Program Outcomes

Social Responsibility

Delinquency. Data collected from police and court records show that juvenile delinquency was significantly lower for the High/Scope Perry Preschool program group as compared with the control group, including fewer arrests and fewer juvenile court petitions (Schweinhart, Barnes, and Weikart, 1993; Schweinhart and Weikart, 1995). Only 31 percent of the program group had ever been arrested, compared with 51 percent of the control group. In addition to police and court records, data collected from respondents at age 19 were used as an overall indicator of delinquency. When study participants were 19 years old, researchers found significant differences between the program and control groups. The program group had fewer arrests overall than the control group (averages of 1.3 versus 2.3 arrests per person), fewer felony arrests (averages of 0.7 versus 2.0 arrests per person), and fewer juvenile court petitions filed (averages of 0.2 versus 0.4 petitions per person).

Like the criminal record data, a misconduct scale based on teacher-report data and self-report data from the 19-year-old respondents demonstrates a significant difference between the program and control groups, as reflected by the following results for the program group:

• Lower overall scores for total misconduct and serious misconduct at ages 15 and 19.

• Lower incidence of fighting and other violent behavior.

• Lower incidence of property damage.

• Fewer police contacts.

Data collected from respondents at age 27 indicate significant differences between the program group and control group for adult arrests: the control group underwent more than twice as many arrests as the program group (averages of 4.0 versus 1.8 arrests per person). Thirty-six percent of the control group accounted for 98 felony arrests between ages 19 and 27, while 27 percent of the program group accounted for 40 felony arrests during the same period. Thirty-five percent of the control group were considered frequent offenders (defined as five or more arrests), compared with only 7 percent of the program group. In addition, 25 percent of the control group had been arrested for drug-related offenses, versus 7 percent of the program group (see figure 1). The control group also averaged more months on probation (6.6 versus 3.2 months) and had more than twice as many of its members placed on probation or parole for longer than 18 months (20 percent versus 9 percent).

Marital status and pregnancy. Marital status among the males was the same for both groups, with 26 percent married at age 27, although program group males, on average, had been married for a longer period (6.2 versus 3.3 years). Marital status among the females differed significantly, with 40 percent of program group females married, compared with 8 percent of the control group females. Although fewer females in the program group were parents (64 percent versus 75 percent), significantly more of them were married, cohabiting parents (28 percent versus 8 percent). Fifty-seven percent of mothers in the program group gave birth out of wedlock, compared with 83 percent of mothers in the control group. In measures related to family stability, the program group scored significantly higher on a measure of closeness to family and friends (66 percent versus 48 percent) and the ability to maintain persistence at tasks (i.e., work or study hard all day) (47 percent versus 33 percent).

Scholastic Success

Participants in the High/Scope Perry Preschool study were characterized by better academic performance than those in the control group, as measured by higher graduation rates, better grades, higher standardized test scores, and fewer instances of placement in special education classes. In addition, the program group spent more time on homework and demonstrated more positive attitudes toward school at ages 15 and 19. More parents of program group members had positive attitudes regarding their children's educational experiences and were hopeful that their children would obtain college degrees. The program group demonstrated significant academic differences in the following areas:

• Special education for mental impairment. Only 15 percent of the program group had been placed in special education programs for mental impairment, compared with 34 percent of the control group.

• Test scores. Each year from ages 7 to 14, the mean achievement test scores of the program group were noticeably higher than those of the control group (an average difference of 16 percent). The difference in the final achievement test scores of the two groups at age 14 was particularly significant: the program group's scores were 29 percent higher than those of the control group.

• Grade point average. The mean high school grade point average of the program group was higher than that of the control group (2.09 versus 1.68).

• Graduation from high school. Seventy-one percent of the program group graduated from high school, compared with 54 percent of the control group. The difference was largely accounted for by graduation rates among females (84 percent and 35 percent, respectively).

Socioeconomic Success

Data collected at ages 19 and 27 indicate that the program group has been more successful socioeconomically than the control group. The data for age 19 reveal that significantly more program group members were employed (50 percent versus 32 percent) and self-supporting (45 percent versus 25 percent). These data also reflect that fewer program group members received welfare assistance (18 percent versus 32 percent). The data for age 27 reveal a continuation of significant economic differences characterized by more economic stability among the program group members, as measured by the following indicators:

• Public assistance. Fifteen percent of the program group were receiving public assistance, versus 32 percent of the control group.

• Monthly earnings. Twenty-nine percent of the program group had monthly earnings of $2,000 or more, versus 7 percent of the control group (36 percent versus 11 percent, respectively, when comparing only employed members in each group).

• Household earnings. When the income of the spouses of the study participants was taken into account, 47 percent of the program group had household income earnings of $3,000 or more per month, versus 17 percent of the control group.

• Home ownership. Thirty-six percent of the program group owned a home, versus 13 percent of the control group.

• Automobile ownership. Thirty percent of the program group owned a second car, versus 13 percent of the control group.

Cost-Benefit Analysis

A cost-benefit analysis of the High/Scope Perry Preschool study indicates a savings to the public of more than seven times the initial investment per child, with a return of $7.16 for every dollar spent (Barnett, 1993; see figure 2). When adjusted for inflation and a 3-percent discount rate, the investment in early childhood prevention resulted in a taxpayer return of $88,433 per child from the following sources:

• Savings in welfare assistance (prior to welfare reform).

• Savings in special education.

• Savings to the criminal justice system.

• Savings to crime victims.

• Increased tax revenue from higher earnings.

An independent reanalysis is provided in a recent RAND Corporation report (Karoly et al., 1998; see figure 3). This report found that eliminating the largest and least reliable savings category (savings to crime victims) still left a return of more than twice the initial investment. Savings to crime victims make up 65 percent of the total investment return in the earlier analysis (Barnett, 1993). Although victim savings should be considered a significant outcome and societal benefit of early childhood intervention, this factor is also distinct from the other factors that can be estimated based on direct governmental costs and savings. With victim savings factored out of the analysis, the largest savings category is in criminal justice costs (40 percent), followed by increased taxable revenue (26 percent), reduced educational services (25 percent), and reduced welfare costs (9 percent).