over-time changes in adjustment and competence...

Over-Time Changes in Adjustment andCompetence among Adolescents fromAuthoritative, Authoritarian, Indulgent, andNeglectful Families

Laurence SteinbergTemple University

Susie D. LambornUniversity of West Florida

Naney DarlingTemple University

Nina S. MountsUniversity of Illinois

Sanford M. DornbuschStanford University

STEINBERG, LAURENCE; LAMBOBN, SUSIE D.; DARLING, NANCY; MOUNTS, NINA S.; and DORNBUSCH,SANFORD M , Over-Time Changes in Adjustment and Competence among Adolescents from Au-thoritative, Authoritarian, Indulgent, and Neglectful Families. CHILD DEVELOPMENT, 1994, 65,754-770. In a previous report, we demonstrated that adolescents' adjustment varies as a functionof their parents' style (e.g., authoritative, authoritarian, indulgent, neglectful). This 1-year follow-up was conducted in order to examine whether the observed differences are maintained overtime. In 1987, an ethnically and socioeconomically heterogeneous sample of approximately 2,30014-18-year-oIds provided infomiation used to classify the adolescents' families into 1 of 4 parent-ing style groups. That year, and again 1 year later, the students completed a battery of standard-ized instruments tapping psychosocia! development, school achievement, internalized distress,and behavior problems. Differences in adjustment associated with variations in parenting areeither maintained or increase over time. However, whereas the benefits of authoritative parent-ing are largely in the maintenance of previous levels of high adjustment, the deleterious conse-quences of neglectful parenting continue to accumulate.

An extensive literature on socialization rind (1967,1971). Children who are raised inpractices and their effects provides consis- authoritative homes score higher than theirtent evidence that parental warmth, indue- peers from authoritarian, indulgent, or ne-tive discipline, nonpunitive punishment glectful homes on a wide variety of measurespractices, and consistency in child rearing of competence, achievement, social devel-are each associated with positive develop- opment, self-perceptions, and mental healthmental outcomes in children (Maccoby & (Maccoby & Martin, 1983). Several recentMartin, 1983). Since the early 1970s, this studies have applied Baumrind's model toconstellation of practices has come to be explain variations in pattems of adolescentknown as "authoritative" parenting, one of development, including academic achieve-several prototypic styles of parenting identi- ment, psychosocial development, behaviorfied in the seminal studies of Diana Baum- problems, and psychological symptoms

Preparation of this manuscript was supported by grants from the Lilly Endowment and theWilliam T. Grant Foundation. The study on which ^ i s report is based was supported by grantsfrom the U.S. Department of Education, through the National Center on Effective SecondarySchools, and the Spencer Foundation. Address correspondence to the first author at the Depart-ment of Psychology, Temple University, Philadelphia, PA 19122.

[Child Development, 1994, 65, 754-770. © 1994 by the Society for Research in Child Development, Inc.All rights reserved. 0009-3920/94/6503-0005«01.00]

Steinberg et al. 755

(e.g., Dornbusch, Ritter, Liederman, Roberts,& Fraleigh, 1987; Lamborn, Mounts,Steinberg, & Dornbusch, 1991; Steinberg,Elmen, & Mounts, 1989; Steinberg, Lam-born, Dornbusch, & Darling, 1992; Stein-berg, Mounts, Lamborn, & Dornbusch,1991), and these reports find that adoles-cents, like their younger counterparts, bene-fit from authoritative parenting.

In a previous report (Lamborn et al.,1991), we provided evidence for the utility,in research on parental socialization and ad-olescent adjustment, of a fourfold parentingtypology based on Baumrind's frameworkand set forth by Maccoby and Martin (1983).In that earlier study, the families of approxi-mately 4,000 14-i8-year-olds were classi-fied into one of four groups (authoritative,authoritarian, indulgent, or neglectful) onthe basis of adolescents* ratings of theirparents on two dimensions; acceptance/involvement and strictness/supervision.Analysis of adolescents' scores on four setsof outcomes—psychosocial development,school achievement, internalized distress,and problem behavior—indicated numerousdifferences among adolescents from authori-tative, authoritarian, indulgent, and neglect-ful homes. Specifically, adolescents fromauthoritative homes scored highest, and ado-lescents from neglectful homes lowest, onthe majority of indices of adjustment. Ado-lescents in either the authoritarian or the in-dulgent group showed a mixture of positiveand negative traits, with adolescents fromauthoritarian homes scoring reasonably wellon measures of school achievement and de-viance but relatively poorly on measures ofself-reliance and self-conceptions, and ado-lescents from indulgent homes scoring rela-tively poorly with respect to school engage-ment, drug and alcohol use, and schoolmisconduct but relatively well on measuresof social competence and self-confidence. Ingeneral, these pattems did not vary as afunction of adolescent age, gender, ethnic-ity, or family background.

The present article presents data from a1-year follow-up of these adolescents. Theshort-term longitudinal follow-up is impor-tant for several reasons. First, although thecross-sectional findings reported in the ear-lier article are consistent with other researchand theory on adolescent socialization, theobserved correlations between adolescentbehavior and parenting practices could bedue to the influence of young people ontheir parents, rather than the reverse (e.g..Bell, 1968). Although some of the specific

findings are difficult to construe from withinthis causal framework (e.g., it is hard toimagine that parents respond to adolescentdrug use with increased indulgence), manyare not (e.g., it is quite plausible that parentsrespond to high achievement with authori-tativeness). Indeed, Steinberg et al. (1989)report that psychosocial maturity duringearly adolescence is likely to evoke parentalwarmth, rather than the reverse. A short-term longitudinal design permits a morecareful, if still imperfect, assessment of theimpact of parenting practices on adolescentdevelopment.

Second, as will be made clear in subse-quent sections of this article, the short-termlongitudinal design employed here helps torebut claims that observed correlations be-tween adolescent adjustment and parentingpractices are due either to third variables orto common source or method variance—acommon criticism of socialization researchthat employs survey data (in the presentstudy, both the classification of parents andthe assessment of adolescent adjustment de-rive from adolescents' reports). Specifically,by using adolescents' initial adjustmentscores as covariates in analyses designed topredict their later adjustment from their par-ents' practices, we greatly reduce the com-mon method and source variance sharedbetween the adjustment and parenting mea-sures.

Finally, and most important, even if thedifferences among the adolescents reportedin our earlier article were actually due todifferences in parental practices, any differ-ences we observed among the youngsters inour sample might have existed long beforeadolescence and simply remained in placeover time. The short-term longitudinal de-sign of the present study permits us to askwhether there are continuing benefits or lia-bilities of particular approaches to parentingduring the middle adolescent years and toreport on the 1-year developmental trajecto-ries of youngsters from different home envi-ronments. Such infonnation is exceedinglyuseful to both practitioners and parents, whounderstandably want to know if parentingpractices continue to "make a difference"once the child has reached high school.

MethodSample

Our sample is drawn from the studentbodies of nine high schools in Wisconsinand northern California. The schools were

756 Child Development

selected to yield a sample of students fromdifferent socioeconomic brackets, a varietyof ethnic backgrounds (African-, Asian-, Eu-ropean-, and Hispanic-American), differentfamily structures (e.g., intact, divorced, andremarried), and different types of communi-ties (urban, suburban, and rural). Data forthe present analyses were collected duringthe 1987-88 and 1988-89 school years viaself-report surveys filled out by the studentson 2 days of survey administration eachschool year. Because of its length, the surveywas divided into two parts.

ProcedureRecent reports suggest that the use of

"active consent" procedures in research onadolescents and their families (i.e., proce-dures requiring active parental written con-sent in order for their adolescents to partici-pate in the research) may result in samplingbiases that overrepresent well-functioningteenagers and families (e.g., Weinberger,Tublin, Ford, & Feldman, 1990). Althoughgroups of participants and nonparticipantsgenerated through such consent proceduresmay be comparable demographlcally (the di-mension along which investigators typicallylook for evidence of selective participation),the procedure screens out a disproportionatenumber of adolescents who have adjustmentproblems and/or family difficulties. Becausewe were interested in studying "neglectful"as well as more involved families in this re-search, and because many of our outcomemeasures are in the domain of adjustment,we were concerned that employing thestandard active consent procedure (in whichboth parents and adolescents are asked toretum signed consent forms to their child'sschool) would bias our sample toward moreinvolved—and, presumably, better func-tioning—families.

After considering the age of our respon-dents and their ability to provide informedconsent, and with the support of the admin-istrators of our participating schools, theschool districts' research review commit-tees, representatives ofthe U.S. Departmentof Education (our chief funding agent), andour own institutions' human subjects com-mittees, we decided to employ a consentprocedure that requested "active" informedconsent from the adolescents, but "passive"informed consent from their parents. Allparents in the participating schools wereinformed, by first-class mail, of the dateand nature of our study well in advance ofthe scheduled questionnaire administra-tion. (We provided schools with letters

in stamped, unaddressed envelopes to bemailed by school officials in order to protectthe privacy of the families.) Parents wereasked to call or write to their child's schoolor our research office if they did not wanttheir child to participate in the study. Fewerthan 1% of the adolescents in each of thetarget schools had their participation with-held by tiieir parente.

All ofthe students in attendance on eachday of testing were invited to participate inthe study and asked to complete the ques-tionnaires. Informed consent was obtainedfrom all participating students. For eachquestionnaire administration, out ofthe totalschool populations, approximately 5% ofthestudents chose not to participate (or hadtheir participation withheld by parents), ap-proximately 15% were absent from school onthe day of questionnaire administeation {thisfigure is comparable to national figures ondaily school attendance), and approximately80% provided completed questionnaires. Inthe 1987-88 school year, 11,669 studentsparticipated in the study. In the 19^-89school year, 11,248 students participated.Our longitudinal study sample across the2 years included approximately 6,357 stu-dents. Subject attrition over the 1-year pe-riod was due primarily to graduation (the1987-88 sample included 2,538 seniors),dropping out of school, or movement out ofthe school district (17% of the sample re-ported having attended more than one highschool, indicative of the h i ^ degree of mo-bility witiiin the population studied).

The use of this consent procedure hasboth costs and benefits. On the positive side,we have responses from a more representa-tive sample of adolescents, including adoles-cents whose parents are not involved inschool, than one would otherwise have. Onthe negative side, however, our consent pro-cedure does not permit us to obtain informa-tion from an equ^ly representative set ofparents. Rather than limit our study to thewell-fiinctioning parents who volunteer toparticipate in research of this sort, we havechosen to collect information on parentingpractices from the adolescents themselves.

Characteristics of the study sample arepresented in the first column of Table 1. Asthe table indicates, the sample is evenly di-vided Eunong males and females and amongninth, tendi, and eleventh graders. The sam-ple is quite diverse with respect to oiiier de-mographic variables: more than 40% of therespondents are from an ethnic minority


TABLE 1

DEMOGRAPHIC CHARACTERISTICS OF THE TOTAL(N = 6,902) AND STUDY (N = 2,353) SAMPLES

Total StudySample Sample

Ethnicity:African-American 8.6 9.1European-American 60.7 62.2Asian-American 15.1 14.8Hispanic-American 15.5 13.9

Sex:Male 49.9 48.2Female 50.1 51.8

Parental education:< College 28.5 29.3College graduate 71.5 70.7

Family structure:Intact 70.8 70.9Nonintact 29.2 29.1

Graduation year:1989 31.9 33.31990 33.7 32.71991 34.4 34.1

group, nearly one-third are from single-parent households or stepfamilies, andnearly one-third come from homes in whichthe parents have not attended school beyondthe twelfth grade.

MeasuresOf interest in the present analyses are

several demographic variables, two parent-ing indices that were used to construct thefamily types, and the four sets of outcomevariables.

Demographic variables.—Students pro-vided infonnation on their background andcurrent family situation. All respondents in-dicated their sex, year in school (used as aproxy for age), ethnic Identification (ulti-mately coded as African-American, Asian-American, Hispanic-American, non-His-panic white, or other), family structure (twonatural parents, single-parent, stepfamily,other), and the amount of education com-pleted by each parent residing with them.Parental education was coded as a two-leveivariable (less than college completion or col-lege completion and higher). In previous re-ports, we have shown that the relations be-tween parenting style and the aspects ofadolescent adjustment examined here do notvary as a function of child age, socioeco-nomic status, gender, or family structure(Lamborn et al., 1991; Steinberg, Lamborn,Dornbusch, & Darling, 1992; Steinberg et

al., 1991). We have, however, found thatsome ofthese relations vary as a function ofethnicity (Steinberg, Lamborn, Dornbusch,& Darling, 1992; Steinberg et al., 1991). Ac-cordingly, the interactive effects of parent-ing style and ethnicity are examined in thepresent report.

Parenting style.—The index of parent-ing style was developed to approximate thecategorical scheme suggested by Baumrind(1971) and Maccoby and Martin (1983). Thefirst year's questionnaires contained manyitems on parenting practices that were takenor adapted from existing measures (e.g.,Dornbusch, Carlsmith, Bushwall, Ritter,Liederman, Hastork, & Gross, 1985; Pat-terson & Stouthamer-Loeber, 1984; Rodgers,1966) or developed for the program of work.(In view of other research [e.g., Hethering-ton et al., 1992] indicating very high 1-yearstability coefficients for similar measures ofparenting, these questions were not re-peated at the 1-year follow-up.) Adolescentscompleted these measures vis-i-vis bothparents in two-parent households (in whichratings for mother and father were averaged)and vis-^-vis mothers in single-parenthomes. (Baumrind [1991] reports that thereis considerable convergence between moth-ers' and fathers' ratings.) Based on the previ-ous work of Steinberg etal. (1989), a numberof items were selected to correspond withseveral dimensions of parenting identifiedin earlier studies, and these items were sub-jected to exploratory factor analyses using anoblique rotation (we had no reason to as-sume that the dimensions are orthogonal).As in other studies of parenting practices(see Schaefer, 1965; Steinberg, 1990), threefactors emerged: acceptance/involvement,strictness/supervision, and psychologicalautonomy.

Our previous work has indicated thatthe psychological autonomy dimension ap-pears to be important in defining authorita-tiveness but less so in differentiating amongauthoritative, authoritarian, indulgent, andneglectful families. Accordingly, scores onthe acceptance/involvement and strictness/supervision dimensions were used in thepresent investigation to assign families toone of four groups, as outlined below. Theacceptance/involvement scale measures theextent to which the adolescent perceives hisor her parents as loving, responsive, and in-volved (sample items: "I can count on[them] to help me out if I have some kindof problem"; "When he wants me to dosomething, he explains why"; 10 items, a =


.72, M = .81, SD = .11, range = .25 to 1.0),The strictness/supervision scale assessesparental monitoring and supervision of theadolescent (sample items: "How much doyour parents try to know where you go atnight?"; "My parents know exactly where Iam most afternoons after school"; nineitems, a = .76, M = .lA, SD = .13, range =.30 to LO). In this sample, the dimensionsare modestly intercorrelated (r = .34, p <.001). For each ofthese scales, several oftheitems are in a true/false format, while othersare Likert scaled on a three-point scale;in the formation of &e composite indicesfor acceptance and strictness, items wereweighted to ^ just for differences in scaling.A complete listing of the items firom eachscale appears in Lambom et al. (1991).

Historically, researchers have appliedboth typological and dimensional ap-proaches to the study of scKiialization inthe family. As we have argued elsewhere(Darling & Steinberg, 1993), die twoapproaches have different theoretical orien-tations and are based on different assump-tions. In the typological tradition, the gen-eral pattem, organization, or climate ofparenting is of primary interest, and the as-sessment of specific parenting practices ordimensions (such as acceptance or strict-ness) is done for heuristic purposes, as ameans of providing a window on the overallparenting environment. In the dimensionaltradition, in contrast, different aispects oftheparent-child relationship are assessed in or-der to test specific hypotheses about theirrelation (separately and jointly) to child ad-justment. Each tradition has merit, and a de-cision to use one versus the other should bemade on tfieoretical grounds.

We decided to use a typological ap-proach in the present study—that is, to usethe acceptance and strictness scales to assignfamilies to categories, rather &an to treatthese dimensions as continuous variables—for two reasons. First, the decision reflectsour interest in examining the specific theo-retical framework set forth by Baumrind(1971) and later elaborated by Maccoby andMartin (1983). Baumrind's theory, which hashad considerable impact on tfee study of so-cialization for nearly 3 decades, is a theoryabout types, not abcmt specific parentingpractices. Second, the typological approachis more appropriate for charting the short-term developmental tsêctories of adoles-cents reared in different parenting environ-ments.

Four parenting categories were definedby trichotomizing the sample on acceptanceand on strictness and examining femilies'scores on Ae two variables simultaneously.Following Maccoby Mid Martin (1983), au-thoritative families (JV = 817) were thosewho scored in the upper tertlles on both ac-ceptance/involvement and strictness/super-vision, whereas neglectful families (iV =838) were in the lowest tertiles on berth vari-ables. Authoritarian families (N = 451)were in the lowest tertile on involvementbut in the highest tertile on strictness. Indul-gent &milies (N = 251) were in the h ig^s ttertile on involvement but in the lowest ter-tile on strictness. Families who scored in themiddle tertiie on either of the dimensionswere excluded from the analysis, in order toensure that the four groups of^&milies repre-sented distinct categories. This procedurewas followed in the analysis of the cross-sectional data (Lamborn et al., 1991) and is

d here for comparability.

The use of the tertile split procedure toassign families to the parenting groups,rather than assigning families on the basis ofpredetermined cutoflfe, results in a categori-zation of families that is ac&ziittedly sample-specific. While we can be confidît, for ex-ample, diat the &milies in our "indulgent"category are indeed relatively more Indul-gent (i.e., more accepting and less sMct)than the other families in the sample, we donot know whedser tiie fiunilies we have la-bded "indulgent" would be a>nsi<fered "in-dulgent" wi&in a dêrent population or ata different point in historic^ time. Titus, itis important to bear in mind tiiat the designa-tion of families as one type or another rela-tive to their counteiparts is done for heuris-tic, not diagnostic, purposes.

Because each year's survey was dividedinto two portions actministered on separatetesting days, there were some students inthe longitudinal sample who completed onlyone of the two parts of the survey in a givenyear. This occasionally presented a problemin the scoring of composite measures thatdrew on both survey parts. In general, wehandled instances of missing data conserva-tively, calcuiatii^ composite scores onlywhen respondents had answered 80% of tiienecessary items. As a result of this proce-dure, however, our Ns varied from analysisto analysis (d^tending on the variables ex-amined), altiious^ in no case were less tiian1,000 subjects included in any analysis. (Be-cause this is a longitudinal study in whichsubjects' prior scores are used as controls in


the analysis, this attrition is less problematicthan it otherwise might be.) Nonetheless,because absence from school was an Impor-tant cause of missing data, the sample onwhom the main data analyses were per-formed is in all likelihood relatively moreengaged in school (and, presumably, betterftinctioning in other respects) than were thestudent bodies in general. One importantimplication of this is that our estimates ofeffects are likely to be overly conservative:Because variability in our outcome measuresis probably constrained, the observed rela-tions between outcomes and predictors areattenuated.

Although the tertile split procedure per-mits us to be more confident that the fourcategories of families actually represent dif-ferent types (thereby strengthening thestudy's internal validity), the procedureeliminates from the analysis a large numberof families whose parenting is "average"(thereby weakening the study's extemal va-lidity). Thus, while our focus on the extremegroups in our sample may provide a clearertest ofthe theory, the approach we have cho-sen limits the generalizability of our find-ings. However, as Table 1 indicates, thesample of families scoring in the upper orlower tertiles on the parenting variables isdemographically comparable to the overallproject sample, suggesting that we have notselectively excluded any demographic sub-groups through the use of the tertile splitprocedure. Table 2 provides infonnation onthe sizes of each ofthe four parenting groupsas well as each group's mean and standarddeviations on the acceptance and strictnessscales.

Outcome variables.—As in the cross-sectional analyses, four sets of outcome vari-ables were examined: psychosocial develop-ment, academic competence, internalized

distress, and problem behavior. Table 3 pre-sents means and standard deviations on theoutcome variables for the entire sample (in-cluding adolescents who were not classifiedinto one ofthe four parenting groups). Table4 presents intercorrelations among the out-come variables. With the exception of gradepoint average, all outcome variables havebeen scaled on four-point Likert scales, with1 as low (e.g., "never," "strongly disagree,""not like me") and 4 as high (d.g., "fre-quently," "strongly agree," "very much likeme"). In the case of grade point average,scores were converted to the standard 4.0metric and could range from 0 (all F's) to 4.0(all A's). In general, the pattem of intercorre-lations (Table 4) supports our distinguishingamong the four sets of outcomes and be-tween the various indicators within each set,although there is some overlap between cer-tain aspects of psychosocial developmentand certain aspects of school competence.Correlations between the two measures ofinternalized distress and among the threemeasures of problem behavior are amongthe highest in the matrix.

The three indices oi psychosocial devel-opment include the social competence sub-scale of the Adolescent Self-Perception Pro-file (Harter, 1982) and two subscales fromGreenberger's Psychosocial Maturity Inven-tory—work orientation and self-reliance(Form D; Greenberger, Josselson, Knerr, &Knerr, 1974). The social competence mea-sure (a = .78) includes five items that askstudents whether they perceive themselvesas popular, as having many friends, and asmaking friends easily. The participants areasked to read two alternatives (e.g., "Someteenagers feel that they are socially ac-cepted, but other teenagers wish that morepeople their age would accept them") andchoose the one that is more like themselves.The work orientation (a = .73) and self-

TABLE 2

SAMPLE SIZES AND MEANS AND STANDARD DEVIATIONS ON PARENTAL INVOLVEMENT

AND STRICTNESS FOR THE FOUR GROUPS OF ADOLESCENTS

Involvement:XSD

Strictness:XSD

Total

2,353100

7914

7516

Authoritative

81734.7

.93

.03

.88

.05

Authoritarian

45119.2

.70

.07

.89

.05

Indulgent

25110.7

.92

.03

.62

.07

Neglectful

83835.4

.67

.08

.58

.09

760 ChUd Development

TABLE 3

MEAN SCORES, STANDABD DEVIATIONS, AND RANGES OFDEPENDENT VARIABLES AT TIME 1 AND TIME 2

Variable X SD Range

Time 1;Self-reliance 3.02 .53 1-4Work orientation 2.76 .47 1-4Social self-competence 2.99 .47 1-4Grade point average 2.85 .72 0-4School orientation 2.77 .56 1-4Academic self-competence 2.87 .48 1-4Psychological symptoms 2.47 .69 1-4Somatic symptoms 2.14 .54 1-4School misconduct 2.34 .67 1-4Alcohol and drug use 1.58 .70 1-4Delinquency 1.19 .36 1-4

Time 2:Self-reliance 2.99 .34 1-4Work orientation 2.72 .30 1-4Social self-competence 2.96 .41 1-4Grade point average 2.87 .65 0-4School orientation 2.73 .32 1-4Academic self-competence 2.92 .38 1-4Psychological symptoms 2.54 .65 1-4Somatic symptoms 2.20 .52 1-4School misconduct 2.27 .53 1-4Alcohol and drug use 1.55 .58 1-4Delinquency 1.18 .31 1-4

reliance (a = .81) subscales are each com-posed of 10 items. The work orientationscale measures the adolescent's pride in thesuccessful completion of tasks. A sampleitem, reverse coded, is "I find it hard to stickto anything that takes a long time." The self-reliance scale measures the adolescent'sfeelings of internal control and ability tomake decisions without extreme reliance onothers. A sample item, reverse coded, is"luck decides most things that happen tome."

The three measures of school achieve-ment include overall grade point average,the academic competence subscale of theAdolescent Self-Perception Profile (Harter,1982), and a scale developed for this projectthat assesses the adolescent's orientation to-ward school. Respondents provided infor-mation on their current grade point average,on a nine-point scale ranging from "mostlyA's" to "mostly F's"; scores were convertedto correspond to a standard 4.0 grading scale.Self-reported grades are h i ^ i y correlatedwith actual grades taken from official schoolrecords (Donovan & Jessor, 1985; Dom-busch et al., 1987). The academic compe-tence subscale (a = .73) includes five items

asking about the student's perceptions of hisor her intelligence in relation to classmates,ability to complete homework quickly, andcapability in classwork. The measure of ori-entation toward school was derived from aset of items that assesses the student's feel-ing of attachment to school (Wehl^e, Rut-ter, Smith, Lesko, & Fernandez, 1989). Ori-entation toward school is a six-item scale (a= .69) tiiat emerged from a factor analysis ofthe totd set of items. A sample item is "I feelsatisfied with school because I'm learning alot."

The set of three measures tapping prob-lem behavior includes reports of involve-ment in drug and alcohol use, school mis-conduct, and delinquency. The measure ofdrug and alcohol use taps the frequency ofinvolvement with cigarettes, alcohol, mari-juana, and other drugs (five items, a = .86)(Greenberger, Steinberg, & Vaux, 1981).The measure of school misconduct assessesthe frequency of such behaviors as cheating,copying homework, and tardiness (fouritems, a = .68) (Ruggiero, 1984). The mea-sure of delinquency assesses the frequencyof such behaviors as carrying a weapon,theft, and getting into trouble with the po-

in •v n en

» in in to in—I —1 -H —I O

X X o t~ n iO

' cq !>• ro in oI I I I I

en O C! O O

' in to M. O -^ -^

I I I

to — • *W CJ W

u5(p-^tD-^a: iQif ; -^Q-^3'3 'cq-^3-Heicncqinir5ioAMcq--w(N(Mrtrtoooooooo(M—<


lice (six items, a = .82) (Gold, 1970). Al-though self-reports of deviant behavior aresubject to both under- and over-reporting(see McCord, 1990), most researchers agreethat these provide a closer approximation ofyoungsters' true involvement in deviant ac-tivity than do "official" reports (e.g., policerecords), and the practice of using self-reportdata in the study of adolescent deviance iswidely established (see Gold, 1970; Jessor& Jessor, 1977; McCord, 1990).

Two measures of internalized distresswere derived from a 13-item version of theDepression Scale ofthe Genter for Epidemi-ologic Studies (GES-D; Radloff, 1977). Re-sults of a factor analysis suggested a somaticsymptoms scale (seven items, a = .67),which includes items concerning the fre-quency of headaches, stomachaches, colds,and so forth; and a psychological symptomsscale (six items, a = .88), which includesitems concerning the frequency of anxiety,tension, and depression.

Plan of AnalysesThere are a number of acceptable statis-

tical techniques that one may employ in theanalysis of short-term longitudinal data. Be-cause preliminary analyses of the presentdata indicated fairly consistent effects of re-gression to the mean, it was necessary to per-form analyses that took this effect into ac-count. One such procedure is analysis ofcovariance (ANGOVA) in which the depen-dent variable is the change in the outcomeof interest and the covariate is the relevantTime 1 score on the outcome (Laird, 1983).Accordingly, we conducted a series ofANGOVAs, each of which examined the ef-fects of parenting style, ethnicity, and thetwo-way interaction between parenting styleand ethnicity on changes in the outcomevariables of interest. For each analysis, therelevant Time 1 score was used as a covari-ate. In essence, the procedure allows forcomparisons of change in each outcomeacross tlie 1-year period for youth from dif-ferent types of families while taking into ac-count differences in adjustment that wereapparent at the first assessment, as well asthe effects of regression to the mean.

These analyses were followed wi& sim-ple pairwise post-hoc comparisons of thefour parenting style groups. The post-hocanalyses controlled for ethnicity. For vari-ables with significant parenting style x eth-nicity interaction effects, the analyses wererepeated separately for each ethnic group touncover ethnic differences in the effects of

parenting style on change in outcomesacross the 1-year period.

ResultsThe ANGOVAs indicated significant re-

lations between parenting style and changein most of tiie measures of adolescent devel-opment assessed across a 1-year period. Ta-ble 5 presents the Time 1 outcome scoresfor each parenting style group, the adjustedchange scores for each parenting stylegroup, and information on the significanceof the main effect of parenting style and thespecific parenting style group contrasts.

Parenting Style and PsychosocialDevelopment

Pattems of change in adolescents' self-reliance and work orientation, but not socialcompetence, varied significantly over the1-year period as a function of" parentingstyle. Examination of the adjusted changescores and post-hoc contrasts revealed that,whereas authoritatively reared adolescents'self-reliance scores improved slightly overthe 1-year period, tiiose of youth from indul-gent and authoritarian families essentiallyremained unchanged, while those of youthfrom neglectful homes generally declinedsomewhat. Similarly, whereas the work ori-entation scores of adolescents from neglect-ful and authoritarian homes declined overthe year, diose of adolescents from autiborita-tive or indulgent homes remained more orless stable.

The interaction of parenting style withethnicity was marginally significant (p < .10)in the prediction of self'-reliance. Follow-upanalyses indicated that in the case of self-reliance, it was relatively more advanta-geous to be raised by indulgent parentsamong Asian-American youdi than was thecase among youngsters from other ethnicgroups, and relatively more deleterious to beraised by authoritarian parents among Euro-pean-American youth.

Although not central to the focus of thispaper, analyses indicated a significant maineffect of ethnicity on social competence.FoIIow-up analyses revealed that African-American youth had more positive self-perceptions than did youngsters from eachofthe other three ethnic groups.

Parenting Style and Academic CompetenceThere were significant main effects of

parenting style on pattems of change in aca-demic self-conceptions and school orienta-tion, but not grade point average, over the 1

I

I I I I

I I

I I

•-< M op—I O p—Ir ; CO CO


year. Although academic self-concep-tions became more positive over time in allfour groups, the increase was significantlygreater among authoritatively and indul-gently reared youth than their counterpartsfrom neglectful homes (the adolescents fromauthoritarian homes fell between the two ex-tremes). And although all youngsters be-came somewhat more negative about schoolover time, the school orientation of neglect-fully reared youth declined relatively moresharply, especially in comparison to thatof youth from authoritative or authoritarianhomes.

A significant parenting style x ethnicityinteraction effect was detected in the predic-tion botii of grade point average and aca-demic self-conceptions, and a marginallysignificant interaction effect was detected inthe prediction of school orientation. In gen-eral, whereas authoritative parenting wasadvantageous, and neglectful parenting dis-advantageous, among European- and His-panic-American youth, authoritarian parent-ing was relatively more advantageous amongAsian-American youth and relatively moredisadvantageous among European-Ameri-can youngsters. Moreover, parenting stylewas unrelated to academic competenceamong African-American youth.

Although not central to this report, theanalyses also indicated main effects for eth-nicity on grade point average and academiccompetence, with grades of Asian-Americanyoudi increasing over the 1-year period andthose of African-American and Hispanic-American youngsters declining. The patternof results for academic s^lf-perceptions,however, showed that those of African-Americans increased more than did those ofother youngsters.

Parenting Style and Behavior ProblemsPattems of change over time in all three

areas of problem behavior also varied as afunction of parenting style, with significanteffects for delinquency and school miscon-duct, and a borderline effect for drug andalcohol use. Youth with authoritative or au-thoritarian families either became less in-volved in problem behavior or reported nochange across time, whereas rates of prob-lem behavior either increased or remainedmore or less constant among youth fromindulgent families and increased sharplyamong adolescents from neglectful homes.

The impact of parenting style on prob-lem behavior did not vary across ethnicgroups. There were significant ethnic differ-

ences in drug and alcohol use, however, andmarginally significant differences in delin-quency. Specifically, drug and alcohol useincreased more among European-Americanyouth than other youngsters, and schoolmisconduct decreased more among Asian-American youth than among their peers.

Parenting Style and Internalized DistressFinally, the results indicated that pat-

tems of change in reports of somatic distress,but not psychological symptoms, varied as afunction of parenting style. Somatic symp-toms increased most markedly among ado-lescents from audioritarian homes but de-creased over time among indulgently raisedyoungsters. There were not signiflcant two-way interactions between parenting styleand ethnicity in the prediction of internal-ized distress, nor were there main effects forethnicity in the prediction of psychologicalor somatic symptoms.

DiaeussionOur previous report on the concurrent

relation between parenting style and ado-lescent adjustment indicated clear advan-tages for adolescents raised in authoritativehomes, clear disadvantages for their peersreared in neglectful homes, and mixed out-comes for adolescents from authoritarianor indulgent households (Lambom et al.,1991). The results of this short-term follow-up indicate that many of the differences ob-served in the initial cross-sectional analyses^ e either maintained or increase over time.As a result, over the 1-year period studiedhere, the adjustment gap between adoles-cents from authoritative and neglectfulhomes widened.

Let us consider the pattems evinced byeach ofthe four groups relative to their start-ing points. In our cross-sectional report, wenoted that adolescents reared in authorita-tive families had advantages over otheryoungsters on measures of psychosocialcompetence, academic competence, inter-nalized distress, and problem behaviors.The longitudinal analyses indicate over-time stability in most of these domains,with only two exceptions: academic self-conceptions, which improved, and schoolmisconduct, which declined. Hence, itseems to be the case that the benefits of au-thoritative parenting during the high schoolyears are largely in the maintenance of pre-vious levels of high adjustment, rather thanin the continued development of compe-tence.


A similar pattern of stability is seenamong youngsters from authoritarian homes,with one important exception. While the dis-advantages to adolescent self-confidence as-sociated with authoritarian child rearingnoted in our earlier report are maintainedover time, youth from authoritarian homesshowed significant increases in internalizeddistress over the 1-year period. One hypoth-esis is that the increased levels of psycholog-ical and somatic distress reported by theseyoungsters is tied somehow to their contin-ued exposure to a home environment that ispsychologically overpowering and increas-ingly developmentally inappropriate. Al-though the terminology differs, our charac-terization of the family environment ofdepressed or anxious adolescents as authori-tarian is consistent with pictures of "over-controlling" parents that emerge from clini-cal investigations (e.g.. Stark, Humphrey,Gook, & Lewis, 1990).

The adolescents from indulgent homescontinued to display a psychological andbehavioral profile that is mixed. In ourcross-sectional report, we described theseyoungsters as well adjusted but "especiallyoriented toward their peers and toward thesocial activities valued by adolescents"(Lambom et al , 1991, p. 1062). One yearlater, this description is even more apt:indulgently reared youth became morepositive over time in their academicself-conceptions and reported less somaticdistress, but at the same time they showedsignificant declines in school orientation andsignificant increases in school misconduct

It is in the case of neglectfully rearedadolescents, however, where we see theclearest evidence of the impact of parentingon adjustment during the high school years.These youth, already at a psychological andbehavioral disadvantage at the time of firstassessment, showed continued declinesover the 1-year period, with sharp drops inwork orientation and school orientation,and sizable increases in delinquency and al-cohol and drug use. The overall pattem sug-gests a group of youngsters on a downwardand troublesome trajectory characterized byacademic disengagement and problem be-havior.

This study followed adolescents overthe course of 1 year—perhaps not a longenough period of time during which to ob-serve marked changes in the sorts of out-comes assessed in this study. We think it im-portant, therefore, to consider the magnitude

ofthe changes we have observed by examin-ing each group's adjusted change scores inrelation to the relevant variable's standarddeviation. In doing so, we see that many ofthe observed changes are not trivial. In thecase of school orientation and work orienta-tion, for instance, the scores of neglectfullyreared youngsters declined by about one-third of a standard deviation over just 1 year,while their delinquency and substance useincreased by about one-fourth of a standarddeviation (see Figs. 1 and 2). Youngstersfrom authoritarian homes reported an in-crease in internalized distress of approxi-mately one-fourth of a standard deviationover the year (see Fig. 3). And on the posi-tive side, the academic self-conceptionsof authoritatively and indulgently rearedyoungsters increased by approximately one-third of a standard deviation in 1 year (seeFig. 4). If changes of similar magnitude wereto occur each year over the course of adoles-cence, the end result would indeed be note-worthy. Further longitudinal research overa longer time frame will help provide clearerpictures of the different developmental tra-jectories of youngsters who have been ex-posed to different parenting styles.

Our ability to demonstrate that the over-time impact of parenting style on adolescentadjustment holds even after controlling forinitial group differences is important for sev-eral reasons. First, in the absence of a ran-domized experimental design, the analyticstrategy employed here provides at least in-direct evidence that authoritative parentingactually precedes—rather than simply ac-companies or even follows from—ado-lescent adjustment. (Because one can notrandomly assign adolescents to differenthome environments, such indirect evidenceis important.) This is not to say that our anal-yses mle out the possibility that the reversecausal process is operating simultaneously(i.e., that well-adjusted adolescents alsoprovoke authoritativeness in their parents);given past research on the bidirectionalnature of socialization (e.g.. Bell, 1968), italmost certainly is. (Unfortunately, the ab-sence of comparable parenting assessmentsin the second year of the investigation pre-cluded our examining the impact of adoles-cents on their parents.) What the present re-sults provide, however, is evidence that thecorrelation between adolescent adjustmentand parenting style is not solely due to theeffect that children have on their parents. Atleast some of it is due to the impact that par-ents have on their children.


Score

3.3

3.2

3.1

3

2.92.8-|2.72.62.52.42.32.22.1

2

AutHortMlve

AuthotitBrtan

FIG. 1.—Changes in adolescent school orientation as a function of parenting style

A second advantage of our analyticstrategy concems our ability to respond tocriticisms that the observed correlationsbetween parenting style and adolescenta(^ustment are due simply to commonsource or method variance, problems that in-herently threaten the internal validity of re-search designs that rely on self-reports fromone respondent. The source and methodvariance shared in common between adoles-cents' characterizations of their parents inthe first year of the study and their self-

reports of adjustment 1 year later is alsoshared in common between their parentalchar£K;terizations in the first year and theirself-reports of adjustment assessed concur-rently. (Indeed, given the common time ofmeasurement, the shared source and methodvariance is likely to be even more substan-tial for concurrent mes^ures.) Due to thehigh 1-year stability coefficients for the ad-justment measures (most are greater than.60), controlling for subjects' initial swijust-ment self-reports has the effect of partialing

Authorttafive

Autftorttariwi

Indirigant

Nsgtocthil

Soore

1.56

1.5

1.46

1.4

1.36

1.3 i

1.25 i

1.2

1.16

1.1

1.06

1

1987-88 1988-89FIG. 2.—Changes in adolescent delinquency as a function of parenting style


Somatic Symptoms2.46 1 — - Aufliorltatlve

Authoritari€Ui2.4

Indulgent

^ Neglectful

2.3

2.26

2.2

2.16

2.1

2.06

21987-88 1988-89

FIG. 3.—Changes in adolescent somatic symptoms as a function of parenting style

out much ofthe common method and sourcevariance.

Finally, the covariance analyses em-ployed here help rule out many potentialthird-variable explanations of the over-timeassociation between parenting style and ad-olescent adjustment. Any such altemativeaccount would necessarily have to posit theexistence of a confounding variable that iscorrelated with the measures of adjustmentin the study's second year but that was notcorrelated with the same measures taken just

1 year earlier. As a test of this possibilit>', wecompared the correlation between parentaleducation (a likely confounding variable)and the Time 1 versus Time 2 measure ofeach of the outcome variables used in thestudy. In every case, the correlation coeffi-cient was virtually identical at Time 1 andat Time 2. While it is of course theoreticallypossible that there exists some other, un-measured, third variable that is differentiallycorrelated with the outcome measures atTime 1 and Time 2, it seems more parsimo-nious—and far more plausible, given other

Score

3.53.43.33.23.1

32.92.82.72.82.6-12.42.32.22.1

2

Academic CompetenceAuthoritative

Authoritahan

Indulgent

Nefllectful

1987-88 1988-89

FIG. 4.—Changes in adolescent academic competence as a function of parenting style


studies in the socialization literature—to ac-cept the observed association between par-enting practices and adolescent adjustmentas internally valid.

Despite these strengths, the study doeshave several weaknesses. First, as discussedearlier, our decision to focus on the extremequadrants in the sample in order to ensurethat we were contrasting different parentingtypes limits somewhat the external validityof our findings. Although our approach pro-vides a reasonable test of Maccoby and Mar-tin's (1983) framework, we cannot be certainthat contrasts based on a less extreme cate-gorization scheme would yield the sameresults. It is important to bear in mind,however, that the extreme groups studiedhere—and, most important, the "neglect-ful" group—are drawn from a representa-tive community sample and not obtainedthrough clinics or similar nonrepresentativesources.

A second weakness is that all ofthe dataderive from youngsters' self-reports. Whilethe nature ofthe analyses does argue againstexplanations based on common source andmethod variance (as discussed above), ourfindings can be interpreted only to show thatadolescents' adjustment is related to the wayin which they suh(jectively experience theirparents. Although we recognize that young-sters' reports of their parents' behavior maybe colored by a variety of factors, we do notbelieve that "objective" assessments of fam-ily processes (derived in most studies by ob-serving families in contrived and unfamiliarsituations in university laboratories) or pa-rental reports provide an inherently superiormeans of assessing family relationships. In-deed, the few studies that have correlated"objective" assessments of family life withboth adolescents' reports of their parents'behavior and with their parents' reports sug-gest that adolescents, not parents, are moreaccurate (e.g., Schwartz, Barton-Henry, &Pruzinsky, 1985). Perhaps more important,our willingness to use adolescents' reportspermits us to study a more representativesample than would be the case if parents'participation in the study were required.Nevertheless, we recognize that it also is im-portant to investigate the relation betweenparenting and adolescent adjustment usingmultiple methods and different sources ofinformation.

Although not a focus of this report, wefind that patterns of change in adolescent ad-justment vary by ethnicity. Most ofthe find-

ings in this regard are consistent with otherreports of ethnic differences in adolescentdevelopment and behavior, specifically, (1)Asian-American youth report the great-est improvement in school performance,whereas the academic perfonnance of Afri-can-American and Hispanic-American youthshowed the greatest decline (see Steinberg,Dornbusch, & Brown, 1992), (2) African-American youth report the most positivechanges in self-perceptions (see Spencer &Dornbusch, 19^), and (3) European-American youth report the greatest increasesin drug and alcohol use (see Gans, 1990).

More interestingly, several ofthe effectsof parenting style appear to be moderatedby the adolescent's ethnicity. Generallyspeaking, the links between authoritativeparenting and both psychosocial and aca-demic competence appear to be strongestamong European-American youth. In lightofthe fact that authoritativeness is a charac-teristically Westem and middle-class ap-proach to child rearing, the fact that it mayhave greater advantages in these groups isnot surprising, since the style may be moreconstMiant with other elements of family life.In addition, in the contemporary UnitedStates, au&oriiativeness is more prevalentamong European-American and middle-class families, and authoritatively rearedyoui^sters from these backgrounds may bemore likely to have these practices echoed(and, presumably, ampli:l^d) in their neigh-borhoods and social networks than are au-thoritatively reared youngsters from otherethnic or class backgrounds (Steinberg &Darling, 1994).

It is important to note also that the puta-tive deleterious consequences of parentalauthoritarianism are evidently not as severeamong minority youth as among dieir Euro-pean-American counterparts. We note thisfinding in particular because the suggestionis not new that minority youngsters—especially those from economically disad-vantaged back^ounds—may benefit from arelatively more authoritari£ui style of parent-ing (see ^ d w i n &: Baldwin, 1^9; Baum-rind, 1972; Dombusch et al., 1987; but seeClark, 1983, for a contrasting view). One hy-pothesis is that audioritarianism may bemore beneficial unong &mi!ies whose livingcircumstances warrant stricter, more vigilantcontrol. Another, equally tenable hypotiiesisis that the meaning of authtwitarianism—orany parenting style, for that matter—is mod-erated by the cultural context in which it oc-curs and, consequently, interpreted differ-


entially by children from different ethnic orsocioeconomic groups. Accordingly, whatmay be experienced by the adolescent as pa-rental intrusiveness in some cultural groupsmay be experienced as parental concern inothers (see Darling & Steinberg, 1993).

It is important to bear in mind, however,that in some domains—most notably, prob-lem behavior and internalized distress—authoritativeness appears to confer compara-ble benefits across varied ecological niches(see also Steinberg et al., 1991). Moreover,we are hard pressed to find any groupswithin which authoritativeness has deleteri-ous consequences for the adolescent, or anygroups in which neglectful or disengagedparenting is desirable. On the whole, then,it appears that the combination of parentalaloofness and disciplinary laxity appearsuniversally harmful to adolescents. Con-versely, parental authoritativeness—thecombination of responsiveness and de-mandingness—carries many benefits andfew disadvantages for adolescents from dif-ferent walks of life.

References

Baldwin, C, & Baldwin, A. (1989, April). The roleof family interaction in the prediction of ado-lescent competence. Symposium presented atthe meeting of the Society for Research inChild Development, Kansas City, MO.

Baumrind, D. (1967). Child care practices ante-ceding three patterns of preschool behavior.Genetic Psychology Monographs, 75, 43-88.

Baumrind, D. (1971). Current patterns of parentalauthority. Developm,ental Psychology Mono-graph, 4(1), part 2.

Baumrind, D. (1972). An exploratory stndy of so-cialization effects on black children: Someblack-white comparisons. Child Develop-ment, 43, 261-267.

Baumrind, D. (1991), The influence of parentingstyle on adolescent competence and sub-stance use. Journal of Early Adolescence, 11,56-95.

Bell, R. (1968). A reinterpretation of the directionof effects in studies of socialization. Psycho-logical Review, 75, 81-95.

Clark, R. (1983). Family life and school achieve-ment: Why poor black children succeed orfail. Chicago: University of Chicago Press.

Darling, N., & Steinberg, L. (1993). Parentingstyle as context: An integrative model, Psy-chological Bulletin, 113, 487-496.

Donovan, J., & Jessor, R. (1985). Structure of prob-lem behavior in adolescence and young adult-hood./oumo/ of Consulting and Clinical Psy-chology, 53, 890-904.

Dombusch, S., Carlsmith, J., Bushwall, S., Ritter,P., Leiderman, P., Hastorf, A., & Gross, R.(1985). Single parents, extended households,and the control of adolescents. Child Devel-opment, 56, 326-341.

Dombusch, S., Ritter, P., Liederman, P., Roberts,D., & Fraleigh, M. (1987). The relation of par-enting style to adolescent school perfor-mance. Child Development, 58, 1244-1257.

Gans, J. (1990). America's adolescents: Howhealthy are they? Chicago: American Medi-cal Association.

Cold, M. (1970). Delinquent behaviorin anAmer-ican city. Belmont, CA: Brooks/Cole.

Creenberger, E., Josselson, R., Knerr, C , & Knerr,B. (1974). The measurement and structure ofpsychosocial maturity. Journal of Youth andAdolescence, 4, 127-143.

Greenberger, E., Steinberg, L., & Vaux, A. (1981).Adolescents who work: Health and behav-ioral consequences of job stress. Develop-mental Psychology, 17, 691-703.

Harter, S. (1982). The Perceived CompetenceScale for Children. Child Development, 53,87-97.

Hetherington, E. M., Clingempeel, W., Anderson,E., Deal, J., Hagan, M., Hollier, E., & Lind-ner, M. (1992). Coping with marital transi-tions: A family systems perspective. Mono-graphs of the Society for Research in ChildDevelopment, 57(2-3, Serial No. 227).

Jessor, R., & Jessor, S. (1977). Problem behaviorand psychosocial development: A longitudi-nal study of youth. New York: AcademicPress.

Laird, N. (1983). Further comparative analyses ofpretest-posttest research designs. AmericanStatistician, 37, 329-330.

Lambom, S., Mounts, N., Steinberg, L., & Dom-busch, S. (1991). Patterns of competence andadjustment among adolescents from authori-tative, authoritarian, indulgent, and neglect-ful homes. Child Development, 62, 1049-1065.

Maccoby, E., & Martin, J. (1983). Socialization inthe context of the family: Parent-child interac-tion. In E. M. Hetherington (Ed.), P. H. Mus-sen (Series Ed.), Handbook of child psychol-ogy: Vol. 4. Socialization, personality, andsocial development (pp. 1-101). New York:Wiley.

McCord, J. (1990). Problem behaviors. In S. Feld-man & G. Elliot (Eds.), A* the threshold: Thedeveloping adolescent (pp. 414-430). Cam-bridge, MA: Harvard University Press.

Patterson, G., & Stouthamer-Loeber, M. (1984).The correlation of family management prac-tices and delinquency. Child Development,55, 1299-1307.

Radloff, L. S. (1977). The CES-D Scale: A self-


report depression scale for research in thegeneral population. Applied PsychologicalMeasurement, 1, 385-401.

Rodgers, R. H. (1966). Cornell Parent BehaviorDescription—an interim report. Unpub-lished manuscript. Department of HumanDevelopment and Family Studies, ComellUniversity, I&aca, NY.

Ruggiero, M. (1984). Work as an impetus to delin-quency: An examination of theoretical andempirical connections. Unpublished doctoraldissertation. University of Caiifomia, Irvine.

Schaefer, E. (1965). Children's reports of parentalbehavior: An Inventory. Child Development,36, 413-424.

Schwartz, J., Barton-Henry, M., & Pruzinsky, T.(1985). Assessing child-rearing behaviors: Acomparison of ratings made by mother, father,child, and sibling on the CRPBI. ChildDevelopment, 56,462-479.

Spencer, M., & Dombusch, S. (1990). Challengesin studying minority youth. In S. Feidman &G. Elliot (Eds.), At the threshold: The devel-oping adolescent (pp. 123-146). Cambridge,MA: Harvard University Press.

Stark, K., Humphrey, L., Cook, K., & Lewis, K.(1990). Perceived family environments of de-pressed and anxious children: Child and ma-temai figures' perspectives. Joumci of Abnor-mal Child Psychology, 18, 527-547.

Steinberg, L. (1990). Interdependency in the fam-ily: Autonomy, conflict, and harmony. In S.Feidman & G. Elliot (Eds.), A( the threshold:The developing adolescent (pp. 255-276).Cambridge, MA: Harvard University Press.

Steinberg, L., & Darling, N. (1994). The broadercontext of adolescent development. In R.Silbereisen & E. Todt (Eds.), Adolescence incontext (pp. 25-45). New Yoric: Springer-Verlag.

Steinberg, L., Dombusch, S., & Brown, B. (1992).Ethnic differences in adolescent achieve-ment: An ecolt^cal perspective. AmericanPsychologist. 47, 723-729.

Steinberg, L., Elmen, J., & Mounts, N. (1989). Au-thoritative parenting, psychosociai maturity,and academic success among adolescents.Child Development, 60, 1424-1436.

Steinberg, L., L^nbom, S., Dombusch, S., & Dar-ling, N. (19^). Impact of parenting practiceson adolescent achievement: Auflioritativeparenting, school involvement, and encour-agement to succeed. Child Development, 63,1266-1281.

Steinberg, L., Mounts, N., Lambom, S., fie Dom-busch, S. (1^1). Authoritative parenting andadolescent adjustment across various ecologi-cal niches. Journal of Research on Adoles-cence, 1, 19-36.

Wehiage, G., Rutter, R., Smith, G., Lesko, N., &Femandez, R. (1989). Reducing the risk:Schools as communities of support. London:Falmer.

Weinberger, D., Tublin, S., Ford, M., & Feidman,S. (1990)̂ Preadolescents' social-emotionala<^ustment and selective attrition in familyresearch. Child Development, 61,1374-1386.

over-time changes in adjustment and competence...

Documents