PSCH 242 Lecture Notes - Lecture 1: Random Assignment, Demand Characteristics, Internal Validity
Document Summary
Randomized post test only control group design(rpocgd): random assignment, group a and b, apply treatment to a, not to b, post test measurement, compare. Random assign controls for selection, confounds controlled by post test comparing, trying to find size of improvement. Randomized pretest-postest, control-group design: randomly assign sample, group a and b, pre test measurement, apply or not apply, post test measure. Adding pretest allows to quantify the amount of change. Verify equality at initial , cannot assume equal based on random assignment. Advantages and disadvantages of between subject designs: as all experiments allows for strong conclusions about causality, main advantage over within subject designs. Demand characteristics: main disadvantage of between subject designs. More difficult to pick up small effects: a pre-test reduces error variance and provides more information about the size of the change, a pre-test can introduce new potential confounds. Self-presentational concerns: adding several different control conditions can help us rule out alternative explanations (confounds)