(a) If the entity of life satisfaction was measured with perfect accuracy and reproducibility and then dichotomized into "satisfied" and "dissatisfied" (with the intermediate group discarded), and assuming no losses to follow-up, the following sample of data would result at the end of 1995. Consider this the truth for this study population:

Exposure	Outcome
Life satisfaction	Death	No Death
(Exposed) Dissatisfied	67	1673
(Unexposed) Satisfied	187	9412

Calculate the preferred measure of association between life dissatisfaction and mortality that can be calculated using these data, and explain your choice. (2 pts)

(b) When the study was actually conducted, measurement of life satisfaction was performed by self-report by using 4 short questions.Assume that a team of behavioral scientists assessed the validity of the 4 questions (versus a gold standard qualitative interview) and found it to be 90% sensitive for detecting dissatisfaction and 85% specific. How would this change the cells in the 2 x 2 table? Please fill in the "observed" data table below. (1 pt)

Exposure	Outcome
Life satisfaction	Death	No Death
(Exposed) Dissatisfied
(Unexposed) Satisfied

(c)Calculate the measure of association in the new table. (1 pt)

(d) Is this differential or non-differential misclassification? (1 pt)

(e) What is the direction of this bias on the estimate of the measure of association? (1 pt)

(f) Using the box and stick diagrams described in class, depict how measurement bias may be occurring. (1 pt)