Table of Contents | Previous | Next |
CHAPTER 10 - CHANGES IN FAMILY WELL-BEING, STRENGTHS, AND CHALLENGES
INTRODUCTION
The family services component of the National Transition Demonstration Program was designed to help strengthen families to realize personalized goals. To this end, family service workers were employed to help families deal with crises, secure basic services (medical, dental, child care), set intermediate and long-term goals for personal and family growth, and participate in relevant educational and training programs. This chapter concerns the extent to which families reported receiving additional services and supports, on a yearly basis, and how they gained access to these. Analyses concerning differential service patterns for the demonstration and comparison groups are reported first, followed by subsequent analyses about use patterns among different family types.
SOURCES OF INFORMATION
Information about family well-being was obtained from family informants in the fall and spring of kindergarten, and in the spring of each year thereafter for three consecutive years. Two instruments served as the primary sources of information, including:* The Family Background Interview, developed for the National Transition Demonstration Study, included questions about family income and a wide array of support services including educational programs, job counseling and training, financial assistance, health services, and other support programs.
* The Supplemental Interview for Families, also developed for the National Transition Demonstration Study, concerned the assistance that families received in gaining access to services for children and adults in the family.
CHANGES IN FAMILY WELL-BEING
|
Changes in family income. Families were asked in the spring of each year whether or not their family’s income had changed in the past 12 months, and, if so, whether the increase or decrease was large, modest, or small. Families reported an annual increase in family income from the time their child went to Head Start (as a 4-year old) to the kindergarten year. In each of the next three years, at least 25 percent of these former Head Start families reported earning higher incomes (see Figure 10.1). Reports of increased family income were comparable for demonstration and comparison groups.a
|
Job/employment training. In the spring of each year, families were asked about participation in job training or assistance programs. These programs included direct job training programs (including skills training in the form of courses or on-the-job training), job placement, employment counseling, and JOBS (Jobs, Opportunities, and Basic Skills Training Program), among others. About 25 percent of families reported participating in some employment training or assistance program during their children’s first year in school (see Figure 10.2)b. In subsequent years, however, the percentage dropped markedly to just under 10 percent of families. There were no significant differences between demonstration and comparison groups at baseline or in subsequent years.
|
Educational programs. Family informants indicated which, if any, educational programs they had received during the past year. Examples of these programs include GED preparation, literacy or adult reading, English as a Second Language, government training, vocational technical school, community college or junior college, and four-year college or university programs or courses. Each year, between 20 to 25 percent of families reported participating in some adult education program (see Figure 10.3).c In all years, slightly more of the demonstration families reported participating in educational programs (2-3 % higher).
|
Family Support Programs. Families reported about receipt of a variety of other family support programs, including cash or cash-alternative programs (e.g., AFDC, SSI, food stamps), counseling (e.g., substance abuse, mental health), housing-related assistance (e.g., public housing), energy programs, and others. In the first year, more than 75 percent of the former Head Start families received at least one type of family assistance, although by the end of the fourth year, participation rates dropped to approximately 65 percent (see Figure 10.4).d This trend is consistent with the interpretation that these families are becoming more self-sufficient over time. It is also noted that slightly more demonstration families report receiving benefits of a variety of family assistance activities each year. This suggests that the Transition Demonstration Program, probably through the efforts of the Family Service Workers, assisted and/or encourages families to participate in such programs.
Specific types of services reported by families. Much of the decline in the percentage of families who reported receiving assistance is attributable to declines in AFDC and/or SSI payment received to supplement household income (see Table 24). Similarly, there was a large decline in nutritional services (WIC and/or food stamps) in families.
| SERVICES | First year | Second year | Third year | Fourth year |
|---|---|---|---|---|
| Employment
assistance |
||||
| Demonstration | 25.80% | 9.10% | 8.90% | 9.20% |
| Comparison | 15.40% | 8.10% | 9.90% | 8.80% |
| Cash assistance
(AFDC and/or SSI) |
||||
| Demonstration | 44.40% | 41.90% | 39.50% | 35.90% |
| Comparison | 42.90% | 39.10% | 36.70% | 33.30% |
| Nutritional
services (WIC and/or food stamps) |
||||
| Demonstration | 64.50% | 57.00% | 50.00% | 45.30% |
| Comparison | 60.40% | 52.00% | 47.10% | 40.60% |
| Energy
program |
||||
| Demonstration | 13.30% | 13.30% | 11.40% | 8.70% |
| Comparison | 13.10% | 12.80% | 10.00% | 7.90% |
| Literacy
Program |
||||
| Demonstration | 0.30% | 0.60% | 0.70% | 0.40% |
| Comparison | 0.30% | 0.40% | 0.60% | 0.10% |
| Social/Mental
health services |
||||
| Demonstration | 6.30% | 6.10% | 6.70% | 6.30% |
| Comparison | 4.90% | 5.80% | 4.90% | 5.90% |
| Public
Housing |
||||
| Demonstration | 21.20% | 19.10% | 19.30% | 16.90% |
| Comparison | 19.60% | 17.20% | 17.00% | 15.70% |
| Unmet needs
(no services provided) |
||||
| Demonstration | 29.10% | 18.20% | 17.30% | 17.10% |
| Comparison | 26.00% | 17.30% | 15.50% | 16.40% |
The percentage of families -- less than one percent -- who reported participating in adult literacy programs (to learn to read or to improve reading skills) is surprisingly low, particularly because the Transition Demonstration Programs were aware of a high need in this area. It is possible that parents may not have labeled these as “literacy programs” and thus under-reported their participation, or they may have been uncomfortable reporting this to an interviewer. It may also be that local programs were not successful in engaging parents to participate in reading programs.
The percentage of families reporting receipt of mental health services, including counseling and treatment for substance abuse, was slightly higher for those in the demonstration than the comparison group, and this remained steady over time. The percentage of both demonstration and comparison families living in public housing showed a steady decline from kindergarten to third grade.
Each year, families were asked if they had family needs that were not being met. In the kindergarten year, between 25 to 30 percent of families indicated they had unmet needs, which they further specified as including money for basic needs, safer and more affordable housing, medical and dental care for children and/or adults, employment, improved literacy and education programs for both children and adults, improved parenting skills, assistance with child care, violence prevention, and substance abuse counseling and treatment. By the end of their child’s fourth year in school, many families reported unmet needs -- between 16 to 17 percent. Thus, unmet needs declined, at the same time families received fewer services.
Assistance in obtaining services. It was hypothesized that more demonstration than comparison families would report that school or project personnel assisted them in obtaining services. Based on interviews in the spring of their child’s fourth year in school, demonstration families were significantly more likely to report that school personnel helped them obtain services, while comparison families reported they found these services on their own initiative.e This suggests that many former Head Start families have the skills to find and obtain needed services without having special service coordinators. However, when there are supportive personnel available to help make referrals, families take advantage of this assistance.
DIFFERENCES IN SERVICE UTILIZATION PATTERNS BY FAMILY TYPE
Additional analyses explored whether different types of families participated to a greater extent in job, educational, and family support programs. Logistic regression analyses were used to predict program participation by different family types (see Table 25).f
Job training programs. Single Parent Welfare families were more than two and a half times more likely than Resourceful families to participate in job training. This finding likely relates to efforts to move families from welfare to work, although this study was conducted prior to the full implementation of the 1996 Welfare Reform Act. Interestingly, Highly Mobile families also were nearly 50 percent more likely to receive job training, in contrast to ESL families, who were approximately 50 percent less likely to participate in job training than were the Resourceful families.
Educational programs. Adults in the Single Parent Welfare, Highly Mobile, and Homeless families were 50 to 60 percent more likely to participate in educational programs than were those in the Resourceful families. Not unexpectedly, Mother Absent families were the least likely to participate in educational programs, in part because the primary caregiver typically was older (e.g., a grandparent) and perhaps less able to participate in formal educational opportunities.
| Family
Type |
Odds ratio | ||
|---|---|---|---|
| Job Training Programs | Educations Programs | Income and Family Support Programs | |
| Resourceful | 1.0 (Reference group) |
1.0 (Reference group) | 0.1* (0.1-0.2) |
| Single Parent Welfare | 1.9* (1.7-2.2) |
1.6* (1.4-1.8) |
1.0 (Reference group) |
| Foreign Language | 0.5* (0.4-.06 ) |
1.5* (1.2-1.9) |
0.1* (0.1-0.2) |
| Highly Mobile | 1.5* (1.1-2.0) |
1.6* (1.2-2.2) |
0.6 (0.3-1.2) |
| Recently Homeless | 2.7* (1.8-3.9) |
1.5 (1.1-2.2) |
3.8 (.5-27.5) |
| Mother Absent | 1.0 (0.7-1.2) |
0.6* (0.4-0.8) |
0.5 (0.3-1.0) |
| Chronically Ill | 1.1 (0.8-1.6) |
0.9 (0.7-1.3) |
0.4 (0.2-0.9) |
Family support programs. As expected, the Resourceful families were the least likely to use income and family support programs, especially when compared to Single Parent Welfare families. Similarly, the Foreign Language families were less likely than others to use income and other family supports. As noted earlier, these families are much like the Resourceful families except for the language they speak at home. That is, the foreign language families have a lower percentage receiving financial assistance (AFDC and/or SSI), more fathers who are actively involved in their children's lives, and higher rates of parental employment.
Overall patterns of usage. Both the Homeless and the Single Parent Welfare groups are the most likely to participate in assistance programs of all types -- job training, educational, and income and family support programs. Service use by Mother Absent families is quite low, while there is some variation across other family types.
SUMMARY FINDINGS
Former Head Start families participated in a variety of educational, employment, and family support programs. Income trends for these families, both demonstration and comparison, were generally positive over the early elementary school years, and reliance on AFDC and SSI and other income supports declined steadily for these families. No statistically significant differences between groups were noted regarding quantity or types of programs or services received. Single Parent Welfare families were the most likely to participate in job training, educational, and family support programs. Families in the comparison group were more likely to access services on their own, while those in the Transition Demonstration group reported that they received referral assistance from family service workers who were hired by the Transition Demonstration Program. Over time, families in both groups experienced a decline in the challenges or stressors they experienced. Similarly, the rates at which families reported they had unmet needs declined considerably over the first four years in elementary school.
aChi-square analyses were completed cross-sectionally to review the relationship between initial treatment condition (demonstration or comparison) and reported change in family income (large increase, modest increase, small increase, small decrease, modest decrease, large decrease, no change). The initial analysis sample included the 7.515 former Head Start families in the baseline sample. Cross-sectional samples included all families who completed family interviews and responded to the question, with the following sample sizes resulting:
| Spring, first year: | 6,126 total [1,389-- without interview; 3,274 - demonstration; 2,852 - | comparison] |
| Spring, second year: | 5,457 total [2,958 - without interview; 2,892 - demonstration; 2,565 - | comparison] |
| Spring, third year: | 5,252 total [2,263 - without interview; 2,772 - demonstration; 2,480 - | comparison] |
| Spring, fourth year: | 5,176 total [2,339 - without interview; 2,739 - demonstration; 2,437 - | comparison] |
bIn the spring of each year of program participation, families were asked if they had participated in one or more of five job training or employment counseling programs during the past year. For each year (cross-sectional data, all families with valid family interviews) a variable was created to indicate how many programs the family participated in that year. Distributions were highly skewed to the right each year. Efforts to treat summed variables as continuous and enter them as outcome variables in repeated measures ANOVA or MANOVA models failed because the distributions so badly violated the normality assumptions underlying those models. Therefore, for each of the families who responded to the question in each of the four interviews (i.e., had complete data for that question; n = 4,050 total; 2,168 = demonstration; 1,859 = comparison), a variable was created to indicate the number of years in which the family reported having participated in at least one job training or employment counseling program. Chi-square analyses were completed to assess the association between initial treatment condition (demonstration, comparison) and (1) participation in at least one program during the year (cross-sectional analyses) and (2) number of years of participation in job training programs (longitudinal analyses). No significant differences were noted. Additional analyses compared number of years of participation in job training programs with years in demonstration school (range of values, 0 to 4) but did not reveal significant differences. (back)
cIn the spring of each year of program participation, families were asked if they had participated in one or more of nine educational programs (i.e., GED, Head Start training, literacy/reading, ESL, government training, vocational technical school, community or junior college, college or university, other) during the past year. For each year (crosssectional data, all families with valid family interviews) a variable was created to indicate how many programs the family participated in that year. Distributions were highly skewed to the right each year. Effort to treat summed variables as continuous and enter them as outcome variables in repeated measures ANOVA or MANOVA models failed because the distributions so badly violated the normality assumptions underlying those models.
Therefore, for each of the families who responded to the question in each of the four interviews (i.e., had complete data for that question; n = 4,050 total; 2,186 = demonstration; 1,864 = comparison), a variable was created to indicate the number of years in which the family reported having participated in at least aneducational program. Chi-square analyses were completed to assess the association between initial treatment condition (demonstration, comparison) and (1) participation in at least one program during the year (cross-sectional analyses) and (2) number of years of participation in educational programs (longitudinal analyses). No significant differences were noted. Additional analyses compared number of years of participation in educational programs with years in demonstration school (range of values, 0 to 4) but did not reveal significant differences. (back)
dIn the spring of each year of program participation, families were asked if they had participated in one or more of fifteen family assistance programs (AFDC, SSI, food stamps, public housing assistance, medical assistance, unemployment insurance, WIC, energy program assistance, home visits, mental health services or counseling, nutrition services, parenting education, literacy education, social services, other) during the past year. For each year (crosssectional data, all families with valid family interviews) a variable was created to indicate how many programs the family participated in that year. Distributions were highly skewed to the right each year. Effort to treat summed variables as continuous and enter them as outcome variables in repeated measures ANOVA or MANOVA models failed because the distributions so badly violated the normality assumptions underlying those models.
Therefore, for each of the families who responded to the question in each of the four interviews (i.e., had complete data for that question; n = 4,050 total; 2,168 = demonstration; 1,859 = comparison), a variable was created to indicate the number of years in which the family reported having participated in at least a family support program. Chi-square analyses were completed to assess the association between initial treatment condition (demonstration, comparison) and (1) participation in at least one program during the year (cross-sectional analyses) and (2) number of years of participation in family support programs. No significant differences were indicated. Additional analyses compared number of years of participation in family support programs with years in demonstration school (range of values, 0 to 4) but did not reveal significant differences. (back)
eSupplemental interviews were optional on the part of the local sites. A total of 23 sites administered Supplemental Interviews for Families to at least some percentage of the families in their final year of study participation. A total of 2,827 supplemental family interviews were available for analysis (2,001 demonstration and 1,826 comparison). No specific statistical analyses were completed to test hypotheses about associations because the data were incomplete and sampling plans varied substantially by site. Thus, the outcomes reported on these data are descriptive only. (back)
f In kindergarten, there were 7,078 families (94% of the 7,515 Head Start families in the analysis data set) for whom a family interview was available. Of the 7,078 families, 784 (11.1%) were deleted from these analyses because their interviews were missing values on two or more key challenge index variables. Therefore, there were 6,294 families (3,318 demonstration and 2,976 comparison) for whom a kindergarten family challenge index score was created. (back)
Logistic regression analyses were completed using the analysis sample outlined in notes b through d above. Logistic models were constructed separately for the three program types - job training, educational, and family support. All models included participation in at least one program during the four years as the outcome variable and family type entered as the predictor variable (entered as six dummy variables indicating cluster membership, with the Resourceful family group serving as the reference group for job training and educational program models and the Single Parent Welfare group serving as the reference group for the family support model).* SAS Proc Logistic procedures were utilized. Odds ratios and 95% confidence intervals were computed and reported in the text.
* The Single Parent Welfare group was chosen as the reference group for the family support programs model because a large percentage (85%) of those families reported receiving AFDC, one of the family support programs used to establish the dichotomous participation variable. Using the Resourceful families group as the reference group caused a quasi-complete separation of the data (since virtually all of the Single Parent Welfare families fell into the yes-participation group), resulting in an unstable model. Use of the Single Parent Welfare group as the referent group yielded a stable, interpretable model.
| Table of Contents | Previous | Next |