PUP-3002 Lecture Notes - Lecture 4: Type I And Type Ii Errors, Internal Validity, Observational Error
Document Summary
The process or method of investigation in which the analyst develops a theory to produce empirically testable or falsifiable hypotheses and then tests these hypotheses with data. We theorize that x->y now we see of the data confirms this relationship. Known as the fundamental problem of causal inference. Can"t fix the levy to see what would happened. May be the best way to actually get a causality. Differences in y between control group and treatment group show the causal effect of x on y. For an experiment to work, there can be no differences between the subjects in the treatment group and the control group that are also correlated with our potential outcome. When we have a set of variables that are correlated with both our treatment and our potential outcome, (confounding variables), our experiment will be unable to reveal a valid causal effect.