This paper reports three experiments concerning methodological issues in studies on incidental learning performance which use verbal and nonverbal procedures and which appear to be hampered by differences in stimulus materials, learning opportunities, and dependent measures. The first study, using 128 children from grades 3, 5, 7, and 9, attempted to determine if differences in developmental trends using paired associate and memory tasks can be attributed to methodological differences between the tasks. The second study, using 120 children from the same grades, attempted to see whether invariance in incidental performance across age with paired associate tasks is peculiar to a criterion level of intentional performance. The third study, using a sample of 80 children, examined whether observed developmental trends in incidental learning may be affected by the degree of difficulty of the task as defined by memory load. General findings are examined to determine if the decline in incidental performance around 11 or 12 years is related (1) to the nature of the intentional task, (2) to minimal learning conditions, (3) to relatively unrelated intentional and incidental independent measures, and (4) to a moderately difficult or demanding intentional task. It is suggested that the decline can be eliminated by criterion levels of learning and by use of a very easy intentional memory task. A viable theoretical explanation for the decline is that at this age children focus on intentional material at the expense of incidental information. (GO)