Nonexperimental and Quasi- Experimental Designs Distinction is the degree of control over internal validity
True Experiment Manipulation of a variable Control of threats posed by confounding variables NE and QE compare scores from diff groups or conditions but there is no manipulated variable
Nonequivalent Group Designs Groups / conditions created by differences in individuals before experiment No control over assignment of individuals to groups Examples: IQ, race, gender
Differential Research Design NE research design because no attempt to control for threat of assignment bias Researcher interested in differences between groups on certain variable Ex post facto looks at differences after the fact
NE Control Group Design Nonequivalent groups used in applied research settings in which goal is to evaluate treatment to a preexisting group of ind Second, control group, used as comparison
Posttest-only nonequivalent control group design “static group comparison” One group of p’s given treatment and then measured after treatment (posttest). Scores then compared to control group X 0 (treatment group) 0 (control group)
Caution! Remember the problem with threat to internal validity with these designs Assignment bias differences between groups could reflect individual differences
Pretest-Posttest NE Control Group Design Phase 1 – Observe (measure) both groups Phase 2 – Administer treatment to Experimental group – No treatment to control Group 0 X 0 (Experimental group) 0 0 (control group) Purpose of Phase 1 to determine if groups equal prior to Phase 2 treatment Quasi-Experimental threats minimized
Threats still possible Differential effects – Differences between groups may be result of different histories effects Similar to influences such as maturation, instrumentation, testing effects etc
Pre-Post design One group of Ps measured before and after a treatment Goal is to evaluate influence of intervening treatment No control group Series of measurements over time on same Ps Caution: time-related threats