An adaptation of a technique devised by Mischel (1958) was used in the longitudinal study to measure delay of gratification. Adaptations involved (1) asking the child to identify the larger of two pieces of candy to facilitate comprehension of the rewards, (2) specifying a standard time limit for receipt of the delayed reward which would be comprehensible to a young child, and (3) asking for a rationale for the choice in order to further understand the dynamics of preferences. Scores were obtained for correctness of the child's identification of the large piece, for his choice, the reason for this choice, and for memory of the instructions. Results of the analysis of the study data show that: 95% of the Year 1 sample and 98% of the Year 2 sample correctly identified the larger piece; in the Year 1 sample 35.1% chose immediate reward and 62.5% chose delayed reward; Year 2 data showed 39.7% chose immediate reward and 59.3% chose delayed reward. Analysis of variance for the longitudinal subjects showed no significant differences in choice behavior for children above and below the median age. An analysis of variance (sex x age x SES) showed no significant differences in choice behavior for sex or SES. In Year 1, egocentric choice rationales were most commonly given, and there was a tendency for test-defined responses to be given more by the older age group. Approximately 72% produced a relevant rationale. Less than one-half of the Year 1 sample could recall the examiner's instructions, while slightly more than one-half of the Year 2 sample correctly recalled them. Nine tables provide the study data. (For related documents, see TM 003 174-185, 187-198.) (DB)