Programs that Work
New Hope Project
Healthy and Safe Children
Children Succeeding in School
Students performing at grade level or meeting state curriculum standards
Families increasing economic self-sufficiency
Children and youth not engaging in violent behavior or displaying serious conduct problems
Age of Child
Early Childhood (0-8)
Middle Childhood (9-12)
Type of Setting
Child Care / Preschool
Community-Based Service Provider
Health Care Provider
Type of Service
Type of Outcome Addressed
Cognitive Development / School Performance
Poverty / Welfare
Evidence Level (What does this mean?)
The New Hope Project operated from 1994 to 1998 in two inner-city areas of Milwaukee, Wisconsin. New Hope offered low-income individuals and families the opportunity to use a comprehensive set of integrated program services designed to increase income, financial security, and access to full-time employment. In the two target locales, all adults whose earnings were below 150 percent of the federal poverty level and who were willing to work full-time were eligible to apply for enrollment in the program. Applicants need not have been welfare recipients nor have children.
New Hope offered two program benefits to all participants:
- Community service-based full-time job opportunities for participants unable to find full-time work (or part-time job opportunities to supplement an existing part-time job) in the private job market
- Personalized services assisting participants in job searches, child care, and other employment-related needs.
For participants who worked full-time (30+ hours per week) New Hope also offered:
- A monthly earnings supplement designed to raise participants’ income to exceed the poverty threshold for the household
- Subsidized health insurance
- Child care subsidies.
New Hope operated outside the existing public assistance system, and many participants continued to receive various types of public assistance while participating. The program was funded by a consortium of local, state, and national organizations interested in work-based antipoverty policy, as well as by the state of Wisconsin and the federal government.
The program targeted low-income individuals ages 18 or older, including both male and female adults with and without children. Eligible participants had earnings at or below 150 percent of the federal poverty level and were willing and able to work a minimum of 30 hours per week.
The New Hope Project enrolled 1,367 low-income adults drawn from two inner-city areas in Milwaukee. Applicants who met the eligibility criteria (i.e., had earnings which did not exceed 150 percent of the federal poverty level and were willing to work at least 30 hours per week) were randomly assigned to participate in New Hope or serve as a control group. From the initial 1,367 participants, 10 cases lacked baseline data, resulting in a total of 678 individuals in the treatment (New Hope) group and 679 individuals in the control group. Very few statistically significant differences were found between the treatment and control groups on a range of baseline measures, suggesting that randomization had been successful.
Individuals who were randomized to the treatment group were eligible to receive New Hope's benefits upon providing verification that they met the 30-hour minimum work requirement. Program benefits were available for three years, and treatment group participants were eligible to access any and all New Hope benefits and services for any portion of that time, subject to the requirements outlined above. Applicants randomized to the control group received no program benefits but were enrolled in the study to serve as a comparison group and remained eligible for other programs in the community that were not related to New Hope. Evaluations of the New Hope program collected data on both treatment and control group participants during and after the program.
Three assessments of the New Hope Project have been conducted by the Manpower Demonstration Research Corporation and members of the MacArthur Foundation on Successful Pathways Through Middle Childhood. The studies were conducted two, five, and eight years after program applicants were randomized into treatment and control groups (Bos et al., 1999; Huston et al., 2003; Miller et al., 2008). The two-year evaluation was conducted when treatment group participants were still eligible to receive program benefits. The five- and eight-year evaluations were conducted only for families with children ages 1-11 at the time parents applied for the program. These assessments occurred after the program services had been terminated. The evaluations investigated numerous short- and long-term outcomes among treatment and control group adults as well as their children. The treatment group was defined as those adults (and their children) who were randomly assigned to be eligible for the New Hope program benefits, whether or not they actually fulfilled the eligibility requirements (30 hours of verified work per week) or accessed the benefits available. Most treatment group members (87 percent) did use at least one program benefit, but the treatment can most precisely be interpreted as eligibility for (rather than use of) New Hope benefits.
Information used in the evaluations came from five sources:
- a baseline enrollment form
- administrative records (i.e., unemployment insurance earnings, public assistance benefit records, tax records, and New Hope administrative records)
- in-person surveys with parents and children conducted two, five, and eight years after parents were randomly assigned to the program or control group
- teacher reports of children's school performance and social behavior collected two, five, and eight years after parents were randomly assigned
- additional data gathered for 44 families taking part in an ethnographic study.
Applicants to the program, i.e., the full study sample, were 72 percent female. The racial/ethnic breakdown was 51 percent black, 27 percent Hispanic, 13 percent white, and 9 percent other ethnicities. Approximately 70 percent of the full study population lived in a household with children. Thirty-eight percent of the full sample was currently employed, 97 percent had earned less than $15,000 in the 12 months prior to program intake, and 71 percent had received Aid to Families with Dependent Children (AFDC) benefits or some other type of welfare or Medicaid within the same 12-month period. Approximately 57 percent of the sample had received a high school diploma or GED. At the beginning of the study, 418 individuals (roughly one-third) were employed 30 or more hours per week, while 935 worked part time or not at all.
An expanded survey was conducted in the homes of participants who had children. The Child and Family Study (CFS) targeted 745 adult sample members who were living in households with at least one child ages 1-11 (55 percent of the total sample). The sample excluded 67 Asian-American families—most of whom were Southeast Asian refugees—because of language barriers and because many of the measurement instruments were culturally inappropriate for them. Overall response rates at each assessment ranged from 76 to 82 percent. The five-year survey was completed by 591 adults and 927 of their children. In addition, teachers of children who were living in New Hope households were sent a questionnaire to assess school behavior and performance. Data from teachers were collected on 420 children who were between the ages of 6 and 16 at the time of the survey. The eight-year survey was completed by 595 families and 866 focal children between the ages of 9 and 19. Teacher reports on 540 children were also received.
The evaluation findings reported below present outcomes for the CFS sample only, since these results focus on outcomes related to children and families. If a family had more than one child in the survey age range, two children were identified as focal children. If there were more than two eligible children per family, the focal children were randomly selected, with the restriction that opposite-sex siblings were given preference over same-sex siblings. The evaluation included a total of 1,140 focal children. The parents in the sample were in many respects similar to those in other studies in which samples were drawn from individuals receiving welfare. When they applied for New Hope, over half were not employed, and about 80 percent were receiving AFDC, general assistance, food stamps, and/or Medicaid. The majority had never been married. Slightly over 10 percent were married and were living with their spouse, and almost half had three or more children. Slightly over half were African-American, and over one-quarter were Hispanic.
Results, for the most part, are presented for the sample of families who responded to the two-year, five-year, and eight-year follow-up surveys; however, some findings are presented for the entire set of 745 families in the sample. Children's ages at the two-year follow-up ranged from 3 to 12; at the five-year follow-up, they were 6 to 15; and at the eight-year follow-up, they were 9 to 18. The child-reported outcomes presented are based on interviews with the roughly 900 children who responded to the surveys. Teacher-reported outcomes are based on the reports of teachers in the teacher-survey samples.
Key Evaluation Findings
Poverty, parental earnings and employment
- During the two years following the intervention, the treatment group worked an average of 3.0 quarters per year, compared with 2.6 and 2.7 quarters for the control group. More than two years after randomization, statistically significant differences in the number of quarters worked were no longer observed (Huston et al., 2005).
- Employment effects varied depending on the services participants utilized. During the two-year period following randomization, the only New Hope benefit that was associated with an increase in the number of quarters worked per year was placement in a community service job. Adults in the treatment group who were placed in community service jobs worked on average 2.6 more quarters (for an increase of 11 percent) relative to the comparison group. In contrast, New Hope's benefits that were intended to increase employment through indirect means (i.e., through child care vouchers, health insurance, and/or supplemental income) were found to have no effect on the number of quarters beneficiaries worked (Gibson, 2003).
- Two years after randomization, treatment group participants earned 19 percent more than control group participants. This gap could be traced to a one-time increase in earnings in the first quarter after randomization, which was attributable to job placement and earnings supplements. In the subsequent quarters, trends in earnings growth did not differ between groups, but the initial advantage of the treatment group was maintained (Dunifon, 2005).
- Five years after randomization, the percentage of families with records-based earnings and transfer income below the poverty line was 59.3 and 64.6 percent in the treatment and control groups, respectively. Eight years after randomization, the gap had narrowed to a statistically insignificant 63.1 and 67.1 percent, respectively (Duncan et al., 2008).
- A study by Huston et al. (2003) concluded that New Hope was responsible for a 20 percent reduction in the number of families living below the poverty line five years after the intervention.
- Huston et al. (2003) found that the program had no significant impact on welfare receipt but did increase earnings-related income by 10.5 to 23.3 percent one and three years after randomization, respectively. However, no such significant increases were observed four and five years after randomization.
- During the three years in which program benefits were available, significantly more treatment group families (42 percent) than control group families (34 percent) had incomes above the poverty threshold (Duncan et. al., 2009).
Children's academic achievement, cognitive skills and social behavior
- Children in New Hope families had significantly higher academic achievement than those in control families two years after the intervention. Of note, boys in the program group families scored one-third of a standard deviation higher than boys in control group families on the SSRS Academic subscale (Huston et al., 2001).
- Boys in treatment group families received teacher ratings of classroom behavior (e.g., attention, independent work) and social skills that were 10 percent higher than their peers in control group families two years after randomization (Bos et al., 1999).
- Similarly, Huston et al. (2001) found that boys in treatment group families scored higher on classroom behavior by 0.38 of a standard deviation two years after the intervention.
- Two years after randomization, boys in treatment group families had fewer behavior problems (including aggression) relative to their counterparts in control group families, and they had more positive behaviors as measured by the Positive Behavior Scale (PBS). The PBS is an assessment tool that was developed to assess behavior (e.g., social competence and sensitivity, compliance and self-control, and autonomy) among children from economically disadvantaged families (Epps and Huston, 2007).
- For both problem behavior and positive behavior, treatment group boys scored about one-third of a standard deviation better than did control group boys two years after randomization (Huston et al., 2001).
- Five years after randomization, boys in treatment group families continued to score 0.3 standard deviations higher on teacher-reported overall achievement compared with control group boys. Significant effects were not found among girls (Huston et al., 2005; Duncan et. al., 2009).
- The New Hope study included measures of the families' and children's environments that might account for program effects on children. Among treatment group families, increases in income improved parents' effective child management, which in turn was associated with improvements in children's test scores and teacher- and parent-rated school performance, as well as children's behavior (e.g., internalizing problems, externalizing problems, delinquent acts) five years after the intervention (Mistry et al., 2006).
- The significant effects on teacher ratings of boys' positive social behavior remained apparent at the five-year assessment. The treatment group boys scored about one-fourth of a standard deviation higher than the control group boys, but there was not a significant difference on behavior problems. Teachers rated treatment group girls about one-fourth of a standard deviation lower (at the 10 percent significance level) than control group girls on positive behavior (Huston et al., 2005).
- As part of the program evaluation, children in treatment and control group families completed several assessments of their psychological well-being. One such test required children to listen to vignettes about peers' physical or social actions and interpret the characters' intentions. Both five and eight years after randomization, boys in treatment group families were significantly less likely than boys in control group families to perceive hostile intent in vignettes. There was no program effect for girls (Huston, et al., 2005).
- A study by Huston, Gupta, et al. (2008) found only modest effects on academic performance—including reading scores, parental ratings of literacy, and overall achievement—eight years after randomization.
- Another study by Huston, Walker, et al. (2008) found slightly more pronounced effects on boys' reading scores and overall achievement and math scores among younger children (under 13 years of age) eight years after randomization.
- Boys in treatment group families averaged 2.6 percent higher on reading scores eight years after the intervention (Miller et al., 2008).
- Eight years after the intervention, youth of both genders in New Hope families had significantly higher parent ratings of positive social behavior than did those in control group families. However, there were no differences in teacher ratings of positive behavior or problem behavior, nor were there differences in self-reported delinquent behavior (Miller et al., 2008).
- Similarly, Huston, Gupta, et al. (2008) found no significant differences in the delinquent behaviors of children from treatment and control group families by age, gender, or on the whole, eight years after the intervention.
Charitable and health and human service organizations, foundations, and state and local governments
New Hope was funded by a large consortium of local, state, and national organizations. Donors who contributed $250,000 or more to the program included the State of Wisconsin Departments of Workforce Development and Health and Human Services, the U.S. Department of Health and Human Services, the National Institute for Child Health and Human Development, the John D. and Catherine T. MacArthur Foundation, the Mott Foundation, the Helen Bader Foundation, the Rockefeller Foundation, the Ford Foundation, the William T. Grant Foundation, Ameritech/Wisconsin Bell, Northwestern Mutual Life, the Wisconsin Energy Corporation, and the Annie E. Casey Foundation.
The five-year and eight-year evaluations were funded by the National Institute of Child Health and Human Development.
The annual taxpayer cost of the program was approximately $7,250 per family (in 2009 dollars).
New Hope required that participants work a minimum of 30 hours per week. Participants were required to verify fulfillment of the work requirement by submitting wage stubs. For those individuals who fulfilled the work requirement, the following benefits were available:
- Earnings Supplement: The supplement was designed to complement the state and federal Earned Income Tax Credits (EITCs), that is, the refundable tax credits for low-income working families, so as to raise the income of full-time workers above the poverty level. The Earnings Supplement was implemented to ensure that additional work effort by participants or higher wages earned would lead to a true increase in overall income.
- Subsidized Health Insurance: Health insurance was available to any New Hope participant who did not already have access to coverage through an employer or a government-provided health plan, or who wished to supplement an employer-provided plan. New Hope required a small co-payment by participants that increased with income.
- Child Care Subsidy: A child care subsidy was available to parents with children under age 13. The program allowed parents to select their own licensed child care facility and then paid a rate comparable to the state-provided subsidy, but differed from the state subsidy by paying providers directly. New Hope participants were required to pay a small co-payment, which increased with income.
- Community Service Jobs: Job opportunities were made available to participants who were unable to find full-time employment after two months. The community service jobs provided wage-paying positions with local nonprofit organizations. Participants were required to apply for these positions and could be removed from the positions if job performance was unreliable or unsatisfactory. Each placement in a community service job was limited to a six-month stint, but participants could take up to two community service jobs over the course of the program. The hours worked in these jobs were counted toward New Hope’s full-time work requirement, on which eligibility for other program benefits was based.
In addition, New Hope provided limited traditional employment services. New Hope participants were assigned a project representative who assisted the participants in identifying their interests and aptitudes, job searching, employment maintenance, and acquiring job training, as well as finding child care services and other job-related needs.
The direct service providers came from a variety of backgrounds (e.g., some were trained in social work) and had professional or volunteer experience in human services or community action. In addition to the direct service providers, New Hope was staffed by management personnel, persons responsible for accounting and information services, and individuals providing administrative support.
Issues to Consider
This program received a "proven" rating. The evaluations employed rigorous standards, including a randomized sample of nearly 1,400 individuals and an analysis sample of 561 to 745 families (depending on the particular outcome measure). Furthermore, the evaluations were conducted by the Manpower Demonstration Research Corporation (MDRC), an organization that was independent of the implementing agencies, and members of the MacArthur Foundation. The evaluations found significant benefits for the New Hope group when compared with the control group in the realms of employment, earnings, poverty, and children's cognitive and social skills and academic achievement.
Outcomes suggest that the majority of the income and employment effects were concentrated in the early years of the study and tended to fade out within five years after randomization. The New Hope program operated for three years, during which time participants were eligible to receive a range of work supports (i.e., earnings supplement, child care subsidies, and employment). It is therefore not entirely surprising that the program's effects were largest during the early years and faded after these supports ended. There is some evidence of lasting effects for the sample as a whole, such as the statistically significant reductions in poverty that occurred throughout the eight-year period and the higher wages earned by the treatment group after eight years, particularly among vulnerable subgroups.
Several studies compared results among ex-ante sub-groups. Participation in New Hope was found to have had a greater impacts on individuals who were not employed full-time at baseline (see Bos et al., 1999), faced one or more barriers to employment (see Yoshikawa et al., 2003), or had children (see Duncan et al., 2009). Among the subgroups defined by the type of New Hope benefit utilized, positive effects on future employment were observed only among participants who had been placed in community service jobs.
The impacts of New Hope were larger and more consistent for boys of New Hope parents than for girls, especially when teacher reports were considered. In fact, teachers rated girls from New Hope families more negatively than they did control group girls, on a range of measures including appropriate classroom behavior, positive social behavior, and problem behaviors.
The most recent study of the effects of New Hope was completed by MDRC and members of the MacArthur Foundation in 2008. It confirms that many of the program effects continue to fade, but several remain as much as eight years after the intervention. The most widespread lasting effects are related to children's academic achievement.
Since 2005, New Hope has administered a new program called Supporting Families, which provides services to help individuals who are among the most difficult to employ (e.g., those who have minimal work histories, criminal records, or no high school diploma) find and retain employment. Supporting Families is not a demonstration project like the earlier New Hope program.
New Hope was implemented in Milwaukee, Wisconsin, from 1994 to 1998.
For information on the New Hope program:
The New Hope Project, Inc.
2620 West North Avenue
Milwaukee, WI 53210
Tel (414) 270-7580
Fax (414) 270-7579
For information on the evaluations of New Hope:
Manpower Demonstration Research Corporation
16 East 34th Street, 19th Floor
New York, NY 10016-4326
Tel (212) 532-3200
Fax (212) 684-0832
University of California-Irvine
Aletha C. Huston
University of Texas-Austin
The New Hope Project website (http://www.newhopeproject.org) provides more detailed information about the New Hope program, links to available publications and reports, and a listing of recent news and events.
Bos, J. M., A. C. Huston, et al. (1999).
New Hope for People with Low Incomes: Two-Year Results of a Program to Reduce Poverty and Reform Welfare.
Duncan, G. J., A. C. Huston, and T. Weisner. (2009). Higher Ground: New Hope for the Working Poor and Their Families. New York: Russell Sage Foundation.
Duncan, G., C. Miller, et al. (2008). New Hope's Eight-Year Impacts on Employment and Family Income. MDRC.
Dunifon, R. (2005). "The Labor-Market Effects of an Anti-Poverty Program: Results from Hierarchical Linear Modeling," Social Science Research, 34(1): 1-19.
Epps, S. R., and A. C. Huston (2007). "Effects of a Poverty Intervention Policy Demonstration on Parenting and Child Behavior: A Test of the Direction of Effects," Social Science Quarterly, 88(2): 344-365.
Gibson, C. M. (2003). "Privileging the Participant: The Importance of Sub-Group Analysis in Social Welfare Evaluations," American Journal of Evaluation, 24(4): 443-469.
Huston, A. C., A. E. Gupta, et al. (2008). New Hope's Effects on Social Behavior, Parenting, and Activities at Eight Years. MDRC.
Huston, A. C., C. Miller, et al. (2003). New Hope for Families and Children: Five-Year Results of a Program to Reduce Poverty and Reform Welfare. MDRC.
Huston, A. C., G. J. Duncan, et al. (2001). "Work-Based Antipoverty Programs for Parents Can Enhance the School Performance and Social Behavior of Children," Child Development, 72(1): 318-336.
Huston, A. C., G. J. Duncan, et al. (2005). "Impacts On Children Of A Policy To Promote Employment And Reduce Poverty For Low-Income Parents: New Hope After 5 Years," Developmental Psychology, 41(6): 902-918.
Huston, A. C., J. T. Walker, et al. (2008). Long-Term Effects of New Hope on Children's Academic Achievement and Achievement Motivation. MDRC.
Miller, C., A. C. Huston, et al. (2008). New Hope for the Working Poor: Effects After Eight Years for Families and Children. MDRC.
Mistry, R. S., E. D. Lowe, et al. (2006). "What Earnings and Income Buy—The 'Basics' Plus 'A Little Extra': Implications for Family and Child Well-Being." In H. Yoshikawa, T. S. Weisner, and E. D. Lowe, ed., Making It Work: Low-Wage Employment, Family Life, and Child Development. New York: Russell Sage Foundation, 173-205.
Yoshikawa, H., K. A. Magnuson, et al. (2003). "Effects of Earnings-Supplement Policies on Adult Economic and Middle-Childhood Outcomes Differ for the 'Hardest To Employ.'" Child Development, 74(5): 1500-1521.