![[Table of Contents]](/icons/up.gif){kind=icon}
In this section we describe some of the ways in which Even Start has changed over the past decade: changes in program size and administration, changes in the types of families served, changes in project services, and changes in program costs.
The administration of Even Start changed substantially between 1989 and 1998, and the program significantly expanded its outreach. Exhibit 1 shows how the numbers of projects and families served as well as federal appropriations for Even Start have grown over the past decade. The first few years of the program saw a steady increase in the number of projects funded, as well as in the total number of families served. While the total number of families served by Even Start continues to grow each year as a function of the larger numbers of projects being funded, the average number of families in a project has decreased from a high of 62 in 1991 to the current average of 54 in 1998.
When Even Start was federally administered, programmatic guidance was centralized; it was provided through the Even Start Program Office, housed in the Department's Office of Elementary and Secondary Education. Hence, decisions and responses to questions about program operations came from a single, centralized source. A consistent message about Even Start was heard by all grantees. When state agencies were given control of Even Start, in 1992, significant variability was introduced into program administration, for understandable reasons. For example, Even Start differs across states due to decisions about where to house the program (typically, either the state agency in charge of child development or the agency in charge of adult and vocational education), states' concerns about serving their own areas of greatest need, states' historical level of involvement in administering federal pass-through funds, and states' depth of experience in family literacy.
One benefit of moving Even Start's administration from the federal to the state level is that local projects receive more direct and more regular technical assistance on programmatic issues from Even Start State Coordinators. Department staff and State Coordinators coordinate technical assistance issues and maintain a close relationship through written guidance, attendance at meetings, and electronic communication.
Another consequence of the shift from federal to state control is increased variability in the funding period for Even Start projects. Federally-administered Even Start projects were funded for one or two four-year grant periods. Under state administration, the Even Start statute continues to allow project periods up to four years, with individual projects being limited to a maximum of eight years of funding. However, some states fund projects one or two years at a time, while other states have maintained the federal four-year grant cycle. Regardless of the length of the project period, the statute requires that all projects be assessed for continuation funding on an annual basis. State Coordinators and local project staff have noted that uncertainty in future funding poses challenges to establishing healthy long-term collaborative relationships with school districts and other providers.
States also vary in the amount of funding they award to local projects. When states were given control of funding decisions, some states decided to award a larger number of grants with decreased amounts of funding, thereby spreading funds across their states in what they perceived to be a more equitable manner. The statute requires states to award local projects at least $75,000 per year, although each state has the discretion to award one smaller grant. Data for 1998 show that the average federal cost per project is approximately $169,000 (Exhibit 1), and most state Even Start grants are between $100,000 and $200,000 per year (St. Pierre & Noonan, 1998, p.6).
There also is variation in the level of state involvement in local evaluation. The Even Start legislation requires that local projects participate in a national evaluation as well as conduct their own local evaluation. Some State Coordinators organize local evaluations so that each project within the state collects the same kind of data about operations, participation, and outcomes. Other states allow each project complete autonomy to conduct their own local evaluation, and as a result, projects within a single state often conduct widely different kinds of studies. While no single local evaluation model is appropriate across all local projects, many State Coordinators refer local projects to the Guide to Local Evaluation (Dwyer & Frankel, 1998) and the Even Start Quality Indicators developed by RMC Research Corporation (Dwyer, undated).
One of the Department's objectives for the Even Start program is that projects will reach their target population of families that are most in need of services. Over the years, Even Start has widened its reach to include more diverse populations, and compared to earlier years, Even Start is now serving more families with greater evidence of disadvantage, including more teen parents and greater proportions of Hispanic/Latino families. Data in this section address the Even Start performance indicator on recruitment of most in need families.
Recruitment of Most in Need. Family income has been consistently low. At least 90 percent of Even Start families in 1996-97 had incomes at or significantly below the federal poverty level. The need for basic and literacy skills education for parents has increased since the early years of the program. Forty-five percent of 1996-97 new enrollees had reached at a maximum only 9th grade before enrolling; 42 percent had reached 10-12th grades but had not graduated.
The percentage of parents under the age of 20 has grown from 9 to 13 percent in the past three years, reflecting a programmatic emphasis on serving teenage parents. At the same time, there has been an increase in the percentage of infants and toddlers in Even Start (from 27 percent in 1989-90 to 30-32 percent in 1995-97) reflecting a statutory change requiring service to children across at least a three-year age span. Among parents under age 20 who are new to Even Start, the vast majority (85 percent) have children who are under two years old—more than double the percentage for parents who are older than 20. The largest group of children continues to be those between three and five years of age (Tao, Gamse & Tarr, 1998, p.32-34).
The proportion of Hispanic families in Even Start has more than doubled over the past 10 years, from about 17 percent of all enrollees in 1989-90 to 39 percent in 1996-97 (Tao, Gamse & Tarr, 1998, p.44-46). This increase is substantially greater than the 10 to 11 percent change observed in the U.S. population over the same time period (U.S. Bureau of the Census, 1998). Increases in the percentage of Hispanic families have led to increases in the proportion of non-native English speakers, whose participation in adult education focuses upon English-as-a-Second-Language.
Some of the changes in the demographic characteristics of participants have consequences for the nature of services offered and received. Teenage parents, whose children tend to be younger than the children of other participants, have different needs both for early childhood educational services and for adult education services. Non-native English speakers have different needs for adult literacy education—in their own languages as well as in English—than native English speakers.
Another objective included in the Department's performance indicator plan for Even Start is the extent to which local projects provide comprehensive instructional and support services of high quality and intensity to all families in a cost-effective manner. Even Start families are required to participate in each of the three core service areas, except where there is a temporary absence for reasons such as illness or childbirth, or when a parent or child is participating under continuing eligibility. While the mandate for the three core instructional components has remained constant since the program began, federal guidelines have become more detailed and explicit, to provide local projects with clearer definitions and guidance.
The Nature of Core Instruction. In part because each project defines its own approach to providing instruction, there is considerable diversity in the content of the core services provided through Even Start. Projects borrow from and/or adapt existing curricular designs and materials as well as develop their own approaches. Even Start projects often use existing materials in their adult education and early childhood programs; they are more likely to develop their own materials for providing parenting education. This reflects the fact that adult education and early childhood education are more commonly offered by other collaborating agencies, while parenting education services, which are not widely available, generally are provided by Even Start staff.
Who Provides Instruction. The Even Start staff who provide instruction are well-educated and experienced; 76 percent have at least a Bachelor's degree, a stable percentage for the past four years, and over half have at least six years of post-high school education (Tao, Gamse & Tarr, p.66). Most projects recognize that continuing professional development for their own staff is critical, and as required by the statute, they routinely provide in-service training to their staff members. Among the most frequently offered topics for such training are parent and child activities, program planning and improvement, team building, child development, and parents' roles as teachers. Only about one-quarter of projects provide in-service training to most staff on adult education topics, while well over half offer training on child development and parenting education topics. Approximately half of local projects offer training to most staff on support service areas, including adapting services to participants' needs and working with other collaborating agencies (Tao, Gamse & Tarr, 1998, p.67).
Service Hours Offered. The Department's performance indicator plan for Even Start also includes the following indicator: Increasing percentages of projects will offer at least 60 hours of adult education per month, at least 20 hours of parenting education per month, and at least 65 hours of early childhood education per month. The amount of instruction offered by projects is important for several reasons. First, the federal statute requires projects to offer intensive high-quality instructional services, although there is no numerical definition of intensive. Second, if greater amounts of services are available, then participants have greater flexibility in choosing when to participate. Third, some participation and outcome data suggest outcomes are more positive for families who participate more intensively, or for a longer duration (St. Pierre et al., 1995, p.253). And fourth, because so many states are undergoing changes in welfare requirements, and state policy changes have repercussions for educational participation of welfare recipients (including a substantial proportion of Even Start participants), Even Start projects are recognizing that increased amount and flexibility of instructional offerings may represent the only way to maintain enrollment and participation.(5)
Recent years have seen increases in the amount of instruction offered in all three core areas. The amount of instruction offered in beginning and intermediate adult basic education, and adult secondary/GED preparation each increased by 25 to 30 percent over the past four years, from roughly 325 hours to 415 hours per year, while the amount of English as a Second Language instruction increased by 12 percent, from 300 to 335 hours (Tao, Gamse & Tarr, 1998, p.82). Over the same period, the amount of early childhood education offered has increased by about 10 percent, from roughly 480 hours to 525 hours per year (Tao, Gamse & Tarr, 1998, p.83). By contrast, the amount of parenting education offered appears to have stabilized at about 200 hours per year (Tao, Gamse & Tarr, 1998, p.82).
Over the past several years, projects have offered roughly the same ratio of center- to home-based hours in each core service area –home-based hours comprise about 1/3 of the total hours offered (Tao, Gamse & Tarr, 1998, p.87-88). The amount of home-based early childhood services appears to be increasing modestly, perhaps in response to the rising enrollment of infants and toddlers for whom there may be fewer existing center-based programs with an educational component.
Participation, Retention, and Continuity. Another Even Start performance indicator is that projects will increasingly improve retention and continuity of services. The extent to which families receive the services offered by Even Start projects has varied over the years. The average number of hours of participation in adult education services rose in the early 1990s from 68 hours to 107 hours per year (St. Pierre et al., 1995, p.140), and has hovered around 95 hours a year for the past several years (Tao, Gamse & Tarr, 1998, p.107). This is equivalent to about 2 hours per week for a year, and is somewhat more than the national average of 80 hours for participation in adult basic education programs (Development Associates Inc., 1994, p.60). Participation in parenting education services has diminished over time, from an average of 58 hours a year in 1992-93 to about 35 hours a year, less than 1 hour per week, for program years 1994-95 through 1996-97 (Tao, Gamse & Tarr, 1998, p.110). Children's participation in early childhood educational services is calculated in terms of months, rather than hours, and approximately half of all children participate for more than 6 months (Tao, Gamse & Tarr, 1998, p.113). Finally, within any given year there is a clear relationship between the amount of instructional service offered and the amount received – when participants have more hours to choose from, they take advantage of this available instruction. This holds for each of the core service areas (Tao, Gamse & Tarr, 1998, p.117).
About 60 percent of all families that participated in Even Start during 1994-95 and 1995-96 remained in the program for a year or less; conversely, about 40 percent participated for longer than a year (Tao, Gamse & Tarr, p.120). This represents a shortening of the average period of participation since data from 1989-90 showed that about 50 percent of all new families remained in the program for more than a year (St. Pierre et al., 1995, p.135). However, these statistics mask the fact that there is great project-to-project variation in when a family is considered to have officially enrolled in Even Start.
Changes in Federal Program Costs
Federal Even Start funds are used to provide core services and support services, to administer projects, to conduct evaluations, and for staff training. Federal Even Start spending grew dramatically from 1989 to 1995, was level from 1995 through 1997, and increased again in 1998 (Exhibit 1). Over time, rising funding has been accompanied by corresponding increases in the number of projects funded and in the total number of families served.
Expenditures Per Project. Average project-level federal expenditures were calculated by dividing the total federal funding for Even Start by the total number of projects funded (see the sixth column in Exhibit 1). Doing so shows that federal per-project expenditures grew during the first three years of Even Start, from $195,000 in 1989 to $208,243 in 1991. During this period the program was administered at the federal level and overall program expenditures were growing each year. After the states began administering the program, annual federal per-project expenditures declined steadily, from $205,882 in 1992 to $155,721 in 1997.(6) Per-project expenditures rose again in 1998, due to the substantial increase in program funding.
The size of local Even Start grants vary. In the 1995-96 program year, the majority of projects (55 percent) had federal grants between $75,000 and $175,000, while 36 percent had federal grants between $175,000 and $275,000 (St. Pierre & Noonan, 1998, p.6). The average size of Even Start grants also varies among states and regions (St. Pierre & Noonan, 1998, p.6-7). In 1995-96, most states made average grant awards between $100,000 and $200,000, but there are four states in which the average grant award was less than $100,000, and ten other states in which the average grant was greater than $200,000. States in the Northeast ($189,000 average) and Midwest ($172,000 average) tended to make larger grants than states in the South ($155,000 average) and West ($150,000 average).
Federal Expenditures Per Family. Exhibit 1 shows that the average federal expenditure for a family participating in Even Start declined quickly from a high of $6,024 in 1989 and has stabilized at roughly $3,000 in each of the past several years. In addition, the number of families served by the average project (about 54 families) has stabilized.
Some Even Start projects spend relatively few federal dollars per family while other projects spend much more per-family. In 1995-96, almost 40 percent of the projects spent $1,000 to $3,000 in federal funds per family each year, while an additional 35 percent were in the $3,000 to $5,000 range. About 3 percent of all projects spent less than $1,000 in federal funds per family, and another about 10 percent of all projects spent over $7,000 in federal funds per family (St. Pierre & Noonan, 1998, p.10). Such wide variation in federal expenditures per family reflects different approaches to organizing and implementing Even Start services. It also reflects project-to-project differences in the use of resources outside of Even Start. For example, some projects use locally-available adult education and early childhood education services, while others provide those services using their federal Even Start funds.
Comparison to Head Start. On a per-family basis, federal Even Starts costs have declined over the past decade, from about $6,000 per family in 1989 to about $3,000 per family from 1993 to the present. At the same time, Head Start's federal per-family costs have increased from about $2,900 per family in 1990 to $4,600 per family in 1996 (ACYF, 1997). The decline, and subsequent stabilization, in Even Start's federal cost per family can be attributed to a variety of factors – projects may have matured and become more efficient, projects may be doing a better job of using existing services, or perhaps projects are being pushed by states to "do more with less." The increase in federal Head Start costs per family is attributable to increased program funding during the early 1990s, which was used primarily to improve the quality of existing programs and services and secondarily to serve more children.
Comparison to Adult Education. In 1992, the federal government spent about $235 million on adult education programs (adult basic and secondary education, ESL programs). These funds were augmented by state expenditures of $810 million, so that total federal and state expenditures for adult education topped $1 billion (Moore & Stavrianos, 1995), making adult education about 10 times the size of Even Start. On a per-participant basis, federal basic adult education programs cost relatively little. Pugsley (1990) estimated annual per-student basic adult education expenditures to be $160, while Development Associates (1994) estimated the average per-student cost to be $258, with most states spending between $100 and $500 per participant. The cost of JOBS training programs are estimated to range from $100 to $1,000 per participant for less comprehensive programs (Gueron & Pauly, 1991) up to $1,400 to $3,900 per participant for more comprehensive versions (Maynard, 1993; Burghardt & Gordon, 1990). Finally, GED programs implemented by community colleges are estimated to cost an average of $250 per student, with ESL programs costing $154 per student (Hershey & Silverberg, 1993).
Percentage of Funding Spent on Different Services. Detailed cost data were collected from five Even Start projects in the early 1990s. These data show that more than half (55 percent) of all federal Even Start funds were spent in the provision of core services: 31 percent for early childhood education, 15 percent for adult education, and 9 percent for parenting education. An additional 9 percent was spent on the provision of support services which are designed to enable families to participate in core service activities. Thus, almost two-thirds (64 percent) of projects' federal funds were spent on the direct provision of services. Remaining federal funds were spent for program administration and coordination (14 percent), evaluation (10 percent), case management and recruiting (4 percent), and for a variety of other functions (8 percent) such as field trips, staff meetings, clean-up, and errands (St. Pierre et al., 1995, p.233).(7)
The finding that Even Start projects use about two-thirds of their federal funding for direct service provision raises the issue of whether Even Start projects are able to comply with the legislative requirement to build on existing services. This is a complicated issue for Even Start projects, and the extent to which a project uses existing services depends on whether those services are available locally, whether the services are easily accessible by program participants, whether the Even Start project considers the services to be of sufficiently high quality, and whether the Even Start project is able to make the necessary collaborative arrangements with other service providers.
Given this management challenge, Even Start projects may rely more heavily on direct service provision early in their life cycle, switching to a greater reliance on existing services once they have had the time to build collaborative arrangements. The cost data presented above were collected in the early 1990s, when Even Start was new. It is possible that the percentage of Even Start funds that are spent on direct services as opposed to coordination of existing services may look quite different a decade later. This will be investigated as part of the third national Even Start evaluation.
Footnotes:
(5) The effect of welfare reform on Even Start is addressed at greater lenghth later in this report.
(6) Even Start participants also benefit from matching funds and in-kind contributions that local projects are required to make, and from participating in programs to which Even Start has referred them. This discussion does not consider those sources of funding. The amount of the match is 10 percent in the first year, grows to 40 percent in the fourth year of a grant, and is 50 percent in any subsequent year. Because Even Start now has a well-distributed mix of projects in their first, second, third, fourth, and subsequent years of funding, the mandated growth in the amount of matching funds is unlikely to account for the decline over time in average per-project federal expenditures.
(7) Even Start's two-thirds/one-third split between direct service costs and other costs is close to what Head Start provides. In 1991, the overall distribution of federal costs in Head Start projects was 70 percent for direct service provision and 30 percent for other costs, with the largest categories being education (41 percent) and occupancy (13 percent).
-###-
[What Difference Has Even Start Made to Families?]