POLI 210 Lecture Notes - Lecture 5: Experiment, Internal Validity, Social Capital

• School: McGill University
• Department: Political Science
• Course: POLI 210
• Professor: Dietlind Stolle

Document Summary

Types of design: pre and post-test o1 x o2 compare a case to itself (automatically controls for lots of other factors) e. g. political knowledge -> civics course -> political knowledge. # 3,6,9 modules -> political knowledge or association memebers of different lengths of membership (have to account for drop out rate, other external effects of the peoples lives, age, *learn these designs and be able to recognize which ones are being used. Key point: use experimental logic also for observational studies (nearly identical control group, time order etc) Control group as similar as possible to your test group: 5. Need causal mechanism (why iv causes dv?: 6. As-if randomization": more or less random but the researcher does not manipulate/have any affect on it, natural phenomenon that is behind the randomization. Lab experimetns: most controlled environment, researcher in full control, complete randomization, good for internal validity, external val.