APSY2216 Lecture Notes - Lecture 13: Analysis Of Covariance, Random Assignment, Internal Validity
Document Summary
Hold off on the survey assignment for the week after break - the syllabus is being re-arranged. Check your email for updates to the syllabus. Standardize procedures/implementation (ex: if treatment teachers need pd, give control teachers some other type of pd) Hold certain variables constant (ex: all low income) Build the variable into the design (ex: use it as a moderator) Using subjects as their own controls - get data on subjects before intervention and after intervention. Using analysis of covariance (ancova) to statistically equate groups and adjust scores accordingly. Essential ingredient to experimental design is having comparison groups with random assignments. Randomized posttest-only control group design - both groups receive different treatments and are assessed/compared. Very similar to randomized posttest-only control group design, only difference being that a pretest is given to each group following the random assignment. Benefit: ensure that random assignment did create relatively equal groups (this is especially important for smaller samples)