Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
96 INTEGRATING FEDERAL STATISTICS ON CHILDREN Shereshefsky, P., and L. Yarrow 1973 Psychological Aspects of a First Pregnancy and Early Postnatal Adaptation. New York: Raven Press. Sigel, I.E., ed. 1985 Parental Belief Systems. Hillsdale, N.J.: Erlbaum. Simmons, R.G., and D.A. Blyth 1987 Moving into Adolescence: The Impact of Pubertal Change and School Context. New York: Aldine de Gruyter. Smeeding, T.M., and B.B. Torrey 1988 Poor children in rich countries. Science 242:873-877. Snow, C.E. 1983 Literacy and language: relationships during the preschool years. Harvard Educa- tional Review 53(2):165-189. Spencer, M.B., G.K. Brookins, and W.R. Allen 1985 Beginnings: The Social and Affective Development of Black Children. Hillsdale, N.J.: Erlbaum. Spencer, M.B., and S. Dornbusch 1990 American minority adolescents. In S. Feldman and G. Elliot, eds., At the Thresh- old: The Developing Adolescent. Cambridge, Mass.: Harvard University Press. Stafford, F.P. 1987 Womenâs work, sibling competition, and childrenâs school performance. Ameri- can Economic Review 77:972-980. Stinnett, N., and Defrain 1985 Secrets of Strong Families. Boston: Little, Brown, and Co. Swithart, J. 1988 Characteristics of Strong Families. Unpublished paper, International Family Cen- ter, Logos Research Institute. Teachman, J.D. 1991 Who pays? receipt of child support in the United States. Journal of Marriage and the Family 53(3):759-772. Thomson, E., and S. McLanahan 1993 Family structure and child well-being: Economic resource versus parental behav- ior. Presented at the Annual meeting of the American Sociological Association. Washington, D.C., August. Timmer, S., J. Eccles, and K. OâBrien 1985 How children use time. In F. Thomas Juster, and F. Stafford, eds., Time, Goods and Well-being. Ann Arbor, Mich.: Institute for Social Research. U.S. Bureau of the Census 1989 Transitions in income and poverty status: 1984-1985. Current Population Re- ports, Series P-70, No. 15-RD-1. Washington, D.C.: U.S. Department of Com- merce. U.S. Department of Health and Human Services 1992 Child Health USA â92. DHHS Publication No. HRSA-MCH-92-6. Washington, D.C.: U.S. Government Printing Office. Watts, H.W., and D. Hernandez 1982 Child and Family Indicators: A Report with Recommendations. Report of the Advisory Group for Child and Family Indicators of the Advisory and Planning Committee on Social Indicators. Washington, D.C.: Social Science Research Council.
CHILD DEVELOPMENT AND FAMILY AND COMMUNITY RESOURCES 97 Weinraub, M., and B.M. Wolf 1983 Effects of stress and social supports on mother-child interactions in single- and two-parent families. Child Development 54:1297-1311. Werner, E.E., and R.S. Smith 1982 Vulnerable But Not Invincible: A Longitudinal Study of Resilient Children and Youth. New York: McGraw Hill. Wilson, J.B., D.T. Ellwood, and J. Brooks-Gunn 1995 Welfare to work through the eyes of children: the impact on parenting of move- ment from AFDC to employment. In P.L. Chase-Lansdale and J. Brooks-Gunn, eds., Escape from Poverty: What Makes a Difference for Children? New York: Cambridge University Press. Wilson, W.J. 1987 The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy. Chicago: University of Chicago Press. Wirth, L. 1956 The Ghetto. Chicago: University of Chicago Press. Yamaguchi, K., and D.B. Kandel 1984 Patterns of drug use from adolescence to young adulthood: II. Sequences of progression. American Journal of Public Health 74(7):668-672. Zaslow, M.J., K.A. Moore, N. Zill, and M.J. Coiro in press Implications of the JOBS Program for Children. Washington, D.C.: Urban Insti- tute Press. Zigler, E., and S. Muenchow 1992 Head Start: The Inside Story of Americaâs Most Successful Educational Experi- ment. New York: Basic Books. Zill, N. 1983 Divorce, marital conflict, and childrenâs mental health: research findings and policy recommendations. U.S. Senate, Committee on Labor and Human Resources. Broken Families: Hearings Before the Subcommittee on Family and Human Ser- vices, Washington, D.C.: U.S. Government Printing Office.. Zill, N., ed. 1989 Research and Policy Uses of Federal Data on Children and Families. Recommen- dations from the Second Interagency Conference on Child and Family Statistics, Washington, D.C., March. Zill, N., J.L. Peterson, and K.A. Moore 1984 Improving national statistics on children, youth, and families. A report on recom- mendations made at the Interagency Conference on Child and Family Statistics, September.
98 INTEGRATING FEDERAL STATISTICS ON CHILDREN Childrenâs Transition to School Sandra L. Hofferth The importance of caring for and preparing children well during their preschool years so that they can learn and grow emotionally into happy, healthy, and productive adults has been well documented (Copple et al., 1993). Although particular policy concerns seem to come and go like the latest movie, there are several that are not new and appear to be lasting. Aggressive, violent, and antisocial behaviors by children shock us almost daily and crime and violence constitute the number one public issue. The increasing proportion of unmarried teenagers bearing children while still in high school is an anomaly in our highly technological society and sets us apart from the European nations. The proportion of children who do not develop the skills to be able to function adequately in this increasingly complex world continues to concern us. We suspect, although we do not know, that antisocial and precocious behavior among children living in poor families, including delinquency, violent behavior, early sexual activity and out-of-wedlock childbearing, and school problems do not suddenly emerge full-grown at adolescence. Rather, many of the diverse problems mani- fested in adolescence are suspected to have common antecedents in child- hood behavior problems and may have common origins. In addition, skills developed in childhood can make the difference between productive, self- sufficient adults and adults who remain dependent. Because skills are cu- Sandra L. Hofferth is at the Institute for Social Research, University of Michigan. 98
CHILDRENâS TRANSITION TO SCHOOL 99 mulative, failure to develop these skills early makes it hard to keep up with oneâs classmates in school. Childrenâs readiness for school has been conceptualized as âthe capac- ity to engage actively in the learning processâ (Copple et al., 1993) and as âan emerging facility to experience and shape oneâs environmentâ (National Task Force on School Readiness, 1991). Whether children arrive on their first day and every day thereafter at their fullest potential depends on fac- tors that predate their birth, including prenatal care and maternal health habits, access to health care and exposure to high-quality care and pre- school programs during their preschool years, a warm and stable family, a safe and supportive community, as well as an engaging and responsive school environment (Copple et al., 1993). Thus the transition from pre- school to school encompasses the period from gestation through the primary grades. This paper takes as its focus, therefore, children from birth (or during gestation) to about age 12 or 13, the end of the elementary school years. The paper first outlines the basic scientific issues that need to be ad- dressed by those concerned with school readiness and the transition to school. Second, it lists the federal statistical data sources that are currently avail- able and how they address these basic issues. Third, it reviews efforts in the planning stages. Fourth, it addresses gaps in federal data collection efforts, taking into account efforts either under way or planned. A set of approaches to addressing the gaps follows. Finally, the potential role of the National Research Council in integrating and coordinating federal statistical efforts is discussed. BASIC SCIENTIFIC ISSUES/THE BIG QUESTIONS How Well Are Children Doing? Documenting how well or poorly children are doing is valuable, be- cause it raises public awareness about these problems and leads to support for interventions. Probably the most commonly addressed question is: How well are children doing? Developmental outcomes are often categorized into three groups: (1) cognitive, (2) socioemotional, and (3) physical health (see for example, Entwisle, 1992; Powell, 1992). Goals 2000 has added (4) approaches to- ward learning and (5) language usage to the set of categories (National Education Goals Panel, 1993).
100 INTEGRATING FEDERAL STATISTICS ON CHILDREN Cognitive Cognitive domains include physical knowledge about objects, relational knowledge, and social conventional knowledge (National Education Goals Panel, 1993). Under this rubric I also include verbal and written language and approaches to learning. Some of the more common measures of cogni- tive domains include the Peabody Picture Vocabulary Test, a test of recep- tive language, scores on math and reading achievement tests, and scores on tests measuring memory and attention. Learning predispositions may differ by gender, temperament, and cultural patterns and values. Learning styles include curiosity, task persistence, reflection, and imagination. Tests mea- suring temperament are available (Baker et al., 1993). Although I am not aware of tests, I have seen teacher ratings used to measure learning styles. Finally, as an alternative measure of cognitive outcomes, behavioral mea- sures of progress and success in school, such as retention in grade, special education placement, grades, and school dropout, are used. Socioemotional Socioemotional dimensions of development can be divided into emo- tional development and social development (see National Education Goals Panel, 1993). Self-concept is a key aspect of social and emotional develop- ment. âSelf-concept consists of the traits, habits, abilities, motives, social roles, goals, and values that define how we perceive ourselves (National Education Goals Panel, 1993:14).â The ability to express feelings construc- tively is an important aspect of emotional development. Social interactions with adults and peers are crucial aspects of social development. One widely used measured of socioemotional development, the Child Behavior Prob- lems Index, captures both social adjustment and emotional problems and has been widely used. Many of these behaviors are negative; positive as- pects of childrenâs behavior need to be defined and measured as well. Re- searchers have recently become interested in the development of moral judgments in children. However, such concepts are not yet well defined. Health Health includes physical development and physical abilities (National Education Goals Panel, 1993). Young childrenâs height and weight can be measured against national norms on growth charts. Measures of birthweight, gestational age, and birth length relative to the norm are also important indicators of health (Korenman et al., 1994). Physical abilities include gross motor skills, fine motor skills, oral motor skills, and functional per- formance. Tests have been developed to measure the development of chil-
CHILDRENâS TRANSITION TO SCHOOL 101 dren from birth through school entry. This should become an even more important area. In order to obtain federal funds, special needs children are required to be identified and served at earlier and earlier ages. However, since motor development proceeds at varying and uneven rates, it may be difficult to determine the long-term implications of relatively small delays in mastering these skills. Good health is generally identified by the lack of poor health: the lack of activity limitations, accidents, untreated chronic conditions. More work is needed to define physical fitness and good health. For example, good health habits and knowledge about risky behaviors such as smoking, drink- ing, and unprotected sexual intercourse are important. Assessment Age An important question is the earliest age at which one can first assess childrenâs development. The Peabody Picture Vocabulary Test, for ex- ample, a measure of receptive language, is not designed to assess children before age 3. Although there are tests to measure language skills earlier, there is evidence that early cognitive assessments may not be as reliable, valid, or stable as later ones (Baker et al., 1993). Health and nutritional status may be better measured before age 3. For example, recent work (Korenman et al., 1994) finds that measures of early child health are related to cognitive assessment scores and to behavior problems. Consequently, researchers often limit assessments prior to age 3 to physical health, motor skills, and nutritional status. What Are the Inputs Children Receive? In order to be able to succeed, children need access to resources and servicesâgood preschools and child care providers, good schools and teachers, developmentally appropriate curricula, families with adequate incomes, good health care, and so on. Thus the next question is: What are the inputs/ resources children receive? One concern is whether children have access to safe, nutritious, and healthy environments. What are maternal habits and life style before and after birth? Does the child receive the proper immunizations and well-child visits prior to school (Zill and Schoenborn, 1990)? A second concern is about the decline in exclusive care by parents and increased use of substi- tute care for even very young children (see Howes, 1992). As mothers are increasingly employed outside the home, is substitute care adequate in qual- ity for childrenâs optimal development? A third concern is about the de- cline in care by both mothers and fathers during the crucial early years of a childâs life due to increasing out-of-wedlock childbearing and high levels of
102 INTEGRATING FEDERAL STATISTICS ON CHILDREN divorce and separation (see, for example, McLanahan, 1985, 1988; Bumpass, 1984; Hofferth, 1985). Do children receive adequate attention from both parents? A fourth concern is the increased proportion of children raised by very young, abusive, immature, unprepared, or poverty-stricken mothers with poor prospects (see, for example, Moore and Snyder, 1991; Wolock, 1981; Zill et al., 1995). Do such mothers raise children who will repeat the cycle of early childbearing, poverty, and dependence? A fifth concern is growing up in poor neighborhoods with dismal and dangerous schools (see, for example, Nettles, 1992; Karweit, 1992). To what extent do neighbor- hoods affect the ability of children to develop into competent, self-suffi- cient adults? One line of research is to examine factors that increase the risk of failure. These, often called risk factors, are associated with lower success, but the mechanisms are not described. Research, again, is only suggestive, but outcomes may be linked to the economic and social conditions in which families live and the level and instability of resources available to them. These include family characteristics such as low income, minority race/ ethnicity, being raised by a single mother or no parent, low birthweight, having a teenage parent, low level of maternal education, English not spo- ken at home, large family size, later birth order, low maternal ability, and low child ability (Hofferth et al., 1994). They also include neighborhood characteristics, such as the proportion of families in the neighborhood who are poor or female-headed and the proportion who are middle class (Duncan et al., in press). How Do Early Experiences Affect Later Well-Being? In order to be able to recommend solutions, we need to address the key scientific issue: How do these inputs relate to later success or failure in achieving normal physical, cognitive, and psychosocial development? Ba- sic scientific researchers are concerned with two processes: (a) the mecha- nisms or paths (mediating factors) whereby children become successful or fail and programs work or do not work (Entwisle, 1992) and (b) the moder- ating or contextual factors, such as race/ethnicity, family and kin factors, and neighborhood (Nettles, 1992), that interact with the characteristics of children to produce varied outcomes. Mediators Mediating factors explain the relationship between low income, for ex- ample, and child outcomes. They include family parenting style, communi- cation, attitudes, and beliefs (Beavers and Hampson, 1992; Powell, 1992); early childhood program participation and program characteristics (Howes,
CHILDRENâS TRANSITION TO SCHOOL 103 1992); adequate nutrition, health care, and immunizations (Shonkoff, 1992); and school engagement, including appropriate attendance, completion of homework, and interest in school (Karweit, 1992). Moderators Moderating factors affect the relationship between inputs and outputs; they are also known as interacting factors and define subgroups for separate analyses. For example, the relationship or process may differ between boys and girls, blacks and whites, and older and younger children (Mott and Menaghan, 1993). The process may vary depending on individual parent characteristics such as warmth, on family characteristics and resources, on child temperament, on family and kin support, and on neighborhood and school characteristics. Some subgroups in which there is special interest are children cared for by foster parents or other nonparent relatives, children with disabilities, and children of non-English-speaking parents. The latter are especially impor- tant given the debate over Chapter I funding for schools with disadvantaged students (Fix and Zimmerman, 1993) and the new requirement for schools to serve children with special needs. Do these funds make a difference? What are the appropriate ways to serve such children? Another important group consists of children not living with a natural or adoptive parentâ those either in foster care or in the care of other family members. Again, this is an important group for policies such as Family Preservation and Support. Is it better to keep families together or remove them from dys- functional homes? Intervention/Prevention Research Once we think we know what resources children should be receiving (in that they have been linked to later well-being), we can design interven- tion programs. There are three important questions: 1. What interventions should be tried? This type of question is directly linked to the scientific questions raised earlier. Research shows a relation- ship between early behavior problems and later school problems and antiso- cial activity. The conclusion is that early aggressive behavior is associated with later school problems; therefore, develop programs to reduce aggres- sive behavior (Kellam, 1993). Another example is that early childhood programs develop childrenâs cognitive and social skills, improve health, and help parents; therefore, develop more early childhood education programs (Schweinhart et al., 1993). 2. How effective are they? This type of question is rarely addressed.
104 INTEGRATING FEDERAL STATISTICS ON CHILDREN The Perry Preschool Project is one of the few to measure effectiveness in dollar terms (Schweinhart et al., 1993; Select Committee on Children, Youth, and Families, 1985). This permits a comparison of the returns from other types of investments. Thus the final question: 3. What is the relative cost-effectiveness of different policy approaches/ interventions? For example, what is the relative payoff of: a. Increasing direct monetary investments in children, such as immuni- zations and Head Start, b. Increasing direct time investments in children, through tutoring and mentoring, c. Increasing indirect monetary investment in children through schools and services, d. Increasing direct monetary investment in families, such as the Earned Income Tax Credit, Aid to Families with Dependent Children (AFDC), and the Job Opportunities and Basic Skills (JOBS) programs, e. Changing parent behavior, such as education, social services, family planning, and f. Changing teacher behavior, through promoting developmentally ap- propriate activities? The only way to address these issues is through scientific analyses that compare the relative effects of these different factors in one model. Conse- quently, it is important to have models that are as comprehensive as pos- sible, that is, that include all the potential mediators. FEDERAL STATISTICAL DATA SOURCES Although the need for information on children has been increasing, data systems that collect needed information on children have been deteriorat- ing. The National Maternal and Infant Health Survey program of National Center for Health Statistics will not be continued, and funds for a new Child and Family Health Supplement to the Health Interview Survey are in dan- ger. Table 1 shows the major scientific databases on children and the scientific questions to which the data are addressed: outputs, inputs, the link between the two, and program evaluation.
CHILDRENâS TRANSITION TO SCHOOL 105 Basic Scientific Issues How Well Are Children Doing? The major efforts at large-scale assessments of childrenâs cognitive abilities occur under the auspices of the National Assessment of Educa- tional Progress (NAEP). Budgeted at $29.3 million per year, NAEP under- takes major testing efforts to determine the competency levels of U.S. youth in reading, writing, math, science, U.S. history, and world geography. Whereas some information on time spent studying, reading, and watching television is collected, family demographic data are limited to those reported by stu- dents, and no information on school functioning has been obtained. Heretofore there has been no major effort to develop a major assess- ment of childrenâs socioemotional development. The National Institute of Mental Health, however, has conducted pilot studies in four sites to exam- ine the prevalence of disorders in children from age 4 to 17 (Methodologi- cal Epidemiological Study for Children and Adolescents or MECA). This effort is planned to lead to a major set of studies in several sites around the country of the prevalence of psychiatric disorders among children and the availability and use of services in the community. Assessment of child health has been conducted by the National Center for Health Statistics. As part of the National Health Interview Survey (NHIS), health data are collected routinely on children in the family. In 1993-1994, the National Health Interview Survey conducted a major Dis- ability Survey, including children as well as adults (Simpson et al., 1992). In 1981 and 1988 a complete child health supplement was collected. The 1988 supplement included data on some 17,000 children from birth to age 17. Several reports have been produced using these data, and they have served as a major source of information on child health in the United States (e.g., Zill and Schoenborn, 1990). A new child health supplement has been tentatively planned for 1996; however, its funding situation is precarious. What Inputs Are Children Receiving? Because of government monitoring of programs, surveys for purposes of identifying who is receiving specific services are common. Of these services, early childhood programs are probably the most adequately sur- veyed. Studies that have obtained information on enrollments of children and about their preschool and early childhood settings include the National Child Care Survey 1990 (Hofferth et al., 1991), the 1991 National House- hold Educational Survey (Brick et al., 1992), the Profile of Child Care Settings (Kisker et al., 1991), the National Transition Study (Love et al., 1992), the Observational Study of Early Childhood Programs (Layzer et al.,
TABLE 1 Databases Relevant to the Transition to School 106 Outputs Link Inter- Socio- Outputs/ vention/ Database Health Cognitive Emotional Inputs Inputs Evaluation National Assessment of Educational X X X Progress (NAEP) Methodological and Epidemiological Study X for Children and Adolescents Child Health Supplements (81,88) to the Health X X X X Interview Survey (CHS) National Maternal & Infant Health Survey (NMIHS) X X X X National Household Education Survey (NHES) X X X X National Child Care Survey 1990 (NCCS) X A Profile of Child Care Settings 1990 (PCS) X Observational Study of Early Childhood Programs - X X X X X Vol. 1 Observational Study of Early Childhood Programs - X Vol. 2 National Study of Before & After School X INTEGRATING FEDERAL STATISTICS ON CHILDREN Programs (NSBASP)
Survey of Income & Program Participants (SIPP) X Child Care Supplement X Child Well-being Module X X X X NCHS Immunization Surveys X WIC X Child Care Food Programs (CCFP) X Michigan Time Use Study 1975-76 & 1982 (MTSU) X X X CHILDRENâS TRANSITION TO SCHOOL Consumer Expenditure Survey (CES) X NICHD Early Child Care Study X X X X X National Longitudinal Survey of Youth (NLSY) X X X X X Mother Child Supplement Panel Study of Income Dynamics (PSID) X X X National Survey of Families & Households (NSFH) X X X X X JOBS Child Outcomes Study (JOBS) X X X X X X Prospects X X X X X X Head Start - Public School Transition Study X X X X X X National Transition Study X 107
108 INTEGRATING FEDERAL STATISTICS ON CHILDREN 1993), the National Study of Before and After School Programs (Seppanen et al., 1993), and the Survey of Income and Program Participation child care supplement (OâConnell and Bachu, 1992). The National Household Educa- tion Survey is planning an early childhood education component in its 1995 data collection wave. I do not know of a survey comparable to the Profile of Child Care Settings or the National Study of Before and After School Programs that would obtain detail on what happens in schools (although the âProspectsâ study of the Department of Education has a small observational component). A prototype would be the Schools and Staffing Survey, but with observations in classrooms. The National Center for Health Statistics has been conducting state-by-state immunization surveys to monitor level of immunizations and regular surveys to monitor health promotion and dis- ease prevention efforts for national health goals. The National Maternal and Infant Health Survey also asked questions about services (Zill and Daly, 1993). Studies have been conducted on coverage of the Supplemental Food Program for Women, Infants, Children (WIC), for example (Kotelchuck et al., 1984), and on the Child Care Food Program (Glantz et al., 1988). Another important set of inputs comes from parents. What are children receiving in terms of care and attention from parents? What amount of time and how it is spent are questions that can adequately be addressed only through a time-budget interview that obtains a complete accounting of time spent during the day. Stylized time accounting, which asks separate ques- tions about time spent in specific activities, is not as accurate. The 1975 and 1976 Michigan Time Use Studies are landmarks in obtaining detailed infor- mation on family membersâ time use. In 1981-1982, family members were once again interviewed and a time budget diary for children ages 3 to 17 was obtained (the information for children ages 3 to 5 came from parents) (Juster and Stafford, 1985; Timmer et al., 1985). The activity categories in these diaries are comprehensive. A teacher was contacted and assessments of the childâs cognitive development were obtained in 1982 (Stafford, 1987). There is no comparable study currently being planned, as far as I know. The National Household Education Survey is planning a parent component in the 1996 wave. It is unlikely to use a diary approach due to its multiple objectives and limited time per household. Several studies have attempted to define the amount of financial contri- bution parents make to their childrenâthe âcostâ of children. Data gener- ally come from the Consumer Expenditure Surveys (e.g., Espenshade, 1984). Finally, the Census Bureau is planning on adding a child well-being topical module to wave 6 of the 1993 panel of the Survey of Income and Program Participation (to be fielded in winter 1994). This primarily obtains information on parental inputs to children (child care, time spent). How- ever, for older children it proposes to ask parents about grade repetition and school suspensions, teacher conferences due to a problem the child was
CHILDRENâS TRANSITION TO SCHOOL 109 having, and whether the child ever stayed out later than permitted. Health information will be obtained for all children. What Is the Link Between Inputs and Outputs? Although detailed data collection efforts have been made to monitor child well-being and to obtain information on inputs to children, the number of efforts that permit linking inputs with outputs is limited. Although all agencies benefit, only the National Institutes of Health, the National Sci- ence Foundation, and the National Center for Education Statistics (in the Department of Education) have these basic scientific questions as their mandate. Both the National Institutes of Health and the National Science Foundation depend on investigator-initiated projects and have severely constrained bud- gets. The Department of Education is limited primarily to contractual work; consequently, its flexibility is constrained, since contract work requires that the government define the scope of work as specifically as possible prior to requesting bids. Although this makes sense for some things, it does not meet the needs of scientific inquiry. There are currently only two national surveys that follow children from preschool-age into the school years: the mother-child supplement to the National Longitudinal Surveys of Youth (NLSY-MC) and the Panel Study of Income Dynamics (PSID). The National Institute of Child Health and Human Development (NICHD) funds the mother-child supplement to the NLSY, a Department of Labor survey. The mother-child supplement is a biennial survey, beginning in 1986, of the children of a nationally representative sample of women ages 14 to 21 when they were first interviewed in 1979. The children are as- sessed beginning at age 3 and interviewed directly beginning about age 10. Additional information about them is obtained from their mothers and from self-administered questionnaires. The limitations of the study are due pri- marily to the basic design of the Department of Labor survey rather than to the design of the child supplement. First, children become part of the survey as they are born. Thus, the older children are children of early childbearers and the younger children are children of later childbearers. As of 1990, it was estimated that the children represent about three-fourths of the childbearing of this age group of mothers. As the children age, the NLSY will eventually include children of older and younger childbearers at each age. Unfortunately, they will represent considerably different time periods and therefore may or may not be comparable. In addition, since the children are reached through their mothers, there is almost no informa- tion on the fathers of the children; however, there is more information on contact with absent fathers than with those in the household. The richness of the NLSY lies in its assessments, along with an exten-
110 INTEGRATING FEDERAL STATISTICS ON CHILDREN sive set of background variables. The NLSY contains a basic set of assess- ments of cognitive, socioemotional, and physical health measures, which it has kept over the years for comparability with previous waves. Since 1986, researchers have broadened the measurement of social and emotional as- pects of child functioning, but these advances are not represented in the NLSY. In addition, little information on schools and school progress is obtained. The major drawback is that mediating factors are limited. A modified version of the Home Observation of the Environment (HOME) scale was obtained for each wave. The mental health of the mother was assessed at irregular intervals, which makes it impossible to relate to the childâs well- being. In addition, almost no attitudinal information was obtained from the mother. Questions about drug and alcohol use and criminal behavior of the mother were asked only on a few occasions (although they were asked again in 1994). Assessments of the marital relationship were obtained in 1988 and 1992 only. More regular assessments would be quite helpful in linking inputs with child outcomes. As of the 1992 wave, four waves of informa- tion are available on the children, and 13 years of information are available on their mothers. Thus this survey provides a rich source of information on relatively short-term outcomes for children of income, marital status, and other family behaviors. The Panel Study of Income Dynamics is rich at the points the NLSY is poor, in that it has detailed information throughout on both parents in households. In 1968 about 5,000 U.S. households were first interviewed. All members of these households and all the children, grandchildren, and great-grand- children from these original sample members continue to be followed annu- ally up to the present. More than 25 years later, about half the original sample is still being followed. Because of split-offs, in 1988 about 7,000 families were being interviewed. Since there is no mechanism for incorpo- rating immigrants and thereby to reflect the changing composition of the U.S. population, a nationally representative Latino sample was first inter- viewed in 1990 and interviewed annually thereafter. Though children are not the focus of the PSID, information on the number and ages of children in each household is obtained annually. The PSID has rich information on inputs such as family income, wealth, and labor force participation during the time the child is in the family. Moder- ating factors are also rich. Information on parents, family resources, and neighborhoods is available through supplements, such as on kin exchanges in 1980 and 1988, and links to the decennial censuses. Child characteristics are not available before age 16, nor were family process measures obtained. However, at age 16 and after, school enrollment and completion informa- tion becomes available. Long-term outputs after the child becomes an adult are also available, since sample children who leave home become respon-
CHILDRENâS TRANSITION TO SCHOOL 111 dents and report on their own households and well-being. The richness of the data lies in its intergenerational linkages. That is, detailed information on the families of each child is available for 25 years. Thus the data are ideal for the analysis of long-term consequences of family income and liv- ing arrangements for children. The National Survey of Families and Households was a nationally rep- resentative sample of 13,000 households conducted in 1987. Detailed infor- mation about one focal child in each household was obtained. The first follow-up was begun in 1992. Focal children ages 5 to 18 in 1987 were interviewed by telephone in 1992-1993, when they were 10 to 23. Very little retrospective information was obtained about experiences prior to age 5. Thus this study will not be useful to follow the transition from preschool into the school-age years. However, given the richness of information col- lected on all household members, it may be worthwhile to create a child- based file to follow focal children over the five years of the survey to date. Several studies have collected data longitudinally, but only during the first several years of a childâs life. For example, the National Maternal and Infant Health Survey represents a national sample of women who had a live birth, fetal death, or infant death in 1988. The 1991 follow-up interviewed the families about their children 3 years after the first interview. Substantial family demographic information was obtained in this survey, including variables of interest to researchers who wish to link inputs and outcomes over the first three years. The NICHD Study of Early Child Care originally intended to follow children for the first 2.5 years (NICHD Early Child Care Network); it now intends to follow them into first grade. This will be an enormous benefit to researchers; however, since this study is located in various sites across the country, it will not be nationally representative. Evaluation/Intervention Studies A considerable number of intervention programs have been evaluated. The best-known example is the Perry Preschool Project (Schweinhart et al., 1993). The data from most studies, generally small and select populations, are rarely made publicly available for secondary analysis. However, three new studies promise to provide information on the transition to school and may become available to researchers: (1) Prospects, (2) the JOBS child outcomes study, and (3) the evaluation of Transition to School programs by the Administration for Children and Families. A major assessment of the cognitive skills and abilities of children, particularly those from low-income families, is being conducted by the De- partment of Education (Puma et al., 1993). This study, called Prospects: The Congressionally Mandated Study of Educational Growth and Opportu-
112 INTEGRATING FEDERAL STATISTICS ON CHILDREN nity, will collect detailed assessments and other information for 5 years on some 30,000 students across the United States in grades 1, 3, and 7 for the purpose of evaluating the long-term effects of exposure to Chapter I ser- vices. Consequently, students in low-income schools have a higher prob- ability of being included. Immigrant and non-English-speaking students are included in this study. An achievement test âSABEâ is available in Spanish for those who prefer it, and questionnaires are also available in Spanish. Though tests and questionnaires in non-Hispanic languages are not avail- able, the study is collecting transcripts and other information on all stu- dents. Unfortunately, because grade 1 is the first contact, very little infor- mation will be available on the transition to school. Although I do not know the actual cost of Prospects, about $9 million is budgeted for the evaluation of Chapter I programs in fiscal 1995 (Office of the President, 1994). The JOBS child outcomes study, being conducted at three sites of the JOBS evaluation, interviewed children ages 3 to 5 at baseline, when their mothers, at intake into the program, were randomly assigned to one of three intervention groups: (1) basic educational activities, (2) employment ac- tivities, and (3) a control group. Information on the mothers and their children was obtained at baseline and will be collected 2 years after baseline when the children are ages 5 to 7. In both waves, information on the parent- child relationship, maternal well-being and characteristics, child care ar- rangements, and family and home characteristics, is being obtained, as well as measures of school readiness and achievement. An observational study is being conducted at one of these sites as well. The Administration for Children and Families has a set of demonstra- tion projects providing Head-Start-like services to assist in the transition of former Head Start children to school. Sharon and Craig Ramey, Civitan International Research Center, Birmingham, received the contract to evalu- ate these transition projects, the Head Start/Public School Transition Study; there are 32 demonstration projects in 31 states and the Navajo nation (Ramey and Ramey, 1993). The national evaluation project has conducted inter- views with parents, teachers, and principals, along with assessments of chil- dren and classroom observations in the fall of the kindergarten year of 2 cohorts of former Head Start and non-Head Start children in randomly as- signed demonstration schools and comparison schools (fall 1992 and fall 1993). All children in demonstration schools are eligible to receive ser- vices. All former Head Start children, non-Head Start children in demon- stration schools, and a subsample of non-Head Start children in comparison schools in the first cohort are being followed up through at least the end of first grade in spring 1994. The original plan was to follow these children through the end of third grade; however, funding has not been provided for the third grade follow-up. Reauthorization is currently being considered.
CHILDRENâS TRANSITION TO SCHOOL 113 The cost of the 32 demonstrations plus the evaluation is approximately $25- 26 million per year, with about $24 million for the demonstrations and the remainder for the evaluation. EFFORTS IN THE PLANNING STAGES There are four studies that will potentially fill some of the gaps identi- fied above: (1) the Program on Human Development and Criminal Behav- ior of the National Institute of Justice and the McArthur Foundation, (2) the Multisite Study of Mental Health Services of the National Institute of Men- tal Health, (3) the Early Childhood Longitudinal Survey of the Department of Education, and (4) the Survey of Program Dynamics of the U.S. Bureau of the Census. The Program on Human Development and Criminal Behavior plans to follow 8 age cohorts of children and youth for 8 years each (Earls, 1993). The youngest cohort will consist of 1,000 females and 1,000 males begin- ning at age 2 or 3 and will be followed for 8 years, to age 10 or 11. A second cohort of 500 males and 500 females will begin at age 3, and again be followed for 8 years. Thus these two cohorts will provide a good picture of transition to school. The study will be conducted in 77 areas of Chicago. The outcome variables are antisocial behavior, criminality, and substance abuse. The types of explanatory variables are quite comprehensive and include most of those mentioned above. Applications for a cooperative agreement for a Multisite Study of Men- tal Health Service Use, Need, Outcomes, and Costs in Child and Adolescent Populations have been requested. It is expected that the study will be con- ducted in several sites, including a national site, with potential follow-ups for several years. Its purpose is, first, to assess the mental health of chil- dren, as do studies of adults. Its second objective is to obtain information on the services such children receive. The National Institute of Mental Health (NIMH) conducted methodological studies of diagnostic assessments of children 4-17 (MECA); however, the assessment of children under age 4 is untested, as far as I can tell. When these assessments have received widespread validation, they should become quite useful for other surveys. The question is how they are to be validated. What are the outcomes of interest? Do these tests predict substantial problems later on? This is especially important if very young children are to be assessed, since there is apparently considerable instability of these tests measured at an early age. This series of studies links well to the Individuals with Disabilities Educa- tion Act, which requires services be provided to young children before entry into school. However, it is important to assess the relationship be- tween scores on diagnostic tests at an early age with outcomes. Some of these measures may have a relationship to later problems; others may not.
114 INTEGRATING FEDERAL STATISTICS ON CHILDREN The usefulness of these diagnostic tests is clearly important in something as potentially costly as this new legislation appears to be (Woodward and Weiser, 1994). The Department of Education has awarded a planning contract for an Early Childhood Longitudinal Study (ECLS), with the full-scale survey planned for 1998 at the earliest. This study would follow a kindergarten cohort of children as they move into regular school to assess factors associated with normal development; it would focus on all the outcome areas described above. Substantial retrospective information prior to kindergarten would be obtained, but assessments would begin only in kindergarten.1 The Census Bureau has proposed a new survey: the Survey of Program Dynamics. Its objectives are to provide data showing dynamic changes in participation in welfare, health, education, and employment and training programs for 1993-2002 and in childrenâs well-being, and to provide infor- mation to analyze causes and consequences of program participation. The sample would be drawn from the 1993 panel of the Survey of Income and Program Participation, and, when combined with SIPP data collected begin- ning in 1996 and ending in 2002, would provide a 10-year panel. GAPS IN DATA COLLECTION EFFORTS Only one national study currently in the field collects data concurrently from parents and their children both before and after entry into school and has the potential of addressing the scientific questions listed aboveâthe NLSY-MC. This dataset has been very heavily used for this reason. A number of gaps can be identified in the data collection efforts cur- rently under way or planned: First, there is currently no longitudinal study of a nationally representative sample of children starting prior to school entry. Each of the studies that is planned or under way has either a limited age range or a sampling design that limits its generalizability. The Depart- ment of Education Early Childhood Longitudinal Study (ECLS) starts with children already enrolled in kindergarten. The National Institute of Justice Study will be conducted in a single urban area. The NICHD study of Early Child Care was conducted in several sites across the country, but is not nationally representative. In 1991, the follow-up of the National Maternal and Infant Health Survey obtained detailed information on the first three years of a birth cohort of children; however, there are no plans to continue this study. Second, no study will permit state-level measures, which are becoming increasingly important as much of the policy focus shifts to the states. Third, no survey proposes to collect time-use data. This is a critical need that no proposed effort will address. Fourth, although surveys have incorporated the capacity to follow, in-
CHILDRENâS TRANSITION TO SCHOOL 115 terview, and assess Spanish-speaking students, no survey has the ability to assess and follow Asian-language students over time. Fifth, the studies currently in the field are relatively small. A survey is needed with a large enough sample to be able to identify and follow chil- dren with special needs and children not living with a parent. APPROACHES TO ADDRESSING DATA GAPS There is no single ideal dataset. Different researchers have different purposes. Whereas some overlap is possible and desirable, the level of financing necessary is likely to preclude one dataset suiting everyoneâs needs. In addition, there is no such thing as unlimited resources; tradeoffs must be made, and the mission of the funding agency is always of prime concern. Consequently, the idea of multiple data collections appears desir- able. Thus my first suggestion is to build on datasets already in existence by supplementing them with additional questions and assessments. Supplement Ongoing Data Collections Given the availability of 25 years of detailed information on families contained in it, supplementing the PSID would be my first priority. Prelimi- nary estimates place the cost of a one-time data collection effort that in- cludes age-appropriate child assessments at about $2.5 million. Collecting more limited data about children on a regular basis (a 8-minute supplement with questions asked of the parent) would cost on the order of $700,000 per year. My second priority is to supplement the NLSY with more regular infor- mation on the mental health, criminal behavior, and drug and alcohol use of the child and parents. More information on fathers is needed on a regular basis. And it is important to make sure that the same questions are asked on every supplement. For example, many items were omitted from the 1990 survey because of funding problems; this produces gaps in the data and nightmares for analysts. A rough estimate of the cost of supplementing the NLSY in 1994 dollars is about $55,000 per minute for questions for all respondents. Asking questions of mothers only, such as about her child, would be about half this cost. Asking questions directly of children age 10 and older is considerably less expensive (perhaps one-quarter of the cost) because of the smaller sample size. Given the enormous amount of information collected in the National Maternal and Infant Health Survey, my third priority is either to follow the National Maternal and Infant Health sample of children into their school years or to start a new cohort at birth; it could become a birth cohort study comparable to the British longitudinal study (Cherlin et al., 1991). One of
116 INTEGRATING FEDERAL STATISTICS ON CHILDREN the reasons another follow-back to the NMIHS has not been pursued is that the response rate in the first year was only 74 percent, although the follow- back response rate was 90 percent. Another 90 percent response rate would produce a total response rate of about 60 percent. Apparently some analysis was conducted and on that basis additional follow-back was rejected.2 A new cohort that is smaller in size but followed for more years may be the cost-effective way to go. Coordinate Existing Data Collection Efforts Much more coordination and collaboration among institutions, so that there is at least some overlap between datasets, would allow more cross- study comparisons and validation than was previously possible. Asking several similar questions would be useful. For example, if the MECA study used the behavior problems index that is contained in the NLSY and com- pared the results in this shortened form with the results of other assess- ments, it would serve the field quite well; this type of cross-agency and disciplinary work would be very helpful. It would serve the NLSY well to include a high-quality diagnostic assessment of adults that has been used by NIMH regularly in its surveys. Regular assessment of substance abuse by parents in the NLSY would also be useful. Almost nothing is collected now on the mental health of mothers, which is a real omission, and there is only an occasional assessment of substance abuse. These measures could be ob- tained from NIMH, the National Institute on Drug Abuse, and the National Institute on Alcohol Abuse and Alcoholism. The four new surveys (men- tioned in an earlier section) would do well to meet and discuss potential collaborations and cross-survey work. Plan New Data Collections National Survey of Children A new survey of a representative sample of children of ages 0-17 is needed. Such a study would include a birth cohort and would permit analy- sis of all other age groups as well. The Census Bureau is interested in mounting such an effort, and its proposed Survey of Program Dynamics is a promising approach; however, I am not convinced that the planning and data collection of a new National Survey of Children should be a federal effort. Much of federal data collection does not allow enough planning time. In addition, a more flexible funding mechanism is neededâeither a request for applications or a collaborative agreement with an agency. The model might be the Health and Retirement Survey, which is being con- ducted by a consortium of private institutions with funding from the Na-
CHILDRENâS TRANSITION TO SCHOOL 117 tional Institute on Aging. The Census Bureau could be one of the members of the consortium rather than the major actor. (The Census Bureau does not permit the release of information on neighborhoods, for example, and ap- proval by the Office of Management and Budget is required.) A large chunk of such an effort could be conducted by the private sector with high response rates at considerable savings to the government. A National Time-Use Survey of Children and Their Parents Only one time-use study has been conducted, and results from this study are still being widely used (Carnegie Council, 1992). An entire sur- vey should be conducted in this format. Given the trade-offs in other survey efforts, there is almost no possibility of conducting a modified time- use study along with other types of questions. The time-use portion must be central. A ROLE FOR THE NATIONAL RESEARCH COUNCIL? The final question I address is the possibility of a role for the National Research Council in examining/promoting more effective strategies for col- lecting and assembling national data on children and their families. Disseminating the results of this workshop to all federal agencies and to the Office of Management and Budget is quite important. Agencies should be made aware that these interesting new studies are being planned. Al- though they may individually come to the attention of the Office of Man- agement and Budget and/or Congress during the budget process, seeing them together as a package makes it easier to see overlaps and gaps. In terms of promoting collaborative meetings, there is an immediate opportunity to get the relevant federal agencies together to discuss the four surveys currently being planned: the Program on Human Development and Criminal Behavior, the Multisite Study of Mental Health Services, the Early Childhood Longitudinal Study, and the Survey of Program Dynamics. Another set of meetings could be held to discuss supplementing exist- ing datasets, including the Michigan PSID, the NLSY mother-child sample, and the NMIHS. For a new National Survey of Children, there should be a collaborative effort originating from the field, rather a single agency effort. The first step might be a set of meetings with the interested agencies, re- searchers in the field, and funders. The second step would be to hold meetings with potential data collection firms, such as the National Opinion Research Center, WESTAT, the University of Michigan, and Temple Uni- versity. It is extremely important to have the scientific community involved in and committed to the design and development of such a study from the beginning.
118 INTEGRATING FEDERAL STATISTICS ON CHILDREN NOTES 1. An embedded substudy would interview a sample of Head Start children during the spring of the year before entering kindergarten. 2. The answer is not whether the response rate is low but whether the sample is in some way biased. Originally I recommended doing additional work to see whether the original sample is biased in a way that would reduce its value were it to continue to be followed. If it could be shown either to have minimal biases or if sampling weights could be constructed to reduce the known biases, then continuing to follow this cohort would be valuable. However, the statisticians do not agree with this and believe that a low response rate biases a survey, regardless of whether the responses are missing systematically or at random. The cost of one follow-up is on the order of $4 to 5 million, compared with the start-up cost of a new survey at an estimated $6-7 million. REFERENCES Baker, Paula C., Canada K. Keck, Frank L. Mott, and Stephen V. Quinlan 1993 NLSY Child Handbook, Revised Edition. Columbus, Ohio: Center for Human Resource Research, Ohio State University. Beavers, W. Robert, and Robert Hampson 1992 Family Variables Related to Childrenâs Capabilities and Interests in Education. Planning Paper for the National Center for Education Statistics Longitudinal Stud- ies of Young Children. Southwest Family Institute. Brick, Michael, Mary Collins, Carin Celebuski, Mary Jo Nolin, Theresa Squadere, Peter Ha, Jacqueline Wernimont, Jerry West, Kathryn Chandler, Elvie Hausken, and Jeffrey Owings 1992 National Household Education Survey of 1991: Preprimary and Primary Data Files Userâs Manual. Washington, D.C.: U.S. Department of Education. Bumpass, Larry 1984 Children and marital disruption: a replication and update. Demography 21(Febru- ary):71-82. Bumpass, Larry, and James Sweet 1993 Longitudinal Data on Families: The National Survey of Families and Households and the Design for the 5-year Reinterview. Madison: The University of Wiscon- sin. Carnegie Council on Adolescent Development 1992 A Matter of Time: Risk and Opportunity in the Nonschool Hours. New York: Carnegie Corporation. Cherlin, A.J., F.F. Furstenberg, et al. 1991 Longitudinal studies of the effects of divorce on children in Great Britain and the United States. Science 252: 1386-1389. Child Trends, Inc. 1993 The JOBS Child Outcomes Study, Overview Briefing. Washington, D.C.: Child Trends, Inc. Copple, Carol, Sharon Deich, Lorelei Brush, and Sandra Hofferth 1993 Learning Readiness: Promising Strategies. Washington, D.C.: U.S. Department of Health and Human Services. Duncan, G., J. Brooks-Gunn, and P. Klebanov in press Economic deprivation and early childhood development. Child Development.
CHILDRENâS TRANSITION TO SCHOOL 119 Earls, Felton 1993 The Program on Human Development and Criminal Behavior: Executive Sum- mary. Washington, D.C.: National Institute of Justice. Entwisle, Doris R. 1992 Recommendations for the NCES Longitudinal Study of Early Childhood/Early Education. Baltimore, Md.: Johns Hopkins University. Espenshade, Thomas 1984 Investing in Children: New Estimates of Parental Expenditures. Washington, D.C.: The Urban Institute. Fix, Michael and Wendy Zimmerman 1993 Educating Immigrant Children: Chapter 1 in the Changing City. Washington, D.C.: The Urban Institute Press. Glantz, Frederic B., Judith Layzer, and Michael Battaglia 1988 Study of the Child Care Food Program. Cambridge, Mass.: Abt Associates. Hofferth, Sandra 1985 Updating childrenâs life course. Journal of Marriage and the Family 47:93-115. Hofferth, Sandra, April Brayfield, Sharon Deich, and Pamela Holcomb 1991 National Child Care Survey, 1990. Washington, D.C.: The Urban Institute. Hofferth, Sandra, Jerry West, Robin Henke, and Phillip Kaufman 1994 Access to Preschool Programs among At-Risk Children. Washington, D.C.: U.S. Department of Education. Howes, Carollee 1992 Preschool Experiences. Paper prepared for the National Center for Educational Statistics. Los Angeles: University of California at Los Angeles. Juster, F. Thomas, and Frank Stafford, eds. 1985 Time, Goods, and Well-Being. Ann Arbor: University of Michigan. Karweit, Nancy 1992 Elementary School and Classroom Characteristics Related to Student Preparation and Progress Through School. Baltimore, Md.: Johns Hopkins University, Center for Social Organization of Schools. Kellam, Sheppard 1993 School-Based Prevention Research on Early Risk Behaviors: Implications for Head Start and Beyond. Presented at the 2nd National Head Start Research Conference, Washington, D.C., November. Kerckhoff, Alan C. 1992 Characteristics of Elementary Schools. Paper for the National Center for Educa- tion Statistics, Longitudinal Studies of Young Children, Duke University. Kisker, Ellen, Sandra Hofferth, Deborah Phillips, and Elizabeth Farquhar 1991 A Profile of Child Care Settings: Early Education and Care in 1990. Washing- ton, D.C.: U.S. Department of Education. Korenman, Sanders, Jane Miller, and John Sjaastad 1994 Long-term Poverty and Child Development in the United States: Results from the NLSY. Minneapolis: University of Minnesota. Kotelchuck, M., Schwartz, J., Anderka, K., and Finison, K. 1984 WIC participation and pregnancy outcomes: Massachusetts statewide evalution project. American Journal of Public Health 74(October):1086-1092. Layzer, Jean I., Barbara D. Goodson, and Marc Moss 1993 Life in Preschool: Volume One of an Observational Study of Early Childhood Programs for Disadvantaged Four-Year-Olds. Cambridge, Mass.: Abt Associ- ates.