PA 784    
Fall 2002

Due Oct. 17

Using Regression Analysis for Impact Analysis

 Open the file exp.sav.  (Note: you will be prompted to save this file.   You must have SPSS loaded on the computer you are working with to work with these data).

In order to reduce the error rate among mechanics working in a particular shop, a training program was developed.  Mechanics were randomly assigned to experimental (those who received the training) and control (those who didn’t receive the training) groups.  This data file contains the pre-training program and post-training program scores both groups.  Please do the following:

 1. Assume no pre-intervention measure had been taken.  Regress the post-training rate on the program variable in order to determine the impact of the program.  

a)      Discuss the impact of the program on error rates. 

b)      What kind of evaluation design is this? 

c)      What potential threat(s) to validity are there?

2. This time, include the pre-intervention rates in your regression model.  Now how would you assess the impact of the program on the error rate? 

a)      How do you explain the difference between your results here and your results in #1?

b)      Now what would you say about the impact of the program?  

c)      What kind of evaluation design is this? 

d)      What potential threat(s) to validity are present with this model? 

e)      Which of the two models is better (more sensitive)?