HLTB15H3 Lecture 9: HLTB15 lecture9
Document Summary
Confounding- an important potential source of bias. When something other than the exposure of interest is responsible for the observed relationship. A threat to the (internal) validity of a research study. Ex- when assessing risk of disease, factors like age and ses can impact risk and therefore should be controlled. Confounding can be controlled through study design. Independent variable- manipulated variable: can also be something that is already there and fixed, something you are evaluating with respect to how it affects something else (dv) Dependent variable- measured variable: what is affected during the experiment, dv responds to or is dependent on the iv. Follow the basic experimental steps but no control groups: single group, post intervention design, single group, pre/post intervention design, common when using education interventions. Useful for quality improvement studies and as a precursor to a randomized control study of a complex intervention. Cofounding prevents establishing true cause and effect.