Outcome of early experience studies as affected by between-litter variance

Studies of early experiences in animals in which the mother or the entire litter is manipulated often involve significant levels of between-litter variance. Such variance, if not accounted for either by analysis which considers litter as an independent source of variance, or by use of mean litter scores, may contribute to misleading treatment effects. This phenomenon is demonstrated for experiments utilizing analysis of variance procedures for data analysis.