long-term effects of early childhood programs on social ...€¦ · preventing chronic delinquency:...

51

Long-Term Effects of EarlyChildhood Programson Social Outcomes andDelinquencyHirokazu Yoshikawa

Abstract

The search for ways to prevent juvenile crime in the United States has become a mat-ter of national urgency, as the incidence of serious offenses continues to rise. Most pre-vention initiatives focus on late childhood or adolescence. Such initiatives may be miss-ing an important additional opportunity to intervene earlier in children’s lives. Thisreview of literature from criminology, psychology, and education shows that there existkey early childhood factors which are associated with later antisocial or delinquentbehavior and that early childhood programs which seek to ameliorate the effects ofthose factors can prevent later antisocial or delinquent behavior. In particular, thereview focuses on programs which have demonstrated long-term effects on antisocialbehavior or delinquency. These programs have in common a combination of intensivefamily support and early education services, and effects on a broad range of child andfamily risk factors for delinquency. Moreover, there is promising evidence of their cost-effectiveness. As one element in a comprehensive plan to address poverty and otherenvironmental causes of crime, programs combining family support with early educa-tion show promise in lessening the current devastating impact of delinquency onAmerica’s children and families.

The call for effective ways to fight juvenile crime echoes across theUnited States as the incidence of serious offenses continues to rise.The juvenile arrest rate for murder and nonnegligent manslaughter

rose 122.7% between 1982 and 1992.1 Arrests of juveniles between 1984 and1993 rose 39.6% for robbery, 98.1% for aggravated assault, and 105.7% formotor vehicle theft.2 Pressures to imprison are great, and efforts to preventare rare. When crime prevention initiatives are put forward, most target latechildhood or adolescence rather than early childhood.3

A review of the literature from criminology, psychology, and educationsuggests that focusing crime prevention efforts on older children or teensmay cause policymakers to miss an important opportunity to intervene

Hirokazu Yoshikawa,M.A., is a researcher inthe Adolescent PathwaysProject in the Departmentof Psychology at NewYork University.

The Future of Children LONG-TERM OUTCOMES OF EARLY CHILDHOOD PROGRAMS Vol. 5 • No. 3 – Winter 1995

earlier in children’s lives. The literature reviewed in this article indicatesthat there exist key early childhood factors which are associated with latercriminal or delinquent behavior and that early childhood programs whichseek to ameliorate the effects of those factors can prevent later criminal ordelinquent behavior.

This article begins with a description and definition of chronic delin-quency and then summarizes the early risk factors associated with delin-quency. The implications of risk factor research for the design of crimeprevention programs are discussed, and the effects of early childhood pro-grams on delinquency and associated risk factors are reviewed to see ifprograms that are designed as suggested by the research do indeed pro-duce anticipated benefits. The article concludes with a discussion of theimplications of these results for policy.

Juvenile Delinquency andConduct DisorderJuvenile delinquency is a legal term whosedefinition varies from state to state.4Generally, however, the term is used todescribe minors whose behaviors have beenadjudicated as illegal by a juvenile court.Delinquency usually refers to behavior thatwould be criminal if the child were an adult.The legal system terms behavior that is ille-gal only if committed by a minor, such asrunning away, a status offense or unrulybehavior.5

In the educational and mental healthfields, some or all of those behaviors mightbe called “antisocial behaviors,” and chil-dren or youths who demonstrate repeatedepisodes of such behaviors might be diag-nosed as suffering from a “conduct disor-der.” According to the Diagnostic and Statis-tical Manual of Mental Disorders, the standardmanual used by psychologists and psychia-trists, a diagnosis of conduct disorderrequires the commission of at least three dif-ferent antisocial acts over a six-month peri-od. Qualifying antisocial behaviors includeinitiating fights, bullying or physical crueltyto people or animals, the use of weapons,stealing, rape, fire setting, chronic truancy,running away or lying, breaking into some-one else’s home or car, and destruction ofproperty.6

Chronic DelinquencyNo matter which terms are used, researchon delinquency shows three key findings:(1) a small group of chronic offenders isresponsible for committing the majority ofserious juvenile offenses; (2) there are two

groups of youthful offenders, distinguishedby when their antisocial behavior begins;and (3) youths whose delinquent careersbegin early tend not to specialize in any par-ticular type of antisocial act.

Studies indicate that a few chronicoffenders commit the vast majority ofoffenses. In a study of 411 working-class boysin London, for example, those childrenrated by teachers and peers as “most trou-blesome” at ages 8 to 10 represented 22% ofthe whole sample, but 70% of future chron-ic offenders.7 Closer to home, an examina-tion of 13,150 men born in 1958 inPhiladelphia demonstrated that, while thosewith five or more contacts with the justice sys-tem comprised only 7.5% of the group, theywere responsible for 61% of all recordedoffenses (including 61% of homicides, 75%of rapes, and 65% of aggravated assaults).8

History of Antisocial Behavior Many longitudinal studies show that severeantisocial behaviors in childhood, such asfrequent fighting, hitting, stealing, destroy-ing or vandalizing property, or lying, are thestrongest predictors of chronic delinquen-cy.7 Both criminological and psychologicalresearch converge on a distinction betweentwo groups of youths: one whose antisocialbehavior or delinquent “career” is limited toadolescence and one whose antisocialbehavior or delinquent career starts early—often in early childhood—and persists intoadulthood.9

Diversity of Delinquent Behaviors Youths whose antisocial behavior persistsinto adulthood are more likely to engage ina range of antisocial behavior rather than to

52 THE FUTURE OF CHILDREN – WINTER 199552

53Long-Term Effects of Early Childhood Programs on Social Outcomes and Delinquency

specialize in any particular type of antisocialact. A study of 195 boys 10- to 17-years-old,for example, indicated that the boys whocommitted different types of crimes were atmuch higher risk for chronic delinquencythan were the boys who specialized in a par-ticular sort of antisocial behavior: half hadthree or more contacts with the police ascompared with fewer than 10% of the boyswho specialized.10

Preventing ChronicDelinquency: The Search forChildhood Risk FactorsTogether, these findings on the characteris-tics of chronic delinquency suggest that oneimportant way to decrease overall crimerates among youths is to prevent chronicdelinquency, and that early childhood maybe an important developmental period totarget for its prevention. The remainder ofthis article explores how and whether chron-ic delinquency can be prevented. Thisrequires answering three interrelated ques-tions: (1) Are there risk factors in early child-hood which increase the probability of laterchronic delinquency? (2) Do these factorscause chronic delinquency or are they onlyassociated with it? (3) Can early childhoodprograms that lessen the impact of these fac-tors prevent chronic delinquency?

Researchers have long sought factorsthat are regularly associated with chronicdelinquency. The strongest factor, as men-tioned above, is a history of antisocial behav-ior in childhood, but many other early riskfactors have also been linked to chronicdelinquency. These factors, listed in Table 1,include perinatal difficulties, neurologicaland biological factors, low verbal ability,neighborhoods characterized by social dis-organization and violence, parental crimi-nality and substance abuse, inconsistentand/or harsh parenting practices, lowsocioeconomic status, and exposure tomedia violence.11

The most important of these factorsappear to be low socioeconomic status, hav-ing parents who have been convicted ofcrimes, the child’s low cognitive ability(espe-cially poor verbal ability), poor parentalchild rearing, and the child’s own history ofantisocial behavior, conduct disorder, ortroublesomeness.12 In one study of boys in

London, for example, the 8- to 10-year-oldswith four or more of these predictors includ-ed 15 of 23 future chronic offenders (the 23were to be responsible for fully half of theconvictions in the cohort of 411 youths).12

The following sections explore evidenceconcerning two of the risks that have beenconsistently associated with later delin-quency and that have most frequently beeninvestigated in outcomes of early childhoodprograms.

Parenting and Social Support Longitudinal evidence from many studiessuggests that hostile or rejecting parentingand lack of parental supervision is associatedwith children’s later antisocial behavior anddelinquency. In more than two decades ofresearch, Gerald Patterson and his col-leagues at the Oregon Social LearningCenter have proposed and developed sup-portive evidence for a model of how parent-

ing behavior can lead to antisocial behaviorin children. They suggest that parents ofantisocial children first reinforce common-place, low level aversive behaviors such asnoncompliance, teasing, or tantrums. Then,as the child learns to respond to aversive actsthrough aversive counterattacks, increasing-ly severe coercive interchanges occur.13

Interventions involving parent training toreduce such coercive interactions havedecreased antisocial behaviors up to 4.5years after treatment.14

If harsh or poor parenting can lead toantisocial behavior, one would expect thatnurturant parenting might protect againstthe development of such behavior. There isevidence that a good relationship with oneparent, marked by warmth and the absenceof severe criticism, can have a substantialprotective effect against the development oflater antisocial behavior.15

One might also expect that factors whichpromote good parenting might indirectlyhelp prevent antisocial behavior. There is

One important way to decrease overall crimerates among youths is to prevent chronicdelinquency.


some evidence that providing social support(emotional, material, or informational assis-tance) for parents can, in fact, operate inthat fashion. Social support, from partnersand from community members, helpedmothers of newborns in one study respondmore positively and attentively to their chil-dren.16 Conversely, low social supportappears to be associated with subsequentbehavior problems: a longitudinal study of83 poor inner-city African-American andPuerto Rican teen mothers found that lowsocial support from friends when childrenwere one year of age predicted behavior

problems when children were three yearsof age.17

Verbal/Cognitive Ability Low scores on measures of children’s cogni-tive ability such as school achievement, gen-eral intelligence quotient (IQ), and verbalability are associated with delinquency.7,18

While there is some disagreement, most ofthe evidence suggests that cognitive deficitslead to antisocial behavior and not viceversa. For example, a longitudinal study of837 children on the Hawaiian island ofKauai indicated that age-appropriate lan-

Relationship Between Relationship BetweenRisk Factor and Risk Factor and

Risk Factors Delinquency Is Delinquency IsStrengthened by Weakened by

Presence of Presence of

Perinatal difficulties Low socioeconomic —status

Family adversity

Neurological and — —biological factors

Child’s low cognitive ability — —

History of antisocial behavior — —

Single parenthooda

— —

Insecure attachment of Poor parenting —child to parent Unplanned birth

Life stressLow social support

Parental criminality Early family conflict —

Parental substance abuse — —

Poor or harsh parenting Marital discord Emotional supportCommunity support

Low socioeconomic status — Good parentingAge-appropriate verbal

ability

Violent or socially disorganized — —neighborhoods

Media violence — —

aEvidence indicates that single parenthood is associated with antisocial behavior, but the relationship isprobably explained by low socioeconomic status or poor supervision, rather than by single parenthooditself.

Table 1

Early Childhood Risk Factors for Delinquentand Antisocial Behavior


guage development at 2 and 10 years pro-tected high-risk children against later delin-quency.19 Another longitudinal study of1,037 children from New Zealand indicatedthat IQ deficits tended to precede the devel-opment of serious antisocial behavior andthat the effects of low IQ on behavior wereindependent of the effects of factors such aslow socioeconomic status, ethnicity, academ-ic attainment, and motivation.20

Are the Risk Factors Causal?Just because a factor is associated with laterchronic delinquency does not mean, ofcourse, that it caused the delinquency. Mosthuman behavior develops through the com-plex interplay of multiple factors across mul-tiple settings (such as home, school, andneighborhood), and delinquent behavior isno exception. Identifying its cause thereforerequires sophisticated analyses designed todisentangle the effects of multiple risk fac-tors. This task is even more difficult thanmight be supposed because there is consid-erable research evidence that the risk factorsoperate differently when multiple risk factorsare present. For example, children exposedto multiple risk factors are much more proneto later delinquency than are those exposedto just one or even two of these factors.21

Evidence also indicates that the potencyof a single risk factor can be increased by thepresence of a second risk factor. For exam-ple, children whose parents are criminalsare more likely to become delinquent them-selves, but that association is strengthenedstill further if children are exposed to earlyfamily conflict.

Finally, a risk factor may exert an indirectrather than a direct influence on develop-ment of delinquent behavior. For example,children who grow up in single-parenthouseholds tend to have higher rates of laterdelinquency, but this appears to be due todifficulty in providing adequate supervision,not single parenthood per se.13,22

Research studies have identified severalexamples of these sorts of complex interre-lationships among early causal factors forchronic delinquency, and some key exam-ples are depicted in Table 1.11

If a given factor is causally linked to delin-quency, then one would expect that buffer-

ing a child against the effects of that factorwould help prevent later delinquent behav-ior. Research indicates that this is so for atleast some risk factors. For example, asmentioned earlier, studies indicate thatproviding emotional and community socialsupport to the parent is associated with con-sistent, nurturing child rearing, which inturn is associated with lower levels of antiso-cial behavior among low-income children.In this instance, social support appears tobuffer children and families from theeffects of low socioeconomic status.

Implications for Preventive ProgramsLongitudinal evidence on the developmentof delinquency behavior suggests severalpromising directions for prevention. First,the evidence suggests that early childhoodprograms which buffer the effects of a givendelinquency risk factor should also be effec-tive in preventing chronic delinquency.

Second, because multiple risk factorsappear to have such a pronounced negativeeffect, early childhood programs that reducemultiple risks may be more successful in pre-venting chronic delinquency than are thosethat target only a single risk factor.

Third, the research implies that the con-tent of preventive early childhood programsshould be such that they attempt to enhanceparents’ social support, foster positive par-enting and family interactions, facilitatechild cognitive development (especially ver-

bal skills), and reduce family level and com-munity level poverty. In other words, crimeprevention programs should seek to reduceor eliminate the risk factors associated withdelinquency.

The next section of this article reviewsearly education and family support pro-grams which have attempted to improve thelives of children and families, to determine ifthe programs either decreased delinquencyor antisocial behavior, or lessened the

Early childhood programs which buffer theeffects of a given delinquency risk factorshould also be effective in preventing chronicdelinquency.


impact of the factors that are hypothesizedto lead to such behavior.

Early Education and FamilySupport ProgramsEarly education and family support pro-grams provide a range of emotional, infor-mational, instrumental, and/or educationalsupport to families with infants andpreschool-age children. Early education pro-grams are usually center based, and theircore service is usually to provide an educa-tional curriculum to groups of preschoolersor infants and toddlers, but they can alsoprovide services as varied as basic preventivehealth care, informational support regard-ing parenting and child development, andemotional support.

In contrast, most family support pro-grams focus primarily on the parents, notthe children, and emphasize providingsupport of various kinds to parents, oftenthrough home visits. Family support pro-grams can help parents in their roles asparents or educators of their children, orsupport the parents’ own educational or

occupational goals. These types of pro-grams are increasing in popularity, and the1993 Family Preservation and Support Actprovides federal funding for them (seeBox 1).

The two models of early intervention arenot mutually exclusive. Some family supportprograms have an educational child care orpreschool component, and some predomi-nantly child-focused educational programsalso offer supportive services for parents orservices to enhance parenting skills. Basedon the literature regarding risk factors fordelinquency, it is those combination pro-grams that address multiple risk factors andthat blend aspects of both family supportand early childhood education which aremost promising in the prevention of chron-ic delinquency.

Scope of the ReviewA computer and manual search of the litera-ture from the fields of psychology and educa-tion identified 40 evaluations of interventionsthat (1) served populations which displayedthe risk factors associated with later delin-quent or antisocial behavior (for example,low household income, single parent, lowparental educational level, low birth weightand/or preterm birth); (2) provided servicesbetween the prenatal period and entry intoprimary school; (3) assessed possible effectson risk factors for chronic juvenile delin-quency and/or possible effects on antisocialbehavior or delinquency; (4) were carriedout in the United States or Canada; and (5)had adequate research design.23–61 When asingle program was evaluated in both ran-domized trials and less well-controlleddesigns, only the results for the randomizedtrial are reported. When multiple evaluationsexist for a single project, only the most recentis cited. (Barnett reviews many of these sameprograms in his article in this journal issue,and the reader may wish to consult Table 1 inhis article.)62

These 40 programs and their effects arelisted and described briefly in Tables 2–4.Each table describes programs that repre-sent one of three general models of services:Table 2 reviews eight child-focused earlyeducation programs,23-30 Table 3 describes23 parent-focused family support pro-grams,31–53 and Table 4 covers 11 programsthat provided both kinds of services.35, 53–61

(Two programs appear in both Table 3 andTable 4).63

Most of the 40 evaluations explored theeffects of the programs on factors many ofwhich have been discussed earlier as risk fac-tors for chronic delinquency. For this review,these risk factors were grouped into threebroad categories: early cognitive ability(including early IQ, school achievement,and language development or verbal abili-ty), early parenting factors (including assess-ments of mother-child interaction, parent-ing behavior, attachment, and child welfareindicators), and life-course variables thatcould be expected to influence familysocioeconomic status (maternal educationand employment, childbearing, and familyeconomic self-sufficiency). Only four evalua-tions of programs actually reported or in-vestigated long-term effects on antisocial

Programs that address multiple risk factorsand that blend aspects of both familysupport and early childhood education are the most promising.


Family Preservation and Support Services Program

PurposeThe purpose of the Family Preservation and Support Services Program is to improve well-being for vulnerablechildren and their families by providing in-home and community-based services.

Brief DescriptionThe Family Preservation and Support Services Program was enacted as part of the Omnibus BudgetReconciliation Act of 1993 (Public Law 103-66). It provides federal funds to state child welfare agencies for pre-ventive services (family support) and services to families at risk or in crisis (family preservation). The programrequires states to develop a comprehensive plan that goes beyond child welfare to include housing, mentalhealth, primary health, education, juvenile justice, and community-based programs for children and families.

Definition of Family Support Services Family support services are defined as “community-based services to promote the well-being of children and fam-ilies designed to increase the strength and stability of families (including adoptive, foster, and extended families),to increase parents’ confidence and competence in their parenting abilities, to afford children a stable and sup-portive family environment, and otherwise to enhance child development.”

Examples of family support services are home visits and parent support groups, respite care, structured activitiesto strengthen parent-child relationships, drop-in family centers, information and referral services, and earlydevelopmental screening of children to assess their need for special services.

Definition of Family Preservation Services Family preservation services are defined by the law as “services for children and families designed to help fami-lies (including adoptive and extended families) at risk or in crisis.”

Family preservation services include programs to help reunify children with their families, aftercare services forchildren returned home, respite care, and services to improve parenting skills.

FundingA capped entitlement program, the Family Preservation and Support Service Program provides $930 million overfive years, from 1994 through 1998 as follows:

FY 1994—$ 60 million

FY 1995—$150 million




Use of FundsEach state decides on the balance of funds used for family support and family preservation services. A “signifi-cant portion” of service funds must be spent for each service. If either allocation is below 25%, the state’s ratio-nale must be especially strong.

AdministrationAt the federal level, the Family Preservation and Support Services Program is administered by the Administrationfor Children, Youth, and Families (ACYF) in the Department of Health and Human Services (HHS). In eachstate, the public child welfare agency responsible for the Child Welfare Services Program (Title IV-B of the SocialSecurity Act) administers the program.

Evaluation of Family Support ProgramsAbt Associates, together with Yale University, is conducting a five-year, three-phase national evaluation of familysupport programs. The evaluation includes a comprehensive review of what is currently known about family sup-port programs and their effects; the design and implementation of a series of new evaluation studies to deter-mine the effectiveness of family support programs for children and families with different characteristics and indifferent communities; and the integration of results into a comprehensive report.

For further information about the National Evaluation of Family Support Programs, contact:

Jean LayzerAbt Associates, Inc.55 Wheeler Street

Cambridge, MA 02138-1168(617) 492-7100

Box 1


Outcomes

Study, Age of Child Intensity Cognitive Parenting Maternal Life AntisocialRelated Endnote Ability Course Behavior

Number,bProgram Name

Abelson, 197423 4 to 5 years Half-day preschool E > C at New Haven Head kindergartenStart E > C at first

grade (boys only)

Beller, 198324 4 to 5 years Half-day preschool E > C at age 10 Aggression: North Philadelphia and kindergarten; (grade 4) E > C at

full-day first grade kindergarten and first grade

Campbell and 6 weeks to Full-day educational E > C at age 12 Mother-child At 54 months, Aggression: Ramey, 199425 3 months at child care interaction: teen mothers E > C atCarolina enrollment to (See also Table 1 E > C at 36 months, only: school entryAbecedarian 5 years at exit in the article by but E = C at 6, 20, E > C for high only; effect

Barnett in this and 60 months school graduation, faded laterjournal issue) postsecondary

Home environment: training, per-E = C at 6, 18, 30, centage self-and 42 months supporting

Deutsch et al., 4 to 9 years Home visits; part- E > C at grade 4198326 day preschool;Institute for school (grades 1-3)DevelopmentalStudies

Hebbeler, 198527 4 to 5 years Half-day preschool E = C, most Maryland Head measures, but Start E > C, grade 11,

but only for oneof three cohorts

Lee et al., 199028 4 to 5 years or Half-day preschool E > C in grade 1ETS Head Start 5 to 6 years

Miller and Bizzell, 4 to 5 years Preschool (6.5 E > C, grade 1, E = C on 198329 hours/day) for one of two aggression,Louisville Head measures kindergartenStart through

E = C, grade 2 grade 2

Reynolds, 199330 3 to 4 years at Half-day preschool; E > C for schoolChicago Child enrollment to full-day kindergarten achievement inParent Center grade 3 grade 6

Table 2

Early Education Programs—Child-Focuseda

Notes a To obtain a more detailed version of the table summarizing each of the 40 programs listed in Tables 2–4, contact the author, Hirokazu Yoshikawa, M.A., Department of

Psychology, New York University, 6 Washington Place, 4th Floor, New York, NY 10003.b See the related endnotes at the end of this article for complete citation of reports and/or studies in which the outcomes of these programs are described.


behavior and/or delinquency. Tables 2-4briefly describe each of the 40 programs andits effects on the broad categories of outcomes.

In general, the review of these 40 pro-grams leads to two main conclusions, both ofwhich are consistent with the research find-ings about risk factors: (1) the programs thatdemonstrated long-term effects on crimeand antisocial behavior tended to be thosethat combined early childhood educationand family support services, in other words,the programs that addressed multiple riskfactors; and (2) among the more specializedprograms, those designed primarily to serveadults tend to benefit adults more than chil-dren, and those designed primarily to servechildren tend to benefit children more thanadults. Barnett and St. Pierre and colleaguesdraw similar conclusions in their articles inthis journal issue.

The next sections describe the four com-bination early education/family supportprograms and their long-term effects onantisocial behavior and delinquency.

Long-Term Effects on AntisocialBehavior and/or DelinquencyFour evaluations, all focusing on programsthat combined early childhood educationwith family support services, assessed long-term (more than five years postprogram)effects on parent or teacher ratings of anti-social behavior and/or actual delinquencyrecords. These programs offered both homevisits and center-based educational childcare or preschool. All four demonstratedpositive effects.58–61

High/Scope Perry Preschool Project In the Perry Preschool Project,60 conductedfrom 1962 through 1967, some 123 three-and four-year-old African-American childrenin Ypsilanti, Michigan, were randomlyassigned to a program or to a control group.The intervention consisted of two and one-half hours of preschool experience five daysa week for seven and one-half months eachyear for two years (except for one smallgroup of children who received only oneyear of services). In addition, teachers visitedeach mother and child at home for 90 min-utes once per week during the school year.

The project decreased rates of self-reported delinquency at age 14, official

chronic delinquency at age 19, and, in themost recent follow-up at age 27, adult crimi-nality.60,64,65 Generally, results indicated thatthe program participants committed fewerdelinquent or criminal acts, the acts theycommitted were less severe, and they wereless likely to be chronic offenders than werecontrol group members: “As compared withthe no-program group, the program groupaveraged a significantly . . . lower number oflifetime (juvenile and adult) criminal arrests(2.3 vs. 4.6 arrests) and a significantly lowernumber of adult criminal arrests (1.8 vs. 4.0arrests). According to police and courtrecords collected when study participantswere 27-32 years old, significantly fewer pro-gram-group members than no-program-group members were frequent offenders—arrested 5 or more times in their lifetimes(7% vs. 35%) or as adults (7% vs. 31%). Ascompared with the no-program group, theprogram group had noticeably fewer arrestsfor adult felonies, significantly fewer arrestsfor adult misdemeanors, and noticeablyfewer juvenile arrests. As compared with theno-program group, the program group hadsignificantly fewer arrests for drug-making ordrug-dealing crimes (7% vs. 25%). . . .”66

Syracuse University FamilyDevelopment Research Program The Syracuse University Family Develop-ment Research Program59 provided educa-tional, nutrition, health and safety, andhuman service resources to 108 low-income,primarily African-American families, begin-ning prenatally and continuing until chil-dren reached elementary school age.Families received weekly home visits andquality child care (one-half day five days aweek for children 6 to 15 months of age, andfull-day care five days a week for children 15to 60 months of age).

Results for the Syracuse program weresimilar to those obtained by the Perry proj-ect: the program decreased the total num-ber, severity, and chronicity of later involve-ment with the juvenile justice system amongparticipants. At follow-up, when childrenwere 13 to 16 years old, four program groupchildren (of 65 who were identified at fol-low-up; the original program group includ-ed 108) had probation records. Three werestatus offenders who had been deemedungovernable, and the fourth was a one-time juvenile delinquent. In contrast, 12


Outcomes

Study and Age of Child Intensity Cognitive Ability Parenting Maternal Life AntisocialRelated Endnote Course Behavior

Numberb

Achenbach et al., From 0 to 3 Seven in-hospital E > C at age199031 months sessions; four 7 years

home visits

Badger, 198132 From 0 to 12 44 classes E > C at posttest Maternalmonths employment:

E > C at 3 yearsPregnancies:E < C at 3 years

Barrera et al., From 0 to 12 Home visits weekly E = C at 4 to 16 E > C at 16198633 months for 1 to 4 months, months months

biweekly for 5to 8 months,monthly for 9 to 12months; averageof 23 visits

Barth et al., 198834 From pregnancy Two home visits E < C on proxiesto 6 months, per month for for abuse (ERapproximately six months visits,removal of

child from home/child cared forby neighbor)E > C on well-baby care

Field et al., 198235 From 0 to 6 Biweekly home E = C at 1 year, Mother-child Maternalmonths visits but E > C at interaction: employment,

2 years E > C at 4 months schooling: E > Cin first year

Home environmentE = C at 6 months Repeat

pregnancies:E < C at 2 years

Gray et al., 197936 From 0 to 2 Weekly home visits; Hospitalization foryears biweekly visits to serious injuries:

pediatrician; E < Cbiweekly calls topediatrician

Gray and Ruttle, Beginning at 17 Weekly home visits E = C at end of E > C at 10 and198037 to 24 months, program and 20 months after

and continuing 10 months later. end of programfor 9 months E > C 20 months

after end ofprogram

Gutelius et al., Seventh month 18 home visits in E > C at 3 years, E > C at posttest Mothers in school: E = C on197738 of pregnancy year 1; 12 in but decreasing but only for some E > C at 3 and 4 behavior

to 3 years year 2; 8 in year 3 later measures years problems at4 years

Fathers’ jobstability: E > Cfor first 3 years

Table 3

Family Support Programs—Parent-Focuseda




Outcomes


Numberb

Hardy and Streett, From 0 to 2 years 10 home visits Hospital admis-198939 sions, diaper rash

(proxies for abuse):E < C at posttest

High/Scope From 3 to 5 years One to two home E > C after 7 Parent-child inter- E = C in use of E = C onEducational visits per week months of action, parental AFDC, food hostility after 7Research services involvement: stamps, Medicaid, monthsFoundation, E > C after job training197440 7 months programs

Jacobson and From pregnancy 30 home visits Attachment: E > CFrye, 199041 to 1 year at 14 months

Jester and From 3 months to From 3 months to E > C through E = C onGuinagh, 198342 3 years 2 years: weekly grade 5 classroom

home visits behaviorVarying lengths problems inof participation; Age 2 to 3: twice grades 2, 3,results reported weekly part-day and 4only for children preschoolwith 2 to 3 yearsof participation

Lambie et al., From 3, 7, or 11 Weekly home Mixed effects at E > C at posttest197443 months to 16 visits posttest

months laterE = C one yearafter end ofprogram

Larson, 198044 E1: Seventh month Seven visits from 6 E1 > E2 = C at 18of pregnancy to 15 weeks to 6 months; monthsmonths postpartum; three visits from 6E2: 6 weeks to 15 to 15 monthsmonths postpartum

Lieberman et al., From 1 to 2 years Weekly home visits Empathy, mother- E less avoidant199145 child interaction: and angry than

E > C at posttest C at posttest

Lyons-Ruth et al., Beginning at 0 to Approximately E = C at posttest Mixed effects at199046 9 months, ending weekly home posttest

at 18 months visits

Madden et al., From 2 to 3 years Home visits twice E > C, but only for E > C at posttest198447 to 4 to 5 years each week one of three and two-year

cohorts on one of follow-uptwo measures

Olds et al., 198848 From pregnancy One home visit per E = C at 12 E > C at 10, 22, At 48 months—to 24 months week for the first 6 months and 46 months for Repeat pregnan-

weeks postpartum, poor, unmarried cies: E < C for poor,gradually slowing teens unmarried mothersto one visit every6 weeks; average Participation inof 31 visits work force: E > C

Table 3 (continued)




control group youths (of 54 found at follow-up; the original control group included 74)had probation records. Five of the 12 con-trol group youths were chronic offenders.Among the offenses committed by the 12were robbery, burglary, sexual assault, andphysical assault.

Yale Child Welfare Project Between 1968 and 1970, 17 pregnant, low-income, primarily African-Americanwomen were recruited to participate in theYale Child Welfare Project,61 an intensiveprogram that began during pregnancy andcontinued until the children reached 30months of age. Each family received freepediatric care, social work, child care (anaverage of 13.2 months), and psychologicalservices as needed. Each family interacted

with a four-person team: a pediatrician, ahome visitor, a primary child care worker,and a developmental examiner. The teammembers remained constant over thecourse of the family’s enrollment in theproject.

The Yale project decreased boys’ antiso-cial behavior as rated by teachers andincreased the number of children with goodschool adjustment for both boys and girls 10years after program services ended.67

Teachers rated boys who had been in theprogram group as being socially well adjust-ed. Most of the comparison group boys weredescribed as disobedient or not gettingalong well with other children, and slightlymore than half were also described as havingproblems with lying or cheating.


Outcomes


Numberb

Osofsky et al., From 0 to 18 21 home visits E = C at posttest198849 months

Ross, 198450 From 0 to 12 15 home visits E > C at posttest E > C at posttestmonths

Seitz, Rosenbaum, Variable, begin- Daily classes for Mothers whoand Apfel, 199151 ning during teen mothers attended longer

pregnancy (ages 14 to 19 had fewer subse-years); duration quent pregnanciesranged from oneto four academic For low-achievingquarters mothers; those who

attended longerwere higher achiev-ing at two yearspostpartum

Siegel et al., 198052 From 0 to 3 Three home visits E = C at one yearmonths per month for

three months afterbirth, with orwithout extendedcontact betweenmother andinfant in hospital

Wasik et al., 199053 From 0 to 5 years 107 home visits E = C through 54 E = C through 54Project CARE months months(home visitgroup only)

Table 3 (continued)




Houston Parent Child DevelopmentCenterThe Houston Parent Child DevelopmentCenter (PCDC)58 was designed to promotesocial and intellectual competence in chil-dren from low-income Mexican-Americanfamilies. It required approximately 550hours of participation over a two-year peri-od. Mothers received 25 home visits for oneyear, beginning when their children wereone year of age. Weekend sessions involvingthe whole family focused on issues such asdecision making in the home or family com-munity. During the second year of the pro-gram, mothers attended classes to learnabout child development, home manage-ment, and other family-related topics. Theirchildren attended educational preschoolfour half days per week.

Results indicated that the HoustonPCDC decreased children’s antisocial be-havior as rated by parents in a one- to four-year follow-up and as rated by teachers in afive- to eight-year follow-up. In the five- toeight-year follow-up, for example, teachersrated control group children as more obsti-nate, impulsive, disruptive, and involved infights than program group children.Program group children were rated as moreconsiderate and less hostile. A more recentfollow-up did not find significant effects onantisocial behavior,68 but attrition rates werequite high.

Magnitude of EffectsIn the research literature, a shorthandmethod of assessing the magnitude of theeffects of human service programs involvescalculating what is called an effect size. Thistranslates results of different studies into acommon metric (the standard deviation),which then permits comparisons amongstudies of the strength of the relationshipbetween an intervention and an outcome.In the studies reviewed in this article, theeffect size measures the strength of the rela-tionship between participation in a programand antisocial behavior or delinquency.

Generally, in the social sciences, an effectsize of 0.2 standard deviation is defined assmall, 0.5 as moderate, and 0.8 or greater aslarge.69 Measured by these yardsticks, thefour programs described had moderate tolarge effects on antisocial behavior anddelinquency: 0.48 standard deviations for

the Houston PCDC, 0.48 for the Syracuseprogram, 0.42 for the Perry PreschoolProject, and 1.13 for the Yale program.

Suggestions about CausationPosttest and short-term follow-up evalua-tions of the four programs provide someclues as to what led to these differences inlater antisocial or delinquent behavior.Positive effects on cognitive and/or verbalability58,59,61,64 and parenting70,71 precededlong-term effects on delinquency and anti-social behavior. This observation is consis-tent with the view that long-term effects ondelinquency occurred through prior effectson early risk factors such as cognitive abilityand parenting ability.

In addition, it is important to note thatthree of these four programs (Yale,Houston, and Perry) assessed effects in twoseparate domains of risk and found somepositive effects in both domains (the cogni-tive effects were mixed for the Yale and

Houston programs). These findings bolsterthe notion that risk factors for delinquencycan have a cumulative effect such that chil-dren who are buffered from multiple risksare less likely to engage in later delinquencythan children buffered from just one risk.

These four studies are relatively atypicalin the literature. As mentioned above, mostof the 40 studies included in this literaturereview did not investigate program effectson long-term delinquent or antisocial behav-ior. Instead, most focused on effects on out-comes found to be risk factors for long-termdelinquent or antisocial behavior (as re-viewed earlier in this article).

Effects on Risk Factors forDelinquencyThe results of the 40 studies are summarizedin Tables 2–4. Although only the four studiespreviously reviewed investigated programeffects on long-term delinquent or antisocialbehavior, many of the 40 investigated effects


Positive effects on cognitive and/or verbalability and parenting preceded long-termeffects on delinquency and antisocialbehavior.


Notesa To obtain a more detailed version of the table summarizing each of the 40 programs listed in Tables 2–4, contact the author, Hirokazu Yoshikawa, M.A., Department of

Psychology, New York University, 6 Washington Place, 4th Floor, New York, NY 10003.b See the related endnotes at the end of this article for complete citation of the reports and/or studies in which the outcomes of these programs are described.

Outcomes

Study, Related Age of Child Intensity Cognitive Ability Parenting Maternal Life AntisocialEndnote Number,b Course BehaviorProgram Name

Andrews et al., From 3 to 5 months Year 1: three to four E > C at 36 and Mother-child Likelihood to return198354 to 3 years half days per week 48 months interaction: E > C to school: E > CBirmingham PCDC with mothers and at 36 and at posttest

infants together in 48 monthscenter; 15 to 36 E > C on problemmonths: four half solving at 36days per week, monthsmothers asunderstudies toteachers; fifthday in classes

Andrews et al., From 2 months to Two half days per E > C at 36 and E > C at 36 and198354 3 years week in center; 48 months 48 monthsNew Orleans child care andPCDC parenting groups

Brooks-Gunn et al., From hospital Home visits weekly E > C at 3 years E > C at 3 years Maternal199455 discharge to in year 1, biweekly on home employment:Infant Health and 3 years in years 2 and 3 environment E > C at 3 yearsDevelopmentProgram At least five half

days of preschoolper week in secondand third years

Bimonthly parentgroup meetings insecond and thirdyears

Field et al., 198235 From 0 to 6 months Five half days per E > C at 2 years Home environment Repeat pregnan-Teenage Preg- week with mothers E = C at end of cies: E > C at 1nancy Intervention and children program and 2 yearsProgram (E2 group together atonly) preschool; mothers Interaction: Return to

employed as E > C at 4 months work/school: E > Cteachers’ aides

Garber, 198856 From 0 to 5 years Full-day child care, E > C at age 10 E > C at posttestMilwaukee Project five days per week

Job counselingand training forparents

Gray et al., 198357 From 4 to 5 years Summer part-day E > C at first andEarly Training at enrollment to 6 preschool; third gradesProject years at exit weekly to two-

times per month E = C at age 17home visits duringthe rest ofthe year

Table 4

Combination Programs with Both Family Support and Early Education Elementsa


Notesa To obtain a more detailed version of the table summarizing each of the 40 programs listed in Tables 2–4, contact the author, Hirokazu Yoshikawa, M.A., Department of

Psychology, New York University, 6 Washington Place, 4th Floor, New York, NY 10003.b See the related endnotes at the end of this article for complete citation of the reports and/or studies in which the outcomes of these programs are described.

Outcomes

Study, Related Age of Child Intensity Cognitive Ability Parenting Maternal Life AntisocialEndnote Number,b Course BehaviorProgram Name

Johnson and From 1 to 3 years Year 1: 25 home E > C at 36 months E > C at posttest AggressiveWalker, 198758 at enrollment to visits behavior: E <CHouston PCDC 3 to 5 years at exit Year 2: four half E = C, most mea- at 1 to 8 years

days per week of sures, at grade 5 postprogram;educational child E = C at 8 tocare plus classes 11 yearsfor parents postprogram

(but high attrition)

Lally et al., 198859 From 0 to 5 years Weekly home visits; E > C at 3 years AggressiveSyracuse FDRP full-day child care behavior: E > C

from 6 months in grade 1to 5 years

E < C in numberand severity ofjuvenile offens-

esat 10 years post-program (ages13 to 16 years)

Schweinhart et al., From 3 to 4 years Weekly home visits; E > C on either IQ, Child rearing Delinquent199360 at enrollment to 5 preschool: 2.5 school achieve- attitudes: E > C at behaviors: E < CHigh/Scope Perry years hours per day, five ment, grades age 4 to 5 at age 14Preschool days per week through high school;

E < C on special Arrests (numbereducation or grade and severity ofretention through offenses): E < Chigh school at ages 19

and 27

Seitz & Apfel,199461 From pregnancy Average of 28 home E > C at posttest At 10-year follow- Aggression:Yale Child Welfare to 30 months visits; optional up: E < C at 10-yearProject educational 10-year follow-up: Family size: E < C follow-up

child care; well- School attendance: On AFDC: E < C (boys only)baby exams E > C Maternal educa-

IQ: E = C tion: E > CRemedial services: Delay to nextE boys < C boys pregnancy: E > C

Siblings: E < Cgrade retention,special educationat 10- to 12-yearfollow-up

Wasik et al., 199053 From 0 to 5 years E1: 110 home visits E1 > C at 12 to 30 E = C through 54Project CARE and full-day child months months(E1 group only) care five days per

week E1 = C > E2 at 31to 54 months

(E2 only receivedhome visits; seeTable 3)

Table 4 (continued)

Combination Programs with Both Family Support and Early Education Elementsa

on outcomes that roughly reflect the riskcategories of early cognitive ability, parent-ing behavior, maternal life course, and short-term antisocial behavior. Table 5 collapsesthe information contained in Tables 2–4 tosummarize the results of these studies byprogram type (that is, by early education,family support, and combination early edu-cation and family support programs).

The data in Table 5 illustrate the follow-ing points:

1. In contrast to the programs thatcombined early education and family sup-port elements, relatively few of the single-focus early childhood education or fami-ly support programs actually assessedeffects on antisocial behavior. Only 3 of 8early education programs and 4 of 23family support programs assessed the

effects of the programs on antisocialbehavior.

2. Instead, most early education pro-grams assessed effects on children’s earlycognitive ability, and most family supportprograms assessed effects on parenting ormaternal life course.

3. All 8 of the early education pro-grams reviewed measured programeffects on variables such as IQ, schoolachievement, or children’s languagedevelopment. Most were effective in pro-moting children’s early cognitive ability.Early education programs appeared todemonstrate positive results more con-sistently in the early cognitive abilitydomain than in parenting, maternal lifecourse, or antisocial/delinquent behav-ior domains.


Type of Program

Outcome Early Education Family Support Early Educationand FamilySupport

Early Cognitive Ability Of 8 measured: Of 14 measured: Of 11 measured:(IQ, school achievement, 5 positive 5 positive 8 positivelanguage development, or 3 mixed 5 mixed 3 mixedverbal ability) 4 no difference

Parenting Of 1 measured: Of 19 measured: Of 8 measured:(mother-child interaction, 1 mixed 13 positive 6 positiveparenting behavior, 3 mixed 1 mixedattachment, child welfare) 3 no difference 1 no difference

Maternal Life Course Of 1 measured: Of 6 measured: Of 4 measured:(maternal education and 1 positive 5 positive 4 positiveemployment, childbearing, 1 no differencefamily economic self-sufficiency)

Antisocial/Delinquent Behavior Of 3 measured: Of 4 measured: Of 4 measured:(parent/teacher ratings, 2 negative 1 positive 4 positiveofficial delinquency, or 1 no difference 3 no differencecriminal reports) (Short-term effects) (Short-term effects) (Long-term effects)

This table summarizes the information contained in Tables 2–4 for 8 early education programs, 23 family support programs, and 11programs that delivered both early education and family support services. Numbers in the cells in the table indicate (1) the numberof studies that demonstrated statistically significant benefits of the program group over the control or comparison group on that par-ticular group of outcomes (“positive”), (2) the number of studies that demonstrated mixed results (“mixed”—some positive or somenegative and some not statistically significant—either at different points in time or for different groups of participants), (3) the num-ber of studies in which the comparison group outperformed the intervention group (“negative”), (4) the number of studies for whichthere were no statistically significant differences measured between the program and comparison group (“no difference”), and (5)the number of studies that actually measured that particular set of outcomes.

Table 5

Number of Studies Reporting Positive Findings Across Outcomes,Categorized by Program Type

4. Family support programs measuredoutcomes for both children and parents butwere most effective in affecting parental out-comes such as parenting behaviors or mater-nal life course, rather than outcomes associ-ated with the children.

5. Only combination early educa-tion/family support programs affected abroad range of outcomes for both childrenand parents. All of the 11 combination pro-grams identified benefits for children’s cog-nitive ability (8 consistently, 3 only at certainfollow-up points). Six of 8 combination pro-grams which sought to measure parentingbenefits found positive effects, and all 4 ofthose which sought to measure maternal lifecourse outcomes found benefits.

6. It is primarily combination programsthat produced long-term declines in antiso-cial behavior and delinquency.

In sum, the literature review indicatesthat the most effective programs with respectto preventing antisocial behavior and delin-quency were also the programs which com-bined early education and family support ser-vices and had the broadest range of positiveeffects on children and their parents.

Characteristics of EffectiveProgramsThe four programs (Yale, Houston PCDC,Syracuse, and Perry) that demonstrated

long-term effects on delinquent or antisocialbehavior shared some common features thatmay help explain their success.

Scope and IntensityThe programs provided quality educationalchild care and/or preschool as well as sup-port to adults in peer group and family set-tings. They assessed and achieved long-term results affecting both children andparents.

Each of the individual components wasalso intensive. Visits were made to the homesof the families weekly to monthly, depend-ing on the program, and ranged from a totalof 25 to 60. By comparison, only 12 of the 23family-support-only programs offered 25 ormore home visits. The early childhood edu-cational component ranged from half-daysummer sessions to full-day sessions, usuallyfour or five days a week.

The combination of early educationaland family support models of interventionmay have been crucial to obtaining effectson multiple risks for chronic delinquency.Although an adequate test of the effects ofearly-education-only programs on both par-enting and child cognitive ability has yet tobe carried out, Table 5 suggests that family-support-only programs appear to be lesslikely than combination programs to affectrisks in both cognitive and parentingdomains.


© S

tep

ha

nie

Ra

uss

er/

FPG

Inte

rna

tion

al

Of course, it is possible that multiplecomponents are not necessary for long-termeffects on chronic delinquency and thatsome of the single-component programsreviewed have had or will have long-termeffects on children and families (none hasyet carried out a long-term follow-up). Thereis evidence, however, from the evaluation ofa two-component intervention to reduceantisocial behavior in middle childhood thatboth child- and parent-focused componentswere necessary for clinically significanteffects.72

QualityThe four programs with long-term effects onantisocial behavior and/or delinquencywere quality programs. They had strong the-oretical bases for their center-based andhome visiting curricula; most curricula

emphasized the initiation and planning ofactivities by the child rather than the teacher(Houston, Syracuse, and Perry programs);home visitor-to-family ratios were generally 1to 10 or better for full-time home visitors;staff-child ratios in infant/toddler educa-tional child care were in the range of oneadult to three or four children, and 1 to 6 inpreschool programs; preservice and in-ser-vice training was extensive; and supervisionwas ongoing. (See the article by Frede in thisjournal issue for further discussion of qualityand curricula.)

Population ServedAlthough none of the four programs had theprevention of antisocial behavior and crimeas their stated purpose, the areas that, in fact,have highest crime rates—urban low-incomecommunities73—were targeted in all fourprograms with long-term effects. These areasand participants were not selected based onrisk for delinquency, but rather on the moregeneral principle that disadvantaged familieshave fewer resources to spend on qualityearly childhood care and education than domiddle- or upper-class families.74,75

Duration and TimingIn general, duration did not appear to berelated to the likelihood or magnitude oflong-term effects on antisocial behavior anddelinquency: none of the programs withlong-term effects was shorter than two years,but length of intervention ranged from twoto five years. With respect to timing, the fourprograms were all implemented during thechild’s first five years. Two of the programsbegan at or before birth,59,61 one began atage one,58 and the other at age three.60 Mostfamily support interventions reviewed herehave been implemented during the prenatalor early infancy periods. This is a time ofheightened stress for parents, when they maybe particularly open to outside support.75

Single or adolescent parents, parents of lowbirth weight infants, and parents with alreadylow levels of social support may benefit par-ticularly from support during the perinatalperiod.48,55 Beginning a program beforebirth would increase utilization of prenatalcare in this high-risk population, which mayhelp reduce the incidence of perinatal riskfactors for chronic delinquency.

This does not mean that a parent-focused intervention begun later in child-hood could not also decrease children’searly antisocial behavior. 76 However, it maybe that the magnitude of the benefits may beenhanced with earlier services.

While family support may be particularlyimportant during the first few years of life,results of the early education studiesreviewed here and in the article by Barnett inthis journal issue are mixed as to when is thebest time to deliver early education services.

SummaryIn sum, this review demonstrates that, first,there are early risk and protective factors forchronic delinquency. Second, research onpossible causal mechanisms for chronicdelinquency suggests that providing supportfor early nurturing, parenting, and verbalability, as well as ameliorating both familyand community level poverty and their cor-relates, are promising prevention strategies.Third, family support programs are quitelikely to reduce risks by improving maternallife course and parenting, but are less likelyto improve early child cognitive ability.Fourth, early education programs, converse-ly, are quite likely to increase early child cog-


The programs with long-term ef fects onantisocial behavior and/or delinquency werequality programs. Most curricula emphasizedthe initiation and planning of activities bythe child rather than the teacher.

nitive ability but seem less likely to affect thematernal life course and parenting. Fifth,combination programs of sufficient intensityand quality are more likely to reduce risks inboth areas. Several of these programs, tar-geting urban, low-income families, have pro-duced long-term decreases in antisocialbehavior or chronic delinquency.

ConclusionsThe findings reviewed above provide somecompelling suggestions about where effortsshould be concentrated in the future.

Research ImplicationsAlthough the research strongly suggests thatcombination early childhood and familysupport programs can prevent delinquency,there still remain many questions about howbest to design a preventive intervention.Further research is required to identify thespecific program characteristics that con-tribute to the effectiveness of preventiveinterventions. Planned variation studies, inwhich different combinations of services atdifferent levels of intensity are compared,are particularly needed. Effects on other out-comes which share risk factors with chronicdelinquency, such as early substance abuse,teenage childbearing, and depression,should be investigated. Finally, research ondiverse populations, especially those neglect-ed thus far in early childhood care and edu-cation research, should be encouraged todetermine if effectiveness of services variesacross different communities. Most researchto date has focused on white or African-American low-income families.

Program and Policy ImplicationsThe economic rationale for governmentprograms for low-income families has beendescribed as governmental investment inhuman capital for those families with fewerresources available to invest in their chil-dren.74,77 The costs to government of provid-ing quality early childhood programs, in thisview, are balanced against the value to soci-ety of increased productivity and decreasedsocial problems. Providing child careresources enables poor parents to work andto increase their education and job skills.Providing poor children with better parent-ing and better education yields more pro-ductive workers in a market which increas-ingly values highly skilled workers.74 Earlychildhood programs that prevent delin-

quency and crime represent at least twopotential sources of savings to society: (1)reductions in crime and in justice systemcosts, and (2) gains in work force participa-tion when youths who are less delinquentthan their peers participate more in thelegitimate economy.78

Economic AnalysesThree of the four programs that producedlong-term effects on crime and delinquencyhave also reported information about costsand benefits. Two (the Perry PreschoolProject and the Syracuse study) report costsor benefits associated with crime or delin-quency. The third, the Yale study, primarilyfocused on costs and benefits associated witheducational outcomes, and it will thereforenot be reviewed here.61,67

The Perry Preschool Project’s analysis isthe most sophisticated of any of the three(see also the article by Barnett in this journalissue). Monetary values were estimated forthe program costs, as well as for benefits inareas such as elementary and secondary edu-cation, adult secondary education, postsec-ondary education, employment-relatedcompensation, public welfare assistance,

and delinquency and crime. Results indicat-ed that the program, which cost about$12,356 per family, yielded benefits totaling$108,002 per family. The net present valueof the program’s benefits was $95,646 (allamounts in 1992 dollars, adjusted for infla-tion, and calculated with a 3% discountrate). Of the benefits, $12,796 was due to sav-ings in the justice system, and $57,585 wasdue to savings for crime victims.60 (For addi-tional details, see Table 3 in the article byBarnett in this journal issue.)

In reports concerning the Syracuse pro-gram, researchers estimated the costsincurred by control group and programgroup participants due to court processing,probation supervision, placement in foster


The costs to government of providing quality early childhood programs are balanced against the value to society of increased productivity and decreased social problems.

care, nonsecure detention, and securedetention. The four youthful offenders inthe program group were judged to haveincurred costs from these sources of $12,111as compared with costs from these sources of$107,192 for the 12 offenders in the controlgroup.59 These data are difficult to interpret,however, without information on the cost ofthe program and the timing of costs andbenefits.

In summary, although only a few studieshave calculated the costs and benefits ofthese programs, it is interesting to notethat in one of the best economic assess-ments conducted to date, the largest per-centage of the total economic benefits wasassociated with decreases in crime anddelinquency.

Relevance for Public Policy Today The studies reviewed in this article representyears of accumulated experience and clearlysuggest that programs combining earlychildhood education and family support ser-vices have helped to prevent delinquencyand antisocial behavior. It is less clear, how-ever, that similar programs launched todaywould generate the same results. The fourprograms with long-term effects were car-

ried out in the early to late 1970s; numerousdemographic, social, and economic changeshave occurred since then which might affectthe outcomes of early intervention. Forinstance, increases in the rate of employ-ment among women, including low-incomewomen, have resulted in greater need forfull-time, quality child care, rather than thehalf-day services provided in most of theseprograms. Frequent home visiting may nowbe less attractive to employed parents withalready busy schedules. In addition, thesurge in youth involvement with the drugtrade and with handguns suggests that fami-ly-focused interventions alone, withoutbroader efforts to attack these neighbor-hood level causal factors, may not have theirintended impact.

Given the limited number of studies thatdemonstrated changes in delinquent, crim-inal, or antisocial behavior, it may be tooearly to bring combined early educationand family support initiatives to nationalscale based solely on their promise to pre-vent delinquency. However, there are othercompelling rationales to combine earlyeducation and family support. Theseinclude the multiple needs of many oftoday’s families and children and the recog-nition that services for children are toooften fragmented and uncoordinated.79

The resultant calls for centralizing and inte-grating child-focused and family-focusedservices parallel the approach of programsfound promising here in the prevention ofdelinquency.

New Head Start initiatives and the imple-mentation of two-generation programs suchas those reviewed by St. Pierre and col-leagues in this journal issue exemplify thesorts of programs that are suggested by thisreview. Head Start, for example, is seeking tostrengthen its family support component.Since its inception in 1965, the program hassought to combine comprehensive familysupport services with a quality preschooleducation program,80 but the family supportcomponents of the program are in need ofimprovement. More than one-third of pro-grams in 1993 had social service worker case-loads of more than 250; in response, the1993 Advisory Committee on Head StartQuality and Expansion called for a 1 to 35ratio for all staff who work with families.81

Proposed improvements in the mentalhealth component82 and the parent involve-ment component83 may contribute to theprogram’s potential as a comprehensivefamily support program, as well. Efforts toestablish a national Head Start for infantsand toddlers may also help improve the pro-gram’s likelihood of decreasing early risksfor chronic delinquency.

Two-generation programs combine thegoals of economic self-sufficiency with thoseof family support and preschool education.They provide a mix of child care, family sup-port, parental educational and job training,and preschool education, and have been dis-tinguished from family support programswith less emphasis on job training andparental education. As the article by St.Pierre and colleagues in this journal issue


As one element in a comprehensive plan,early education may lessen the currentdevastating impact of chronic delinquencyon America’s children and families.

points out, these are in reality three-com-ponent programs, providing adult-focused,parent-focused, and child-focused services,in contrast to the primarily two-componentprograms reviewed in this article. Two-gen-eration programs of sufficient quality andintensity may address risk factors for delin-quency in three important areas: familysocioeconomic status, parenting, and childcognitive development.

Combining quality early education andfamily support services holds great promisefor preventing delinquency, both on theo-retical grounds, based on what is known

about risk factors for antisocial behavior, andon empirical grounds, based on the resultsof the program evaluation studies reviewedhere. However, even if such combined pro-grams are widely implemented, they will noteliminate juvenile crime, and the early inter-vention community should not overstatetheir potential effect in that domain. Nev-ertheless, as one element in a comprehen-sive plan to address poverty, drugs, guns, andother environmental causes of crime, earlyeducation and family support programs maylessen the current devastating impact ofchronic delinquency on America’s childrenand families.

1. Maguire, K., and Pastore, A.L., eds. Sourcebook of criminal justice statistics: 1993. Washington,DC: U.S. Government Printing Office, 1994.

2. Federal Bureau of Investigation. Uniform crime reports for the United States: 1993. Washington,DC: U.S. Government Printing Office, 1994.

3. Violent Crime Control and Law Enforcement Act of 1994, Public Law 103-322, 108 Stat. 1796,103rd Congress.

4. Binder, A., Geis, G., and Bruce, D. Juvenile delinquency: Historical, cultural, legal perspectives. NewYork: Macmillan, 1988.

5. National Council of Juvenile and Family Court Judges. Child development: A judge’s reference.Reno, NV: National Council of Juvenile and Family Court Judges, 1993.

6. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed.Washington, DC: American Psychiatric Press, 1994.

7. Farrington, D.P. Early precursors of frequent offending. In From children to citizens: Families,schools, and delinquency prevention. J.Q. Wilson and G.C. Loury, eds. New York: Springer-Verlag, 1987.

8. Tracy, P.E., Wolfgang, M.E., and Figlio, R.M. Delinquency careers in two birth cohorts. New York:Plenum Press, 1990.

9. Moffitt, T.E. Adolescence-limited and life-course-persistent antisocial behavior: A developmen-tal taxonomy. Psychological Review (1993) 100:674–701.

10. Loeber, R., and Schmaling, K.B. The utility of differentiating between mixed and pure formsof antisocial behavior. Journal of Abnormal Child Psychology (1985) 13:315–36.

11. Yoshikawa, H. Prevention as cumulative protection: Effects of early family support and educa-tion on chronic delinquency and its risks. Psychological Bulletin (1994) 115:27–54.

12. Farrington, D.P. Predicting self-reported and official delinquency. In Prediction in criminology.D.P. Farrington and R. Tarling, eds. Albany, NY: SUNY Press, 1985.

13. Patterson, G.R., Reid, J.B., and Dishion, T.J. Antisocial boys. Eugene, OR: Castalia Press, 1992.

14. Baum, C.J., and Forehand, R. Long term follow-up assessment of parent training by use ofmultiple outcome measures. Behavior Therapy (1981) 12:643–52.

15. Werner, E.E., and Smith, R.S. Vulnerable but invincible: A longitudinal study of resilient children andyouth. New York: Adams, Bannister, Cox, 1982.

16. Crnic, K.A., Greenberg, M.T., Ragozin, A.S., et al. Effects of stress and social support on moth-ers and premature and full-term infants. Child Development (1983) 54:209–17.

17. Leadbeater, B.J., and Bishop, S.J. Predictor of behavior problems in preschool children of inner-city Afro-American and Puerto Rican adolescent mothers. Child Development (1994) 65:638–48.

18. McGee, R., Williams, S., Share, D.L., et al. The relationship between specific reading retarda-tion, general reading backwardness and behavioral problems in a large sample of Dunedin


boys: A longitudinal study from five to eleven years. Journal of Child Psychology and Psychiatry(1986) 27:597–610.

19. Werner, E.E. Vulnerability and resiliency in children at risk for delinquency: A longitudinalstudy from birth to adulthood. In Primary prevention of psychopathology: Vol. 10. Prevention ofdelinquent behavior. J.D. Burchard and S.N. Burchard, eds. Newbury Park, CA: Sage, 1987,pp. 16–43.

20. Moffitt, T.E. The neuropsychology of conduct disorder. Development and Psychopathology (1993)5:135–52.

21. Kolvin, I., Miller, F.J.W., Fleeting, M., and Kolvin, P.A. Social and parenting factors affectingcriminal offense rates. British Journal of Psychiatry (1988) 152:80–90.

22. McCord, J. Some child-rearing antecedents of criminal behavior in adult men. Journal ofPersonality and Social Psychology (1979) 37:1477–86.

23. Abelson, W.D. Head Start graduates in school: Studies in New Haven, Connecticut. In A reporton longitudinal evaluations of preschool programs: Vol. 1. Longitudinal evaluations. S. Ryan, ed.Washington, DC: U.S. Department of Health, Education, and Welfare, 1974.

24. Beller, E.K. The Philadelphia study: The impact of preschool on intellectual and socioemo-tional development. In As the twig is bent . . . lasting effects of preschool programs. Consortium forLongitudinal Studies, ed. Hillsdale, NJ: Erlbaum, 1983, pp. 333–76.

25. Campbell, F.A., and Ramey, C.T. Effects of early intervention on intellectual and academicachievement: A follow-up study of children from low-income families. Child Development(1994) 65:684–98.

26. Deutsch, M., Deutsch, C.P., Jordan, T.J., and Grallo, R. The IDS Program: An experi-ment in early and sustained enrichment. In As the twig is bent . . . lasting effects ofpreschool programs. Consortium for Longitudinal Studies, ed. Hillsdale, NJ: Erlbaum,1983, pp. 377–410.

27. Hebbeler, K. An old and a new question on the effects of early education for children fromlow income families. Educational Evaluation and Policy Analysis (1985) 78:207–16.

28. Lee, V.E., Brooks-Gunn, J., Schnur, E., and Liaw, F.R. Are Head Start effects sustained? A lon-gitudinal follow-up comparison of disadvantaged children attending Head Start, nopreschool, and other preschool programs. Child Development (1990) 61:495–507.

29. Miller, L.B., and Bizzell, R.P. The Louisville Experiment: A comparison of four programs. InAs the twig is bent . . . lasting effects of preschool programs. Consortium for Longitudinal Studies,ed. Hillsdale, NJ: Erlbaum, 1983, pp. 171-99.

30. Reynolds, A.J. One year of preschool intervention or two: Does it matter for low-income Black childrenfrom the inner city? Paper presented at the Second National Head Start Research Conference.Washington, DC, November 1993.

31. Achenbach, T.M., Phares, V., Howell, C.T., et al. Seven-year outcome of the VermontIntervention Program for low-birthweight infants. Child Development (1990) 61:1672–81.

32. Badger, E. Effects of a parent education program on teenage mothers and their offspring. InTeenage parents and their offspring. K.G. Scott, T. Field, and E.G. Robertson, eds. New York:Grune and Stratton, 1981.

33. Barrera, M.E., Rosenbaum, P.L., and Cunningham, C.E. Early home intervention with low-birth-weight infants and their parents. Child Development (1986) 57:20–33.

34. Barth, R.P., Hacking, S., and Ash, J.R. Preventing child abuse: An experimental evaluation ofthe Child Parent Enrichment Project. Journal of Primary Prevention (1988) 8:201–17.

35. Field, T., Widmayer, S., Greenberg, R., and Stoller, S. Effects of parent training on teenagemothers and their infants. Pediatrics (1982) 69:703–7.

36. Gray, J.D., Cutler, C.A., Dean, J.G., and Kempe, C.H. Prediction and prevention of child abuseand neglect. Journal of Social Issues (1979) 35:127–39.

37. Gray, S.W., and Ruttle, K. The Family-Oriented Home Visiting Program: A longitudinal study.Genetic Psychology Monographs (1980) 102:299–316.

38. Gutelius, M.F., Kirsch, A.D., MacDonald, S., et al. Controlled study of child health supervision:Behavioral results. Pediatrics (1977) 60:294–304.


39. Hardy, J.B., and Streett, R. Family support and parenting education in the home: An effectiveextension of clinic-based preventive health care services for poor children. Journal of Pediatrics(1989) 115:927–31.

40. High/Scope Educational Research Foundation. The national Home Start evaluation: InterimReport V. Summative evaluation results. Ypsilanti, MI: High/Scope Press, 1974.

41. Jacobson, S.W., and Frye, K.F. Effect of maternal social support on attachment: Experimentalevidence. Child Development (1991) 62:572–82.

42. Jester, R.E., and Guinagh, B.J. The Gordon Parent Education Infant and Toddler Program. InAs the twig is bent . . . lasting effects of preschool programs. Consortium for Longitudinal Studies,ed. Hillsdale, NJ: Erlbaum, 1983, pp. 103–32.

43. Lambie, D.Z., Bond, J.T., and Weikart, D.P. Home teaching with mothers and infants. The Ypsilanti-Carnegie Infant Education Project: An experiment. Ypsilanti, MI: High/Scope EducationalResearch Foundation, 1974.

44. Larson, C.P. Efficacy of prenatal and postpartum home visits on child health and develop-ment. Pediatrics (1980) 66:191–97.

45. Lieberman, A.F., Weston, D.R., and Pawl, J.H. Preventive intervention and outcome with anx-iously attached dyads. Child Development (1991) 62:199–209.

46. Lyons-Ruth, K., Connell, D.B., Grunebaum, H.U., and Botein, S. Infants at social risk:Maternal depression and family support services as mediators of infant development andsecurity of attachment. Child Development (1990) 61:85–98.

47. Madden, J., O’Hara, J., and Levenstein, P. Home again: Effects of the Mother-Child HomeProgram on mother and child. Child Development (1984) 55:636–47.

48. Olds, D.L., Henderson, C.R., Tatelbaum, R., and Chamberlin, R. Improving the life-coursedevelopment of socially disadvantaged mothers: A randomized trial of nurse home visitation.American Journal of Public Health (1988) 78:1436–45.

49. Osofsky, J.D., Culp, A.M., and Ware, L.M. Intervention challenges with adolescent mothersand their infants. Psychiatry (1988) 51:236–41.

50. Ross, G.S. Home intervention for premature infants of low-income families. American Journalof Orthopsychiatry (1984) 54:263–70.

51. Seitz, V., Rosenbaum, L.K., and Apfel, N. Effects of an intervention program for pregnantadolescents: Educational outcomes at two years postpartum. American Journal of CommunityPsychology (1991) 19:911–30.

52. Siegel, E., Bauman, K.E., Schaefer, E.S., et al. Hospital and home support during infancy:Impact on maternal attachment, child abuse and neglect, and health care utilization.Pediatrics (1980) 66:183–90.

53. Wasik, B.H., Ramey, C.T., Bryant, D.M., and Sparling, J.J. A longitudinal study of two earlyintervention strategies: Project CARE. Child Development (1990) 61:1682–96.

54. Andrews, S.R., Blumenthal, J.B., Johnson, D.L., et al. The skills of mothering: A study ofParent Child Development Centers (New Orleans, Birmingham, Houston). Monographs of theSociety for Research in Child Development. Serial No. 198 (1982) 47,6.

55. Brooks-Gunn, J., McCormick, M.C., Shapiro, S., et al. The effects of early education inter-vention on maternal employment, public assistance, and health insurance: The InfantHealth and Development Program. The American Journal of Public Health (1994)84:924–31.

56. Garber. H.L. The Milwaukee Project: Preventing mental retardation in children at risk. Washington,DC: American Association on Mental Retardation, 1988.

57. Gray, S.W., Ramsey, B.K., and Klaus, R.A. The Early Training Project: 1962–1980. In As the twigis bent . . . lasting effects of preschool programs. Consortium for Longitudinal Studies, ed. Hillsdale,NJ: Erlbaum, 1983, pp. 33–69.

58. Johnson, D.L., and Walker, T. Primary prevention of behavior problems in Mexican-Americanchildren. American Journal of Community Psychology (1987) 15:375–85.

59. Lally, J.R., Mangione, P.L., and Honig, A.S. The Syracuse University Family DevelopmentResearch Project: Long-range impact of an early intervention with low-income children and


their families. In Parent education as early childhood intervention: Emerging directions in theory,research and practice. D.R. Powell, ed. Norwood, NJ: Ablex, 1988.

60. Schweinhart, L.J., Barnes, H.V., Weikart, D.P., et al. Significant benefits: The High/Scope PerryPreschool Study through age 27. Ypsilanti, MI: High/Scope Press, 1993.

61. Seitz, V., and Apfel, N. Parent-focused intervention: Diffusion effects on siblings. ChildDevelopment (1994) 65:677–83.

62. Many of these programs are also reviewed in Olds, D.L., and Kitzman, H. Review of researchon home visiting for pregnant women and parents of young children. The Future of Children(Winter 1993) 3,3:53–92.

63. The study by Field and colleagues of the Miami Teenage Parent Intervention Project and thestudy by Wasik and colleagues of Project CARE appear in both the family support and combi-nation categories because they present the results of both kinds of programs. The mono-graph by Andrews and colleagues presents short-term results of the Birmingham, NewOrleans, and Houston Parent-Child Development Centers.

64. Berrueta-Clement, J.R., Schweinhart, L.J., Barnett, W.S., et al. Changed lives: The effects of thePerry Preschool Program on youths through age 19. Ypsilanti, MI: High/Scope Press, 1984. TheHigh/Scope Perry Preschool Project will be referred to simply as the Perry Preschool Projectin the remainder of this article.

65. Schweinhart, L.J., and Weikart, D.P. Young children grow up: The effects of the Perry PreschoolProgram on youths through age 15. Ypsilanti, MI: High/Scope Press, 1980.

66. See note no. 60, Schweinhart, Barnes, Weikart, et al., p. 83.

67. Seitz, V., Rosenbaum, L.K., and Apfel, N.H. Effects of family support intervention: A ten-yearfollow-up. Child Development (1985) 56:376–91.

68. Johnson, D.L., and Blumenthal, J.B. A follow-up of the Parent-Child Development Centers. Paperpresented at the Second National Head Start Research Conference. Washington, DC,November 1993.

69. Cohen, J. Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Erlbaum, 1983.

70. Johnson, D.L., Breckenridge, J.N., and McGowan, R.J. Home environment and early cognitivedevelopment in Mexican-American children. In Home environment and early cognitive develop-ment. A.W. Gottfried, ed. New York: Academic Press, 1984.

71. Weikart, D.P., Bond, J.T., and McNeil, J.T. The Ypsilanti Perry Preschool Project: Preschool years andlongitudinal results through fourth grade. Ypsilanti, MI: High/Scope Press, 1978.

72. Kazdin, A.E., Siegel, T.C., and Bass, D. Cognitive problem-solving skills training and parentmanagement training in the treatment of antisocial behavior in children. Journal of Clinicaland Consulting Psychology (1992) 60:733–47.

73. Earls, F. Violence and today’s youth. The Future of Children (Winter 1994) 4, 3:4–23.

74. Barnett, W.S. New wine in old bottles: Increasing the coherence of early childhood care andeducation policy. Early Childhood Research Quarterly (1993) 8:519–58.

75. Slaughter, D.T. Programs for racially and ethnically diverse American families: Some criticalissues. In Evaluating family programs. H.B. Weiss and F.H. Jacobs, eds. New York: Aldine deGruyter, 1988.

76. Patterson, G.R., Dishion, T.J., and Chamberlin, P. Outcomes and methodological issues relat-ing to treatment of antisocial children. In Handbook of effective psychotherapy. T.R. Giles, ed. NewYork: Plenum, 1993.

77. Becker, G.S. A treatise on the family. Enlarged edition. Cambridge, MA: Harvard UniversityPress, 1991.

78. Ehrlich, I. Participation in illegitimate activities: An economic analysis. In Essays in the econom-ics of crime and punishment. G.S. Becker and W.M. Landes, eds. New York: Columbia UniversityPress, 1974.

79. Schorr, L.B. Within our reach: Breaking the cycle of disadvantage. New York: Doubleday, 1988.

80. U.S. Department of Health, Education, and Welfare, Office of Child Development. Recommendationsfor a Head Start program by a panel of experts (February 19, 1965). Washington, DC: U.S. Department ofHealth and Human Services, Administration for Children, Youth, and Families, 1965.


81. Advisory Committee on Head Start Quality and Expansion. Creating a 21st century Head Start:Final report of the Advisory Committee on Head Start Quality and Expansion. Document #1994-517-593. Washington, DC: U.S. Government Printing Office, 1993.

82. Task Force on Head Start and Mental Health. Strengthening mental health in Head Start:Pathways to quality improvement. New York: American Orthopsychiatric Association, April 1994.

83. Zigler, E., and Styfco, S. Head Start and beyond: A national plan for extended childhood intervention.New Haven: Yale University Press, 1993.


long-term effects of early childhood programs on social ...€¦ · preventing chronic delinquency:...

Documents