BIO 301D Lecture 21: Bio 301D – Controls – 3.5
Document Summary
E. g. there might be an imbalance in how red cars and associated with these other variables the third variable would go away if we could destroy any association between car color and the third variable. Want to make third variables the same across red and non-red. Example in this comparison, we refer to: red cars as the treatment group: non-red cars as control root. 1) eliminate it- so it is no longer variable. 2) leave it variable, but spread it evenly across control and treatment group. Say we want to control for car type as a cause of the correlation. Type of driver between accident rate and color: C- risky type of car, not red. D- safe type of car, not red. Which will eliminate the third variable: answer: a&c or/and b&d. What would the following pattern mean: table entries are accident rate. 10/24/16: answer: type of driver impacts accident rates, not color.
