POL222H1 Lecture 5: 222-5
Document Summary
Gerber and green 2000: field experiment in usa. External validity: representativeness of sample: the experimental participants are the real subjects of interest of our theory, generalizability beyond the participants is another issue; however, it is more plausible then laboratory experiments. Internal validity: random assignment- by doing this we can reach the level of the confounding variable. In field experiments, a completely random assignment of treatments may be difficult and internal validity maybe somewhat compromised. If the number of experminetal subjects/participants is small, randomisation may not achieve its goal- a balance in the subjects/ participants" characteristics (covariates) between treatment and control groups. Internal validity: random assignment: field experiment examples, gerber and green 2000: 29,380 registered voters, (treatment: 5,794 (20%), control: 23,586 (80%), beath et al. 2013: 500 villages: (treatment: 250 (50%), control: 250 (50%), wantchekon 2003: in each of the 8 districts, 1 village was in the clientelism treatment another in the public policy treatment, and the rest as control.
