Testing equality of two mean vectors with monotone incomplete data.

In this paper, testing the equality of two mean vectors is considered when each dataset has a monotone pattern of missing observations. A simplified Hotelling's T2-type test statistic in a two-sample problem under monotone missing data is given. Transformed test statistics of this simplified T2 statistic based on Bartlett adjustment to improve the chi-squared approximation are derived. Furthermore, the accuracy of the approximations is investigated using a Monte Carlo simulation. Finally, the numerical power of the proposed test is presented. [ABSTRACT FROM AUTHOR]