Intercohort and Cross Generational Research

NLSY79 Mothers

Table 1 presents a synopsis of some of the attitudinal and behavioral questions that have been asked of mothers, and of their children when they were both pre- and early teenagers, and then again when they were young adults. When these mothers were in their middle to later adolescent years they were asked about their own educational, employment and family aspirations. The mothers' records in the NLSY79 main Youth also contain information about their progression through school, their employment experiences and their marriage and childbearing experiences, all systematically updated at every survey round. Comprehensive annual income and individual earnings information is also collected, along with household structure updates, and county of residence information (available on a separate geocode file). Attitudinal information and self-reports on mental health and wellbeing (for example attitudes towards women's roles, the Rosenberg self-esteem scale and the CES-D depression scale) are available in selected years. A battery of tests measuring maternal cognition, plus measures of educational achievement are also available. For example, the Armed Services Vocational Aptitude Battery, or ASVAB, was administered to NLSY79 main Youth respondents in 1980 and scores are available for about 94% of the cohort's members. PSAT, SAT I, or ACT scores were also coded for about a quarter of NLSY79 respondents from Transcript Surveys between 1980 and 1983. A cognition module was added to the NLSY79 main Youth health section in 2006; this same module is included in the expanded health module introduced into the Young Adult survey in 2010. Additionally, there is a short series of questions on the mother's religious affiliation and upbringing, plus information pertaining to some of the less normative activities she might have been involved in. Finally, the mother's first survey round in 1979 included a series of questions measuring the occupation and earnings of her parents (i.e. one set of the children's grandparents) that provides information on the socio-economic background she experienced while she was growing up. This background input was then greatly enhanced by a detailed childhood residence history that was collected retrospectively for all NLSY79 women in 1988. We will incorporate some of this information above into our discussion of research possibilities that follows.

NLSY79 Children 10-14

Once NLSY79 children reach age 10, they were asked about their own behaviors and experiences, and, as shown in Table 1, some of the information obtained was comparable with that obtained some years earlier from their mothers. These self-reports for children 10 and over included details about joint activities and interactions with parents, parental rules about activities and behaviors, household decision-making processes, peer pressures, attitudes about school, after-school employment and other extra-curricular activities, incipient substance use, involvement in a variety of non-normative activities, religious activities, dating relationships, sex education and expectations about future marriage and family building activities. Starting in 1992 the content of the child self-report was expanded to include details about within-family interactions, and--for those age 13 and over--items about sexual activity. Mothers were asked to report on their children's schooling success (e.g., grade repetition, reasons for success or failure), and each child ages 10-14 was asked to rate the quality of the school they attended and how they felt about their life in general. Questions about parental supervision activities were also directed to the mother.

NLSY79 Young Adults (age 15+)

All children who reach their 15th birthday during the survey year are interviewed as young adults. Reports by these young adults often have considerable overlap with questions asked of their mothers at similar ages. For example, scores are available for the Pearlin Mastery, Rosenberg Self-Esteem and CES-D scales for young adults and their mothers during their late teens. Beginning in 2016, children ages 12 to 14 have also been fielded as part of the Young Adult data collection. The 14 year olds answer the same questions as the 15-16 year olds, but the 12 and 13 year olds answer far fewer questions, including some previously asked in the CSAS.

• Introduction
• Exploring Cognitive and Socio-Emotional Trajectories
• Validity of Repeat Measures
• Early Family and Relationship Transitions and Behaviors
• Young Adult Schooling and Work Outcomes
• Geographic Moves, Location, and Employment
• Non-Normative Behaviors across Generations
• How Children Affect Mothers
• Variation in High School and College Attendance
• Within-Family Differences in Outcomes
• Child Health and Child/Young Adult Outcomes

Introduction

In this section we present examples of a number of research options for which the current data set is highly appropriate. Again we consider both within- and cross-generational research possibilities as we emphasize that a strength of these data is that not only can they be used to explore connections between childhood, adolescence and adulthood, but also to examine connections across generations going back to maternal ages that closely match the ages many of our young adults are now. Additionally, the NLSY79 has large samples at many ages, a large sibling sample, and substantial minority oversamples.

The primary limitation to using the NLSY79 Child and Young Adult data is that these young adults cannot be generalized to the overall US population of similar ages. Although their mothers are representative of youth ages 14-21 who were living in the United States in 1979, the sample has not been refreshed to reflect changing population characteristics resulting from immigration. However, a subset of these children can be compared with members of the NLSY97 cohort as there is overlap in the years of birth between the two data sets, and the NLSY97 is nationally representative. Additionally, as pointed out earlier, the older young adults are primarily born to younger mothers. However, with every passing wave of data collection, this issue of heterogeneity at the older child ages becomes less of a limitation as the young adult sample is not only increasing in size, but the older young adults are becoming more representative of a broader spectrum of individuals in their twenties who have been born to a more representative group of women. And importantly, the older children at this time represent an ideal sample for exploring a wide range of programmatic and policy issues related to the adjustment process and mainstream economic and social assimilation of disadvantaged youth, while at the same time permitting comparisons with children who have been born to relatively younger but middle class mothers. Samples at the younger ages, however, are quite heterogeneous, especially when children from different birth cohorts are pooled.

We present topical information pertaining to mothers and to their children in both childhood and young adulthood in Table 1. Information from this table will be selectively utilized in our research examples below.

Exploring Cognitive and Socio-Emotional Trajectories

The NLSY79 Child and Young Adult study is one of the few large-scale data sets that permit researchers to examine the same children repeatedly over time, allowing them to explore linkages between intellectual and emotional development, or the stability of scores across multiple survey waves. For example, it is possible to examine connections between infant and early childhood cognitive batteries such as the Kagan Memory for Location assessment, the Body Parts assessments, or the Temperament Scale, and their later intellectual development. Further, one can explore possible causal connections between these early cognition measures and not only levels of subsequent cognition (as measured by PPVT or PIAT scores, for example), but also with intellectual trajectories over the early childhood to middle adolescent period. Researchers can also consider variations in trajectories across different socio-economic or racial/ethnic groups. Variations in trajectories may be linked with differences in life cycle paths and these data contain a range of information relating to family structure and family social and economic well-being over time. As of the current survey round, more than 80 percent of children born to the NLSY79 women have been followed up through age 15, or through all of their "testing years." More than two-thirds of the children ever reported by interviewed NLSY79 mothers have been interviewed at least once as a young adult.

Validity of Repeat Measures

These data may be used to explore the face validity of selected assessments over time. As already noted, some children have taken the PIAT assessments as many as five times, so it is possible to examine the stability of normed scores over time (bearing in mind that the sample cases were normed against national standards). Similarly, mothers have repeatedly completed a 28-item behavior problems scale for all of their children as they have aged from 4 to 14 years. One can also consider how assessment trajectories may be linked to other family factors, including scores on the HOME scale (described in detail in the child assessment sections of this users guide). Repeated measures on this HOME scale provide comprehensive information about the quality of the child's cognitive and socio-emotional home environment for the first 14 years of a child's life. Taken in conjunction with related family information, the repeated measures on these scales allow researchers to explore relevant connections in a more comprehensive and methodologically refined manner than is typically possible. The availability of many sibling pairs (see Tables 5 and 6 in Sample Design) also permits a more careful exploration of how within-family transitions can alter intellectual and socio-emotional developmental processes. Finally, as we describe in sections that follow, selected assessments may also be of value as intervening measures when exploring various within- and between-generational processes.

Early Family and Relationship Transitions and Behaviors

The NLSY79 child and young adult surveys provide excellent data to explore potential predictors of relationship timing, and the types of relationship trajectories that youth follow from mid adolescence and on into adulthood. Beginning at age 13, children report their early sexual activity and fertility experiences. Starting at age 15, in-depth information is collected about relationships in which they have been involved. Beginning with the 1998 survey round, a sequence of items that describe the full range of heterosexual relationships that sexually active youth can be part of are included in the young adult data collection. Family and maternal inputs, as well as over a decade of assessments from childhood, can be used as predictors of transitions into early relationships and also to explore a variety of questions such as why some early relationships last and others do not; why some cohabitations ultimately turn into marriages; and why some young adults have children early while others delay parenthood. It is also possible to investigate why some youth from relatively disadvantaged families effectively avoid some of the hurdles posed by early parenthood whereas others are less successful, and to explore intergenerational tendencies to form early relationships and begin childbearing at young ages.

Young Adult Schooling and Work Outcomes

The young adult data provide an excellent vehicle for examining educational progressions and transitions into the work force. And again, we emphasize that these outcomes can be linked to a rich range of child and early young adult data available from many prior survey rounds (for example, how might child cognitive or behavioral trajectories be linked to greater or lesser success in later educational progressions, or how might early relationship patterns impact transitions into the workforce). Schooling and work outcomes among these young adults can also be compared with the trajectories that their mothers took a generation earlier.

In addition to collecting detailed information about their post-secondary college experiences, when these youth were juniors/seniors in high school, they were asked about all the colleges they applied to. This information can be compared against actual college attendance. Data on specific colleges applied to and attended are available on the restricted geocode files.

In terms of employment, information is available about summer and school year employment from early adolescence onward. During the late adolescent and early adult years a lot of data is gathered regarding the nature and intensity of employment and training experiences. These data are described more fully in topic specific sections of this users guide and here we just point out that young adult educational trajectories, family transformations, and dimensions of employment can all be carefully linked in order to provide a better understanding of the associations, and perhaps causal linkages among these major life cycle events.

Geographic Moves, Location, and Employment

A special geocode file is available for all survey rounds for the main respondents in the NLSY79, and for 1994 to the current survey round for all young adults regardless of their residence. This data file can be linked with all of the NLSY79 main, younger child, and young adult data. Because many of these youth have been effectively tracked since birth, first in their mother's home and then increasingly in their own homes, these data offer unique opportunities to investigate the inherently complex connections that exist between geographic moves, early family structure transitions, and human capital. Migration of young people is an important feature of their early life course, and these geographic data permit a careful study of residential move patterns, their determinants, and consequences. The migration literature points to the importance of family connections in helping to explain migration. The residence information available for young adults can also be linked to data on their past locations, as well as to data on the current location of selected family members.

The availability of matched county-level information over most of these children's lives also allows potential connections between the residential dislocation of children and their intellectual and academic outcomes to be explored. For example, it is possible to examine whether or not younger children and young adults are academically disadvantaged if they are subject to repeated geographic movements throughout their formative years. These locational data also support explorations of how geographic trajectories might impact earnings during the early young adult years, particularly when this kind of geographic information is augmented by knowledge of school quality.

Non-Normative Behaviors across Generations

As the information we present in the section on Life Cycle Profiles for the NLSY79 Children shows, this data set presents many opportunities to look at cross-generational tendencies for family members to repeat non-normative behaviors (see the NLSY79 User's Guide for details about the mother's record). For example, the child/young adult data can be used to explore substance use trajectories over time and then to investigate possible linkages between these trajectories and a wide range of parallel child and family behaviors/attitudes. Since many of these data elements are available for both the children and their mothers at comparable life cycle points, it is possible to not only examine whether certain behaviors tend to be repeated across the generations, but also to consider the kinds of family units that appear more protective and hence evidence a better ability to break intergenerational connections of activities or behaviors that might typically be considered non-productive. Use of the young adult geocode file can further enhance these kinds of studies as the county-level identifiers that are included in the geocode file provide data to help sort out the potential relevance of area-level factors that may either additively or interactively affect connections within and between generations.

As illustrated in the section on Life Cycle Profiles for the NLSY79 Children and Table 1 above, there are many data elements that are available for both mothers and their children including information on early sexual activity, the Rosenberg self-esteem scale, the Pearlin mastery measure, depression measures, items measuring attitudes towards the roles of women in society, school satisfaction, and early (during adolescence) expectations regarding family, education and work. All of these inputs, along with a wide range of behavioral information, permit researchers to sort out cross-generational socio-economic predictors of non-normative child behaviors from other, perhaps more subtle, non-normative mother-to-child connections. Further, this research area can be enhanced by the large number of sibling pairs available in the child/young adult generation.

How Children Affect Mothers

For those interested in exploring possible feedback effects, the availability of repeated child measures of several behaviors and attitudes, in conjunction with detailed information about maternal behaviors, permits one to investigate the effect of changing child behaviors on subsequent actions of their mothers. For example, just as one can examine the impact of maternal employment on a child's behavior or changes in a child's behavior, one can also explore how changes in a child's behavior or health can affect subsequent maternal employment decisions.

Variation in High School and College Attendance

The NLSY79 data file includes in-depth information relating to both maternal/family and youth priors that are typically considered as significant analytical predictors of young adult schooling decisions and behaviors. As noted earlier, there has been a substantial increase in the number of youth who have passed through the primary college ages and hence censuring issues associated with very early school leaving have been reduced. As of the current survey round, the data set offers large samples of young adults past adolescence including fairly substantial minority samples. The child data collection includes a wide range of information asked of the mothers and children about school progression patterns, parent-school interaction processes, and child and parental satisfaction with the child's schooling. The young adult survey then collects a range of information about the high school completion process as well as college attendance. If a young adult has attended college since the date of last interview, the name of the current or most recent college attended is asked as well as the start and stop dates for attendance. Young adults in college are asked their major, whether they are part or full-time students, the cost of attending college, and if they have received degrees of any type. The young adult surveys ask questions about financial aid: whether or not a loan was received and, if so, the amount and how much of the years expenses it covered; whether or not the young adult received work study, scholarships, grants, fellowships, assistantships, as well as various other forms of help from government, institutions, friends, or family. Beginning in 2000, an additional series of questions has been included that ask about all colleges that have been applied to and whether or not the youth was accepted. Given that the college attendance questions have been asked repeatedly since 1994, the data now permit one to construct college trajectories for a relatively large population. These college profiles can be linked with early employment success, differentiating between white, black, and Hispanic youth.

These educational profiles can also be linked with early adult family-related activities. It is also possible to investigate the extent to which cognitive and socio-emotional tests administered in earlier waves are useful predictors of early career or family success, independent of the host of family factors known to be associated with child development.

Within-Family Differences in Outcomes

Because the NLSY79 Child sample is comprised of all children born to female respondents, many sibling clusters have been interviewed over time making it possible to explore the origins of differences between siblings in cognitive, emotional, and particularly behavioral outcomes. It is also possible to clarify the independent impact of differential family behaviors reflecting the reality that children from the same parents may nonetheless encounter different family processes due to variations in their parents' life cycle stages or sibling placement or gender.

A substantial number of children have one to three siblings, and the bulk of these siblings are now of young adult age (see Tables 5 and 6 in Sample Design). In addition to sharing many common data elements, siblings also have unique background characteristics. In the NLSY79, the HOME scale can provide insights into variations in child raising patterns by child parity, gender, or other characteristics such as health status. Many of the older children and young adults are within two or three years of each other in age so their outcomes, such as employment or family attributes can be measured at approximately similar life cycle points. Of course, with every additional survey round, the heterogeneity of the sample increases, bringing more children who have been born to a wider age range of mothers into the young adult sample. It is therefore possible, for example, to explore how child-raising practices for individual young adults may be sensitive to the age of their mother at birth and how this may translate into considerable variation in adolescent and young adult behaviors for children in the same family.

Child Health and Child/Young Adult Outcomes

Over the years, the NLSY79 main, child, and young adult interviews have included a number of questions designed to measure various dimensions of the physical and emotional development of the children. Mothers have reported on limiting health conditions, utilization of health services, illness and accidents. This type of health related information can be linked with all the other developmental and behavioral information available about the child and his or her family. Table 2 summarizes the types of health variables found in the Child data files. Key health items from the Young Adult surveys are listed in Table 3.

For each child, the collection of health data essentially begins prior to their birth and encompasses a wide range of prenatal inputs about the mother's behavior and attributes (including weight gain during pregnancy and key facts about each pregnancy and birth), as well as summary health measures for the mother, including periodic body weight and height measures over her own NLSY79 time line.

During childhood, health information is obtained through maternal reports. As children age into the young adult years, they provide self reported health information. During the first year of life, a relatively intensive body of child health information is collected that consists of birth/early infant-connected health problems, as well as medical visitations for both sick and well care, including the nature of the contact and the need for hospitalization. During the first four years of life, mothers complete a motor and social development scale about each child (described in detail in the Child Assessments section of this users guide) that assesses the maturational development of the child compared to other children of the same age. Height and body weight are reported or measured at each interview point. Then continuing over the child/young adult's life, reports are given biennially on accidents, injuries, and hospitalizations. One advantage that this data collection provides over institutionally collected health data is that, because it is self-reported by mother or child, it avoids biases due to the fact that different race, class, and ethnic groups are differentially likely to actually contact appropriate health individuals or institutions.

Across the surveys, considerable ancillary information is also collected about each child/young adult that details limiting health conditions (with respect to school, peers, and work); emotional problems; and the utilization of specialized medical, dental, and psychiatric services. Much of this information has been collected for each child repeatedly, plus collected for each child in a family unit. Additionally, introduced in 2004, a detailed retrospective on asthma incidence and prevalence was asked of all mothers, children and young adults. This makes it possible to link asthma histories with factors such as body weight and height over the years, a variety of potential illnesses, maternal and youth smoking, maternal activity limitation due to asthma, and a detailed geographic residence profile, all of which can help sort out possible determinants of child/young adult asthma.

The overview of research applications is not intended to be all-inclusive, but rather to suggest possible ways to approach the data, and to illustrate that there are many research questions waiting to be explored. Researchers interested in using the data are encouraged to draw on a variety of other NLSY79 informational sources including the other sections of this users guide, the NLSY79 User's Guide, public use Child documentation, and in particular, the Child and Young Adult questionnaires.