Fall, 2007 (Roos, Soc. 311)
Assignment 4: Hypothesis construction, take 2 (due Thursday,
October 25th)
This assignment is designed to give you another chance at Assignment 1. Choose the three variables from the General Social Survey that you hope to use for your final project. You may use the same two variables you used in Assignment 1 (adding on an appropriate "test variable"). As you did for Assignment 1, use the web to access the General Social Survey codebook. Include a copy of the "codebook" description for each of your three variables with your writeup. [You can get this by typing the variable name in the box on the top left of the GSS website and clicking "view."] MAKE SURE ALL YOUR VARIABLES HAVE DATA FOR 2006 (you may choose an earlier year if all your variables aren't available in 2006). As a rule of thumb, you should have at least 1800 cases, and all the variables must be available in the same year. As we discussed in class, you can combine years to increase your sample size (e.g., in the selection filer, include "year (2004-2006"), but all your variables must be available in all the combined years. Check this, it will save you time later!
As in Assignment 1, select an independent and dependent variable and develop a theory and a hypothesis that describes your expectation about how these two variables are related. [Remember, learn from your mistakes in Assignment 1: BE SPECIFIC.] Include a brief statement about how you came up with your theory and hypothesis--why do you have the expectation you do? Then, follow through the research process to conceptualize and operationalize your variables. Recall that you are operationalizing your concepts using the GSS variables.
Now that you're familiar with creating tables, make up hypothetical data, and present them in an original, bivariate table (titled and labelled correctly; take your percentages to the 1/10th decimal place). These data should confirm your hypothesis. Describe these hypothetical results, and then describe the third ("test") variable you have chosen. Operationalize and conceptualize the test variable (this is your reformulation). DO NOT INCLUDE THE TEST VARIABLE IN YOUR TABLE AT THIS POINT; YOUR TABLE SHOULD SHOW ONLY THE HYPOTHETICAL DATA FOR THE ORIGINAL, BIVARIATE RELATIONSHIP. MAKE SURE YOUR HYPOTHETICAL TABLE CONFIRMS YOUR HYPOTHESIS.For this assignment, you should decide how you want to collapse your variables for presentation in your final tables (assuming you need to collapse). Give substantively compelling reasons for why you collapse the variables the way you do. Rule of thumb: for a dependent variable you can have about 6 to 8 categories, but for independent and test variables, you'll probably need no more than 3, unless you have a big sample size (e.g., from combining years).
YOUR PAPERS MUST BE TYPEWRITTEN DOUBLE SPACE (12 font, about 2-4 pages MAX). Your table should be in Word or Excel.