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.
IV Future Directions
9 Recommendations for Data Collection and Research r The panel has reviewed a broad array of data sets, academic research studies, and program evaluations for their contribution to understanding issues related to child care policy. In previous chapters we have summarized what is known about the consequences of supplemental care for children's health and development, described existing child care services and related public policies and programs, and analyzed the adequacy of existing services, policies, and programs to meet current and increasing needs for out-of- home child care and the effects of alternative proposals to meet those needs. Although researchers have made significant advances in knowledge about child care in recent years, we have repeatedly noted that many questions remain unanswered, and those questions suggest priorities for future data collection and research. Many of the panel's recommendations reiterate and expand on the work of previous panels and study groups of the Committee on Child Development Research and Public Policy (see Committee on Child Development Research and Public Policy, 1981; Hayes, 1987; Hayes and Kamerman, 1983; and Kamerman and Hayes, 1982~. Essential to framing an agenda for research is an underlying concept of the possible applications of increased knowledge. What do concerned policy makers, program administrators, advocates, employers, and parents need to know? How would particular information make a difference for public or private efforts to develop responses to parents' growing needs for out-of-home child care? The relationship between empirical study, scientific theory, and policy and program development is interactive and continuously evolving. Advances in one domain inevitably influence new initiatives in others. Implicit in the research questions that have emerged in previous chapters is the need to link data collection, analyses of the developmental consequences of different forms of child care, and studies of 269
270 WHO CARES FOR AMERICA'S CHILDREN? program costs and effects to underlying theoretical constructs, for example, theories of child development, theories of social structure and adaptation, and theories of human ecology. What is the meaning of mothers' employment and child care in the context of parents' and children's psychosocial, cognitive, and physical de- velopment? How do mothers' employment and child care relate to race, family structure, and socioeconomic status? What do they mean in different cultural communities and neighborhood environments? And what do they mean in terms of national productivity, public welfare, and public costs? In light of such questions, data needs can be specified, measures can be derived, hypotheses concerning the relationships among relevant variables can be tested, and programmatic approaches can be developed with some logical connection between often distinct and unrelated activities. Within this framework, the rest of this chapter presents the panel's recommenda- tions for data collection, research on child care and child development, and policy and program analysis. DATA COLLECTION The panel recommends that data systems that monitor par- ents' employment and use of child care, as well as indicators of children's health and well-being, be maintained and strengthened. Such data are essential for understanding trends and correlates of employment, their effects on child and family well-being, the demand for and supply of child care services, and the availability and affordability of child care as a basis for policy and program development. Data concerning levels and variations in parents' employment, income, family structure, and the availability, use, and costs of child care services of different types were the basis for much of the panel's deliberations. In addition, the panel relied on data concerning the health and well-being of children and their adaptations to the social and economic changes in U.S. society over the past decade and a half. Such data will continue to be essential for future research and analysis of child care issues. Relevant information is available from several sources, including large-scale surveys, federal and state administrative reporting systems, and service providers. Each of these sources has particular strengths and weaknesses, and individ- ual data sets vary in their underlying purposes and special emphases as well as their specific characteristics (e.g., definitions, sampling, data collection intervals). Thus, the panel concludes that a multidimensional strategy for data collection is essential. Several general issues are relevant to the collection of information in large-scale data sets that affect their usefulness in studies of child care.
RECOMMENDATIONS FOR DATA COLLECTION AND RESEARCH 271 First, in many cases the definitions of key concepts (such as types of child care programs or arrangements) are not uniform across data sets, making it difficult for researchers studying particular phenomena or relationships to use or compare information from different sources. Within individual data sets, standardized information is often unavailable in sufficient detail to support the desired analyses: for example, data on the age of children in licensed child care facilities; data on children's gender, race, and ethnicity; or data on parent's work patterns and sources of income. Second, there are no national data on the supply of child care and early childhood education programs. Information concerning the availability of different types of child care programs and arrangements and their regulation and financing is not available from any central source. And although states collect some of these data, they are not centrally available even at the state level: for example, although state licensing agencies collect data on the number of licensed child care facilities and their capacities, they do not keep information on early childhood programs operated by schools or Head Start. Third, data necessary to match supply of and demand for child care of different types, within a relevant range of costs and within and across relevant geographic units, is currently unavailable. Efforts to compare the demand for services to the available supply are stymied by the unavailability and inconsistency of data at the local, county, state, or national level. Finally, some important information on employment, attitudes about work and child care, and indicators of child health and well-being are not collected on a routine basis. Consequently, researchers and policy analysts cannot track changes over time that may have significant implications for the development and implementation of policies and programs. Our discussion of priorities for data collection is organized according to the types of relevant data sources: large-scale surveys, national and state reporting systems, and special surveys. Large-Scale Sunveys The major large-scale surveys that provide cross-sectional information on parental employment and child care include general population surveys, surveys of income and program participation, and youth surveys, many of which have had long-standing federal support. The panel endorses the protection and maintenance of these data sets and highlights several specific ways in which their usefulness in studies of parental employment and child care could be enhanced.
272 General Population Surveys WHO CARES FOR AMERICA'S CHILDREN? General population surveys contain a broad array of descriptive infor- mation on characteristics of a population. Because they provide lengthy time series, they permit analyses of population trends, such as employment, over time. Because of their very large sample sizes, they support analyses of small population subgroups that are difficult to study using other data sources. Two of the most relevant U.S. general population surveys for the study of child care issues are the decennial census and the Current Population Survey (CPS). The decennial census provides the largest sample and most complete information on general characteristics of the U.S. population of any avail- able data source. In addition to identifying patterns of change in household and family composition, racial and ethnic composition, age composition, geographic distribution, and employment and personal income, it is invalu- able for tracing trends and making estimates at the state and local levels. It is especially useful for analyses of small geographic areas, such as towns and neighborhoods, within larger metropolitan areas. Data on employment status, family structure, marital status, and fertility among small popula- tions, such as small ethnic groups and recent immigrants, allow researchers to examine the trends and patterns of diverse population subgroups. More- over, because of its broad coverage of the population, decennial censuses frequently provide the basis for sampling designs for other data collection efforts. Unfortunately, however, census data are not detailed in many areas of interest to researchers studying child care. For example, they do not contain information about the type of child care arrangement that a family uses, the costs of care, or the hours care is provided. Nor do they contain separate information on the incomes of husbands and wives when both parents are earners. Although there is a great detail of resistance to further expansion of the census, these data would be extremely useful to child care researchers and policy analysts. The CPS is the source of monthly estimates of employment and unem- ployment, including extensive detail on population characteristics. Through the regular addition of supplemental questions, the survey also provides both annual and one-time information on a broad spectrum of subjects, such as family and personal income, poverty, receipt of noncash transfers, annual work experience, school enrollment, and migration. Among the many supplements that have been included periodically on the CPS are questions concerning child care use and attitudes about work and child care. Information about these issues is related to age of child, family struc- ture,--and household income. However, these data only cover the child care arrangements for the youngest child under age 6 in a household. Therefore,
RECOMMENDATIONS FOR DATA COLLECTION AND RESEARCH 273 the data provide a limited view of parents' use of multiple arrangements and especially the packaging of arrangements when their are several children in the family of different ages with different child care needs. In addition, information concerning school enrollment, employment, and child care patterns among very young parents who are not living on their own are not reported separately from those of the head of household. Therefore, it is difficult to determine the need for and the use of child care by this population subgroup. Such data would greatly facilitate analyses of public income transfers and child support to teenage mothers as well as their patterns of labor force participation and their use of child care outside the home. Survey of Income and Program Participation The Survey of Income and Program Participation (SIPP) is a major source of information on the demographic and economic circumstances of U.S. individuals and families. It covers a stratified sample of the U.S. civilian noninstitutionalized population. It is a particularly useful tool for understanding the effects of government transfer and service programs. SIPP gathers detailed data on earned, unearned, and asset income, and it measures monthly variations in contributing factors such as household structure, the determinants of program eligibility, and actual program par- ticipation. It is a continuous survey in which overlapping panels are added and existing panels are rotated out periodically. In addition to its fixed questions, covered in "topical modules," SIPP also contains variable "topi- cal modules," one of which in 1984-1985 covered child care arrangements. These data have been extremely useful in examining parents' use of al- ternative child care arrangements as they relate to income and program participation, as well as other demographic factors. This special topical module should be continued in order to allow researchers, policy makers, and program managers to track the dynamics of social change and the effectiveness of public policies and programs designed to address the child care needs of working parents. National Longitudinal Survey Youth Cohort The National Longitudinal Survey Youth Cohort (NLSY) provides data on social, educational, occupational, and other aspects of the lives of adolescents and young adults. Because the NLSY collects detailed data on the youth experiences of males and females, it permits comparisons of patterns of family formation and parenting in conjunction with education and labor market experiences. A recent supplement to the NLSY, the Mother-Child Assessment, includes detailed information on the health and
274 WHO CARES FOR AMERICA'S CHILDREN? development of the children of young mothers in the NLSY. These home assessments of children's social, emotional, and cognitive development, as well as information on child care arrangements and other parenting supports and services, will greatly facilitate researchers' ability to link specific child development outcomes to those factors for a nationally representative sample of the U.S. population. We applaud plans to replicate the first wave of the Mother-Child Assessment, which will provide researchers and policy analysts with time patterns of family structure, income, parental employment, and child care arrangements as they relate to children's health and behavior. National and State Reporting Systems As we have noted, there are no national data sources on the overall supply of and demand for child care. The demand for services is gener- ally inferred from national survey data. Data on supply are provided by state departments or agencies responsible for the regulation of child care programs and by child care resource and referral agencies. They are also available at the federal level for specific programs such as Head Start and the Child Care Food Program. State regulatory agencies usually collect data on the programs they regulate, including licensed child care centers and family day care homes. They do not, however, collect data on unregulated facilities, which means that their estimates of the supply of child care are certainly less than the actual supply. For example, in many states, programs that operate only part day, those that are administered by the schools or as a part of the Head Start program, and those that are operated under the auspices of religious organizations are exempt from regulation and therefore not included in states' estimates. In addition, unlicensed family day care homes are not included in state reports of the supply of care. State regulatory agencies usually can provide detailed information on regulatory requirements and enforcement and on the licensed capacity- distinct from the actual number of children being served of centers and family day care homes. However, they usually do not collect information on the population served or on program characteristics, such as the profit status of the program or facility, or on financing or program fees. Therefore, it is impossible to determine the extent to which current licensed programs serve children of different ages, from families with different characteristics, and those with special needs. Moreover, it is difficult to trace the flow of public funding from the federal government, from states, and from local revenues to particular child care programs. Local resource and referral agencies do collect more complete infor- mation on child care programs than do state agencies, including information
RECOMMENDATIONS FOR DATA COLLECTION AND RESEARCH 275 on program characteristics (e.g., profit status, hours of operation, caretaker qualifications) and on the population served. Many resource and referral agencies also collect information on unlicensed and unregulated facilities in jurisdictions in which certain types of facilities and programs are exempt from state regulation or licensing. Because they attempt to match con- sumers and providers of care, they generally also have data on the demand for particular types of care. However, because resource and referral agen- cies may or may not be located in any community, their coverage is uneven. The information they collect may provide a relatively complete picture of the supply and demand in a particular community, but the data from all resource and referral agencies do not by any means add up to a complete national picture. In addition to presenting an incomplete picture of the available pro- grams and the children and families they serve, these three data sources- national and state reports and resource and referral agencies suffer from noncomparability. For example, there is no common definition across states of what constitutes a child care center, a family day care home, or a group home. In addition, age categories of children specified in state regulations also vary: infants may be defined in one state as newborn to 12 months of age whereas in another they are children up to 22 months. In the absence of common terminology and definitions, it is difficult to accurately estimate the demand for and supply of licensed care (let alone unlicensed care) or to assess the factors that affect supply. At the national level, uniform information from each Head Start grantee is reported to the Administration for Children, Youth, and Fami- lies (ACYF), which compiles it in a computerized data base that is updated annually. These data include a wide range of information on program characteristics and costs, as well as the characteristics of children who are enrolled. The data permit program comparisons by type of sponsoring agency and geographic region, as well as projections of costs and enroll- ments. Similarly, information collected by the local sponsoring agencies for the Child Care Food Program is reported to the U.S. Department of Agriculture, which compiles it in a computerized data base. Information from the food program is valuable because it presents a partial picture of family day care in addition to center care. Both Head Start and the Child Care Food Program data bases are extremely useful for program planning and evaluation purposes. Unfortunately, they are not matched with other comparable data sets that would facilitate analyses across programs and types of providers of regulated and unregulated care. The panel concludes that there is a critical need for better and more systematic information on the supply of child care and early childhood education programs and on the characteristics of the children and families that are served by these programs. These data should include information
276 WHO CARES FOR AMERICA'S CHILDREN? On financing, costs of operation, and fee scales. They should be collected at the state level and reported and compiled at the federal level. Special Surveys In addition to data on the supply of child care and early childhood education programs that should be collected regularly at the state level and compiled at the national level, there is an urgent need for special surveys to provide current information on the demand for and supply of child care and the experiences of children in child care. Such data are needed as a basis for decisions by policy makers and program planners who are responsible for responding to changing conditions and needs. Information is also needed on families' child care arrangements, including preschool and before- and after-school programs; how child care affects work patterns and household responsibilities; and how parents make their choices of care for their children. Data on the demand for child care should be matched with information on the supply of programs and arrangements. Detailed information is needed about the national supply of child care options and how they are distributed among families in different social, economic, and cultural circumstances and among different regions of the country and community settings. The panel applauds a joint initiative by the ACYF within the U.S. Department of Health and Human Services and the National Association for the Education of Young Children to undertake the National Child Care Survey. This survey of a nationally representative sample of 5,400 parents with children under the age of 13 will be conducted in early 1990. It will be complemented by a Profile of Child Care Settings Study, sponsored by the National Center for Education Statistics in the U.S. Department of Education, that will survey directors of child care centers, preschools, and licensed family day care providers. Although unlicensed providers will not be included, these two surveys, taken together, will provide the most comprehensive picture of child care supply and demand yet available. RESEARCH ON CHILD CARE AND CHILD DEVELOPMENT The panel recommends the continued support of a broad- based research program on the relationship between child care and child development to enhance understanding of the conse- quences of children's experience in out-of-home care for their social and cognitive development as well as for their physical health and safety. The results of such studies will have continu- ing value in the development of policies and programs related to children and families.
RECOMMENDATIONS FOR DATA COLLECTION AND RESEARCH 277 Over the past decade and a half, research has added significantly to the knowledge of trends, correlates, and consequences of children's experience in supplemental child care programs and arrangements. These research findings have provided an essential basis for the panel's work. Numerous studies have examined the effects of child care on children's growth and development. As child care research has become more theoretically and methodologically sophisticated, researchers have refined their questions and designs to explore the specific features and characteristics of child care programs and settings that affect psychosocial, physical, and cognitive de- velopment and the practices that can safeguard children's health and safety and promote positive outcomes. Despite advances in knowledge about child care and child development, many questions remain unanswered. In some cases, gaps reflect issues that have not been adequately studied because of methodological problems; in other cases, new issues have emerged from the findings to date. In Chapters 3 through 5, we highlighted a variety of salient research issues and questions. They are presented here under five general headings: dimensions of child care quality; the relationship between child care quality and family characteristics; participation in child care during the first year of life; family day care, care by relatives, and use of multiple forms of care; and health in child care settings. Those chapters also pointed to a need for studies using new research strategies and focusing on emerging questions; these are summarized in the final part of this section. Dimensions of Child Care Quality As we discussed in Chapter 4, widespread reliance on global or sum- ma~y measures of quality in child care research is not simply a function of convenience or simplicity. Rather, it reflects the fact that the individual components of quality have often been found to be intercorrelated. Sev- eral researchers have observed that "separation of . . . various dimensions of care quality may be difficult, if not impossible, as they seem to occur naturally in clusters" (Anderson et al., 1981:60) and that "good things" in child care seem to go together (McCartney et al., 1985~. Although research has documented the interrelatedness of structural features of child care that constitute quality, additional work is needed to clarify the implications of those links for child outcomes. If dimensions of quality tend to cluster, then policies or programs may well need to be designed around clusters of features as well as individual features. Such studies will need to consider the assumption that improvements in one quality feature may have implications for others. 1b give just one example, it may well be that improving staff/child ratios has a meaningful effect on
278 WHO CARES FOR AMERICA'S CHILDREN? children only if caregivers are well trained or only if groups are of limited size. On the basis of our review of the existing literature, the panel concludes that studies of the interrelatedness of quality dimensions should include attention to several issues: (1) What is the nature of the correlation among quality variables? Do consistent clusters of variables emerge across studies of different types of child care programs and settings? (2) Does manipulation of one variable (or cluster of variables) have ramifications for others? (3) Does the relatedness of the quality dimensions depend on a program's level of funding or on the philosophy of the program director? For example, do more generously funded centers show higher quality across dimensions? Or does a director with more training or a particular philosophy struggle to maintain quality across dimensions? Understanding of the dimensions of quality and their implications for development would be improved by the use of research designs involving random assignment or manipulation of quality dimensions. At present, there appear to be two virtually segregated approaches to the study of quality: intervention studies in high-risk populations, which rely heavily on random assignment and manipulation of program features, and natural- istic studies of community-based child care, which are vulnerable on the grounds that they do not isolate characteristics of the children who are served, the characteristics of the programs, or quality variables. A decade after the publication of the National Day Care Study, which set forth the methodological and conceptual basis for using randomization and manipu- lation in studies of quality dimensions in community-based child care, it is surprising that subsequent research on quality has not complemented the body of naturalistic studies with more experimental ones. Future research should examine the implications of (1) initiating change in particular qual- ity features (e.g., providing training for caregivers, improving the staffIchild ratios, disrupting or permitting continuity of peer groups) in a random subset of classrooms; and (2) randomly assigning children to child care settings that vary on key quality dimensions (or clusters). The absence of such approaches is particularly glaring in the study of caregiver training, when it is clearly possible that self-selection factors may influence levels of training and education. The panel's review suggests that understanding of child care quality needs to be expanded to include dimensions that have not yet received a great deal of attention by researchers and to include consideration of acceptable and unacceptable ranges on the traditional quality variables. Features of quality that may well be linked to development but that have not been adequately studied include size of center (as opposed to size of group); affirmation of children's racial, ethnic, or cultural group identity; parental involvement; stability of the peer group; and, for family day care, the age
RECOMMENDATIONS FOR DATA COLLECTION AND RESEARCH 279 mix within groups. There is a need for further study of child care curricula, of both the content and the process of learning. The guidelines for group sizes and staff/child ratios at different child ages proposed in professional standards should form the basis of research focusing on ranges on these variables and the extent to which they are associated with developmental outcomes. Our review of the studies on quality also pointed to a lack of research assessing magnitude of effects: that is, what is the magnitude of improve- ment on child outcomes for measured improvements in quality? Such data would be extremely important in evaluating the benefits to children of selected quality improvements. Finally, our review points to the need for further study of the longer term implications for children's development of participation in child care of high versus low quality, of quality dimensions that may be uniquely important in the care of children with disabilities, and of variation in the quality of care according to auspice of care. Links Between the Quality of Child Care and Family Characteristics In Chapter 4 we summarized the growing body of evidence that the quality of child care and family characteristics are linked, and we noted the "double jeopardy" of children from stressed families being placed in poor- quality child care. Future studies should attempt to clarify the nature of the association between the quality of child care and family characteristics. As a first step, researchers need to ask whether existing studies have captured the full range of family variables that may be related to child care quality. In addition to socioeconomic variables, it now appears that family stress and social support are important. Other variables, such as marital discord, marital status, job characteristics of one or both parents, motivation for parenting, and the quality of parent-child interactions, also merit attention. Future studies should examine the process by which parents in different circumstances choose child care. Do more stressed families allocate less time to search for child care? Are they less informed about alternative arrangements, about the significance of choosing high-quality care, or about what constitutes high quality? Are they equally knowledgeable but less able to persevere (given such factors as long waiting lists) in obtaining higher quality care? Or are they simply less able to afford care of higher quality? In this regard, future studies should explore the everyday decision- making process regarding child care. How do parents weigh various di- mensions of quality in judging and choosing child care settings for their children? What are the relevant folk beliefs or cultural norms that influ- ence their decisions? For example, how important is it to parents that their children are cared for by kin or others from their own community or social group rather than by strangers? How important is it to them that child care
280 WHO CARES FOR AMERICA'S CHILDREN? includes moral training, that gender differences are managed in a particular way, or that a particular type of discipline is used? Indeed, parents may want quality arrangements for their children, but their concept of quality may be shaped more by culturally determined fold views of what is impor- tant for child development than by scientific research. Cultural beliefs and norms may or may not be related to race and ethnicity; therefore, they need to be studied and understood separately from racial and ethnic differences in child care. In the future, decisions concerning the organization of child care programs and the mix of public child care policies should be much more explicitly linked to the results of research on what parents prefer and what they are really choosing in child care. The links between family characteristics and child care quality may provide an exemplar of what we have referred to as transactional processes in development: the mutual influence of the child (and family) on the child care environment and of the child care environment on the child (and family). A longitudinal study of this association could explore patterns of mutual influence over time. For example, while the level of family stress may influence choice of child care quality, the choice of child care may also subsequently influence stress within the family and affect the child's development. As noted in Chapter 3, studies are needed not only of the direct effects of child care on children, but also of the indirect effects on children of the influences of child care on parents. Consistent with our suggestion concerning studies of the relatedness of dimensions of quality, knowledge of the family-quality association would be improved by studies that systematically alter quality variables in a ran- domly selected set of families that are similar on social, economic, and psychological factors. Participation in Child Care During the First Year of Life As discussed in Chapter 3, children who participate in child care for more than 20 hours a week during the first year of life show higher rates of behaviors that are categorized on a frequently used laboratory measure as "anxious avoidant" in their attachment to their mothers (see Chapter 3 for definition and discussion). Although there is agreement on this finding, its bases and its implications are still subjects of heated debate. This debate will only be resolved through rigorous examination of several sets of questions. First, are differences in security of attachment rooted in ongoing features of the mother-child relationship rather than in the timing and amount of exposure to child care? Are there relevant self-selection factors, that is, are mothers who resume employment (early and more than part-time) different from those who do not? Are there differences in their responsiveness to their infants that both antedate choice of care
RECOMMENDATIONS FOR DATA COLLECTION AND RESEARCH 281 and underlie later differences in security of attachment or other indices of development? Second, what are the implications for the development of infant-mother attachment (as well as for mothers' commitment to their infants) of child care experiences in the first year of life of families with employed mothers? For example, do variations in mothers' subjective sense of stress, of over- load, or of role conflict influence the emergence of mutual attachment or the way in which the child's attachment influences other relationships and behavior? Can anxious-avoidant attachment be modified? Would such factors as a daily visit to the child care setting (perhaps during lunch time) or counseling parents on structuring evening reunion time to maximize parent- infant interaction time affect the incidence of anxious-avoidant attachment in infants? How does the assessment of security of attachment relate to other indices of socioemotional development and well-being? Is avoidant at- tachment in infants indeed associated with less optimal development both contemporaneously and over time? Or is it a reflection of adaptive be- havior? Future research needs to move beyond this single measure of emotional functioning and question whether it predicts subsequent devel- opment equally well in children with markedly different early experiences. It would be particularly useful to include recognized clinical measures that help distinguish between variations in child functioning within the normal range and disturbed functioning (e.g., assertive versus hostile behavior). Studies need to question and examine the developmental implications of higher rates of anxious-avoidant attachment in infants in child care rather than assume they are negative. 1b what extent do findings of anxious-avoidant attachment in infants with a history of full-time child care attendance in the first year reflect the use of poor-quality infant day care rather than the use of infant day care per se? Throughout the earlier chapters of this report, we have questioned whether infant care is of adequate quality given the developmental needs of these very young children and the cost of providing that care. Is there a difference in security of attachment to mothers among infants who have experienced high-quality infant care? In short, scholars of child development agree on the observed behavior, but they disagree about what it means and whether it necessarily has negative implications for children's future development. Resolving this disagreement should be a high priority for further research because it has significant implications for the role of public policies in establishing standards for the quality of infant care and in the debate about parental leave.
282 WHO CARES FOR AMERICA'S CHILDREN? Family Day Care, Care by Relatives, and Use of Multiple Forms of Care Most research on the effects of child care has studied children in center care. Yet most of the children in out-of-home care are in family day care. The disproportionate research focus on center care undoubtedly reflects the greater difficulty in finding and gaining research access to family day care homes, particularly those that are unlicensed and unregulated. As a result, less is known about the development of children in this type of child care setting and about the specific features of family day care that risk or support children's development. Future child care research needs to examine systematically the experi- ences of children in family day care (especially unlicensed family day care). Studies should go beyond the variables included in studies of center care to examine features unique to family day care. For example, do older and younger children form stable friendships in family day care groups? Do children of certain ages experience problems in mixed-age groups, as some studies suggest? Is family day care associated with greater concordance of values and cultural practices from home to care setting than is the case with center care? If so, what are the implications for children's development? We note that there are few curricula or other materials to guide family day care providers. Such materials should be developed, tested, and evaluated. Similarly, study is needed of the effective ways to provide training and technical assistance to family day care providers. Because of the primary research focus on center care, there is also little knowledge about care by relatives. Although a substantial proportion of infants, toddlers, and preschoolers whose mothers are employed are cared for by relatives (including fathers), almost no research in the United States has included examination of the nature or effects of such care. Research on the effects of child care needs to include comparisons of care by relatives (including care by parents in split-shift arrangements) with other forms of care and with care by at-home mothers. Such studies could inform the continuing debate about the appropriateness and desirability of public policies that encourage or discourage care by relatives for children with employed mothers. As discussed in Chapter 4, there is evidence that children experiencing a sequence of caregivers over time (unstable care) differ in their develop- ment from children experiencing more stable care. Do such developmental differences also exist for children who experience multiple caregiving ar- rangements in the course of a day or week? At present, data suggest that there are families for whom a single child care setting does not suffice. What are the implications for children when they are placed regularly in more than one care setting?
RECOMMENDATIONS FOR DATA COLLECTION AND RESEARCH Health in Child Care Settings 2~3 Our review of existing research on illness and injury in child care also highlighted directions for future study. Of particular importance are two issues that have not been carefully studied: middle-ear infections among children in group care and their possible implications for language development, and rates and circumstances of injury, abuse, and neglect in child care settings. With regard to both issues, future studies should contrast the experiences of children in different types of care, including care by parents. In addition, with regard to both issues, researchers should employ prospective as well as retrospective research designs. Of significant concern is the possibility that reports by parents, teachers, and physicians of children's past injuries and illnesses are biased by their attitudes about mothers' employment and the lack of availability of emergency services and child care for sick children. Additional research is also needed to evaluate the health and safety implications of peer contacts of children diagnosed with HIV infection. New Research Strategies and Issues Throughout our review of the evidence there were indications that knowledge about child care and development would be strengthened by the use of particular research strategies and by addressing issues that have been neglected to date. For example, there is a need for long-term longitudinal research. Such research is needed both to understand the changing needs of children in child care with increasing age (using more developmental demarcations than that between infancy and the remaining preschool years) and to expand knowledge concerning the long-term implications of early child care experiences. Research is needed on the implications of child care experiences for children of different racial, ethnic, or cultural backgrounds. For example, what are the implications for children in programs that stress multicultural sensitivity, both regarding their own group identity and their attitudes about children of other groups? Does such exposure have implications for later adaptation to school? Are there differences in development according to whether the child care environment is consonant or dissonant with the cultural orientation of the home? Research is also needed on the implications of child care participation for children's mental health. For example, does child care result in early identification of family and child mental health problems? Is there follow- through on such problems? We pointed to evidence in Chapter 3 that maternal participation in Head Start has implications for mothers' mental health. Are there similar effects in other child care settings or only for
284 WHO CARES FOR AMERICA 'S CHILDREN? those that stress parental involvement? What are the implications of child care participation for children from families stressed by divorce, mental illness of parents, or parental tendencies toward abuse or neglect? In general, what are the implications of participation in child care for the mental health of children as well as parents? As noted above, there have been few studies that examine the perspec- tives of parents on child care choice and child care quality. User surveys are needed that ask on what bases parents choose child care. 1b what extent are parents aware of dimensions of quality? How important are particular dimensions of quality in the choice of a child care setting? The generalizability of results could be improved through use of na- tionally representative data sets that include questions regarding child care. It would be valuable to incorporate within national surveys particular ques- tions, using subsamples of the survey. For example, observational data are not readily obtainable in a large representative sample, but such data could be obtained for a small subsample with selected demographic characteris- tics. Finally, studies that focus on school-age children are needed. The evidence regarding child care for school-age children is extremely limited. Studies should examine the extent of need for child care in this age range and the implications of self-care for school-age children. And studies should examine the dimensions of quality for child care for school-age children. RECOMMENDATIONS FOR POLICY AND PROGRAM ANALYSES The panel recommends that policy and program analyses to measure the costs, ejects, and effectiveness of alternative pro- posals for the provision, financing, and regulation of child care be an essential component of child care research. Federal and state funding agencies, along with private foundations and corpo- rations, should support policy analyses and program evaluations to inform public- and pr~vate-sector decision making. Between the late 1970s, when the National Day Care Study and the National Day Care Home Study were completed, and the late 1980s there was a dearth of national policy studies of child care issues. Throughout this period mothers of very young children entered the labor force in unprecedented numbers, and the need for and supply of out-of-home child care programs and arrangements expanded significantly. Knowledge of the costs and effects of government policies, employer policies and practices, and community services and programs has not kept pace with social change. As a result, there is a sparse base of rigorous scientific knowledge to guide future policy and program development. Accordingly, the panel concludes
RECOMMENDATIONS FOR DATA COLLECTION AND RESEARCH 285 that greater investment in evaluating the direct and indirect outcomes of existing and proposed policy and program initiatives is urgently needed. Federal, state, and local government policies play an important role in influencing the nature and extent of social and economic change in society as well as the responses to them. Federal labor and wage policies affect employment and unemployment rates; income tax policies may affect decisions, especially married mothers' decisions, to work; income transfer policies may affect the employment decisions of single as well as married mothers by providing incentives and disincentives to work; and federal policies toward employers, through direct legislation, tax incentives, and regulation, influence the extent and ways in which employers structure their employee policies and benefits (Kamerman and Hayes, 1982~. The federal and state governments have been the major fenders of publicly subsidized child care and related services since World War II. The federal government and some states provide direct subsidies through the dependent care tax credit to offset the child care expenses of employed par- ents. They supplement the funding available to schools for early childhood programs and before- and after-school programs through such means as direct grants for special programs and funding for compensatory education. All states regulate child care services, and some support the development and maintenance of resource and referral agencies and other supportive services. Both the federal and the state governments invest in the training and certification of child care providers, and the federal government also provides some subsidies to employers who develop child care policies and programs for their employees. As we have shown, however, patterns of government funding have shifted over the past decade. Direct support to providers for the provision of child care services, through programs such as the Social Services Block Grant, has declined as more public resources, es- pecially at the federal level, have given way to consumer subsidies through the tax system. Proponents of both approaches debate the effects of these changes on the quantity and quality of child care services. Except in the most general sense, there is little systematic knowledge of the consequences of these policies for children, parents, employers, or child care providers. For example, these patterns of support have clearly led to the development of a diverse array of child care programs and providers, but what is not known is the extent and ways in which they have altered the behavior of providers and consumers and whether they have improved or decreased the quality of care. As Congress and state legislatures now consider a range of proposals for new child care initiatives, there is little empirical basis for making choices. Accordingly, the panel concludes that studies should be launched to assess the effects of different types of government child care
286 WHO CARES FOR AMERICA'S CHILDREN? policies. In particular, the panel has identified several sets of questions that merit attention: · What are the effects of regulation on the supply and mix of child care? Do more restrictive requirements discourage center care providers or family day care providers from entering the market? Do they affect the cost of providing care and the fees that are charged? Do they affect the location of child care facilities and, as a consequence, their accessibility to families living in different areas? · What are the effects of alternative financing mechanisms on the supply and mix of child care and on the behavior of consumers? Do particular types of financing (e.g., direct provider subsidies, tax benefits to consumers, vouchers) foster or discourage the development of different types of programs provided under different institutional auspices? Do they cause parents to prefer or select one type of care or another? How does the growth of public school programs for 3- and 4- year-olds affect the supply and mix of child care services for children of these ages provided under other auspices? Has the growth of school-based programs diminished the demand for center-based or family day care? Has it affected the costs or quality of care provided under other auspices (e.g., through competition for a limited pool of qualified staffs? · How have various employer policies and programs (e.g., on-site child care, flexible spending programs, child care subsidies, resource and referral services, parental leave) affected staff recruitment and retention in different industries and geographic regions? How has it affected employee productivity and firm profitability? . What are the effects and effectiveness of policies and programs to improve the qualifications and wages of child care workers? Do investments in education and training lead to increases in the supply and quality of child care workers? What effects do wage subsidies have on the quality of staff and retention rates? What are the effects of alternative interventions to provide preservice and in-service training for family day care providers on the supply of home-based providers and on the quality of care they offer? . 1b what extent and in what ways does the availability of affordable high-quality child care influence parents' decisions to work? ~ what extent and in what ways is the lack of adequate care a barrier to labor force entry or retention? Do the effects differ for mothers (and fathers) in different social, economic, and cultural circumstances, among those in different occupational categories, those of different ages, and those with different educational backgrounds?
RECOMMENDATIONS FOR DATA COLLECTION AND RESEARCH 287 The Family Support Act of 1988 offers an important opportunity to examine the direct and indirect effects of a fundamental shift in the U.S. approach to income security policy. The new act requires mothers of preschool-age children to work, attend school, or participate in an employ- ment training program as a condition of receiving welfare support. The act requires the states to provide child care services for children of dependent mothers, and it further requires them to provide "transition" child care services for up to one year after mothers find jobs and become economi- cally self-sufficient As the states move to implement the provisions of the new law, analyses of the changes and families' adaptations to them could offer valuable insights into many child care issues. The findings from this (or other) natural experiment could provide the basis for formulating more refined hypotheses for subsequent demonstration and experimentation and for future policy. This chapter outlines a broad agenda for future data collection and research aimed at filling the gaps in the current knowledge basis. Idgether, such work to expand the existing body of knowledge will significantly strengthen the basis upon which decisions concerning the care of the nation's children are made. REFERENCES Anderson, C., R. Nagle, W. Roberts, and J. Smith 1981 Attachment to substitute caregivers as a function of center quality and caregiver involvement. Child Development 52:53-61. Committee on Child Development Research and Public Policy 1981 Senaces for Children: An Agenda for Research. Commission on Behavioral and Social Sciences and Education, National Research Council. Washington, D.C.: National Academy Press. Hayes, C., ed. 1987 Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing. Panel on Adolescent Pregnancy, Committee on Child Development Research and Public Policy, Commission on Behavioral and Social Sciences and Education, National Research Council. Washington, D.C.: National Academy Press. Hayes, C., and S. Kamerman 1983 Children of Working Parents: Erperiences and Outcomes. Panel on Work, Family, and Community, Committee on Child Development Research and Public Policy, Commission on Behavioral and Social Sciences and Education, National Research Council. Washington, D.C.: National Academy Press. Kamerman, S., and ~ Hayes, eds. 1982 Families That Work: Children in a Changing World. Panel on Work, Family, and Community, Committee on Child Development Research and Public Policy, Commission on Behavioral and Social Sciences and Education, National Research Council. Washington, D.C: National Academy Press. McCartney, K., S. Scarr, D. Phillips, and S. Grajek 1985 Day care as intervention: Comparisons of varying quality programs. Develop- mental Psychology 6:247-260.
