Retention

This section provides sample sizes for the number of times younger children and young adults of different ages have been interviewed over the life course of the survey. Given that the child interviewing process began with the 1986 interview round and continued on a biennial basis through 2014 for the Child Supplement and 2016 for the Mother Supplement, the maximum number of child interviews a respondent could have is eight, since children age into the Young Adult survey the survey round in which they turn 15 or older (12 or older since 2016) and the Young Adult interviews began in 1994. Clearly, the content varies considerably between the Child and Young Adult interviews, partly because in most rounds only younger children were assessed, and partly because many of the questions are life-cycle specific. Many questions that might be relevant for an eight- or a fourteen-year old might not be appropriate for an older adolescent, much less someone in their twenties, thirties, or forties. Additionally, there have been some changes in questions and question wording over time, so researchers who are using these data in a longitudinal manner need to carefully review the content of the questions they are using.

The Child and Young Adult surveys are characterized by inherently different question structures, modes of data collection, and indeed potentially different research agendas. It is therefore useful to present separately the sample sizes for younger and older children, even though the ultimate research agenda in many instances may join these two sample types.

• The Number of Child Interviews
• The Number of Young Adult Interviews

The Number of Child Interviews

Table 1 presents the number of child interviews ever completed by the NLSY79 children as of the most recent survey round. This table references age at the end of the survey year (December 31) rather than the survey date because the age determination for inclusion as a young adult rather than a younger child was the age as of the end of the calendar year. This method of computing age avoids a need to split the fourteen-year old age group between a younger child and a young adult component. Whether one uses a survey date or end-of-year age typically has little impact on the magnitude of age-specific sample sizes. Most estimates of sample size for younger children in this users guide use child age as of the survey date. This is the reason why sample sizes by age presented across tables may not always be identical.

Table 1 indicates the total number of child interviews reported for each NLSY79 child who has ever been interviewed, regardless of current age. For this table, a child is defined as interviewed if the sampling weight (CSAMWTyyyy or CSAMWT_REVyyyy) is greater than zero for a given survey year. The greatest number of possible child interviews would be eight biennial interviews. Only 727 of the respondents at the end of 2018 fall into that category. However, Table 1 shows much larger numbers of children in all the other interview frequency categories. Children who fall into the older age categories as of the current round, but who have completed only a small number of interviews (e.g., 11 year olds with only one or two interview points), have missed some interviews. For example, an 11 year old could potentially have six completed interviews. The implications of repeat interviewing for these younger children are expanded on in the documentation on the child assessment data, where the extent of interview repetition is connected with the specific cognitive and socio-emotional assessments that the children complete at various ages.

Table 1. NLSY79 Children: Number of Child Interviews by Age as of December 31^st, 2018