PSY 250 Lecture Notes - Lecture 7: Internal Validity, Random Assignment, Construct Validity
8 Dec 2016
School
Department
Course
Professor
Document Summary
Quasi-experimental design: not quite an experiment, but still helpful. Measurement of dv: constrained by several factors. Quasi-experimental design: characteristics of quasi-experiments (1) low sample size (low n) (2) no random assignment to groups (3) inferring causality is difficult. One-group design: does not have a comparison condition. One-group posttest only design: mostly everything you do to affect people uses this design! One-group posttest only design: example: paper on the table study. Hypothesis: people at restaurant less likely to sit at a table with a paper on it. Dv: time it takes for someone to sit at a table. We know something about the group before testing: risks to interval validity. Regression toward the mean: extreme scores for a participant, measured at one time, tend to change toward a central value for the group at another time. Non-equivalent control group: slight improvement over one-group design, but, must explain pre-existing group differences, still threats to internal validity.
