The Memory for Locations assessment was completed by age-eligible NLSY79 children in 1986 and 1988 only. It was developed as a measure of a child's short-term memory and has been extensively used by Jerome Kagan of Harvard University (Kagan 1981). The child, aged eight months through three years, watches as a figure is placed under one of two to six cups. The cups are screened from a child's view for one to fifteen seconds; the child is then asked to find the location of the figure. Items increase in difficulty as the number of cups and/or the length of time during which the cups are hidden from view increases. A child's score is based on his or her ability to select the cup hiding the figure. The Memory for Locations assessment is found in Section 4 of the 1986 Child Supplement and the 1988 Child Supplement, available on the Questionnaires page.

Scoring Memory for Locations

The number of individual items that a child can potentially answer in this assessment is contingent on the age of the child. Children between the ages of 8 and 23 months start with item 1, the easiest question; children who are at least two years of age begin with item 4, and children age three start with item 7. A child's score is based on the highest (most difficult) question answered. A child who cannot answer the entry item receives a raw score of zero regardless of where he or she enters. Otherwise, if Q.1 is the highest item answered correctly, the child receives a score of 1. The maximum score is 10, if the tenth or final item is correctly answered. A child under two years of age is eligible to receive a score between zero and ten; a child age three, by virtue of the fact that he or she enters at item seven, can only receive a raw score of 0, 7, 8, 9, 10. Because external norms were not available, internally normed standard and percentile scores were developed. The user is still advised to use the normed scores cautiously because of the unusual distribution of raw scores described above.

Because of the complexity of the administration procedures for the Memory for Locations assessment, a number of responses were not coded precisely according to the decision rules. On the advice of the assessment developer, children who followed a sequence that might have led to "extra learning" (as part of the assessment administration process) were still scored. For example, if a child was asked Q.1B after having correctly answered Q.1A, the child was scored and not given an "invalid skip" code, even though, theoretically, the child was supposed to proceed directly from Q.1A to Q.2A. In addition, a careful examination of the individual responses suggests that there were a number of children who began the assessment at an improper entry point but who ended up at a level where they would, in all likelihood, have wound up anyway. In these instances, a score was provided for the children and these cases were "flagged" with a code of "2" on the Memory for Location flag variable (C07977.00 for 1988 and C05782.00 for 1986). A code of "1" on this flag includes all scored cases except those defined as 2s. Researchers who plan to use this assessment extensively should carefully examine the actual response patterns to individual items. Individual researchers may choose to impose scoring criteria that are more or less stringent than those used in computing the raw scores provided in this data file.

Age eligibility for Memory for Locations

In the 1986 and 1988 Child surveys, children aged 8 months to 3 years were eligible to complete Memory for Locations.

Additional information about Memory for Locations

The Body Parts and Memory for Locations assessments were no longer used in the NLSY79 Child surveys following the 1988 Child data collection effort, partly because of funding constraints and partly because of the greater difficulty in administering them to children in a home setting. Interviewers found it difficult to make an unambiguous determination as to whether a child was unable to respond or whether he or she was just shy. It was sometimes difficult to be definitive regarding the direction in which a child was pointing, either toward a cup or toward a body part.

This assessment displays a clear tendency to "top out" for the oldest children in the sample. That is, a very large proportion (63 percent in 1986) of all three-year-olds and 32 percent of two year olds received the maximum raw score of ten on the assessment. A relatively normal distribution may be in evidence only for children below the age of two.

Finally, evaluation of these two assessments in 1986 showed little in the way of significant linkages between a wide range of socio-economic antecedents and these two outcomes. However, more recent research suggests that these two assessments may be useful independent predictors of cognitive development (Mott, et al., 1995) since Body Parts and Memory for Location scores in 1986 are highly significant predictors of Peabody assessments in 1992. It appears that, in standard multivariate analyses, these early child cognitive measures may indeed be useful predictors of aptitude and achievement measures six years later.