Testing Association in the Presence of Linkage检验协会存在的联系.pptVIP

* * * * * * * * * * * * * * /ConfidRR.htm * * chisq.test(matrix(c(2,17,28,153),2,2)) * * * Arvid Sj?lander Traps and pitfalls in medical statistics Arvid Sj?lander * Arvid Sj?lander * Motivating example You are involved in a project to find out if snus causes ulcer. A questionnaire is sent out to 300 randomly chosen subjects. 200 subjects respond: We can use the relative risk (RR) to measure the association between snus and ulcer: Can we safely conclude that snus prevents ulcer? Ulcer Yes No R Snus Yes 2 28 2/30?0.07 No 17 153 17/170=0.1 * Arvid Sj?lander * Outline Systematic errors Selection bias Confounding Randomization Reverse causation Random errors Confidence interval P-value Hypothesis test Significance level Power * Arvid Sj?lander * One possible explanation It is a wide spread hypothesis that snus causes ulcer. Snus users who develop ulcer may therefore feel somewhat guilty, and may therefore be reluctant to participate in the study Hence, RR1 may be (partly) explained by an underrepresentation of snus users with ulcer among the responders. This is a case of selection bias. * Arvid Sj?lander * Selection bias We only observe the RR among the potential responders. The RR among the responders (observed) may not be equal to the population RR (unobserved). Population Potential non- responders Potential responders Sample * Arvid Sj?lander * How do we avoid selection bias? Make sure that the sample is drawn randomly from the whole population of interest - must trace the non-responders. Send out the questionnaire again, follow up phone calls etc. Population Potential non- responders Potential responders Sample * Arvid Sj?lander * Another possible explanation Because of age-trends, young people use snus more often than old people. For biological reasons, young people have a smaller risks for ulcer than old people. Hence, RR1 may be (partly) explained by snus-users being in “better shape” than non-users. This is a case of confo