EFFECTS OF ASSUMPTION VIOLATIONS ON TYPE-I ERROR RATE IN GROUP SEQUENTIAL MONITORING

We compare the degree to which the Type I error probability can be inflated by assumption violations under the different sequential monitoring schemes. For methods that assume equally spaced time intervals between ''looks'' at the data, we examine the effects of unequal spacings. For methods that do not specify the look times in advance, we examine the degree of Type 1 error inflation that can result when future look times are chosen based on data trends. We also evaluate a Bayesian method of monitoring from this non-Bayesian perspective of Type I error inflation.