PA 705
Fall 2004

Homework Assignment # 7

Due Nov. 3

Contingency Tables, Measure of Association and ANOVA

I. Please do ORB, problem 2 on p. 420-421.

II. Lab Work

Please turn in your SPSS output with the assignment.

For the first part of this assignment, please use small_city.sav. This is a subset of variables from the SF Citizen 2004 survey I created to make analysis more manageable. You should also have small_city.doc, which is a data dictionary for the file.

  1. State a hypothesis (and the null hypothesis) for a relationship involving two nominal level variables.
     

    1. Run a crosstab procedure, asking for column percentages, chi square and a measure of association appropriate for nominal level variables. Make sure you set up the crosstab with the independent and dependent variables in the correct places (row and column).
       

    2. Write a paragraph interpreting your results.

       

  2. State a hypothesis (and null hypothesis) involving a relationship between two ordinal level variables
     

    1. Run a crosstab asking for column percentages, chi square and a measure of association appropriate for ordinal variables.
       

    2. Write a paragraph interpreting your results.

       

  3. State a hypothesis (and null) involving how a 3rd (control) variable might affect the relationship between the variables you used in question 2. (To make interpretation easier, I suggest using a nominal level variable with few categories as your control variable).
     

    1. Re-run your crosstab, this time including the 3rd variable.
       

    2. Write a paragraph interpreting your results.
       

  4. For this question, open employee.sav.  This is a database that has information on the agency's 475 employees.
     

    1. State a hypothesis (and null) where your dependent variable is measured at the interval level. The hypothesis should be stated as a comparison of the means.
       

    2. Run a one-way ANOVA, asking for descriptive statistics.
       

    3. Write a paragraph interpreting your results.