HSCI 330 Chapter Notes - Chapter week 8: A Priori And A Posteriori, Relative Risk, Seroprevalence
Document Summary
(cid:449)hethe(cid:396) (cid:448)a(cid:396)ia(cid:271)les othe(cid:396) than the e(cid:454)posu(cid:396)e (cid:448)a(cid:396)ia(cid:271)le a(cid:396)e (cid:272)ont(cid:396)olled in the anal(cid:455)sis. Table 1: 6 green and 10 blue hats; 5/6 (83%) of green fit, 8/10 (80%) of blue fit. Table 2: 10 green and 6 blue hats; 2/10 green (20%) fit, 1/6 (17%) of blue fit. 3 variables: table number (c), hat fitting outcome (d), and hat color (e) Confounding is a distortion in a measure of effect: e. g. Rr that may arise because we fail to control for other variables, for example, smoking history, that are previously known risk factors for the health outcome being studied. Crude estimate of effect (crr): no variables are controlled. Adjusted estimate of effect (arr): one or more variables is controlled. Clinically important difference: 4. 3 vs. 3. 8 no clinically important difference, 1. 3 vs. 1. 2 no clinically important difference, we do not use statistical testing to evaluate confounding. 10% - confounding has occurred with at least a 10% change in the estimated association.