Lecture 6 - Quasi-Experimental Designs Flashcards
Quasi-Experimental Designs
- No Random Assignment and/or no IV manipulation
- Increased Internal Validity Concerns
- Non-equivalent groups (time series)
Building Blocks of Quasi-Experiments
- # of treatment and comparison groups
- # and variations of the IV
- # of pretest & posttest measures
- Participant selection & assignment procedures
Non-equivalent group designs
- No randomization to groups = no equivalency
- Control Group = Comparison Group
- non-equivalent group designs minimize differences
Static-Group Comparison
- No Random Assignment
- Quasi-equivalent to the post-test only experiment
Pretest-posttest Non-equivalent control group
- No Random Assignment
- Quasi-equivalent to the pretest-posttest only experiment
- Increased protection against internal validity threats
One group pretest-posttest
- No Random Assignment
- Measures one group before and after the DV
- Measures change but can’t determine what caused it
Time-Series Designs
Longitudinal Research
* Multiple observations over time
Internal Validity Threat:
* Instrumentation
Interrupted Time Series
Multiple measures of DV before and after IV exposure
Interrupted Time Series with Non-equivalent Comparison Group
Multiple measures of DV before and after IV exposure
* With a comparison group to add additional control
Interrupted Time-series with Removed Treatment
- Adds and removes the treatment
- Can confirm if observed changes are due to IV without a comparison group
Interrupted Time-Series with Switching Replication
- Treatment is implemented on different timelines
- Determines if timing of treatment impacts DV
Non-Experimental Designs
- Designs that do not involve manipulation of the IV, a comparison group, or repeated measures
Ex.
* Cross-sectional descriptive surveys
* Observational studies with no variable manipulation
* Case studies
