1. The relation between bodysize and risk of developing type 2 diabetes was investigated in the Nurses' Health Study. The USbased study was restricted to female nurses aged 3055 years at recruitment in 1976 (Colditz et al., 1990). The table below shows a sample of the results that were obtained after 8 years of followup: Bodymass index (BMI, kg/m2) < 22 2222.9 2526.9 2930.9 > 35 a. RR of diabetes 95% confidence interval 1.0 2.1 5.2 19.0 58.2 (1.43.3) (3.77.5) (13.626.4) (42.479.9) 10pts. What were the benefits of restricting the study to female nurses of a certain age and what effects would this restriction have had on (i) the internal validity and (ii) the generalizability of the results? b. 5pts. Only nurses who were free from cancer and coronary heart disease (CHD) in 1976 were included in the study. Why? 2. Calculation problem. Imagine a hypothetical casecontrol study conducted to investigate the association between high voltage power lines and birth defects that gave the following results: Mother lives near high voltage power lines Yes No Total Babies born with a birth defect Babies born on the same day without a birth defect 18 87 105 35 70 105 Total 53 157 210 a. 10pts. Does there appear to be an association between living near highvoltage power lines and having a child with a birth defect and, if so, how strong is the association? (hint: calculate the appropriate ratio measure for a casecontrol study) b. 10 pts. The following table shows the results separately for women of high and low socioeconomic status (SES). Is the association between highvoltage power lines and birth defects confounded by socioeconomic status? Mother lives near high voltage power lines High SES Yes No Low SES Yes No Babies born with a birth defect Babies born on the same day without a birth defect Total 7 45 8 74 15 119 28 25 10 13 38 38 1 2