CH&E 4010 Lecture Notes - Lecture 15: Observational Error, Headache, Socioeconomic Status
Ch&e 4010 lecture 15 march 8, 2021 50 ways to introduce bias?! (and what to do about it) Identify opportunities for the introduction of bias in epidemiologic studies. Identify ways to minimize bias in epidemiologic studies. Selection bias: during wwii, a statistician was asked to help the british decide where to add armour to their bombers. Upon analysis of returned bombers, he recommended adding more armour to the places where there was no damage. The planes that got shot in other places did not come back to be analyzed. Systematic error introduced from the manner in which subjects are selected or retained in a study. Potential for selection bias in all study designs: rct adequate randomization, allocation are essential, cohort exposed/unexposed identical but for exposure, case-control select controls from the same population as cases. The nature of the research: access to eligible participants. Lifestyle factors: values, competing demands for time. Impute missing data, repeat analysis big differences indicate selection bias.