Analysis of incomplete data in a two-period crossover design with reference to clinical trials

SUMMARY A procedure for analysing a simple two-period crossover design is proposed when some observations at the second treatment period are missing. Using all available data, the maximum likelihood estimators of the parameters are obtained and approximate tests proposed under the assumption that the data come from bivariate normal populations. The performances of these tests are compared with those of Grizzle's tests which require complete pairs of observations, using Monte Carlo methods.