CRM 2303 Lecture Notes - Lecture 14: The Control Group, Informed Consent, Causal Inference
Document Summary
Assess the causal impact of the independent variable (treatment the one whose impact we wish to study) on the dependent variable (outcome the one we measure to see whether any impact has occurred) Conditions necessary for determining causation: the two variables vary together (correlated, the cause occurs before the effect, the correlation between cause and effect is not due to some other spurious factor. 3 key elements to an experiment: cause-and-effect hypothesis, modify a situation or introduce a change, compare outcomes with and without the treatment/modification. Has all the key elements necessary to strengthen internal validity. Valid = the extent to which evidence supports the inference or conclusion. If a treatment/program is given, an outcome occurs. To show that there is a causal relationship, you must simultaneously address two propositions: If the program is given, then the outcome occurs. And if the program is not given, then the outcome does not occur.