HLTH 200 Lecture Notes - Lecture 14: Block Design, Covariance, Covariate
Document Summary
Observational: don"t involve anything that you are manipulating. Experimental/quasi-experimental: involves an intervention, experimental: has random assignment, quasi-experimental: does not have random assignment. Basic: randomized pre-post design with an experimental group and either a control or comparison group. Switching replications (who gets the treatment and who doesn"t switches) Randomized block design (repeats the experiment for different subgroups) The same exact design but in different types of groups (e. g. children, young adults, older adults) Covariate: another variable that affects your potential outcome. Two independent factors: coffee (yes/no), stress (some, none) Basic: non-equivalent or non-randomized pre-post design: single group (pre-post is the comparison, multiple groups (experimental group and either a control or comparison group) Interrupted time series (event or policy to be evaluated) Regression discontinuity (need to give the treatment to severe cases) Proxy-pre-test (measure the baseline after the intervention) Used when there is a specific policy or event that you want to study.