I.  Goals.
A.  Designs with more than two groups, one IV.
B.  How these designs relate to simple experiments.

II.  Designs with more than two groups, one IV.  So far, we've only had treatment and control.  A lot of situations call for more than just these two groups.  Let's change our class experiment.  I want to investigate the effects of mood on perception.  My hypothesis is that people will devote more attention to information that matches their present emotional state.  The IV is mood, and I'll have three groups:  sad mood, neutral mood, and happy mood.  I'll manipulate these by playing musical selections that reflect the correct mood.  For a dependent variable, I'll have participants read a story with a neutral beginning, then have them continue that story.  My hypothesis is that people in sad moods will have sad continuations, and people in happy moods will have happy continuations.
 
III.  How these designs relate to simple experiments.  Other than having more than two groups, everything I've said about two groups designs still applies.  Some issues to think about:
A.  Which design format should we use (after only, pre-post, or matched)?  Why?
B.  Assess the internal validity of this design.  Which issues are particularly relevant?
C.  Assess the external validity of the design.

Back to Langston's Research Methods Page