Fall, 2006 (Roos, Soc. 311)
Assignment 6: Univariate Frequencies on Final Project Variables
Due: Thursday, December 7th
This assignment will move you a little farther along on your final project. You'll be accessing the General Social Survey online for this assignment, and your final project. For the current assignment:
1) Make sure you have your three variables from the GSS (the ones you used for Assignment 4). If you want to choose new variables, go back to the GSS website (see Ass. 1). But if you do this, make sure you check with one of us to make sure your variables are okay for the final project. As you did in Ass. 1 and Ass. 4, include copies of the "codebook" description for each of your three variables with your writeup.
2) Describe your variables: which is your independent variable, your dependent variable, your test variable (and will your test variable be antecedent or intervening)? State your research hypothesis. So far, this just duplicates what you've already done for Assignment 4.
3) Describe in your writeup how you recoded your three variables (if necessary). We will talk in class about how to do this (and see "Hints" below). For example, if one of your variables is EDUC ("years of school completed"), you will need to recode your 20 education values (where 1 = one year of schooling, 2 = two years of schooling, up to 20 means 8 years or more of college) into approximately three or four recoded groups (i.e., grammar school or less, some high school, high school degree, etc.). In general, to avoid running out of cases you should have no more than two to three categories for each of your independent and test variables (or about eight total columns in your trivariate table). Leave more categories for your dependent variable (probably up to 6 to 8), if possible.
4) Make sure the missing values the GSS assigned are acceptable to you. For example, Don't know (DK) and No answer (NA) responses should in most circumstances be assigned as missing, which would eliminate them from your analyses. In some cases, you may want to code "don't know" responses as a valid code (for example, if a lot of people answer DK). Describe your missing values choices in your writeup (if any).
5) To complete this assignment, you will use the 2004 GSS (or, 2000 or 2002, depending on data availability) to generate univariate frequencies for each of your three variables. You may use combined years to increase your sample size (e.g., to include 2002 and 2004, use "year (2002-2004)" in the selection filter space). If you recode, get frequencies for both the unrecoded and recoded versions of your variables, and turn both in.
Hints for using GSS : from the main GSS site, use the "Frequencies or Crosstabluation Program" on the right.
For univariate frequences: put names of three variables in "row" (if you have trouble, try doing each variable separately)
Don't forget to add the year(s) you are using, in the "selection filter"
A recoding example: "educ(r:1=0-11; 2=12; 3=13-15; 4=16-20)" [this
recodes the 20 educ categories into 4 recoded categories]
Another recoding example: "educ (r:1=0-11 "LT HS"; 2=12 "HS";
3=13-15 "Some Coll"; 4=16-20 "College Grad+")" [this
recodes the 20 educ categories into 4 recoded categories with assigned names)
If you want, you can also click on "Create Variables" at the top left for a program that leads you through the recoding process.
6) Once you do you the analysis, check your work. You should have three basic tables, one for each of your variables (more if you recode variables). Before you get too excited, however, closely peruse your output. Make sure you didn't make any logical errors. Go back to the GSS codebook to compare your unrecoded frequencies with the frequencies in the codebook.
7) Use the data from the computer printout to make typed versions of your univariate tables for each of your recoded variables (through Word, or Excel). Label them sequentially Tables 1 through 3.
8) Make bar charts for each of your three variables (recoded variables only). Label them sequentially Figures 1 through 3. If you know how to use Excel, you can produce nice looking bar charts. Or, use the GSS site and "print the image." We'll talk about this in class, but in the interim play around with inputting your data into Excel to produce various kinds of charts, or use the GSS site. If you use the GSS site to produce your tables, choose either "Bar Chart" or "Pie Chart" for "type of chart."
9) Use these data to write up your results for the recoded variables only. In discussing your results tell us, for example, what percentage of the sample is male, what percentage is female (if sex is one of your variables). This last part of the assignment asks you to interpret your output, and to practice writing about it.
[Note: these frequencies will not yet allow you to test your research hypothesis; rather, it will simply give you percentage distributions for each of your three variables. Your next task (for the final project) will be to actually test your hypothesis, by generating bivariate and trivariate tables.]
HERE IS A LIST OF WHAT YOU NEED TO TURN IN:
1) COPIES OF THE PAGES YOU PRINTED FROM THE GSS CODEBOOK
FOR ALL THREE VARIABLES.
2) GSS-PRODUCED UNIVARIATE FREQUENCIES FOR EACH OF YOUR THREE VARIABLES (BOTH
UNRECODED AND RECODED).
3) EXCEL (OR WORD)-PRODUCED TABLES FOR EACH OF YOUR THREE VARIABLES (IF RECODED,
USE ONLY THE RECODED VERSION).
4) EXCEL (OR GSS)-PRODUCED BAR OR PIE CHARTS FOR EACH OF YOUR THREE VARIABLES (RECODED
ONLY).
5) WRITE IT ALL UP, FOCUSING ESPECIALLY ON INTERPRETING YOUR RESULTS.
Remember, computing is fun! YOUR ASSIGNMENT MUST BE TYPED.