• The Utility of Household Surveys
• Two Studies That Use NHES Data
• References

Conducting a household survey is not a simple enterprise. In general, such surveys are more logistically difficult and demanding than are school-based surveys, or at the least create a whole new assortment of sampling and other technical problems. Household surveys require the screening of large numbers of households to get the smaller number that can provide data on the issues that the researcher cares about. Response rate becomes a particular problem, since the interviewer must first secure the participation of the household and then the participation of the proper respondent within the household. Why, then, would NCES launch a long-term and ambitious series of household surveys?

The answer is simple enough. Much of the most significant decisionmaking and resource allocation having to do with “education” (very broadly construed) takes place in households. Families, even more than schools, provide the locations at which many education policy instruments have their most direct impacts. Further, many educational events and processes are not primarily school based—household library use, noncredit adult learning, and school readiness activities are just three examples. The share of these events and processes that is not inherently school based is almost certainly getting larger. “Out-of-school schooling” takes place as families seek extraeducational opportunities for their children, as postindustrialism fuels the demand for worker reskilling, and as families and children pursue alternatives to traditional schooling. As sociologists would have it, the normative American life course is becoming more fragmented and diverse, and families are developing different strategies to pass on various resources to their children. So if everyone (with some inevitable ambiguities) by definition lives in a household, and if many educationally significant activities take place in households and away from schools, it follows that we need to study these through household surveys.

Household surveys have shown their utility elsewhere. The Current Population Survey conducted by the U.S. Census Bureau produces such important data as the monthly unemployment rate. Sociologists and demographers in particular have for years made great use of such series as the National Survey of Families and Households and the General Social Surveys. The new American Community Survey from the Census Bureau may have a similar impact over time. In fact, the household survey is the research method of choice in much of the development policy literature. As pointed out by Angus Deaton in his valuable The Analysis of Household Surveys: A Microeconometric Approach to Development Policy (1997), household surveys provide a means to examine the microeconomics of household budgets of families who may have little formal link to such institutions as schools or work establishments.

The two studies highlighted in this issue of the Education Statistics Quarterly—Xianglei Chen’s analysis of parental involvement in children’s K-8 education and Sean Creighton and Lisa Hudson’s examination of trends and patterns of participation in adult education—both demonstrate the payoff of NCES’s commitment to NHES. Creighton and Hudson’s study is based on the Adult Education Surveys of NHES (AE-NHES:1991/1995/1999). Chen combines data from the Parent and Family Involvement in Education/Civic Involvement Survey of NHES (PFI/CI-NHES:1996) and the 1996 “Survey on Family and School Partnerships in Public Schools, K-8.” The latter, a survey of schools, is part of the NCES Fast Response Survey System (FRSS). The two surveys used by Chen are independent of one another, meaning that the children in PFI/CI-NHES are not matched to the schools in the FRSS survey. Chen builds on sociologist Joyce Epstein’s useful work on the connections between schools and families (e.g., Epstein 1990).

Chen’s concern is with the extent to which schools and parents agree or disagree about the nature of their relationships with one another. This agreement turns out to be pretty dismal. Schools tend to report that they are providing parents with both numerous opportunities to participate in school activities and adequate amounts of information about what goes on in their children’s school. Parents are more likely to report that schools fall short in these efforts. Similarly, parents describe their own involvement as engaged and regular, while schools report that parents often disregard offers to participate and show far too little interest in what happens in school. The author found these discrepancies in all types of schools, but was struck that they generally increased with school level, school size, and minority concentration.

None of this is necessarily too surprising. The 1966 Equality of Educational Opportunity (i.e., Coleman) report showed that schools and parents often described the same schools in very different ways. One could probably find further evidence well before that, and Chen cites research that shows that parents and teachers (as well as students) give different estimates of parental involvement. What Chen has, of course, are not so much data on parental involvement as data on perceptions of parental involvement. While there is ultimately a real amount of interaction between parents and schools, it may lie somewhere between the perceptions of educators and those of parents, or it may lie beyond these extremes.

The problem of determining the actual extent of parent-school interaction raises a difficulty with household surveys. In PFI/CI-NHES, parents are asked to describe the characteristics of institutions in which they do not work, that they do not own, and often in which they spend little time. How accurately can we expect parents to report the characteristics of schools? In a methodological exercise, Chen shows that parents do make quite substantial mistakes in describing their children’s schools. This by no means diminishes the value of the central findings regarding a disconnect between home and school, but it does help provide some context for these findings.

A similar problem emerges in Creighton and Hudson’s comprehensive analysis of trends in participation in adult education in the 1990s. Like Chen, Creighton and Hudson rely on household members to describe the characteristics of various institutions to which they are more or less strongly linked—work establishments, colleges, and other education providers. We don’t really know how well people can do this, but there is probably some slippage between how AE-NHES respondents would describe these providers and how these providers would (presumably more accurately) describe themselves. In the case of AE-NHES, the problem is less that individuals report inaccurately on the characteristics of institutions than the fact that the survey has no means by which to provide independent estimates describing providers of instruction. This is particularly evident when looking at employer-provided instruction. While we know quite a lot about how to collect good self-descriptive data from work establishments (e.g., Kalleberg et al. 1996), the challenge is still how to gather high-quality information from the employees of these establishments.

Notwithstanding the lack of independent descriptions of adult education providers, Creighton and Hudson’s analysis is an informative one, remarkably attentive to detail and nuance. They provide compelling evidence of the increase in participation in adult education in the space of a single decade. This increase was as broad as it was deep—most social and demographic groups increased their participation. It also cut across most kinds of adult education. For the most part, more Americans were pursuing more kinds of adult learning at the end of the nineties than they were at the beginning of the decade.

However, Creighton and Hudson are careful to point out that not everyone participated equally in adult education at the end of the decade. In both 1991 and 1999,* for example, those with lower levels of education and those with lower status jobs had relatively low rates of participation in adult education overall, in work-related adult education, and in non-work-related adult education. In 1999, rates of participation in work-related adult education also remained relatively low among Hispanics and part-time workers. This pattern presents a particular challenge as we enter a “learning society” in which one’s initial experience in the education system no longer suffices as preparation for the employment and civic demands brought on by rapid technological, economic, and cultural change. If Creighton and Hudson’s main story is an encouraging one of increased participation and engagement in lifelong learning, it is also a warning against the ongoing marginalization of some groups and the troubling polarization of opportunities.

Both Chen and Creighton and Hudson are judicious in their recommendations. Both, too, are candid about methodological or conceptual shortcomings. Chen puts some of the responsibility for the poor relationships between schools and parents on both parties, indicating that schools may need to be more energetic about keeping in touch with parents and that many parents may need to make a greater commitment to their children’s schooling. To my mind, Chen is correct in not reducing the problem to one of “better communication” between schools and families. As considerable research has shown (see Lareau 2000 for a good example), in many cases the relationships between schools and families are rife with cultural or economic conflict that cannot be solved by simply enhancing communication. Such findings are consistent with those reported by Chen and should point policy in the direction of providing both schools and families with the means to interact more effectively. This will probably have more to do with finding ways to permit working parents and overextended teachers the material and logistic means to actually be in the same place at the same time than it will with greater “communication.” Chen’s study goes a long way in describing some of the constraints that have to be overcome for this to take place.

Creighton and Hudson do not comment at any length on the policy implications of their findings, but these are every bit as urgent as those arising from Chen’s report. Adult education—whether job training, English as a Second Language (ESL), basic skills education, or academic or vocational credential programs—is no longer optional for successful participation in a postindustrial economy. While perhaps too often a cliche, the “learning society” is going to require a different set of institutions and expectations as technology and transformed work arrangements draw adult Americans back into the education system. Whether through incentives or sanctions, we need serious attention to policies that target employer involvement in the post-compulsory education of the marginalized groups described by Creighton and Hudson.

As the authors also show, though, not all adult education is driven by the exigencies of making a living. Americans have, and probably always have had, a remarkable attraction to education for personal growth, cultural development, or simply for distraction and amusement. (This, incidentally, is a finding that would have been resistant to discovery by anything other than a household survey.) Based in both community colleges and a range of still-vibrant clubs, civic associations, and assorted institutes, such lifelong learning gives every indication of thriving as much in the next decade as it did in the last. We need to know much more about the motivations that people have for participating in these educational activities and about reducing the barriers that stand in the way of their participation.

* Keep in mind that NHES is not a panel study, so that Creighton and Hudson are not examining the same individuals at these different points in time.