PS295 Lecture Notes - Lecture 13: Time Series, Interrupted Time Series, Quasi
Quasi Experimental Research
-used when researcher is not able to randomly assign participants to conditions and/or cannot
manipulate the causal variable of interest (IV)
-gender, race, ethnicity, socio economic class, diagnoses, personality, laws etc
-these designs generally do not possess the same degree of internal validity as experimental designs
5 types:
1) Pretest-Posttest Design
A) one group pretest posttest design:
-pretest → IV → posttest
-fails to eliminate most threats to internal validity and lacks control, do not use
B) nonequivalent control group design:
-researcher finds a group similar to the group that received the IV to serve as a control group
quasi experimental group → IV → posttest
+
nonequivalent control group → posttest
-caution: biased assignment, groups may not have been similar before IV
-local history: an event might happen to one group and not the other
C) Solomon four group design
-combo of pretest-posttest design and the posttest-only design
-two experimental groups + two control groups
Document Summary
Used when researcher is not able to randomly assign participants to conditions and/or cannot manipulate the causal variable of interest (iv) Gender, race, ethnicity, socio economic class, diagnoses, personality, laws etc. These designs generally do not possess the same degree of internal validity as experimental designs. 5 types: one group pretest posttest design: Pretest iv posttest: pretest-posttest design. Fails to eliminate most threats to internal validity and lacks control, do not use: nonequivalent control group design: Researcher finds a group similar to the group that received the iv to serve as a control group quasi experimental group iv posttest nonequivalent control group posttest. Caution: biased assignment, groups may not have been similar before iv. Local history: an event might happen to one group and not the other: solomon four group design. Combo of pretest-posttest design and the posttest-only design. Allows researchers to evaluate a treatment effect (influence of iv) as well as the possibility of a testing effect.
