On the consequences of overstratification

It is common, in particular in observational studies in epidemiology,to impose stratification to adjust for possible effects of ageand other variables on the binary outcome of interest. Overstratificationmay lower the precision of the estimated effects of interest.Understratification risks bias. These issues are studied analytically.Asymptotic results show that loss of efficiency depends on thetrue effect and on a measure of the average imbalance acrossstrata between exposed and unexposed individuals. Bias dependson the correlation between stratum-specific size imbalancesand event rates in the unexposed. Approximate results are alsogiven. An example is used.