HLTH 200 Lecture Notes - Lecture 10: Analysis Of Covariance, Covariate, Covariance
Document Summary
The trade off to high internal validity can be reduced external validity (generalizability) due to. The real world isn"t perfect, it doesn"t have perfect conditions (internal validity) In addition to knowing whether or not the intervention is effective, you are interested in knowing how different types of people respond to the same intervention. Sample is divided into relatively homogenous subgroups. Experimental design is implemented within each block or subgroup . Pre-program measure (pretest) is called a covariate. Pre-program measure is statistically controlled for in the data analysis to remove variability. The analysis of covariance (ancova) design adjusts posttest scores for variability on the covariate (pretest) Any event other than the program (cause) affects outcome (post-test) Ex: participants pick up math concepts watching sesame street so their math ability improves anyway improvement is not attributable to your program: maturation. Natural development between pre-test and post-test affects outcome.
