Get Updates
2008 Association for Public Policy Analysis and Management Abstracts
"Evaluating the Impacts of Supplemental Educational Services Under the No Child Left Behind Act" Brian Gill and John Deke
The federal No Child Left Behind Act (NCLB) created a large increase in the market for after-school academic services by mandating that chronically low-performing schools make such services available. In particular, NCLB requires districts with federal Title I schools that fall short of state standards for three years or more to offer supplemental educational services (SES) to their students from low-income families who attend those schools. SES are tutoring or other academic support services offered outside the regular school day by state-approved providers free of charge to eligible students. Parents can choose the specific provider from an approved list. SES providers come in all varieties, including national for-profit firms, local nonprofits, faith-based organizations, institutions of higher education, and local school districts (which are permitted to become approved providers unless they are themselves identified for improvement under NCLB). As of the summer of 2007, 1,859 providers across the country had been approved and were serving about half a million students. Despite the rapid growth of SES and their key role in NCLB, little information has been available about their effectiveness. A randomized experiment cannot be conducted to examine SES because depriving eligible students of services based on a lottery would violate the provisions of NCLB. Lacking the opportunity to conduct a randomized experiment, the U.S. Department of Education has commissioned two studies using quasi-experimental methods to assess the effects of SES. The first of these used a within-student longitudinal design (i.e., student fixed-effects) to examine whether the achievement trajectories of participating students improved after receiving SES. This study produced a report (Zimmer et al, 2007) finding achievement gains for participating students in several school districts. With the aim of gaining more-rigorous evidence on SES impacts, the department commissioned a follow-up study that will use a regression discontinuity design to assess impacts. While the regression discontinuity design should permit stronger causal inferences about the impacts of SES, it is possible only in school districts that are oversubscribed for services. In oversubscribed districts, NCLB permits districts to allocate services to the lowest-achieving eligible applicants—thereby creating the opportunity to examine whether there is a discontinuity in subsequent achievement at the cut point used for assignment.
"The Evaluation of Teacher Preparation Models" Daniel Player, Jill Constantine, and Tim Silva
Every year, thousands of new teachers pass through hundreds of different teacher preparation programs and are hired to teach in the nation’s schools. In recent years, “alternative teacher certification” (AC) programs or routes have expanded rapidly, offering an increasingly popular route into teaching that differs from that offered by traditional certification (TC) programs. TC programs generally require trainees to complete their training before they begin teaching full time. In contrast, AC programs typically allow trainees to start teaching full time before they have completed their training. Despite the expansion of these new routes into teaching, there exists little research to provide guidance as to the effectiveness of different teacher training strategies. The study uses a mix of experimental and nonexperimental methods to assess the relationship between preparation models and student and teacher outcomes. The study focuses on the performance of teachers from two alternative models of teacher preparation (one model with less selective recruiting and substantial coursework requirements, and one model with less selective recruiting and minimal coursework requirements) compared with traditionally prepared teachers teaching in the same schools and grades. The sample includes the nearly 3,000 Kindergarten through 5th grade students of 190 traditionally and alternatively certified teachers. Within each school in the study, students in the same grade were randomly assigned either to a novice teacher from one of the AC training programs or to a novice teacher from a TC training program. All students took achievement tests at the beginning and the end of the school year from which we estimate an experimental estimate of the effect of teachers from a particular AC program compared with teachers from a TC program. To supplement the experimental estimates, we collected detailed information about course requirements and teacher practices through interviews with each AC teacher’s program, classroom observations to determine the quality of teacher practices, and principal evaluations of teacher performance. Using nonexperimental analysis, we use these data to assess whether any effect of AC teachers is related to the characteristics of preparation programs or to the characteristics of teachers who choose different programs.
"Impacts of Comprehensive Teacher Induction on Student and Teacher Outcomes: Results from a Large-Scale Randomized Experiment" Amy Johnson, Eric Isenberg, Martha Bleeker, Sarah Dolfin, and Steven Glazerman
The presentation offers impact findings from the first year of a major national experiment on comprehensive teacher induction, an intervention designed to help beginning teachers through their first year on the job. We randomly assigned more than 900 first-year teachers from over 400 elementary schools in 17 urban school districts across the United States to either a treatment group that received comprehensive mentoring and induction services or a control group that received the prevailing level of induction services. The high intensity services were provided by the New Teacher Center at the University of California at Santa Cruz (NTC) and the Educational Testing Service (ETS), who hired and trained veteran teachers in each district to be released full time for mentoring new teachers. We build on prior evidence on the effects of teacher induction programs which have relied primarily on observational data and other nonexperimental methods (e.g. Youngs 2002; Smith and Ingersoll 2004; Fletcher et al. 2008). For each teacher, the research team measured classroom practices through direct observations, “value added” to student achievement growth through test score analysis, and teacher retention through surveys conducted annually for three years. The study achieved high response rates (nearly 90 percent on most instruments), demonstrated faithful implementation of the “treatment,” and found virtually no deviation of teachers from their originally assigned treatment status. This study provides definitive evidence on the effectiveness of these interventions on several dimensions. We examined whether teachers had improved practices, produced greater gains in student achievement, and whether they stayed in their schools, districts, and the teaching profession. We also examined the types of teachers who stay, move, or leave, including their background characteristics like SAT/ACT scores and teacher preparation, and “quality” measures such as their observed teaching quality and their value added to student achievement growth. In addition, we estimated impacts on teacher attitudes (feelings of preparedness and satisfaction). As a randomized controlled trial, the study employs straightforward, transparent methods. The impacts are measured as the differences in average outcomes–with regard to both classroom practices and student achievement–between the treatment and control groups. To improve precision, the study employed regression adjustment based on data collected through an extensive baseline survey. The researchers also accounted for the nesting of teachers within schools in all hypothesis tests.
"Covering Kids and Families Influences on Children's Enrollment and Retention in Public Coverage" Christopher Trenholm
We are assessing the influence of state and local policy and procedural changes on new enrollment in public programs. Using regression methods to analyze a panel data set of new State Children's Health Insurance Program (SCHIP) and Medicaid enrollment taking place at the county level across 10 states from 1999 to 2006. Enrollment data are drawn from Medicaid Statistical Information System and linked with census and labor statistics data to construct covariates that can effectively control for variation in the numbers of eligible but uninsured. State/time interaction variables are used to identify significant gains in enrollment, which are then linked to potential changes in policies and procedures (such as simplification or outreach), as well as linked to specific Covering Kids and Families activities. Preliminary analysis suggests that elimination of face-to-face interviewing and other steps to destigmatize coverage in public programs may be particularly important to expanding coverage of low-income children. Additional results indicate substantial variation in new enrollment across states (and through time) that may be linked to broader simplification and outreach policies. While advocates for children's coverage have identified a long list of potential best practices for enrolling children, this study is one of the first to test empirically the value of these practices. Findings suggest that not all these practices are of equal importance and that continued steps to destigmatize and simplify coverage in public programs may be essential to insuring greater numbers of low-income, uninsured children. Findings provide explicit and statistically grounded recommendations also for state policymakers and children's health advocates on the most effective policies for expanding coverage to low-income children.
"Coverage Advocacy Can Be Sustained After the Funding Ends" Beth Stevens and Sheila Hoag
Foundation-funded programs aimed at improving health insurance coverage run for a finite period. When the funding stops, what happens to the programs the funding supported? This presentation reviews the experiences of state Covering Kids and Families (CKF) projects and coalitions after their four-year grants ended, looking at the legacies this project has left. Surveys of CKF project directors and coalition leaders indicate that 70 percent of projects and 85 percent of coalitions continued in some manner more than 6 months after grant funding ended. States have played a crucial role in sustaining CKF’s work: Half of the CKF project directors, for example, said state leaders were continuing to simplify and improve coordination between Medicaid and SCHIP; CKF project directors said states had committed in-kind support in 10 states, and financial support in 10 states. Interviews with former CKF employees also reveal an expansion of community capacity to deal with problems related to access to insurance and children’s health insurance, as many former CKF staff moved into positions that used knowledge and skills gained or improved at CKF. It is possible to sustain coalition and program activities after funding ends to maintain and expand coverage. Program activities can continue after grants end, but post-grant funding challenges hamper efforts. States were disappointed at the ending of CKF funding as they believed the coalitions were fruitful in sustaining improvements in enrollment and retention practices and in increasing coverage.
"Studying the New Universal Preschool Initiative in Los Angeles County: Measurement and Analytic Challenges in the Context of a Culturally and Linguistically Diverse Community" John Love, Sally Atkins-Burnett, Cheri Vogel, Nikki Aikens, and others
Among the 10 million residents of Los Angeles County, more than 200 languages are spoken. Of the more than 150,000 4-year-olds in the county, 61 percent are Latino, while fewer than 20 percent are white, 8 percent Asian, and 9 percent African American. About 44 percent of public school kindergarteners are considered English Language Learners (ELL). Spanish is the most dominant language, but substantial numbers of ELLs speak Armenian, Korean, and Cantonese. Some 60 percent of all public school students receive free or reduced-price meals. In fall 2007, we selected a representative sample of almost 2,000 4-year-olds in 98 preschool programs from among all the programs operated by the Los Angeles Universal Preschool (LAUP) program. We conducted assessments with the children across the full range of developmental domains important for school readiness, interviewed their parents and teachers, and obtained ratings of children’s behavior by their parents and teachers. We also observed their preschool classroom environments and teacher-child interactions using two observation instruments, the widely used Classroom Assessment Scoring System (CLASS) and a measure of teacher-child language interactions developed for this study, the Language Interaction Snapshot (LISn). We will show how we have addressed various ELL assessment challenges (e.g., differential home language experiences, rates of language and literacy development within each language, exposure to classroom instructional practices, and determination of the most-appropriate language of assessment), as well as limitations of current assessment approaches (e.g., psychometric properties of measures and measurement equivalence analyses). We describe the observations conducted to assess classroom instructional practices and the content of teacher-child interactions, with a particular focus on interactions with Spanish and English language speakers in addition, we present preliminary findings showing children’s developmental gains from fall to spring (2007-2008), the relationships seen between those gains, family characteristics, and the quality and interactions observed in LAUP classrooms. Our conclusions will highlight the importance of obtaining carefully designed and collected data both on children’s development across diverse cultural and linguistic groups and on the classroom environments in which children are learning.
"Evaluation of Jamaica’s PATH Program" Jim Ohls and others
This paper presents the findings of an evaluation of a social safety net initiative, the Programme of Advancement Through Health and Education (PATH), which was undertaken by the Government of Jamaica, beginning in 2001. The main objectives of the initiative, which is operated by the Ministry of Labour and Social Security, are to better target social assistance benefits to the poor and to increase human capital by conditioning receipt of the benefits on participants meeting certain requirements for school attendance and health care visits. The paper assesses both the program's implementation and its impacts on key education and health outcomes. The main impact findings are as follows: (1) PATH has been effective at accomplishing its objective of encouraging households to send their children to school with greater regularity; (2) PATH is estimated to have increased school attendance by approximately 0.5 days per month. The estimated increase is about three percent over the baseline level, and it is statistically significant; (3) PATH was also successful in meeting its objective of increasing the use of preventive health care for children in PATH families. The results of the statistical analysis suggest that health care visits for children 0-6 increased by approximately 38 percent as a result of PATH. While PATH was successful at increasing school attendance and preventive health visits, there is no evidence that it was able to affect longer term outcomes such as grades, advancement to next grade, or health care status.
"Evaluating the Impact of the Teacher Advancement Program Using Experimental and Quasi-Experimental Methods" Allison McKie and Steven Glazerman
This paper presents the study design and early findings from an experimental evaluation of a version of the Teacher Advancement Program (TAP) implemented in the Chicago Public Schools. The district recruited schools to participate in the pilot implementation and Mathematica randomly assigned the schools to implement immediately (in 2007) or delay implementation. The delayed implementation schools serve as the controls for the 2007-2008 school year. The random assignment will be repeated for two more cohorts of schools in 2009. In all, there will be 4 cohorts of 10 schools implementing TAP on a staggered schedule. In addition, Mathematica is gathering data on approximately 300 comparison schools in the district. The study includes teacher and principal surveys, student test data, and qualitative information on program implementation. The presentation will cover findings through the end of the first year of implementation. We expect to have early estimates of TAP impacts on teacher retention, student achievement, and school climate/practices.
"Home Investment: Comparing the Benefits and Costs of a Home Visitation Program for Hard-to-Employ TANF Recipients" Alicia Meckstroth, Andrew Burwick, and Quinn Moore
TANF work requirements have encouraged states to develop new approaches for improving the work and well-being of their most disadvantaged TANF families. Intensive services for this population can be expensive, however, raising questions of whether their benefits outweigh costs. This paper presents findings from a benefit-cost analysis conducted for one innovative program for hard-to-employ TANF recipients, Building Nebraska Families (BNF). BNF used home visits to bring specialized life skills education and mentoring to TANF clients across rural Nebraska. Services were intended to help participants enhance basic life skills and personal and family functioning so that they might transition from welfare to sustained employment and self-sufficiency. The BNF evaluation is part of the Rural Welfare-to-Work Strategies Demonstration Evaluation, which Mathematica is conducting under contract to the Administration for Children and Families (ACF). The evaluation used a rigorous experimental design in which eligible program applicants were randomly assigned to a treatment group or a control group. Our analysis compares the experiences of these two groups (358 treatment and 242 control group members) across a 30-month follow-up period. Particular attention is focused on the subgroup of more disadvantaged, or very-hard-to-employ, TANF recipients (149 treatment and 103 control group members). For the full sample, BNF improved employment toward the end of the 30-month follow-up period but did not affect earnings. For the more disadvantaged subgroup, BNF led to significant, robust impacts on employment and earnings. Impacts were particularly large during the last six months of the follow-up, when program group members’ reported earnings were 56 percent higher than corresponding control group members’ earnings. BNF’s average cost per participant ($7,383) created a high threshold for costeffectiveness. During the 30-month follow-up, BNF’s benefits to society did not outweigh its costs for the full sample or the more disadvantaged subgroup. However, the program’s estimated net costs to society for the more disadvantaged subgroup were over one-third less than those for the full sample. Projecting impacts on income and fringe benefits beyond the follow-up period, we estimate that impacts for the full sample would need to persist for an additional 9.7 years for BNF’s benefits to exceed its costs. For the more disadvantaged subgroup, the time horizon for benefits to exceed costs is much shorter, 1.7 years.
"Can an Intensive, Comprehensive Program Help At-Risk Youth to Succeed? Final Impacts of the Quantum Opportunity Program Demonstration" Allison McKie, Allen Schirm, and others
From July 1995 through September 2001, the United States Department of Labor and the Ford Foundation conducted a demonstration of the Quantum Opportunity Program (QOP) in seven sites. Operated by community-based organizations, QOP was mainly an after-school program that began offering intensive and comprehensive services to youth with low grades when they entered ninth grade in 11 high schools with high dropout rates. QOP’s primary goals were to increase the rates of high school graduation and enrollment in postsecondary education or training. Its secondary goals were to improve high school grades and achievement test scores and to reduce risky behaviors, such as substance abuse, crime, and teen parenting. To evaluate the program, we randomly assigned nearly 1,100 eligible youth to QOP or control groups. The program model specified that the youth in the QOP group were to be served even if they transferred to other schools, dropped out of school, became incarcerated, or became inactive in QOP. In addition to financial incentives, the program provided case management and mentoring, supplemental education, developmental activities, community service activities, and a wide range of supportive services year round for five years to those youth who had not graduated from high school. To estimate QOP’s impacts on high school performance and graduation, postsecondary education or training, and risky behaviors, we conducted four surveys, administered achievement tests in reading and mathematics, and collected high school transcripts. This paper presents and discusses QOP’s impacts on outcomes measured when most sample members were 22 to 25 years old. It also summarizes earlier findings. Impacts are presented for the demonstration as a whole as well as the individual sites and subgroups relevant to targeting and service delivery. In addition to impacts, the paper presents estimates of program participation and costs, and summarizes our assessment of the demonstration sites’ fidelity to the program model.
"Health Insurance and Health Care Access Among Social Security Disability Insurance (SSDI) Beneficiaries Before and After SSDI Entitlement" Gina Livermore, David Stapleton, and others
This study uses National Health Interview Survey (NHIS) data from the years 1994, 1995, and 1996 linked to administrative records from the SSDI and Medicare programs to analyze the characteristics, health insurance coverage, and health care access of SSDI awardees who were observed in the NHIS while in the six-year window surrounding SSDI entitlement—from the 36th month before the SSDI entitlement to the 36th month after. Analyses of this synthetic SSDI awardee cohort illustrate the progression of disability and changes in insurance status and health care access and utilization. We find that SSDI awardees are more likely to be uninsured than the general population of persons ages 18 to 64, even two to three years prior to SSDI entitlement. As expected, the likelihood of having employer-sponsored health insurance through one’s own employment declines markedly during the periods after entitlement, but having such coverage through a family member declines only slightly over the periods before and after entitlement. Health care utilization, as well as problems related to access, increase during the periods closer to SSDI entitlement, peaking during the periods right before and after entitlement. The majority of awardees had entitlement dates that preceded the SSDI decision date. About one-third had used up 12 months or more of the Medicare waiting period by the time they had received their award, and a little over ten percent had used up the entire 24-month Medicare waiting period at the time of their award. The findings suggest that elimination of the Medicare waiting period for SSDI beneficiaries would help a large share of awardees who lack insurance coverage during the period when their health care needs are greatest. The findings also suggest that elimination of the waiting period alone will not address the lack of insurance coverage for a rather large percentage (about 22 percent) who are uninsured during the periods preceding entitlement, and for those whose award decisions are made at long intervals after SSDI entitlement.
"Is the Craze Over Regression Discontinuity Designs in Program Evaluations Justifiable? The Issue of Statistical Power" Peter Schochet
Regression discontinuity (RD) designs are increasingly being used by researchers to obtain unbiased impact estimates of public programs and interventions. These designs are applicable when an exogenous, continuous “scoring” rule is used to assign the intervention to study units (i.e. schools, students, or hospitals). Units with scores above a cutoff value are assigned to the treatment group and units with scores below the cutoff value are assigned to the comparison group (or vice versa). Thus, unbiased estimates can be obtained by comparing the outcomes of treatment and comparison group members using regression models that control for a smooth function of the continuous score. An impact occurs if there is a “jump” or “discontinuity” in the regression line at the cutoff score. RD designs are sometimes touted as a viable alternative to gold-standard random assignment (RA) designs in impact evaluations. However, it is often overlooked that RD designs require much larger sample sizes than RA studies to achieve the same level of statistical precision for the impact estimates. Thus, although RD designs may have operational advantages over experimental designs, they could be more costly to conduct due to larger required sample sizes. This is a particular problem for evaluations that use clustered designs (such as in education where treatments are assigned to entire schools or classrooms), and hence, where statistical power is an important concern even under RA designs. In our proposed paper, we will compare required sample sizes under RD and RA designs. We will examine design effects for different realistic distributions of the score variable used to assign the treatment to study units, for clustered and unclustered designs, for different treatment-comparison group splits, and for fuzzy designs (where some units do not “comply” with their treatment assignment). The paper will provide a mathematical framework for addressing the power issue and provide formulas that can be applied by researchers. The paper will also provide concrete recommendations on appropriate sample sizes in a wide range of policy settings.
"Modernization in Florida: A New Model for Public Assistance" Scott Cody, Emily Sama Martin, and others
Faced with budget constraints, increased attention to fiscal accountability, and pressure to increase program access, public agencies must identify strategies to deliver comparable or enhanced social services with fewer resources. This study of Florida’s welfare intake modernization initiative–“ACCESS Florida”–examines one state’s experience with reducing costs without privatizing. In Florida, the state Economic Self Sufficiency agency restructured staff roles, leveraged technological efficiencies, streamlined eligibility rules, and used partner organizations to cut costs and reduce participation barriers. These changes drastically altered the way that clients in Florida access social services, including TANF, food stamps, and Medicaid. The result was a new administrative structure that was a substantial departure from the old “caseworker” model of public assistance. This case study draws on interviews with staff and clients and tracks changes in costs and participation. The study concludes that ACCESS Florida resulted in sweeping changes and substantial cost savings. While many clients prefer ACCESS Florida to the old caseworker-based system, other clients and some staff expressed concern over the changes. There is no evidence that any one type of client has program access restricted under ACCESS Florida, but there is some evidence that the changes could reduce participation in some locations. The study’s results highlight key questions that public managers should weigh when considering social service modernization.
"Private Management of Public Schools" Brian Gill and others
One of the reforms characterizing public education in the United States over the past two decades has been the creation and growth of regional and national Education Management Organizations (EMOs) that operate public schools under contract. We briefly describe the history and role of EMOs in the United States. We then examine research on student achievement in EMO schools and on the operational characteristics of these schools. We conclude with a discussion of the implications of the research and of the need for new research to address gaps in what is known about the operations and effectiveness of EMOs.
"The Role of Faith Community Liaisons in Charitable Choice Implementation" Ann Person, Liz Clary, and Pam Winston
The Charitable Choice provisions of the 1996 welfare reform law sought to “level the playing field” for sectarian social service providers in their access to certain types of federal funding. Since then, thirty-five states and dozens of cities have formally established a liaison with the faith community to encourage public sector partnerships with faith-based organizations (FBOs) in the provision of social services. This study aims to enhance understanding of the role played by faith community liaisons (FCLs) in promoting states’ and localities’ implementation of the Charitable Choice and equal treatment rules, which govern how they contract with FBOs. It is funded by the United States Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. Through case studies of seven states and one city, this study describes and analyzes the function of FCLs, with particular emphasis on the cultural, political, and organizational contexts that influence their activities. Specifically, the study addresses five research questions: (1) What is the current status of the FCL function within and across states? (2) How do FCL characteristics—such as their mandate, legal status, formal and informal duties, background, and outreach efforts—appear to be associated with accurate understanding and communication of Charitable Choice and equal treatment rules to potential or actual contractors? (3) How have FCLs contributed to observable changes in the ways in which states reach out to and partner with FBOs, and how has this evolved over time? (4) What models of FCL implementation and activity appear to be the most promising in promoting partnerships with FBOs and encouraging effective and appropriate implementation of Charitable Choice rules? (5) What kinds of information or guidance on effective practices do states and localities need to support appropriate and effective public partnerships with FBOs? Data are being collected through interviews with FCLs and their staff, public officials, agency staff, and representatives of multiple FBOs at each site. Additionally, the research team is reviewing program documents and other materials from each site, as well as examining public legal and financial data related to resources available to and leveraged by FCLs and their affiliates. The study identifies several “promising practices” that appear to be associated with effective implementation of Charitable Choice, as well as contextual supports and obstacles that may be of interest to policymakers, practitioners, and researchers interested in public partnerships with faith-based and community organizations.
"Are Public Schools Really Losing Their ‘Best’? Assessing the Career Transitions of Teachers and Their Implications for the Quality of the Teacher Workforce" Daniel Player and others
Most studies that have fueled alarm over the attrition and mobility rates of high-quality teachers have relied on proxy indicators of teacher quality, which recent research finds to be only weakly correlated with value-added measures of teachers performance. We examine attrition and mobility of teachers using teacher value-added measures for early-career teachers in North Carolina public schools from 1996 to 2002. Our findings suggest that the most effective teachers tend to stay in teaching and in specific schools. Contrary to common expectations, we do not find that more-effective teachers are more likely to leave more-challenging schools.
"Variations Across States in Eligibility for the Food Stamp Program" Jim Ohls and Karen Cunnyngham
United States public assistance programs often face tensions between achieving uniform national objectives and accommodating state differences in need. During the 1990s, two important assistance programs took divergent paths on this dimension. The Aid for Families with Dependent Children program was replaced by Temporary Assistance to Needy Families (TANF), a decentralized program that allows greater diversity across states, while the Food Stamp Program (FSP) continued to have largely uniform eligibility criteria in all states. Starting in 2000, however, a series of policy changes in the FSP have given states increased latitude to set their own FSP asset eligibility policies. States now have two tools to streamline eligibility certifications and increase the number of households that are eligible for the FSP. They can (1) expand the types of households that are categorically eligible for the FSP and thus are exempt from the FSP asset test; and (2) align state FSP vehicle rules with the rules in place for certain other programs. All states have taken advantage of this increased flexibility, leading to considerable diversity in asset eligibility policies across states where once there were uniform requirements. This variation in eligibility policies raises two important questions. First, how widely do asset eligibility policies vary across states, in terms of their effects on FSP eligibility? Second, what are the implications for equity, in terms of how similar households are treated in different states? We address these questions with a microsimulation model that uses Survey of Income and Program Participation as the underlying database and incorporates Current Population Survey data, FSP administrative data, and detailed information on state asset eligibility rules. Through the use of a small area estimation methodology, the model is able to estimate FSP eligibility on the state level. We use this capacity to compare households' eligibility across states under the FSP rules in place in 2006 with household eligibility under a simulated return to the old uniform FSP rules. We also examine eligibility under additional simulated FSP asset policies. We estimate that 68 percent of households eligible for the FSP in 2006 would have been eligible regardless of the state in which they reside and 32 percent of households would have been eligible in at least one state and ineligible in at least one state.
"Predictors of Low-Income Families’ Participation and Engagement in the Federal Early Head Start Program" John Love, Cheri Vogel, and others
In 1996 and 1997, 1,513 low-income pregnant women and families with infants enrolled in 17 Early Head Start programs as part of the Early Head Start Evaluation. Programs provided families with opportunities to participate in home visitation, attend group socialization meetings, enroll their children in center-based child care, and be involved in parenting education. The national evaluation documented the experiences of these families as part of an in-depth implementation study and created indices of multiple dimensions of participation and engagement throughout the first 28 months after enrollment. The families provided extensive information at the time of enrollment about their demographic characteristics, some data on family needs, and (for about half the sites) scores on a maternal depression scale. This paper reports on analyses that will inform policymakers and program planners about the parental demographic, needs, and mental health factors that predict various patterns of participation and engagement in program services. Family participation and engagement are complex and dynamic processes, requiring us to measure multiple dimensions. Over 28 months of program enrollment, we measured the frequency of the key services mentioned earlier and the duration of enrollment as reported by parents, along with a staff-reported rating of overall program engagement and involvement. Hierarchical cluster analyses guided us in constructing four groups of participants: (1) sustained strong (22 percent); (2) substantial (29 percent); (3) weak (25 percent); and (4) early leavers (24 percent). Staff ratings, which took into account intensity of participation as well as engagement, showed that 39 percent were highly involved and 8 percent were rated as uninvolved. Others’ involvement was either variable or consistently low. Data showed that most of the parents rated as highly involved were either sustained strong or substantial participants, while most of the variable-involvement families were classified as substantial or weak participants, indicating that staff ratings correspond well with data documenting participation and engagement. Our paper explores the roles of family characteristics and needs at intake in predicting patterns of program participation. For all 17 sites, we predict participation patterns using demographic characteristics and family needs scores. For the eight sites that administered the Center for Epidemiological Studies Depression scale, we also explore the risk-of-depression index as a predictor. We also explore variations by program approach and the role of program implementation of the Head Start Program Performance Standards. Findings will help understand factors affecting program participation and be useful as programs devise strategies for more fully engaging their intended participants in services.
"Family-Friendly Benefits: Helping Women But Hurting the Firm?" Annalisa Mastri
Although women have gained numeric parity with men in law and medical school cohorts and make up approximately 30 percent of Masters of Business Administration recipients, they continue to be disproportionately underrepresented at the upper levels of law firms, medical school faculties and businesses. One common explanation is that combining careers with family is exceedingly difficult, and causes women to drop out of the work force or remain concentrated at the lower levels of professional work. In recent years, many companies have adopted family-friendly policies, such as Employee Assistance Plans (EAPs) and emergency child care, in an attempt to retain these capable female employees. However, research has not yet examined whether these family-friendly policies actually have their intended effect. This paper investigates how the adoption of family-friendly policies is linked to the percentage of female employees and employee turnover. Using longitudinal data on approximately 1,400 law firms over ten years, the author estimates models with firm fixed effects that allow credible estimation of these relationships in the presence of firm heterogeneity. There are three main findings. First, Employee Assistance Plans, dependent care flexible spending accounts, and emergency child care are associated with increases in the percentage of associates who are female. Lending credibility to the results, payment of bar dues and paternity leave, which shouldn't have an effect on the gender composition of associates, are shown to have no relationship to the percentage of female associates. Second, family-friendly policies are linked to reductions in the turnover rates of associates. This, combined with the first finding, indicates that family-friendly policies help retain female employees. Third, by linking a subset of the firm-level data to supplemental data on firm profitability over the same period, evidence shows that adoption of these policies is associated with increases in profitability for the firms that adopt them. Contrary to an argument advanced in some circles, adopting such policies does not hurt firm profitability or lead firms to go out of business. The results of the study are generalizable to firms with similar organizational structures and a similar workforce composition. Overall, these findings suggest that family-friendly policies can play an important role in retaining female employees without hurting firm profitability.