POL SCI 3 Lecture Notes - Lecture 12: Internal Validity, Dependent And Independent Variables, Birth Weight
Document Summary
E. g. entity city of fremont, x police reform, y crime rate: hypothesis: police reform would decrease the crime rate. 6 vulnerabilities/ rival explanations to (x y) (not all, just the common ones: (1) normal fluctuation in y. Maybe crime rates just fluctuate a lot. Typical, change variation, fluke: (2) maturation, (3) regression to the mean/ regression. Maybe crime rates are on a decline as part of a long time trend. People get richer over time less crime. Pretest observation is an outlier extreme observation that is atypical, posttest observation is the ordinary level regression to the mean: (4) history. Between pre and post, potentially confounding event z occurs and z caused the difference: (5) testing. Gathering the pretest data prompted a change in the dependent variable (pretest y) Intervening variable, somehow can trace back to effect of pretest: (6) instrumentation. How you measure the dependent variable: procedures used to collect data difference in measurement between pre and posttest.