PADP 7110 Lecture Notes - Lecture 6: Internal Validity, Randomized Controlled Trial, Random Assignment
Document Summary
Lecture 6 : research design experimental: two design purpose, designs for description. Researcher has little control over events; data is often observational. Ex. cross-sectional studies, case studies: designs for explanation. Internal > causality link (x > y) Draw causal inferences: timing of outcomes, extent to which changes in outcomes occurred after treatment, presence confounding factors. 2: design contamination > diffusion, compensatory rivalry or equalization of treatment, & resentful demoralization. By replicating a program in different settings, personnel, or subjects. Individuals are randomly assigned to treatment & control groups. 3: all groups represent the same population so the groups cannot be regressing toward different means, examples, post-test only, control group design, pre-test, post-test, control group design, stepped wedge design, random assignment. Chance decides whether a person receives the treatment or the control con- dition. Resulting experimental & control group differ from one another only by chance. Need enough cases so the laws of probability work.