Pooling Sample Sizes

The panel dimension of the NLSY79 data collection permits one to cumulate sample cases for children at specified ages across survey points, thus attaining rather substantial sample sizes for those ages. Pooling in this manner also can greatly enhance the heterogeneity of the sample for specific research topics. The trade-off to this methodology is that the ability to follow a particular age cohort across survey years becomes somewhat limited, although it is still doable for selected research topics. Table 1 below highlights potential sample sizes using this approach.

Pooling the Child samples. From the perspective of the younger children, it is possible to attain single year of age samples numbering in the thousands for specific research efforts. For example, if one wishes to examine associations between scores on the digit span assessment and other factors for seven year olds, it would be possible to cumulate a sample of about 4,000 seven-year-old children from the first to the current assessment survey year (see Table 1 below). Because these children would have been born to mothers in all years between 1979 and the current survey round (see Table 2 in the Sample Design section), the mothers of these children would range in age from their teens to their forties. The relevance of this pooling approach for evaluation of the child data, utilizing various assessments, is dealt with in the "Repeat Assessments" section of the Child Assessments--Introduction.

Pooling the Young Adults. Parallel estimates can be found for young adults. Because the young adult interviews have only been ongoing since the 1994 survey round, no single year-of-age cumulative estimate can include more than eleven points. Nonetheless, returning to Table 1 below, the number of cases cumulated in this way for ages 15 to 18 all attain single year of age totals of between 3,417 and 3,560 cases. It is worth noting that this sample cumulation modestly increases the heterogeneity of these young adult samples. For example, cumulating cases at these early young adult ages expands the age of the mothers at children's birth from mid-adolescence to the late forties. Because beginning in 2010 young adults age 31-32, 35-36, 39-40, 43-44, etc., as of December 31 of the survey year are not fielded, pooled sample sizes in these age ranges will be constrained.

Table 1. NLSY79 Child and Young Adult Sample Types for Pooled Age Groups through the Current Survey Round