Ch. 10: Quasi-Experimental Designs

1
Ch. 10 Quasi-Experimental Designs Program
Evaluation

• Context
• The Problem examples
• History the Campbell Stanley book
• The Crucial Issue
• Non-random assignment (why?)
• (Lack of control)
• True experiments (see pp.340-1, 352-3)

2
Threats to Internal Validity

• (A nice little review confounds)
• History (ind. difference variables)
• Maturation (natural change)
• Testing (reactive effects)
• Instrumentation (bias, expectations)
• Regression to the mean (!!)
• Selection (non-random, more of history)
• Subject loss
• Interactions with selection (new ??)

3
Interesting Side-bar(pp.349-352)

• Contamination of Ss (a type of confound)
• Issues of external validity
• Hawthorne effect

4
Quasi-Experimental Designs(a few)

• Non-equivalent control group (lots of examples
  and discussion plus pp. 358-362 repeat much
  of 346-349)
• pre-post without control, p. 354 top
• very common, not very useful
• pre-post with control, p. 354 bottom
• side-bar on ANCOVA

5
Other Quasis

• Time series
• Interrupted (w/o control), p.362
• With control, p. 366
• ABAB (with group)

6
Program Evaluation

• Examples
• Application of Quasi-exp. Designs
• Other Issues
• Needs assessment
• Establishing implementation (process)
• Direct outcome evaluation
• Participant/observer opinion as proxy
• Measuring efficiency

7
The End!!!

• Except for presenting results
• written paper
• oral presentation
• poster