Hi all, Please note that we re-uploaded a newer version of HW 1 at 5:30pm on Jan 9. We made a slight (hopefully helpful) wording edit in q6. If you downloaded the prior version, please make note of the revised wording shown below (underlined).
6. The authors found no association between multivitamin use and total prostate cancer risk (Table 2). Assume the exposure variable was not perfectly sensitive or specific. Please comment on how measurement bias could account for this result. Is this differential or non-differential misclassification? Can you tell the direction of this measurement bias on the estimate of the measure of association? Please use a 4x2 table to create a “box and arrow” diagram and show how measurement bias may be occurring in this study. Please explain the diagram in words. (3 points)
Best wishes,
JMC