CRM 2303 Lecture Notes - Lecture 15: Random Assignment, Quasi, External Validity

• School: University of Ottawa
• Department: Criminology
• Course: CRM 2303
• Professor: Holly Johnson

Document Summary

Advantages of lab experiments maximal manipulative control over the experimental situation. But, the lab is hierarchal, controlled, and artificial, has weak external validity. Used to make causal inferences when you are unable to: manipulate the independent variable, randomly select and assign control and treatment groups, conduct a pre-test. Move from manipulative control to analytic control. Quasi-experiment allows us to address concerns regarding ecological validity in lab experiments: external validity is improved, internal validity is weaker. Apply the basic logic of experimentation: was the independent variable (an experience or intervention) manipulated, is there any evidence of change/difference in the dependent variable, systematically generate, and find the data to test, all rival plausible explanations. There are two types of quasi-experimental designs: Structured like classic experimental design with a treatment and comparison group: but they cannot randomly assign. Use intact groups that are as comparable as possible: e. g.