Conservation of a property is the ability to recognize that the property in question hasn't altered, even though other properties have been varied. It is Piaget's view, confirmed by empirical evidence that children are not able to conserve length until age 7 or 8. Recent concern over confounding variables led to the present study in which the children from three kindergarten classes (N=71) were tested for conservation of length after different training conditions. One class received a pretest, training, and posttest; one class received only training and posttest; and one class received both tests but no training. The training was three 20-minute sessions aimed at an understanding of the property of length. Analyses of variance determined differential effects on performance of the pretest, the training program, age and sex of the subject, and tester bias. Results showed the training program to be a significant variable, and girls made greater gains than did boys. (Author/MH)