Which Test for Comparing a Group Pre and Post

#1
A group of people have participated in a program
That group of people have completed a questionnaire pre and post program
Because it is a questionnaire there are multiple questions that needed to be answered
Each question has multiple choices (using a likert scale) but always the same choices for each question (e.g., almost never, sometimes, often, almost always)
I want to see if there is a significant difference in the group's answers between pre and post program

Am I to use ANOVA?
 

Karabiner

TS Contributor
#2
So you've got ordinal scaled items (NOT Likert scales; Likert scales are composed of 2 or more Likter-type items), measured pre and post. You could use the "nonparametric" sign test for this.

HTH

K.
 

hlsmith

Omega Contributor
#3
I agree with Karabiner, but I will provide a little more information. You will find the difference of the two measures (coded 1-4, and subtracting pre from post value).