- Slides: 18
Other Quasi-Experimental Designs
Design Variations Show specific design features that can be used to address specific threats or constraints in the context
Proxy Pretest Design N N l l O 1 X O 2 Pretest based on recollection or archived data Useful when you weren’t able to get a pretest but wanted to address gain
Separate Pre-Post Samples N 1 N 2 l l l O X O O O Groups with the same subscript come from the same context. Here, N 1 might be people who were in the program at Agency 1 last year, with those in N 2 at Agency 2 last year. This is like having a proxy pretest on a different group.
Separate Pre-Post Samples N N l l l R 1 R 2 O X O O O Take random samples at two times of people at two nonequivalent agencies. Useful when you routinely measure with surveys. You can assume that the pre and post samples are equivalent, but the two agencies may not be.
Double-Pretest Design N N l l l O O X O O Strong in internal validity Helps address selection-maturation How does this affect selection-testing?
Switching Replications N N l l l O O X O O Strong design for both internal and external validity Strong against social threats to internal validity Strong ethically
Nonequivalent Dependent Variables Design (NEDV) N N l l O 1 O 2 X O 1 O 2 The variables have to be similar enough that they would be affected the same way by all threats. The program has to target one variable and not the other.
NEDV Example l l l Only works if we can assume that geometry scores show what would have happened to algebra if untreated. The variable is the control. Note that there is no control group here.
NEDV Pattern Matching l l Have many outcome variables. Have theory that tells how affected (from most to least) each variable will be by the program. Match observed gains with predicted ones. If match, what does it mean?
NEDV Pattern Matching r =. 997 l l A “ladder” graph. What are threats?
NEDV Pattern Matching l l l Single group design, but could be used with multiple groups (could even be coupled with experimental design). Can measure left and right on different scales (e. g. , right could be t-values). How do we get the expectations?
Regression Point Displacement (RPD) N(n=1) O N O l l l X O O Intervene in a single site Have many nonequivalent control sites Good design for community-based evaluation
RPD Example l l Comprehensive community-based AIDS education Intervene in one community (e. g. , county) Have 29 other communities (e. g. , counties) in state as controls measure is annual HIV positive rate by county
RPD Example 0. 07 0 1 Y 0. 06 0. 05 0. 04 0. 03 0. 04 0. 05 0. 06 X 0. 07 0. 08
RPD Example 0. 07 0 1 Y 0. 06 Regression line 0. 05 0. 04 0. 03 0. 04 0. 05 0. 06 X 0. 07 0. 08
RPD Example 0. 07 0 1 Y 0. 06 Regression line 0. 05 Treated community point 0. 04 0. 03 0. 04 0. 05 0. 06 X 0. 07 0. 08
RPD Example 0. 07 0 1 Y 0. 06 Regression pine 0. 05 Treated community point 0. 04 0. 03 0. 04 0. 05 0. 06 X Posttest displacement 0. 07 0. 08