DEVELOPING CHILDREN'S DATA FOR THE

SURVEY OF PROGRAM DYNAMICS (SPD)

Robert Kominski

U.S. Census Bureau ⁽¹⁾

Population Division

U.S. Bureau of the Census

Washington, DC 20233-8800

Email: Robert.a.kominski@census.gov

and

Loretta Bass

780 Van Vleet Oval

Department of Sociology

University of Oklahoma

Norman, OK 73019

Email: Lbass@ou.edu

Prepared for presentation at the
1998 American Statistical Association annual meeting.

Appears in published proceedings of the 1998 ASA meeting.

August, 1998

Population Division, U.S. Bureau of the Census, Washington, DC 20233-8800

Key Words: child, adolescent, welfare reform, SPD

INTRODUCTION

Social Welfare

During this century, the nation's attitude toward children and their welfare and needs has varied considerably. Beginning as the child labor movement at the turn of the century, concern for children and families became more institutionalized in the form of financial assistance programs created in the midst of the Great Depression. Federal involvement increased dramatically during the War on Poverty in the 1960's. By the 1980's, the mood of the country had changed, and concern was routinely voiced about the extent that social service programs had created dependency on federal aid.

Obviously, as with many social events, the mood of the nation is often reexpressed in the form of legislation and government action. Federal involvement during the Great Depression and War on Poverty had set a context in which the actions taken by government were often enacted at the national level. During the 1980's, however, a new sentiment, directed against large government began to mediate many actions of the federal government. More and more political candidates became elected on promises of cutting back "big government," or returning activities operated at the federal level to the states. This mood has prevailed into the 1990's, and continues on some level in the current political environment.

It is out of this context that the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 evolved and was enacted (The Federal Register 1996). Known to most people as the 'welfare reform act,' this law represents one of the largest revisions of the family welfare programs since the Family Support Act of 1998. The fundamental focus of the act is to 'devolve' welfare systems from the federal level down to states, and ultimately, to even lower levels of government.

Measuring Social Welfare

As enacted by the federal government, social assistance programs were funded by federal agencies for the most part. This meant that as programs were put in place, state agencies were charged with both operating them and providing some level of accounting to the federal government. Administrative record systems were typically utilized to monitor the involvement and outcomes of people and families in these programs. Due to the large number of programs, however, there was no systematic database that could be used to monitor and evaluate all of the programs a person might be a part of. Moreover, agencies operating programs often tended to focus their efforts more on the implementation of the program, rather than on the outcomes. Responding to this information gap, the federal government began the Survey of Income and Program Participation (SIPP) in the early 1980's.

The SIPP gave the federal government and research communities a tool with which to follow people, monitor their involvement in the full array of social assistance programs available, and see what happened to these individuals over an extended period of time (U.S. Bureau of the Census 1991). The SIPP is a longitudinal survey in which households in sample are interviewed once every four months for a period of several years.

Extending the Measurement of Social Welfare

As new panels of the SIPP have been fielded, additional subject matter concerns focusing on the well-being of children have been brought forward by various federal agencies. In the mid-1980's, for example, the Census Bureau began to administer "topical modules" - one-time survey supplements - to the SIPP on health and disability conditions for both adults and children, and the utilization of health care services for children. During this same time the Bureau began a series of topical modules which focused on the child care arrangements that were used for children, especially for those with working mothers.

In the 1992 and 1993 panels, the Bureau developed a series of questions on the topic of "child well-being," following on the heels of a supplement administered several years earlier on topics of "adult well-being" (such as ownership of consumer durables, sources of social support and life satisfaction). The child well-being topical module, fielded in the ninth interview (wave) of the 1992 panel and simultaneously in the sixth interview of the 1993 panel, touched on a number of issues related to children - experience with preschool and head start programs, routine day-care involvement, interaction with parents, family activities, television viewing rules, repetition of grades, and extracurricular school activities. The questions were designed to provide general information on the status of the children in the SIPP, viewed in conjunction with the vast array of social, demographic, and economic data already collected in the SIPP for the adult members of the household. These questions were also fielded one year later, in the ninth interview of the 1993 panel.

DEVELOPING THE SPD

The Initial Idea

About the same time as the development of children's well-being data for the SIPP, a small group of researchers at the Census Bureau had begun development on an idea to build a longer longitudinal panel of data than the SIPP could provide. The idea was to take an expired SIPP panel and 'bring it back to life' for a series of additional interviews, extending it for several more years. As with SIPP, this survey was focused on the economic well-being of the households -- with concentrations on the jobs and income and program participation of the adults in those households. Other topics, such as child care, health status and child well-being, mostly used as one-time topical modules in the SIPP, were also considered.

Finding a Sponsor

As the project began, the new longitudinal survey was nothing more than an idea. Neither the Census Bureau nor any other agency or funding source had made a formal commitment to conduct the survey. In time, both the Food and Consumer Service of the Department of Agriculture and the Assistant Secretary of Planning and Evaluation in the Department of Health and Human Services expressed interest (the former sought a tool to measure food sufficiency, the latter to measure the status of children), and provided seed money for development.

By the fall of 1994 developmental work had proceeded far enough that the basic goals of the survey were fairly clear. The working design had two fundamental points of focus:

1. The survey would provide information on actual and potential program participants for some time period by extending an existing SIPP panel for this period.

2. The survey would focus on the consequences of program participation on the well-being of recipients, their families, and children.

At this point the proposed content of the new survey spanned a fairly large domain. In addition to the traditional "jobs, income and program" focus of the SIPP, the involvement of a large community of other parties had yielded a sizable set of content material, much of it focused on children. Meetings held in 1995 with various possible sponsors made it clear that the full array of content could not be fielded easily, and preparations began to reduce the content scope by relating it explicitly to whatever funding source could be identified.

Adjusting the Scope

During 1995 the leading candidate for funding the Survey of Program Dynamics (as it was now being called) was through a new 'welfare reform' law. When the 1995 bill died by Presidential veto, so too did our best hopes for funding. With no other source of funding apparent, work on the project stopped. In mid-1996, however, a new bill had been formulated - one that appeared to have much better chances of passage. On August 22, 1996, President Clinton signed the Personal Responsibility and Work Opportunity Reconciliation Act of 1996, which became Public Law 104-193 (The Federal Register 1996). Part of the language of this law directs the Census Bureau to:

"Continue to collect data on the 1992 and 1993 panels of the Survey of Income and Program Participation as necessary to obtain such information as will enable interested persons to evaluate the impact of the amendments made by Title I of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 on a random national sample of recipients of assistance under State programs funded under this part and (as appropriate) other low income families, and in doing so, shall pay particular attention to the issues of out-of-wedlock birth, welfare dependency, the beginning and end of welfare spells, and shall obtain information about the status of children participating in such panels." [Section 414]

With the passage of the law, the Survey of Program Dynamics - the SPD - became a reality. Ten million dollars was to be appropriated to the Census Bureau for each of the fiscal years 1996-2002, insuring a seven-year extension to the existing SIPP panels. Identification of a sponsor also meant that the survey now also had a clear focus, welfare reform, and the status of children was to be one component (Weinberg et al. 1997).

CHILDREN IN THE SPD

Goals

The general goal of the SPD is established by the legislation which funded it. In this respect, the primary focus of the SPD is to extend the kinds of information gathered in the predecessor SIPP panels as it relates to the general issue of welfare program activity and the impact of these changes on the members of households. Children are well within the scope of these analyses. Since the SIPP's primary focus is on adults, a great deal of detailed information is already in place for these people, and the first challenge of the SPD is to continue to collect and enhance these data (Hess and Rothgeb 1998; Weinberg et al. 1997).

Development of data for children in the SPD is a more involved task. Some information on children has been collected during the life of the SIPP panels. Neither the quantity nor scope of this material matches that collected for adults, however. The task has been to try to extend as much as possible the existing data for children from the SIPP while adding new questions that will allow researchers to measure the outcomes of welfare reform for children. The sum of these data collections can be placed into four groups: core children's data, the self-administered adolescent questionnaire, extended measures of well-being, and a children's residential history.

Core Children's Data

The fundamental data collection instrument of the SPD is an annual Computer Assisted Programming Instrument (CAPI) administered interview of the adult members of the sampled household. The one-hour interview collects data associated with welfare reform: employment, earnings, income sources and amounts, and program participation. Other topics collected in the SIPP are also extended here: assets and eligibility, school enrollment, work training, functional limitations and disability, health care utilization, health insurance, and food security. Many of these topics are not part of the routine SIPP interview, but are part of the topical modules in SIPP -- that is, the one-time data administrations that are done throughout the life of a SIPP panel.

By using these data it will be possible to characterize the conditions of children in many fundamental ways. How many of them live in households with income at or below the poverty level? How many live in households where no one holds a paying job, or where one or more transfer programs are utilized? How many have no health insurance? In addition to these data about fundamental circumstances that affect all members of the household, a set of items focused more specifically on children has also been developed. These questions cut across a wide variety of domains, and touch on topics covered in the SIPP (mostly in topical modules), as well as some topics that are completely new to the survey.

During the early development of the SPD, the Census Bureau was inundated with questions about children. This is not unreasonable - the mere mention of the concept of "well-being" conjures up radically different notions in any collection of researchers interested in children. Given the wider focus of the survey, and the limited amount of time allocated for data collection, the Bureau has tried to adopt data items for the SPD that balance aspects of scope, depth, operational practicality, and compatibility with the interview context. The core children's questions focus on the following issues:

* school enrollment

* changing schools

* school expulsions/suspensions

* advanced classes

* enrichment activities (sports, lessons, outings)

* television viewing rules

* family outings

* general health

* disability

* special education needs

* functional limitations

* hospital/doctor/dentist use

* health expenses

* mother's work schedule

* child care arrangements

* child support arrangements

* contact with absent parent

* marital conflict

* parental depression

Some of these items reflect fairly direct behavioral measures of children that might be influenced by changes in households occurring because of welfare reform, for example, school enrollment, family outings, and extracurricular activities (Corcoran 1995; Duncan and Brooks-Gunn 1997; Havenman and Wolfe 1995; Mayer 1997; McNeal; and Smith et al. 1997). Others, like child care arrangements and marital conflict, are intermediate effects between the child and the parents, brought on by welfare changes (Adams, Schulman, and Ebb 1998; Conger, Conger, and Elder 1997; Sameroff et al. 1993). Together, the items should allow us to paint a general and broad portrait of the status of all children -- both those in households directly affected by welfare reform, and those removed from it.

In the original design the Bureau hoped to administer these 'core' children's questions every year for the period 1998-2002. However, new items are planned and being tested for the 1999 administration. To find additional interview time in the existing survey to accommodate these new children's items, and coupled with a Census Bureau commitment to minimize response burden during the 2000 census, the Bureau may have to curtail the administration of some of these new items to less than every year.

Adolescent Self-Administered Questionnaire (SAQ)

Determining the content of the SPD for children has been a long process, looking at many different kinds of possible data collections. One area of study that arose was the consideration of items that would characterize the daily behavior of children, particularly those in ages who might, in just a few years, be possible participants in the new rules of welfare reform, as opposed to simply occupants of households affected by it. For this reason, there was strong support to ask a series of questions of children in adolescence - defined in the SPD as persons ages 12-17. Because of the sensitive nature of some of the questions, and to reduce overall interviewer time in the household, the Bureau developed a self-administered questionnaire that utilizes a cassette tape player and answer booklet that the adolescent(s) in the household can fill out while the core interview is being conducted. In this way we are able to save time and provide a context of confidentiality (from both the interviewer and parents) for the adolescents answering the questions.

The content focus of this 100-item sequence is:

* family routines and responsibilities

* relationship with male and female guardians

* parental involvement

* household rules

* academic attachment

* problem behaviors

* tobacco use

* alcohol use

* drug use

* knowledge of welfare rules

* personal relationships

* sexual behavior

* interaction with absent parent

* attitudes about absent parent

As this list demonstrates, the content coverage of the Adolescent SAQ is substantial. These sections contain questions to evaluate critical dimensions of interest to policy-makers. Questions were developed and chosen in a process that involved child researchers and analysts from across the country, focusing on a wide variety of research and policy issues. Factors that may influence academic success and the transition to adulthood like substance use and relationships with family and friends can be analyzed (see Conger, Conger, and Elder 1997; Sewell and Hauser 1975; and Shilts 1991). Questions on family routines and responsibilities have also been included at the specific request of Congressional staff. Obviously, every possible topic cannot be explored, but we have attempted to cast the net widely in the content domains covered. Some of the more sensitive topics (e.g. sexual behavior) are not asked of 12-13 year olds.

Even though some of these items were believed to be highly sensitive, the pretest evaluation showed that few respondents objected to the questions (Hess 1997). Perhaps more importantly, few parents, informed of the general content of the Adolescent SAQ, refused their children's participation. Analysis of the pretest results indicated distributions that were well within range of similar data collected in other surveys. We have included the Adolescent SAQ as part of the data collection activity for the 1998 SPD and are planning to repeat it as part of the 2001 administration.

Extended Measures of Children's Well-being

As we have noted, early in the development of the SPD, a wide array of possible measures of children's conditions were solicited and received for consideration and inclusion in the survey. After the development period in 1995, the Census Bureau decided to put a portion of these children's topics to the side until the focus of the survey could be better established and more resources could be marshaled to continue questionnaire development. During this period we came to refer to this assortment of topics as possible content for "track 2," a questionnaire section of up to one half hour in length that would focus predominantly on children's issues.

Once the funding and intent of the survey became clearer, we were able to reevaluate our content needs, as well as the response burden these needs would create. By mid-1997 we were aware that the basic core instrument, including its children's content, was far longer than we had anticipated, testing at about one hour in length. As a result, Bureau management decided to restrict track 2 development to no more than 10 minutes. This decision reflected a group of different critical concerns:

* the basic instrument was longer than expected,

* the survey focus was not solely on children,

* some of the proposed children's items, (i.e. assessment tests), were not practical given the survey format, administration, and period for change measurement, and

* human and fiscal resources to accommodate the full breadth of possible children's content were inadequate.

The reaction to the Census Bureau's decision was not completely positive. Some analysts still see the decision to limit the track 2 content (a decision supported by the Office of Management and Budget's oversight committee) as a reversal of commitment to a children's longitudinal survey. In fact, this commitment had never been made because the funding for doing so had never materialized. Considering the context of the authorizing language of the welfare reform bill, the Census Bureau was committed to the extension of two survey panels focusing on social program participation, and was instructed to add children to this focus. This authorization, not the hopes of a group of well-intentioned researchers of children's issues, had to act as the directing force that determines the overall content of the survey.

The challenge was to try to organize the vast array of other possible children's items into a prioritized list for development and testing. Working with Child Trends, a private non-profit research organization, and the Family and Child Well-Being Research Network, a group of primarily academic researchers providing consultative advice to the National Institutes of Child Health and Development, the Bureau created a list of priority topics for the set of "Extended Measures of Child Well-Being," tentatively slated for first administration as part of the 1999 SPD. This supplement is still being reviewed, but the child-related topics being developed at this time include:

* enrichment activities

* gang/criminal activities

* cognitively stimulating activities

* positive interactions

* positive behavior/ social competence

* grades/ achievement

* parental conflict (co-resident parents)

* parental conflict (not co-residential)

* family conflict

* child care quality

These topics focus on relationships and interactions that act as intermediate forces in households, or that are more generic elements of household dynamics affecting children, for example: poverty and early-childhood development (Duncan and Brooks-Gunn 1997); educational attainment and children's circumstances (Havenman, Wolfe, and Spauding 1991); and conflict between parents and the self-confidence and achievement of children (Conger, Conger, and Elder 1997). The number of questions ranges from a few for some topics, to more than a dozen for some of the others. Cognitive testing has been underway since early 1998, and we expect to be able to field many, if not all, of these items in the 1999 survey, subject to the additional time burden they may bring, which will not be known until later this year.

Residential History

The final component of children's data for the SPD is the development of a residential history to document the living situations of children who are part of the sample households. One important indicator of the security and well-being of children is the extent to which they have been a part of intact families (see Currie 1995 and Sameroff et al. 1993). Part of the aim of some of the authors of the welfare reform legislation has been to try to find ways to keep families together. Data collected through the administration of the SIPP and the SPD, when pieced together, will give us some view of how these families have changed over time. However, little is actually known about the status of the children in these households and their whereabouts. For this reason, we have proposed to conduct a retrospective residential history for the children of these households. The idea is to be able to determine the level of chaos in their living situations, and the situations they have spent the majority of their lives in.

Currently, the Census Bureau is exploring several different ways of conducting this data collection. The first attempts at using a series of recall questions imbedded in the CAPI device have not yielded satisfying results. In order to facilitate field administration and data processing, however, the Bureau is trying not to introduce a separate data collection device that would need to be independently keyed. The Bureau is now developing a calendar-assisted CAPI instrument to aid respondents as they attempt to reconstruct the residential histories of their children, some of whom may be 17 years old. In order to reduce burden in the 1999 SPD, and to provide sufficient development time, the residential history is planned to be fielded during the 2000 SPD.

Current Plans and Time Frame

The SPD continues to be a survey in which all decisions have not been finalized. The plans as expressed in this paper represent our best thinking as to what we believe the children's content in the SPD will ultimately be. Changes in emerging data needs or lessons that we learn after the first full administration of the SPD in June and July of 1998 will have an impact on our final design. Also, the results of cognitive testing underway this spring for the extended measures of children's well-being, and the time necessary to collect this information will shape the final survey design.

If we are able to follow through on our plans as now established, we expect to collect a base of recurring children's data in each of the five years of SPD data collections (1998-2002). In addition, in each of these years, we will administer significant additional information regarding the status of children. The adolescent self-administered questionnaire will be fielded in the '98 and '01 surveys. The extended measures of children's well-being will be implemented for the '99 and '02 surveys. In both cases, the passage of several years time should be long enough to detect change but short enough to assess how quickly things are changing. The residential history for children will be a part of the 2000 SPD. Together with the children's data collected as part of the earlier SIPP panels, and the detailed data on the parents and households these children live in, we should have sizable data on the lives of children as welfare reform takes hold.

Our first longitudinal file, planned for late 1999, will give researchers a unified dataset combining SIPP and SPD children's data. In the interim, researchers will be able to construct their own custom datasets, once the preliminary version of the SPD 1998 data collection is released in late '98 or early 1999.

CONCLUSION

The path in developing children's content for the Survey of Program Dynamics has not always been clear and direct, but the survey is dedicated to providing a significant collection of information about children in the midst of a changing public welfare system. Using these data, researchers should be able to undertake a wide array of investigations and analyses about the conditions of children over roughly a ten-year period, starting with the child well-being modules in 1992 and 1993 and extending through the SPD years, 1998-2002. The SPD focuses on a large series of different topical indicators of children's conditions and provides a major data resource that will allow a multitude of analyses of major changes in the public welfare system, changes that directly affect many adults, and in doing so, affect their children.

REFERENCES

Adams, Gina, Karen Schulman, and Nancy Ebb. 1998. Locked Doors: States Struggling to Meet the Child Care Needs of Low-Income Working Families. Children's Defense Fund. March.

Conger, Rand, Katherine Conger, and Glen Elder. 1997. "Family Economic Hardship and Adolescent Adjustment: Mediating and Moderating Processes." Pp. 288-310 in Consequences of Growing Up Poor, edited by G. Duncan and J. Brooks-Gunn. New York: Russell Sage.

Corcoran, Mary. 1995. "Rags to rags: poverty and mobility in the United States." Annual Review of Sociology. 21: 237-67.

Currie, Janet. 1995. Welfare and the Well-Being of Children. Chur, Switzerland: Harwood Academic Publishers.

Duncan, Greg and Jeanne Brooks-Gunn. 1997. Consequences of Growing Up Poor. New York: Russell Sage.

The Federal Register. 1996. The Federal Register. August 22, 1996. Public Law 104-193. Government Printing Office.

Havenman, Robert, Barbara Wolfe, and James Spauding. 1991. "Educational achievement and childhood events and circumstances." Demography. 28: 133-58.

Havenman, Robert and Barbara Wolfe. 1995. "The determinants of children's attainments: a review of methods and findings." Journal of Economic Literature. 33: 1829-78.

Hess, Jennifer. 1997. "SPD pretest field representative debriefing summary." Unpublished report. Bureau of the Census.

Hess, Jennifer and Jennifer Rothgeb. 1998. "Measuring the impact of welfare reform: issues in designing the survey of program dynamics questionnaire." Proceedings of Statistics Canada Symposium XIV.

Mayer, Susan. 1997. What Money Can't Buy: The Effect of Parental Income on Children's Outcomes. Cambridge, MA: Harvard University Press.

McNeal, Ralph. 1995. "Extracurricular activities and high school dropouts." Sociology of Education. 68: January. 62-80.

Sameroff, Arnold J., Ronald Seifer, Alfred Baldwein, and Clara Baldwin. 1993. "Stability of intelligence from preschool to adolescence: the influence of social and family risk factors." Child Development. 64: 80-97.

Sewell, William and Robert Hauser. 1975. Education, Occupation and Earnings: Achievement in the Early Career. New York: Academic Press.

Shilts, Lee. 1991. "The relationship of early adolescent use to extracurricular activities, peer influence, and personal attitudes." Adolescence. 26: Fall. 613-17.

Smith, Judith, Jeanne Brooks-Gunn, and Pamela Klebanov. 1997. "Consequences of growing up poor for young children." In Consequences of Growing Up Poor, edited by G. Duncan and J. Brooks-Gunn. New York: Russell Sage.

U.S. Census Bureau. 1991. Survey of Income and Program Dynamics User's Guide. Second edition. Washington, D.C.

Weinberg, Daniel H., Vicki J. Huggins, Robert A. Kominski, and Charles T. Nelson. 1997. "A survey of program dynamics for evaluating welfare reform." Statistics Canada Symposium XIV.

1. This paper reports the results of research and analysis undertaken by Census Bureau staff. It has undergone a more limited review than official Census Bureau publications. This report is released to inform interested parties of research and to encourage discussion.

2. This paper reports the results of research and analysis undertaken by Census Bureau staff. It has undergone a more limited review than official Census Bureau publications. This report is released to inform interested parties of research and to encourage discussion.

SPD Working Papers

SPD Publications