1. You are conducting a cohort study to measure whether a sedentary lifestyle is a cause of dementia in older adults. Activity levels (hours standing per day) were self-report by participants at the start of the study. The outcome of dementia was assessed by study nurses using a validated scale at the end of the five years of study follow-up. The dementia scale does a very good of identifying people who have dementia (i.e. it has good sensitivity - more to come on this soon) but it is also known to classify some people as having dementia when they really don't (i.e. lack of specificity), however, this isn't expected to differ based on whether or not people are sedentary.

Is this differential or non-differential misclassification (i.e. information bias) and why?

How will this bias affect the estimated RR reported in this study?

1b. You are conducting a case-control study to examine whether parental smoking is associated with leukemia in children. Your cases are children newly diagnosed with leukemia at one hospital. The controls are non-hospitalized children from the neighborhood immediately surrounding the hospital. After conducting the study, the researchers observed that the children in the neighborhood around the hospital are much more likely than the general population of children to be exposed to parental smoking.

Is there a problem with the control selection? Explain why or why not? If it is a bias, please specify the type of bias.

How would the above affect the OR in the case-control study (if at all)?

Please send me answer ASAP.