1

A short shot of help would be great!! Thanks a lot in advance.

I have asked same probands to rate two groups of people, e.g. unemployed indivduals versus employed individuals with parallel formulated items, e.g.

  • How motivated are unemployed people?
  • How motivated are employed people?

Answers are given on a 6-point likert scale. Everyone rates both groups.

I want to test hypothesis whether unemployed people are rated significantly worse than employed people.

However, I have troubles determining the right kind of test, thougts so far:

  • Independ t-Test is inadequate because one person gives both ratings.

  • Dependend t-Test is inadequate because it is no repreated measures and both items are not at all identical.

Please help me out...

Thank you so much.

Jens
  • 563
  • 1
  • 4
  • 13

1 Answers1

1

My first thought would be to take a new variable which is the difference between the two likert scales, then do some kind of test to determine if it is significantly different from zero.

But upon further research, these people have figured out how to do it all properly; https://www.researchgate.net/publication/266212127_Five-Point_Likert_Items_t_Test_Versus_Mann-Whitney-Wilcoxon

Gordon McDonald
  • 168
  • 1
  • 11