means – useful to compare WORDSUM scores between groups
* Correlation matrix – useful for “pairwise” correlation matrix
* Comparison of correlations – compare correlations when controlling for another variable
* Multiple regression – multiple R

The lower left panel is a topic-sorted index of variables. Pick a variable from the list and click “view” to get a description. Use this panel to build your queries, or type in variables yourself.

useful variables:
* SEX (1=male, 2=female)
* AGE
* RACE RACECOMB (1=white,2=black,3=other)
* WORDSUM – a 10 Q vocabulary test
* EDUC and DEGREE – measures of education
* INCOME – family income
* GOD, IMPCHURH, GODCHNGE
* CHILDS, AGEKDBRN, CHLDIDEL

filtering: to limit an analysis to a subset of the total population
* example: “age(18-50), racecomb(1)” = Whites age 18-50

One Comments

  1. You should use RACECOMB and not RACE. RACE is the old variable. 
     
    Also be aware of the fact that there were “black ovesamples” in the 1980s. Use the black oversamples weighting for any analysis that includes that time period.

a