Biology 2244A/B Lecture Notes - Lecture 3: Recall Bias, Random Assignment, Confounding
12 Dec 2017
School
Department
Course
Professor
Document Summary
Issues with observational studies are mixed up with another explanatory variable (ex. amount of housework) Confounding: when other variables" effects on our response variable (ex. longevity) We can"t know whether its explanatory variable or other ones that are making the effects. Replication: repetition of a treatment on different units within a study. Ex. high heel study: multiple men asked to answer survey. Replication is fwithin one study: different from repeating study in canada vs. france. Want to get a big sample: reduce outliers, get a more accurate representation. Randomization: random assignment of units to treatments, or, the order of treatment experienced. Ex. high heel study: order of which shoes women wore was randomly assigned (people in morning got a random type of shoe vs. afternoon) Treatment groups are homogeneous when randomized (on average) Control: keeping other explanatory variables constant to isolate the effect of variable of interest. Simplest way to have more than one treatment group.
