in analysis of variance, a strategy for handling missing data in which the actual sample size in each of the different conditions or levels of the independent variable is replaced by the average sample size. When participants become ill, withdraw, or otherwise fail to complete a research study, there are no longer equal numbers of cases in each experimental condition. An unweighted means analysis takes this inequality into account by applying an adjusted sample size value across all conditions, thus ensuring that data from all experimental groups contribute equally to the determination of treatment effects. Also called unweighted means procedure.