Changes in Parenting Practices and Adolescent Drug Abuse ... · Changes in Parenting Practices and Adolescent Drug Abuse During Multidimensional Family Therapy Susan E. Schmidt, Howard

Journal of Family Psychology1996, Vol. 10, No. 1, 12-27

Copyright 1996 by the American Psychological Association, Inc.0893-3200/96/S3.00

Changes in Parenting Practices and Adolescent Drug AbuseDuring Multidimensional Family Therapy

Susan E. Schmidt, Howard A. Liddle, and Gayle A. DakofTemple University

The nature and extent of changes in parenting and the link between parental subsystemchanges and reduction in adolescent substance abuse and problem behaviors wereexamined in a sample of 29 parents and their drug-abusing adolescents. Participantscompleted 16 sessions of multidimensional family therapy. Over two thirds of theparents showed moderate to excellent improvement in parenting. Chi-square goodness-of-fit analyses revealed a statistically significant association between improvement inparenting and reduction in adolescent drug use and behavior problems. Results of thisexploratory study provide qualified support for a fundamental tenet of family thera-py—that change in the parental subsystem is related to improvement in the problembehavior of adolescents.

Family therapy is a credible and effectivetreatment for a variety of child and adolescentproblems (Alexander, Holtzworth-Munroe, &Jameson, 1994; Hazelrigg, Cooper, & Borduin,1987). Scientific work carried out during thelast decade demonstrates the efficacy of certainforms of family therapy with adolescent behav-ior problems (Henggeler, Borduin, & Mann,1993; Lebow & Gurman, 1995; Tolan & Loe-ber, 1993), including substance abuse (Liddle &Dakof, 1995a), a disorder known to be amongthe most difficult to treat (Lambert, 1982). Par-ticular forms of family-based intervention canretain adolescents and their families in treat-ment; can significantly reduce drug use in

Susan E. Schmidt, Howard A. Liddle, and Gayle A.Dakof, Center for Research on Adolescent DrugAbuse, Temple University.

Conduct of this research was supported by Na-tional Institute on Drug Abuse (NIDA) GrantsR01DA3714 and P50DA07697, by an AAMFTFoundation Graduate Student Research Grant, and bya National Research Service Award from NIDA(DA05545). We are grateful to Ruth Palmer, KathyVila, and Jodi Johnson for their work on this project;to Frances Sessa for comments on an earlier versionof this article; and to Lawrence Steinberg for com-ments on this project at its early stage of develop-ment.

Correspondence concerning this article should beaddressed to Howard A. Liddle, Center for Researchon Adolescent Drug Abuse, Temple University, TU265-66, Philadelphia, Pennsylvania 19122.

youth; can demonstrate in-session changes ofparent-adolescent conflict; and, in comparativecontrolled trials, can demonstrate greater effec-tiveness than peer group therapy, individualcounseling, and family-based educational pro-grams in eliminating or reducing drug use (Di-amond & Liddle, in press; Henggeler et al.,1991; Joanning, Quinn, Thomas, & Mullen,1992; Lewis, Piercy, Sprenkle, & Trepper,1990; Liddle & Dakof, 1995a; Szapocznik,Kurtines, Foote, Perez-Vidal, & Hervis, 1986).

Multidimensional family therapy (MDFT) isone of the promising, new, empirically based,multicomponent interventions developed for thetreatment of adolescent substance abuse(Lebow & Gurman, 1995; National Institute onDrug Abuse, 1995; Selekman & Todd, 1990;Shalala, 1995). In a controlled clinical trial inwhich adolescent drug users were randomly as-signed to one of three treatments (MDFT, ado-lescent group therapy, or multifamily educa-tional intervention), the general pattern ofresults indicated the greatest and most consis-tent improvement among those who receivedMDFT (Liddle & Dakof, 1995b; Liddle, Dakofet al., 1995). The results indicated that MDFTwas significantly more effective than the othertwo treatments in reducing drug use at termina-tion. This reduced level of drug use was main-tained 1 year later. From pretreatment to follow-up, adolescents who received MDFT alsoshowed more improvement in school gradesthan those who received either of the other two

12

SPECIAL SECTION: CHANGES IN PARENTING PRACTICES 13

treatments. Adolescents who received MDFTimproved their grades from below average topassing (C average) in just over 1 year.

MDFT has four major areas of intervention(Liddle, 1991), and each of these areas hasseveral dimensions of focus. The approach tar-gets (a) the adolescent's intrapersonal and in-terpersonal (i.e., vis-a-vis parents and peers)functioning, (b) the parent's intrapersonal andinterpersonal functioning (e.g., parenting prac-tices and functioning as an adult apart from theparenting role), (c) parent-adolescent interac-tions as observed in sessions and reported byparent and adolescent, and (d) family members'interactions with extrafamilial sources of influ-ence (e.g., school and child welfare personnel,probation officers; Liddle, Becker et al., 1995).Specific intervention modules guide assessmentand intervention within each of these sub-systems (e.g., Liddle et al., 1992) and coordi-nate interventions in one domain with interven-tions in related domains (Liddle, 1995).Currently under refinement and pilot testing inan experimental context (Liddle, 1994), thesemodules are informed by developmental re-search about adolescents, parents, and families(Liddle, Schmidt, Ettinger, & Sessa, in press),and they involve multiple components. This ap-proach is in harmony with the kind of interven-tion packages now recommended for problemssuch as adolescent drug abuse and conduct dis-order (Kazdin, 1987; Newcomb & Bender,1989; Segal, 1986). Individual symptoms areunderstood in the context of the other problembehaviors that accompany them (Newcomb &Felix-Ortiz, 1992).

An underlying assumption of MDFT, and allfamily-based interventions, is that change in anindividual (i.e., decrease in symptoms and in-crease in prosocial functioning) results fromchange in the family system. Parenting has re-ceived an enormous amount of attention fordecades from intervention scientists and basicresearchers (Bornstein, 1995). As a key ingre-dient in child socialization, parenting is a fun-damental aspect of the family system (Maccoby& Martin, 1983). Research reaffirms the impor-tance of parenting as a critical facilitator ofdevelopment throughout the second decade oflife (Steinberg, 1990). Certain parenting prac-tices serve as a buffer against risk factors knownto be associated with dysfunction (Hawkins,Catalano, & Miller, 1992). For example, parent-ing practices are linked to peer group affilia-

tions: "If parents model deviant behavior or failto maintain close relationships with their teen-ager, the child is more likely to drift into deviantpeer crowds and, as a consequence, be moreinvolved in drug use or delinquency" (Brown,Mounts, Lamborn, & Steinberg, 1993, p. 469).Parenting is multifaceted and encompasses awide spectrum of features, strategies, and meth-ods reflecting behavioral, affective, and cogni-tive domains of functioning (Abidin, 1992; Bel-sky, 1984; Dix, 1991; Goodnow, 1988; Sigel,McGillicuddy-DeLisi, & Goodnow, 1992).

Previous research on both adolescent sub-stance abuse and conduct disorder suggests thatthe initiation and continuation of these disordersare associated with family processes generallyand with certain parenting practices in particular(Block, Block, & Keyes, 1988; Farrington et al.,1990; Steinberg, Fletcher, & Darling, 1994).Poor family management, disrupted or omittedparenting, inappropriate discipline, inadequateparent monitoring, parent irritability, and coer-cive family processes (Patterson, 1982) charac-terize the family environments of adolescentswith conduct disorders and substance abuse(e.g., Baumrind, 1991; Hawkins, Catalano, &Miller, 1992; Maccoby & Martin, 1983).

Although a considerable body of work hasestablished the important role played by familyprocesses and parenting in the etiology of ado-lescent substance abuse and although efficacyevidence exists for family therapy with adoles-cent substance abuse (see reviews by Henggeleret al., 1993; Liddle & Dakof, 1995a, 1995b),only a few studies have explored the link be-tween symptomatic change in an adolescent andcorresponding change in the family environ-ment (Barrett, Simpson & Lehman, 1988;Chamberlain, 1990; Dadds, Schwartz, & Sand-ers, 1987; Mann, Borduin, Henggeler, &Blaske, 1990; Szapocznik et al., 1989). A crit-ical step in developing effective treatments in-volves determining whether or not the selectedinterventions actually affect those target pro-cesses that are established or hypothesized to berelated to the dysfunction the intervention istrying to change (Kazdin, 1994). Szapocznik etal.'s (1989) study revealed that family therapycould not only improve the targeted child be-haviors, but also prevent deterioration in thefamily environment over time—a finding notevidenced in the comparison treatment condi-tion (individual psychodynamic child therapy).Changes in parental resistance directly predict

14 SCHMIDT, LIDDLE, AND DAKOF

long-term (i.e., 2 years) outcomes of treatment(Chamberlain, 1990), and a strong relationshipexists between changes in parental resistanceand improvements in parental disciplinary prac-tices (Stoolmiller, Duncan, Bank, & Patterson,1993). Given the role of family environments inbuffering the child from risk factors and thecentral contribution of parenting in this regard,changes in parenting created by family interven-tions that persist over time are noteworthy.

The present exploratory study addressedthese complex issues by focusing on the parentsof adolescents and, specifically, on the qualityand malleability of parenting and their associa-tion with adolescent substance abuse and be-havior problems. Although this study was notan experimental test of whether MDFT couldchange parenting, it suggested possible mecha-nisms that are responsible for the measureddecline in drug use and the improvement inschool performance among adolescents who re-ceived MDFT. The present study examined (a)the nature and extent of change in the behav-ioral, affective, and cognitive features of parent-ing observed in families treated with MDFT and(b) the link between parental subsystemchanges, or lack thereof, and reduction in ado-lescent substance abuse and behavior problems.

Method

Client Characteristics

Families who participated in this study were partof a controlled clinical trial that compared the effi-cacy of three treatments—MDFT, adolescent grouptherapy, and multifamily educational interven-tion—in reducing drug abuse and behavior problemsin adolescents (Liddle & Dakof, 1995b; Liddle,Dakof et al., 1995). Each treatment lasted between 14to 16 sessions and spanned a maximum of 6 months.The current study focused on those adolescents andtheir families who completed a course of MDFT.

The sample included 29 families (out of 33 in theMDFT treatment condition). Four families were ex-cluded from the present analysis because a completeset of data (i.e., videotapes of family therapy ses-sions) was not available. The mean age of the ado-lescents who completed treatment was 16 years(SD = 1.29); 72% were male, and 55% identifiedthemselves as European American (45% were ethnicminorities, primarily African American and Hispan-ic). Twenty-one percent of the adolescents came fromhouseholds with two parents, and 79% came fromother family configurations (single parents, remarriedparents, parents with live-in partners, and other

guardians). Annual family incomes were categorizedas less than $30,000 (50%), $30,000-$50,000 (32%),and $50,000 or more (18%). Eleven percent of themothers did not graduate from high school; 75%completed high school, and 14% completed collegeor beyond.

Measures

Parenting. Parenting was assessed with an obser-vational measure we developed. It consisted of eightcategories reflecting the diversity and complexity ofparenting practices and features found to be related tothe development of problem behaviors in childrenand adolescents. A modified grounded theory (Glaser& Strauss, 1967) approach was adopted to constructthe new measure. Both a priori theoretical and em-pirical work as well as traditional grounded theorymethods were used to develop the parenting codes.Hence, categories were developed from the litera-tures on parenting, attachment, parenting style, andparent social cognitions and from grounded observa-tions of parents' comments and expressed attitudes asobserved on videotaped sessions of MDFT. Thesecodes were generated from clients who were not partof the current study but who shared similar demo-graphic and symptom configurations as study fami-lies. Potential coding categories were generated andrefined by additional observations until the point ofredundancy of categories was reached. This modifiedgrounded theory approach has been used successfullyin other studies (e.g. Dakof & Mendelsohn, 1989;Spitzer, Webster-Stratton, & Hollinsworth, 1991).

Final coding categories for categorizing parents'comments and behaviors observed in videotapedtherapy sessions included the following: (a) powerassertive discipline (e.g., endorsement or expressionof physical or verbal aggression or deprivation ofprivileges); (b) positive discipline and communica-tion (e.g., endorsement or expression of verbal rea-soning, sharing of values, or behavior modificationmethods); (c) positive monitoring and limit setting(e.g., statements about, or evidence of, success inefforts to monitor and set limits); (d) negative mon-itoring and limit setting (e.g., statements about, orevidence of, difficulty or reluctance to monitor andset limits); (e) interparent inconsistency (e.g., state-ments about, or in-session evidence of, mother-fatherconflict in endorsement and application of philoso-phy, methods, or values); (f) negative affect anddisengagement (e.g., statements about, or in-sessiondisplay of, anger, depression, lack of energy, or par-tial or full abdication of parental duties); (g) positiveaffect and commitment (e.g., statements about, or in-session display of, parent warmth, optimism, love,delight in relationship with teen, and commitment tohelping teen); and (h) cognitive inflexibility (e.g.,statements reflecting inappropriate or rigid notions of


autonomy, independence and responsibility, blamingof others, or scapegoating). Interrater reliability, asassessed by the intraclass correlation coefficient(Shrout & Fleiss, 1979), ranged from a low of .62(cognitive inflexibility) to a high of .87 (negativemonitoring and limit setting). The average intraclasscoefficient was .71, indicating overall satisfactoryreliability.

Transcripts of videotaped therapy sessions fromthe beginning phase of therapy (first three sessions)and final phase of therapy (last three sessions) werecoded. Transcripts from the two phases of therapywere coded in separate coding sessions (i.e., codersdid not code all categories in the same coding ses-sion), and mothers and fathers from the same familywere coded on separate occasions. Coders did nothave any specific information about family demo-graphics beyond that expressed or evident in thetranscripts. To facilitate coding, each transcript in thestudy was divided into units of two sentences, threelines, or a rational break in the conversation. Tran-scripts included actual verbalizations and descrip-tions of salient, nonverbal behavior (e.g. weeping,hugging, shoving). The total number of communica-tion units was calculated for each parent for eachphase of therapy. Each unit could receive no code atall or could receive a code on more than one cate-gory. Multiple codes, which reflected the complexnature of actual parent communications, were neces-sary for 30% of the coding units.

A parent's score on a given category was a pro-portion determined by the number of units coded ona particular scale divided by the total number of units.For example, a parent with 5 of 100 units coded aspositive discipline would receive a score of .05. Theuse of proportion scores made it possible to comparescores of parents who differed considerably in theirtotal number of communication units per phase oftherapy (because of individual differences).

The construct validity of the coding instrument wasassessed in a sample of 10 parents and adolescentswho received MDFT at the research clinic of theCenter for Research on Adolescent Drug Abuse,Temple University. Initial therapy sessions were tran-scribed and coded according to the parenting catego-ries described above. Scales from two sets of parentself-report questionnaires that appeared to be relatedto the coding instrument were used in its validation:(a) parental acceptance, psychological control, andbehavioral control were measured by the ParentingQuestionnaire (adapted from Steinberg, 1990), a 35-item checklist, and (b) parent affect was measured bythe Positive Symptom Total (PST) and by the De-pression and Hostility scales, which were all derivedfrom the Symptom Checklist-90 (SCL-90; Deroga-tis, 1977).

Using a clinical rating method with an interob-server agreement of 85% (Schmidt, 1994), parentingstyle was rated as authoritative, authoritarian, or con-

flicted-disengaged (i.e., nonoptimal styles such asindulgent, uninvolved, and confused). The extent ofagreement among these ratings of parenting stylewith the three scales of the Parenting Questionnairewas 80%. Thus, the self-report questionnaire methodand the observational method developed for use inthe present study converged in identifying parentswith optimal and nonoptimal styles.

Using the SCL-90, we found significant positivecorrelations between negative affect and disengage-ment and both PST (r = .85; p < .01) and thecombined Depression and Hostility scales (r = .87,p < .01). We found significant negative correlationsbetween positive affect and commitment and both thePST (r = - .80, p < .01) and the combined Depres-sion and Hostility scales (r = —J5;p< .05). Parentswith high levels of negative affect and disengage-ment reported emotional problems, depression, andhostility, whereas parents with high levels of positiveaffect and commitment reported low levels of emo-tional problems, depression, and hostility. In sum,these analyses supported the general construct valid-ity of the parenting codes used in the current study.

Clinical ratings of parent improvement. Parentimprovement was conceptualized as a global variablethat should take into account a parent's pattern ofchange across various categories, recognizing thatparents could remain unchanged in certain areaswhile improving or declining in other areas. Theclinical rating measure used here combines quantita-tive and qualitative features.

For each coding category, the distribution of theparent change scores was transformed into z scores.This transformation made it possible to determine thenumber of parents who improved by at least onestandardized unit, a conservative criterion of change.Then, to obtain a global measure of overall parentimprovement that would take into account positive aswell as negative change, we calculated a clinicalrating that used (a) each parent's profile of standard-ized change scores across categories and (b) knowl-edge of the actual category scores relative to the restof the sample. Each parent was rated as showing"moderate to excellent improvement" or "no mean-ingful improvement" on the basis of the followingcriteria: Parents were given a rating of excellentimprovement if they maintained excellent status(e.g., very low mention of power assertive disciplineat beginning and end phases of therapy) or if theyshowed significant improvement on three or morecategories without worsening on any categories. Par-ents were rated as having achieved moderate im-provement if their pattern of change fell betweenexcellent and not meaningful. Parents were given arating of no meaningful improvement if progress theymade in some categories appeared to be canceled outby declines or maintaining of lower-than-average sta-tus on other categories or if they made no significantgains. Using these criteria, two independent raters


achieved 97% agreement in their ratings of overallparent improvement as moderate to excellent or nomeaningful improvement.

Adolescent drug use. Using both adolescent self-report and urinalysis data, trained raters (who wereblind to all identifying information) rated severity ofadolescent drug use by considering the followingcriteria: (a) nature of drugs used, (b) frequency ofuse, and (c) number of different drugs used. Therating scale used here was designed to reflect notonly conventional wisdom about severity of adoles-cent drug abuse but also the drug using patterns in thecurrent sample. It is interesting to note in this contextthat all participants who used drugs (typically hallu-cinogens, stimulants, and cocaine) other than mari-juana and alcohol were polydrug users who usedalcohol and marijuana five times per week or more inaddition to other drugs. Not one case used these otherdrugs in the absence of frequent alcohol and mari-juana abuse. Drug use (frequency and number andtype of drugs used) in the course of the previous 30days was, then, rated on 15-point scale in which arating of 1 indicated no drugs used and each subse-quent scale point indicated gradually increasing druguse. For example, a score of 3 indicated alcohol ormarijuana was used 2-3 times during the previousmonth. A score of 2-9 indicated increasing frequencyin alcohol or marijuana use, but no other drugs wereused. At point 10, other drugs (most notably halluci-nogens, cocaine, stimulants) were introduced. Ascore of 10-15 indicated daily, or near daily, alcoholor marijuana use and increasing frequency of otherdrug use, from once during the last month (at a scoreof 10) to more than twice per week (at a score of 15).Interrater reliability was .92 (intraclass correlationcoefficient), indicating excellent agreement amongraters.

The adolescent's profile of drug use across fourassessment times (pretherapy, termination, and 6-and 12-month follow-ups) was then rated as showingmoderate to good improvement if there was an im-mediate and sustained drop in drug use. The profilewas rated as showing no meaningful improvement ifthe drug use remained high, became more severe,was erratic in its course, or dropped only slightly onthe 15-point scale. Two independent raters rated eachadolescent, and their percentage of agreement was90%.

Acting out behaviors. Adolescent behavior prob-lems were measured by the Acting Out Behaviors(AOB) scale (Ben-Porath, Williams, & Uchiyama,1989)—derived from the Devereux Adolescent Be-havior Rating Scale (Spivack, Haimes, & Spotts,1967)—which was administered to the adolescent'sprimary parent. The AOB scale identifies the extentof poor anger control, interpersonal problems, moodswings, and antisocial and aggressive behavior. TheAOB scale has been found to be internally consistent

with an average coefficient alpha of .87 (Ben-Porath,Williams, & Uchiyama, 1989). External and concur-rent validity have also been demonstrated (Williams,Ben-Porath, Uchiyama, Weed, & Acher, 1990) inthat the scale discriminates between adolescent sub-stance abusers and adolescent psychiatric patientsand in that it converges with record reviews andparent ratings.

The adolescent's profile of AOB scores across fourassessment times (pretherapy, termination, and 6-and 12-month followups) was rated as showing mod-erate to good improvement if there was an immediateor gradual drop of at least 10-20 points to a finalscore below 78, which was below the AOB mean foradolescent substance abusers in the validation sample(Williams et al., 1990). The profile was rated asshowing no meaningful improvement if the AOBscale score did not meet the criteria for change (e.g.,remained high, became more severe, or was erratic inits course). Two independent raters rated each profileof AOB scale scores, and their percentage of agree-ment was 93.

The operationalization of clinically meaningfulchange is a perennial methodological concern in psy-chotherapy outcome research, in general, and drugabuse treatment research, in particular (Carroll &Rounsaville, 1991; Liddle & Dakof, 1995a; Moras,1993). Despite concerted efforts (e.g., Beutler &Hamblin, 1986; Hsu, 1989; Jacobson & Revenstorf,1988), "no single method has come close to provid-ing us with a universally acceptable definition" ofthis construct (Jacobson & Revenstorf, 1988, p. 132).Methods for defining meaningful change range fromobjective statistical procedures of various types tothose that may be subjective, arbitrary, value laden,and observer driven. In assessing change, we soughtan approach that combined quantitative measurementand clinical judgement.

We took several steps to reduce potential rater biasand subjectivity (Saal, Downey, & Lahey, 1980).First, the raters were highly trained clinicians andresearchers who made independent ratings of the datafollowing preestablished criteria. Second, theyreached high levels of interrater agreement. Finally,to reduce rater leniency or severity, a middle range of"moderate improvement" was initially provided forraters, although participants were ultimately dichot-omized with respect to improvement.

Results

Positive and Negative Features ofParenting

The content of parents' discussions duringfamily treatment were multidimensional—re-flecting behavioral, affective, and cognitive do-


mains of parenting. During the beginning phaseof treatment, positive and negative features ofparenting coexisted, though negative featuressurpassed positive ones (72% vs. 53%). Parentsexpressed negative affect and doubts about theirability to set limits, and they demonstratedpower assertive techniques, such as punish-ment, restriction, and verbal aggression. How-ever, they also showed positive features ofparenting, such as optimism, sharing of self andvalues, and affection. By the end of therapy,positive parenting features (77%) surpassed thenegative (47%).

Change in Parenting

Paired t tests that compared parenting at thebeginning and end phases of therapy indicated adecrease in the proportion of negative parentingfeatures and an increase in the proportion ofpositive parenting features by the end of treat-ment (See Table 1). Parents improved signifi-cantly on seven of the eight parenting catego-ries. By the end phase of therapy, parent scoreswere significantly lower on power assertive dis-cipline (t = 1.81, p < .04), negative monitoringand limit setting (t = 2.80, p < .005), negativeaffect and disengagement (t = 2.16, p < .02),and cognitive inflexibility (t = 2.27, p < .02).The decrease in interparent inconsistency was inthe expected direction, but it did not reach sig-nificance. The interparent inconsistency cate-gory was excluded from further analyses be-

cause it pertained only to a subsample of 16families in which both a father and motherparticipated in the therapy. At the end phase oftherapy, parent scores were significantly higheron positive discipline and communication (t =2.93, p < .004), positive monitoring and limitsetting (t = 3.82, p < .001), and positive affectand commitment \t = 3.13, p < .003).

The average parent changed significantly on2.2 parenting categories. Seventy-two percentof the parents improved significantly in at leastone category; 62% improved on at least two;and 52% improved on at least three categories.Clinical ratings of overall parent improvementindicated that 69% showed moderate to excel-lent improvement, whereas 31% failed to showoverall meaningful improvement.

Parent Improvement and AdolescentSymptom Reduction: Drug Use andBehavior Problems

The two levels of overall parent improvement(moderate to high and no meaningful improve-ment) were cross-tabulated with the two ratinglevels of reduction in adolescent substance useand acting out behaviors (moderate to goodimprovement and no meaningful improvement).For each set of cross-tabulated frequencies, achi-square goodness-of-fit test was calculated todetermine whether the observed frequencies inthe four cells differed significantly from a hy-

Table 1Beginning Phase and End Phase Means and Standard Deviations onParenting Features

Beginning phase oftreatment

End phase oftreatment

Parenting feature

Power assertive disciplinePositive disciplinePositive monitoring and

limit settingNegative monitoring and

limit settingInterparent inconsistencyNegative affect and

disengagementPositive affect and

commitmentCognitive inflexibility

M

.1755

.3091

.0409

.1880

.0735

.2224

.1779

.1323

SD

.127

.170

.048

.100

.080

.129

.074

.166

M

.1250

.4421

.0971

.1252

.0451

.1615

.2326

.0576

SD

.111

.244

.074

.106

.075

.132

.101

.053Note, n = 29 for all categories except interparent inconsistency {n = 16).


pothetical distribution assigning equal numbersto each cell (which would result if there were noconnection between parent change and adoles-cent outcome). As Tables 2 and 3 show, theresults indicated a significant relationship be-tween parent improvement and reduction in ad-olescent drug use, ̂ ( 3 , N = 29) = 18.31, p <.001, and acting out behaviors, ^ ( 3 , N = 28) =9.99, p < .02.

Fifty-nine percent of the families demon-strated both an improvement in parenting and areduction of adolescent drug use. Interestingly,in 21% of the families, parenting did not im-prove meaningfully, but the adolescent's druguse showed meaningful reduction by the end oftreatment. Ten percent of the families showedimprovement in parenting but no correspondingreduction in adolescent drug use, and another10% showed no improvement in either parent-ing or adolescent drug use.

The results with respect to the AOB scalewere similar to those concerning drug use, afinding that is consistent with other researchthat has demonstrated interrelationships be-tween one problem behavior and another(Donovan, Jessor, & Costa, 1988; Newcomb &Felix-Ortiz, 1992). Fifty percent showed im-provement in both parenting and parent reportof adolescent acting out behaviors; 21% of theadolescents were reported to be improved with-out any corresponding improvement in theirparent's parenting; 18% of the parent's im-proved their parenting and reported that theiradolescents failed to significantly reduce theiracting out behaviors; and 11% did not showmeaningful improvement in either the parentsor the adolescents.

Discussion

These data suggest four major findings. First,parents of drug-using adolescents, in the begin-

Table 2Improvement in Parenting Cross-Tabulatedwith Reduction in Adolescent Drug Use(n = 29)

ParentingAdolescent

drug use No improvement Improvement

Table 3Improvement in Parenting Cross-Tabulatedwith Reduction in Adolescent Acting OutBehavior (n = 28)

Adolescent Parentingacting outbehavior No improvement Improvement

ReductionNo reduction

145

ReductionNo reduction

173

ning phase of treatment, evidenced negativeparenting behaviors (as expected). These obser-vationally derived data are consistent with self-report data from this sample, which indicatedconsiderable disengagement among familymembers. These parents of substance-abusingteenagers were disengaged both emotionallyand in terms of day-to-day interactions withtheir children (Liddle, Dakof et al., 1995).These results are consistent with previous find-ings indicating disconnected parent-adolescentrelationships in families with drug-abusing ad-olescents (Brown et al., 1993; Shedler & Block,1990; Volk, Edwards, Lewis, & Sprenkle,1989).

An aspect of this relationship has been de-scribed in the clinical literature as parental ab-dication (Isaacs, Montalvo, & Abelsohn, 1986).One parent expressed her disengagement in thisway: "If I had a choice, I would rather not be aparent. That's bottom on my list in terms ofpriorities. It's awful to be defeated. I don't wantto be his mother. I want his probation officer toput him somewhere." Other parents expressedconsiderable difficulty in attempts to monitorand set limits. In the words of one parent: "Hewasn't coming home. I say 'come home atmidnight'; he rolls in at 2:00. I say 'don't goout'; he sneaks out. He is smoking marijuanaand I know what kind of people he hangsaround with . . . he is not going to school; he'son the streets. Screaming and shouting, theusual kind of stuff. I set limits and tell him youcan't do this. It doesn't do any good. I'm thetype of person that can't follow through."

The second major finding is that the parents,at the outset of treatment, evidenced certainstrengths and competencies in the parentingrealm. This circumstance is often overlooked byresearchers and clinicians alike. During the be-ginning phase of treatment, one parent ex-pressed her commitment to her son in this way,


"In our home, both of us are willing to bendover backwards. I love my son." Anothermother took responsibility for past mistakes andexpressed a firm commitment to change: "Uh,hum. She [the adolescent daughter] is right. I'vemade some mistakes. I was a single parent. Ihad my own problems. But I bore these kids,you can't forget that. I want it to be differentnow."

The pattern revealed here supports the docu-mented difficulties clinical families have in sev-eral areas (e.g., behavior management and dis-cipline, negative affect, ineffective problemsolving, a wavering commitment to parenting inresponse to personal stress or distress, negativeattribution bias about their adolescent). How-ever, we see also that these same parents havereadily identifiable strengths. In the context ofwhat appears to be considerable dysfunction,there are areas of competence that coexist withthe aforementioned problems in parenting. Thisfinding is reminiscent of Luthar, Doernberger,and Zigler's (1993) study indicating that certainhigh-risk teenagers who were resilient in a num-ber of areas also were found to have areas offunctioning in which they were not only notresilient, but also dysfunctional. The research-ers interpret these data in terms of the need fora differentiated view of resilience in high-riskteenagers. Returning to the present data withthis point in mind, we needed to understand thecoexisting competencies and problem areas thatparents present. There are practical implicationsof these insights. The search for strengthsis a hallmark of family therapy's tradition(Minuchin, 1974; Montalvo, 1986), and it is amajor emphasis of the family-based interven-tion tested in this study. Given the emotionaltoll this kind of work extracts from its practi-tioners (Bank, Marlowe, Reid, Patterson, &Weinrott, 1991), findings about the presence ofstrengths amidst the expected array of problemsshould provide encouragement and guidance fortherapists in meeting challenges presented bydistressed families.

A third finding concerns change in the parent-ing realm. Although the data presented herecannot prove changes in parenting as a result ofthe family intervention, they indicate thatparenting is malleable and that a family inter-vention is associated with positive change inparents and adolescents. This suggestion is par-ticularly important given that most previouswork on the connection of parent change to

change in offspring was conducted with chil-dren, not adolescents. Studies reveal that it ismore difficult both to retain in treatment and tochange the parenting practices of parents whoseadolescents are drug involved or show chronicpatterns of delinquency as opposed to thoseparents whose children have behavior problems(Bank et al., 1991; Dishion & Patterson, 1992;Patterson & Chamberlain, 1994). Chamber-lain's (1990) study with parents in foster fami-lies demonstrated that, under certain conditions,trained foster parents of extremely antisocialadolescents could change their behavior and,thus, affect the teenager's behavior. Althoughmany studies have demonstrated the efficacy offamily therapy, the specific connection ofchanges in parenting practices to changes in anadolescent's symptoms rarely has been exam-ined. In the present study, by the end of treat-ment with MDFT, most parents (69%) showedsignificant improvement in their parenting byachieving decreases in negative features andincreases in positive features. For example, onemother, at the beginning of treatment, expressedher frustrations and difficulties in parenting:"Part of me wants to give up. Life is too short togive myself an ulcer or heart attack . . . I'm nota disciplinarian. It's real hard for me to enforcethings. I'm either not there, or I'm at work. Formonitoring, I'm not always there." By the endof therapy she reported, "I've been keeping upwith the school, calling to see if he is going. Heis being very closely monitored." During one ofthe last few therapy sessions, she turned to herson and said, "Let's talk and not assume whatsomeone else is thinking. I promise I will notassume what you are thinking. Is that a deal?"Given the degree of dysfunction present in clin-ical samples at the outset of treatment (Kazdin,1994), the stability of antisocial behavior of thetype treated in this sample (Loeber, 1991), thedifficulty of demonstrating change in interven-tion research with these difficult problems(Kazdin, 1994), the capacity of negative parent-ing behavior to elicit untherapeutic or nonmodelconsistent behavior from clinicians during treat-ment (Patterson & Chamberlain, 1994), and thenegative cultural stereotypes about adolescentsin general (Offer, Ostrov, & Howard, 1981;Steinberg, 1990), these findings are significant.

The fourth finding relates to different patternsof change.

Parent change—adolescent change. As ex-pected, a significant association was found be-


tween improvement in parenting and reductionin adolescent symptomatology: Most parentsand adolescents improved in tandem. This find-ing supports a fundamental prediction of familysystems theory: Change in a core domain offamily functioning—in this case, parentingpractices—will be associated with change in theadolescent problem behavior (Nichols &Schwartz, 1994). Although parent improvementshowed a significant positive association withadolescent improvement in most cases, therewere exceptions. Examining these exceptionsmay illuminate important issues germane toproducing change in clinical work with adoles-cents and their parents. These efforts are in linewith current treatment development efforts torefine our models of change and enhance exist-ing interventions (Miller & Prinz, 1990).

Exceptions to the Pattern of TandemParent-Adolescent Change

Within the sample, there were treatment suc-cesses and failures. In looking at the cross-tabulations of parent change with reduced ado-lescent drug use, 41% of the families showedpatterns of change that diverged from the prev-alent one of tandem parent-adolescent improve-ment. These families had either a partial change(e.g., adolescents who showed reduced symp-tomatology despite lack of meaningful changein their parents or parents who improved with-out any meaningful reduction in their children'sdrug use) or a lack of meaningful change (fam-ilies in which neither parents nor adolescentsimproved in any meaningful fashion).

Parent change—no adolescent change. Insome cases (10%), significant changes inparenting did not accompany adolescent im-provement in symptomatology. A variety ofvariables may be operating here. Perhaps thetherapy targeted and changed dimensions ofparenting that, in these few cases, were notsufficiently powerful or relevant to facilitateeffects in the adolescent. Furthermore, to under-stand how a parent changes and her or hisadolescent does not, we may need to rememberthe influence of individual and extrafamilialvariables that are always present but are notalways accounted for or amenable to change viatherapy. Peer systems or others with whom theadolescent interacts directly or indirectly (e.g.,the school, juvenile justice system, neighbor-

hood, and community) as well as adolescentpersonality variables can support or deter thefocus, intensity, duration, and success of treat-ment (see Brook, Nomura, & Cohen, 1989;Shedler & Block, 1990). Research indicatesthat, although the impact of authoritativeparenting behaviors produces differentially pos-itive outcomes through the adolescent years, "incertain ecologies overarching forces, outside thecontrol of parents, may entirely overwhelm thebeneficial effects of authoritative parenting inthe home" (Steinberg, Darling, Fletcher,Brown, & Dornbusch, 1995, p. 461).

Parents may be limited in their capacity toinfluence the behavior of older adolescents gen-erally and problem behaviors that may haveparticularly change-resistant and stable features(e.g., drug abuse) given later developmentalprogression of the disorder. Allen and col-leagues' research asserted that severe adoles-cent problem behavior is linked with failures inattachment and with failures to maintain relat-edness with parents, which in turn weaken pa-rental controls over adolescent behavior (Allen,Aber, & Leadbeater, 1990). Allen et al. (1990)hypothesized that a lack of relatedness (i.e.,disengagement) between parents and adoles-cents removes an important behavior-regulatinginfluence within the family—the adolescent'sdesire to please the parent. Similar ideas havebeen offered from a behavioral perspective.Writing about the limitations of a behavior con-tingency approach with parents of delinquentadolescents, Rueger and Lieberman (1984) ad-vised that "In cases where parents have lostreinforcement control over their adolescent, en-gaging the child's peer group or communityagencies (i.e., law enforcement or probation de-partments) in the intervention strategy becomesnecessary (p. 416). Hence, chronic delinquency,serious drug abuse (or perhaps certain kinds ofdrug abuse such as drug addiction), antisocialpersonality, and solidity of connection to a de-viant peer culture could singly or in combina-tion make it less likely that changes in a parentwould be sufficient to influence the adolescent.For example, we know that once patterned druguse begins (or, for that matter, after problembehavior of any sort begins in earnest), peersexert more powerful ongoing influences tomaintain these problem behaviors than do par-ents to stop them (Dishion & Loeber, 1985;Kandel, 1985). For some of these adolescents,intervention simply may have come too late to


reverse these powerful, already well-estab-lished, detrimental influences. So, given what isknown about how extrafamilial contexts exertpowerful negative forces to support deviant be-havior, the cumulative deleterious effects of ne-glectful parenting styles on adolescent develop-ment, and the link of these parenting behaviorsto adolescent problem behaviors (Steinberg,Lamborn, Darling, Mounts, & Dornbusch,1994), our findings challenge family therapiststo address the limits of family intervention.These results give indications of the ways inwhich current theories of change themselvesneed to change (Liddle, 1995). Reexaminationof our beliefs about change in families and othermultiperson systems is related to the new think-ing about our understanding of the characteris-tics, onset, and course of functioning and adap-tation (Kazdin & Kagan, 1994). Our changingtheories of change will be informed greatly notonly by intervention science but also by usingthe knowledge from related specialties, such asdevelopmental psychopathology and clinicalepidemiology.

No parent change—adolescent change. In21% of the families, the adolescent achievedand sustained decreased drug use despite lack ofmeaningful change in his or her parent or par-ents. Positive events outside the family—changes in friendship networks, neighborhood,or school environments—or changes in the ad-olescent self-system can be sufficient or canoverride a lack of change or insufficient changein the parental subsystem. These results supportcontemporary perspectives on problem forma-tion and intervention that argue for multipleperspectives, levels, and kinds of change pro-cesses. This perspective asserts that a narrowadherence to single-level or domain interven-tions fails to consider the multivariate nature ofchange as well as the practical extension of thispossibility—that change facilitation may re-quire accessing different pathways and usingmany methods. In MDFT (the tested interven-tion), if change cannot be effected in a primarytarget area such as the parental subsystem, otherprimary target areas are emphasized (e.g. workwith teenager alone). Each target area involvesdifferent hypothesized mechanisms and path-ways of change. In this sense, when reasonableattempts to achieve desired change in one sub-system failed or, even more basically, when wefailed to materialize a certain way of working orpreferred content in a particular subsystem, we

worked around this situation by emphasizingother, already targeted intervention areas. It ispossible that processes of this nature were atplay in cases in which the adolescent changedand the parent did not. That is, MDFT's workwith the adolescent individually, a situation thatwas emphasized when our alliance with theparent had not been formed, may have ac-counted for the cases in which the adolescentchanged but the parent did not. However, theseideas remain informed speculation until we dofine-grained therapy analyses that investigatemechanisms of change (Pinsof, 1989).

Additionally, all but one of the parents in thiscategory were low income, single parents(mothers without any financial or other supportfrom the child's father) with past or presentextreme life stresses (e.g., serious physicalabuse, alcoholism and drug addiction, death ofclose family members). For some time, psy-chologists have known about the influence ofthese life circumstances on mental health andtheir influence in treatment. Social disadvantageand parental depression were among the firstvariables identified as significant predictors forpoor outcomes in parent intervention programs(McMahon, Forehand, Griest, & Wells, 1981;Patterson, 1974; Wahler & Dumas, 1989). Re-cent findings suggest that the effect of maternalsocial disadvantage on child outcomes is medi-ated through disrupted discipline and parentingpractices (Bank, Forgatch, Patterson, & Fetrow,1993).

No parent change—no adolescent change.Finally, in 10% of the families, neither the ad-olescent nor the parent or parents changed. Sev-eral variables may apply here. For this group offamilies, the interventions may not have beenpowerful enough or sufficiently well matched tothe parent or adolescent. Perhaps a more inten-sive intervention model, such as residentialtreatment or offering families more frequenthome-based therapy, would have been effec-tive. Some researchers are experimenting withhome-based therapy (Liddle, 1994), and otherresearchers have demonstrated positive resultsby adapting family therapy to a home-basedformat (Henggeler, Melton, & Smith, 1992). Athorough understanding of what happened withthis 10% of the sample would require inclusionof previously mentioned variables, such as thestage of problem progression and the individualand familial characteristics alone and in combi-


nation with the particular interventions offeredin the approach under study.

In summary, these data suggest the need for adifferentiated change model. This complex ex-planation of change would account not only fortandem change with the parent and adolescent,but also would account for the fact that changein one family member does not automaticallylead to change in other family members (i.e.,one family member may change and others maynot). Psychologists need to understand morefully those situations in which parent change is,and is not, connected to adolescent change. Inthe past, family therapy theory embraced one ofthe more optimistic systems tenets—the hy-pothesized ripple effect phenomenon. In thisscenario, change in one family member wasthought to serve as an ipso facto prompt ofchange in another family member (i.e., a dom-ino effect of change process; Haley, 1976;Watzlawick, Weakland, & Fisch, 1974). Thepresent data suggest the need to revise thisdeterministic premise about change. Althoughour findings, as well as those of other research-ers, support the connectedness of parent andadolescent change, exceptions (not uncom-monly noticed in clinical work) were also foundin the current data. These data advise againstadopting an undifferentiated, overly general rip-ple effect view of change, and, more basically,they suggest that we need more complex theo-rizing and research about how to define changein family systems. Given the likely multivariatenature of change processes and the correspond-ing need to take into account both family andextrafamilial sources of influence in change andnonchange, these results suggest the need forflexible, individualized, therapeutic models thatcan be adapted to a variety of parenting andpersonality styles as well as to diverse familystructures.

Limitations and Future Directions

Some limitations of the present study need tobe noted. First, these results must be interpretedcautiously because of the absence of a compar-ison group. The data and the analyses werederived from a clinical sample who received aspecific form of family intervention: MDFT.Although we can conclude that parentingchanged in the expected direction, we cannotinfer that this change was caused by MDFT.

Our data cannot rule out the possibility that theobserved change would have occurred amongparents whose adolescents received alternativetreatments, family-based or otherwise, oramong parents whose adolescents received notreatment whatsoever. At this point, we can onlyhypothesize that the change was related to theparents' participation in MDFT.

Second, the study did not measure outside-of-therapy influences that may have impeded orfacilitated the parenting process and its influ-ence on adolescent symptomatology. Thus, wecan only speculate about the nature of theseprocesses, pending a true empirical exploration.Measurement of these processes are importantin future studies, given the need to developtheories of change that are multifactorial (i.e.,theories that, like interventions, take into ac-count multiple levels and aspects of functioningand individual differences in accounting forchange).

Third, although these data reveal that parentsof adolescents who use drugs can change, thedata do not allow us to understand preciselyhow, or whether, parents applied the new un-derstanding or practices developed and dis-cussed in therapy to their everyday lives. De-signs that include multiple observations anddata from inside and outside of treatment areneeded to chart the different kinds of processesand variables likely to be involved in change(see Gottman & Rushe, 1993). These studieshave already begun in the individual treatmentarea but are rare in family therapy (see Fried-lander, Heatherington, Johnson, & Skowron,1994). Through the new levels and kinds ofdetail available, these studies promise to yieldnew, more complex models of human change(Barkham, Stiles, & Shapiro, 1993; Stoolmilleret al., 1993).

Fourth, we must note that, as is always thecase, this study's analyses both reflect and areconstrained by the structure of the measurementdevice. In this particular study, the assessmentof parenting practices, beliefs, and affect wasbased on parent statements about parentingrated in the context of MDFT. This approachmay lead to certain problems in interpreting theresults if parents refrained from sharing certaintypes of information because of social desirabil-ity, defensiveness, or lack of opportunity. Nev-ertheless, both external and internal validitymay temper this criticism: The validation sam-ple used transcripts of only a single session, and


the results agreed with self-reports of parentingpractices and parental emotionality. Moreover,the observational measure of parenting and itsassociation with the outcome variables were inaccord with expectations. These findings sug-gest that, despite limitations, the observationalmeasure of parenting is sensitive to strengths,weaknesses, and change in parenting.

Fifth, although we most frequently postulate asequence of change in which parent changeleads to adolescent change, the reverse is alsoplausible. Change in adolescent symptomatol-ogy and behavior is possible apart from tandemchange in parents' parenting, given findings inthe developmental literature on children's influ-ence on parenting and psychological develop-ment of their parents (Bugental & Shennum,1984; Scarr & McCartney, 1983). These notionsare particularly relevant because the therapymodel intervenes actively and directly withindifferent subsystems simultaneously (parental,adolescent, adolescent-parent interaction, andextrafamilial).

In conclusion, the present study indicates thatnonoptimal parenting of drug-using adolescentscan change. However, many questions remainabout the mechanisms by which parenting prac-tices change, the nature of these changes, andthe relative efficacy of different means to pro-mote such changes. Additionally, in areas withclear overlap in the developmental and devel-opmental psychopathology specialties, we needto delineate those aspects of parenting in whichchanges are most likely to be associated withthe best outcomes in adolescent behaviorchange. This knowledge, of course, must befirmly grounded in accurate developmentalknowledge. Taking the lead from the basic sci-ence done in these areas (Bornstein, 1995), re-searchers still need to examine differences inmothers' and fathers' parenting, differences ac-cording to culture and ethnicity, differences re-lated to socioeconomic circumstances, and dif-ferences between parenting boys versus girls.Other recent work in the adolescent develop-ment field provides clues for the kind of workthat may prove beneficial in this regard. Stein-berg, Elmen, and Mounts (1989) found thateach of three aspects of authoritative parenting(acceptance, psychological autonomy, and be-havioral control) makes an independent contri-bution to the adolescent outcome of schoolachievement. Linver and Silverberg (in press)used a similar "unpacking strategy" (i.e., an

attempt to disentangle unique contributions ofdifferent aspects of complex behaviors) in ar-riving at differential predictions about adoles-cent problem behavior, development, and ado-lescent sense of self. These authors found thatparental monitoring predicted problem behav-ior, whereas adolescent sense of self was bestpredicted by parenting practices that includedpsychological control and parental respect.

The clinical problems family therapists seekto influence are complex, multidetermined, andstable. Their ideas about how to intervene mosteffectively to influence the lives of the peoplethey see in therapy are in flux. The field needsnot only new studies that will, as we haveindicated, help reveal more about the subtletiesand variations inherent in human change (whichwill help to revise our models of change alongmore complex lines), but also revised and re-newed conceptual frameworks (blending ideasfrom developmental and clinical spheres, forinstance) for treatment. Perhaps these integra-tive frameworks will prompt the kind of clinicalinnovation that characterized the family inter-vention area in its earliest days. Unlike thepioneer days, however, these new clinical mod-els will be subjected to rigorous empiricaltests—an activity already in evidence in manycorners of the family intervention field today.

References

Abidin, R. R. (1992). Presidential address: The de-terminants of parenting behavior. Journal of Clin-ical Child Psychology, 21, 407-412.

Alexander, J. R, Holtzworth-Munroe, A., & Jame-son, P. B. (1994). Research on the process andoutcome of marriage and family therapy. In A. E.Bergin & S. L. Garfield (Eds.), Handbook of psy-chotherapy and behavior change (4th ed., pp. 595-630). New York: Wiley.

Allen, J. P., Aber, J. L., & Leadbeater, B. J. (1990).Adolescent problem behaviors: The influence ofattachment and autonomy. Psychiatric Clinics ofNorth America, 13, 455-467.

Bank, L., Forgatch, M. S., Patterson, G. R., &Fetrow, R. A. (1993). Parenting practices of singlemothers: Mediators of negative contextual factors.Journal of Marriage and the Family, 55, 371-384.

Bank, L., Marlowe, J. H., Reid, J. B., Patterson, G.R., & Weinrott. (1991). A comparative evaluationof parent-training interventions for families ofchronic delinquents. Journal of Abnormal ChildPsychology, 19, 15-33.

Barkham, M., Stiles, W. B., & Shapiro, D. A. (1993).The shape of change in psychotherapy: Longitudi-


nal assessment of personal problems. Journal ofConsulting and Clinical Psychology, 61, 661-611.

Barrett, M. E., Simpson, D., & Lehman, W. E.(1988). Behavioral changes of adolescents in drugabuse intervention programs. Journal of ClinicalPsychology, 44, 461-473.

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

Belsky, J. (1984). The determinants of parenting: Aprocess model. Child Development, 55, 83-96.

Ben-Porath, Y. S., Williams, C. L., & Uchiyama, C.(1989). New scales for the Devereux AdolescentBehavior Rating Scale. Psychological Assessment,1, 58-60.

Beutler, L. E., & Hamblin, D. L. (1986). Individual-ized outcome measures of internal change: Meth-odological considerations. Journal of Consultingand Clinical Psychology, 54, 48-53.

Block, J., Block, J. H., & Keyes, S. (1988). Longi-tudinally foretelling drug usage in adolescence:Early childhood personality and environmentalprecursors. Child Development, 59, 336-355.

Bornstein, M. (1995). Handbook of parenting (Vols.1-4). Hillsdale, NJ: Erlbaum.

Brook, J. S., Nomura, C , & Cohen, P. (1989). Anetwork of influences on adolescent drug involve-ment: Neighborhood, school, peer, and family. Ge-netic, Social and General Psychology Mono-graphs, 115, 123-145.

Brown, B. B., Mounts, N., Lamborn, S. D., & Stein-berg, L. (1993). Parenting practices and peer groupaffiliation in adolescence. Child Development, 64,467-482.

Bugental, D. B., & Shennum, W. A., 1984. "Diffi-cult" children as elicitors and targets of adult com-munication patterns: An attributional/behavioraltransactional analysis. Monographs of the Societyfor Research in Child Development, 49(1, SerialNo. 205).

Carroll, K. M., & Rounsaville, B. J. (1991). In L.Onken & J. Blaine (Eds.), Psychotherapy andcounseling in the treatment of drug abuse (NIDAResearch Monograph No. 104, pp.91-104). Wash-ington, DC: U. S. Government Printing Office.

Chamberlain, S. (1990). The new ethnography:"Windmills and giants." Research and FamilyTherapy, 2, 39-45.

Dadds, M. R., Schwartz, S., & Sanders, M. R. (1987).Marital discord and treatment outcome in behav-ioral treatment of child conduct disorders. Journalof Consulting and Clinical Psychology, 55, 396-403.

Dakof, G. A., & Mendelsohn, G. A. (1989). Patternsof adaptation to Parkinson's disease. Health Psy-chology, 8, 355-372.

Derogatis, L. R. (1977). SCL-90—R: Administra-tion, scoring & procedures manual-II for the re-

vised version. Towson, MD: Clinical PsychometricResearch.

Diamond, G., & Liddle, H. A. (in press). Resolvingin-session parent-adolescent conflict in multidi-mensional family therapy. Journal of Consultingand Clinical Psychology.

Dishion, T. J., & Loeber, R. (1985). Adolescent mar-ijuana and alcohol use: The role of parents andpeers revisited. American Journal of Drug AlcoholAbuse, 77(1—2), 11-25.

Dishion, T. J., & Patterson, G. R. (1992). Age effectsin parent training outcome. Behavior Therapy, 23,719-729.

Dix, T. (1991). The affective organization of parent-ing: Adaptive and maladaptive processes. Psycho-logical Bulletin, 110, 3-25.

Donovan, J. E., Jessor, R., & Costa, F. M. (1988).Syndrome of problem behavior in adolescence: Areplication. Journal of Consulting and ClinicalPsychology, 56, 762-765.

Farrington, D. P., Loeber, R., Elliot, D. S., Hawkins,J. D., Kandel, D. B., Klein, M. W., McCord, J.,Rowe, D. C , & Tremblay, R. E. (1990). Advanc-ing knowledge about the onset of delinquency andcrime. In B. B. Lahey & A. E. Kazdin (Eds.),Clinical Child Psychology, (Vol. 13, pp. 283-342).New York: Plenum.

Friedlander, M. L., Heatherington, L., Johnson, B., &Skowron, E. A. (1994). Sustaining engagement: Achange event in family therapy. Journal of Coun-seling Psychology, 41, 438-448.

Glaser, B. G., & Strauss, A. L. (1967). The discoveryof grounded theory: Strategies for qualitative re-search. New York: Aldine-Atherton.

Goodnow, J. J. (1988). Parents' ideas, actions andfeelings: Models and methods from developmentaland social psychology. Child Development, 59,286-320.

Gottman, J. M., & Rushe, R. H. (1993). The analysisof change: Issues, fallacies, and new ideas. Journalof Consulting and Clinical Psychology, 67(6),907-910.

Haley, J. (1976). Problem solving therapy. San Fran-cisco: Jossey-Bass.

Hawkins, J. D., Catalano, R. F., & Miller, J. Y.(1992). Risk and protective factors for alcohol andother drug problems in adolescence and earlyadulthood: Implications for substance abuse pre-vention. Psychological Bulletin, 112, 64-105.

Hazelrigg, M. D., Cooper, H. M., & Borduin, C. M.(1987). Evaluating the effectiveness of family ther-apies: An integrative review and analysis. Psycho-logical Bulletin, 101, 428-442.

Henggeler, S. W., Borduin, C. M., & Mann, B. J.(1993). Advances in family therapy: Empiricalfoundations. Advances in Clinical Child Psychol-ogy, 15, 207-241.

Henggeler, S. W., Borduin, C. M., Melton, G. B.,Mann, B. J., Smith, L. A., Hall, J. A., Cone, L., &


Fucci, B. R. (1991). Effects of multisystemic ther-apy on drug use and abuse in serious juvenileoffenders: A progress report from two outcomestudies. Family Dynamics Addiction Quarterly, 1,40-51.

Henggeler, S. W., Melton, G. B., & Smith, L. A.(1992). Family preservation using multisystemictherapy: An effective alternative to incarceratingserious juvenile offenders. Journal of Consultingand Clinical Psychology, 60, 953-961.

Hsu, L. M. (1989). Reliable changes in psychother-apy: Taking into account regression toward themean. Behavioral Assessment, 11, 459-467.

Issacs, M., Montalvo, B., & Abelsohn, D. (1986). Thedifficult divorce. New York: Basic Books.

Jacobson, N. S., & Revenstorf, D. (1988). Statisticsfor assessing the clinical significance of psycho-therapy techniques: Issues, problems, and new de-velopments. Behavioral Assessment, 10, 133-146.

Joanning, H., Quinn, W., Thomas, F., & Mullen, R.(1992). Treating adolescent drug abuse: A compar-ison of family systems therapy, group therapy, andfamily drug education. Journal of Marital andFamily Therapy, 18, 345-356.

Kandel, D. B. (1985). On processes of peer influ-ences in adolescent drug use: A developmentalperspective. In D. W. Brook, J. S. Brook, D. J.Lettieri, & B. Stimmel (Eds.), Alcohol and sub-stance abuse in adolescence (Vol. 4, No. 3-4, pp.87-135). New York: Haworth Press.

Kazdin, A. E. (1987). Treatment of antisocial behav-ior in children: Current status and future directions.Psychological Bulletin, 102, 187-203.

Kazdin, A. E. (1994). Methodology, design, andevaluation in psychotherapy research. In A. E. Ber-gin & S. L. Garfield (Eds.), Handbook of psycho-therapy and behavior change (4th ed., pp. 543-594). New York: Wiley.

Kazdin, A. E., & Kagan, J. (1994). Models of dys-function and developmental psychopathology.Clinical Psychology, 1, 35-52.

Lambert, M. J. (1982). The effects of psychotherapy(Vol. 2). New York: Human Sciences Press.

Lebow, J. L., & Gurman, A. S. (1995). Researchassessing couple and family therapy. Annual Re-view of Psychology, 46, 27-57.

Lewis, R. A., Piercy, F. P., Sprenkle, D. H., &Trepper, T. S. (1990). Family-based interventionsfor helping drug-abusing adolescents. Journal ofAdolescent Research, 50, 82-95.

Liddle, H. A. (1991). Empirical values and the cul-ture of family therapy. Journal of Marital andFamily Therapy, 17, 327-348.

Liddle, H. A. (1994). Multicomponent interventionfor adolescent drug abuse: A therapy developmentstudy (Grant No. R01DA 07697). Rockville, MD:National Institute on Drug Abuse.

Liddle, H. A. (1995). Conceptual and clinical dimen-sions of a multidimensional, multisystems engage-

ment strategy in family-based adolescent treat-ment. Psychotherapy, 32, 39-58.

Liddle, H. A., Becker, D., Diamond, G., Diamond, G.S., Johnson, C, Pemberton, B., & Taylor, M.(1995, August). Working with juvenile justice in-volved drug abusing adolescents. Paper presentedat the 103rd Annual Convention of the AmericanPsychological Association, New York.

Liddle, H. A., & Dakof, G. A. (1995a). Family-basedtreatment for adolescent drug use: State of thescience. In E. Rahdert & D. Czechowicz (Eds.),Adolescent drug abuse: Clinical assessment andtherapeutic interventions (NIDA Research Mono-graph No. 156, NIH Publication No. 95-3908, pp.218-254). Rockville, MD: National Institute onDrug Abuse.

Liddle, H. A., & Dakof, G. A. (1995b). Efficacy offamily therapy for drug abuse: Promising but notdefinitive. Journal of Marital and Family Therapy,27,511-544.

Liddle, H. A., Dakof, G. A., Parker, K., Barrett, K.,Diamond, G. S., Garcia, R. G., & Palmer, R.B.(1995). Multidimensional family therapy for treat-ing adolescent substance abuse: A controlled clin-ical trial. Manuscript submitted for publication.

Liddle, H. A., Diamond, G., Dakof, G. A., Arroyo, J.,Guillory, P., Holt, M., & Watson, M. (1992). Theadolescent module in multidimensional familytherapy. In G. Lawson & A. Lawson (Eds.), Familytherapy with adolescent drug abusers (pp. 165-186). Gaithersburg, MD: Aspen Publishers.

Liddle, H. A., Schmidt, S., Ettinger, D., & Sessa, F.(in press). Adolescent development research:Guidelines for clinicians. Journal of Marital andFamily Therapy.

Linver, M. R., & Silverberg, S. B. (in press). Parent-ing as a multidimensional construct: Differentialprediction of adolescents' sense of self and engage-ment in problem behavior. International Journal ofAdolescent Medicine and Health.

Loeber, R. (1991). Antisocial behavior: More endur-ing than changeable? Journal of the AmericanAcademy of Child and Adolescent Psychiatry, 30,393-397.

Luthar, S., Doernberger, C , & Zigler, E. (1993).Resilience is not a unidimensional construct: In-sights from a prospective study of inner city ado-lescents. Development and Psychopathology, 5,703-717.

Maccoby, E. E., & Martin, J. A. (1983). Socializationin the context of the family: Parent-child interac-tion. In E. M. Hetherington (Ed.), Social develop-ment, Vol III ofMussen 's manual child psychology(4th ed., pp. 1-101). New York: Wiley.

Mann, B. J., Borduin, C. M., Henggeler, S. W., &Blaske, D. M. (1990). An investigation of systemicconceptualizations of parent-child coalitions andsymptom change. Journal of Consulting and Clin-ical Psychology, 58, 336-344.


McMahon, R. J., Forehand, R., Griest, D. L., &Wells, K. C. (1981). Who drops out of treatmentduring parent behavioral training? BehavioralCounseling Quarterly, 1, 79-85.

Miller, G. E., & Prinz, R. J. (1990). Enhancement ofsocial learning family interventions for childhoodconduct disorder. Psychological Bulletin, 108,291-307.

Minuchin, S. (1974). Families and family therapy.Cambridge, MA: Harvard University Press.

Montalvo, B. (1986). Family strengths: Obstaclesand facilitators. In M. Karpel (Ed.), Family re-sources (pp. 93-115). New York: Guilford Press.

Moras, K. (1993). Substance abuse research: Out-come measurement conundrums. In L. S. Onken, J.D. Blaine, & J. J. Boren (Eds.), Behavioral treat-ments for drug abuse and dependence (Monograph137, pp. 217-248). Rockville, MD: National Insti-tute on Drug Abuse Research.

National Institute on Drug Abuse. (1995). Clinicalreport and administrators guide to drug abusetreatment models (NIDA Monograph). Washing-ton, DC: U. S. Government Printing Office.

Newcomb, M. D., & Bentler, P. M. (1989). Sub-stance use and abuse among children and teenag-ers. American Psychologist, 44, 242-248.

Newcomb, M. D., & Felix-Ortiz, M. (1992). Multipleprotective and risk factors for drug use and abuse:Cross-sectional and prospective findings. Journalof Personality and Social Psychology, 63, 280-296.

Nichols, M., & Schwartz, R. (1994). Family therapy(3rd ed.). Needham Heights, MA: Allyn & Bacon.

Offer, D., Ostrov, E., & Howard, K. I. (1981). Themental health professional's concept of the normaladolescent. Archives of General Psychiatry, 38,149-152.

Patterson, G. R. (1974). Retraining of aggressiveboys by their parents: Review of recent literatureand follow-up evaluation. Canadian PsychiatricAssociation Journal, 19, 142-158.

Patterson, G. R. (1982). Coercive family process: Asocial learning approach (3rd ed.). Eugene, OR:Castilia.

Patterson, G. R., & Chamberlain, P. (1994). A func-tional analysis of resistance during parent trainingtherapy. Clinical Psychology, 1, 53-70.

Pinsof, W. (1989). A conceptual framework andmethodological criteria for family therapy processresearch. Journal of Consulting and Clinical Psy-chology, 57, 53-59.

Rueger, D. B., & Lieberman, R. P. (1984). Behav-ioral family therapy for delinquent and substance-abusing adolescents. Journal of Drug Issues, 14,403-418.

Saal, F. E., Downey, R. G., & Lahey, M. A. (1980).Rating the ratings: Assessing the psychometricquality of rating data. Psychological Bulletin, 88,413-428.

Scarr, S., & McCartney, K. (1983). How peoplemake their own environments: A theory of geno-type environment effects. Child Development, 54,424-435.

Schmidt, S. (1994). Changes in parenting practicesduring multidimensional therapy. Unpublisheddoctoral dissertation, Temple University, Philadel-phia, PA.

Segal, B. (1986). Intervention and prevention ofdrug-taking behavior: A need for divergent ap-proaches. The International Journal of the Addic-tions, 21, 165-173.

Selekman, M. D., & Todd, T. C. (1990). Major issuesfrom family therapy research and theory. In T. C.Todd & M. D. Selekman (Eds.), Family therapyapproaches with adolescent substance abusers (pp.311-325). Needham Heights, MA: Allyn & Bacon.

Shalala, D. (1995). White paper: Effectiveness ofdrug abuse treatment (DHHS Publication). Wash-ington, DC: U. S. Government Printing Office.

Shedler, J., & Block, J. (1990). Adolescent drug useand psychological health: A longitudinal inquiry.American Psychologist, 45, 612-630.

Shrout, P., & Fleiss, J. (1979). Intraclass correlations:Uses in assessing rater reliability. PsychologicalBulletin, 86, 420-428.

Sigel, I. E., McGillicuddy-DeLisi, A. V., & Good-now, J. J. (Eds.). (1992). Parental belief systems:The psychological consequences for children (2nded.). Hillsdale, NJ: Erlbaum.

Spitzer, A., Webster-Stratton, C , & Hollinsworth, T.(1991). Coping with conduct-problem children:Parents gaining knowledge and control. Journal ofClinical Child Psychology, 20, 413-427.

Spivack, G., Haimes, P. E., & Spotts, J. (1967).Devereux Adolescent Behavior Rating scale man-ual. Devon, PA: The Devereux Foundation.

Steinberg, L. (1990). Autonomy, conflict, and har-mony in the family relationship. In S. S. Feldman& G. R. Elliott (Eds.), At the threshold: The devel-oping adolescent (pp. 255-276). Cambridge, MA:Harvard University Press.

Steinberg, L., Darling, N. E., Fletcher, A. C , Brown,B. B., & Dornbusch, S. M. (1995). Authoritativeparenting and adolescent adjustment: An ecologi-cal journey. In P. Moen, G. H. Elder, Jr., & K.Luscher (Eds.), Examining lives in context: Per-spectives on the ecology of human development(pp. 423-466). Washington, DC: American Psy-chological Association.

Steinberg, L., Elmen, J. D., & Mounts, N. S. (1989).Authoritative parenting, psychosocial maturity,and academic success among adolescents. ChildDevelopment, 60, 1424-1436.

Steinberg, L., Fletcher, A., & Darling, N. (1994).Parental monitoring and peer influences on adoles-cent substance use. Pediatrics, 93, 1-5.

Steinberg, L., Lamborn, S. D., Darling, N., Mounts,N. S., & Dornbush, S. M. (1994). Over-time


changes in adjustment and competence among ad-olescents from authoritative, authoritarian, indul-gent, and neglectful families. Child Development,65, 754-770.

Stoolmiller, M., Duncan, T., Bank, L., & Patterson,G. R. (1993). Some problems and solutions in thestudy of change: Significant patterns in client re-sistance. Journal of Consulting and Clinical Psy-chology, 61, 920-928.

Szapocznik, J., Kurtines, W. M., Foote, F. H., Perez-Vidal, A., & Hervis, O. (1986). Conjoint versusone-person family therapy: Some evidence for theeffectiveness of conducting family therapy throughone person. Journal of Consulting and ClinicalPsychology, 51, 881-889.

Szapocznik, J., Murray, E., Scopetta M., Hervis, O.,Rio, A., Cohen, R., Rivas-Vasquez, A., Posada, V.,& Kurtines, W. (1989). Structural family versuspsychodynamic child therapy for problematic His-panic boys. Journal of Consulting and ClinicalPsychology, 57, 571-578.

Tolan, P. H., & Loeber, R. (1993). Antisocial behav-ior. In P. H. Tolan & B. J. Cohler (Eds.), Handbook

of clinical research and practice with adolescents(pp. 307-332). New York: Wiley.

Volk, R. J., Edwards, D. W., Lewis, R. A., Sprenkle,D. H. (1989). Family systems of adolescent sub-stance abusers. Family Relations, 38, 266-272.

Wahler, R. G., & Dumas, J. (1989). Family factors inchildhood psychopathology: A coercion-neglectmodel. In T. Jacob (Ed.), Family interaction andpsychopathology (pp. 381-428). New York: Ple-num Press.

Watzlawick, P., Weakland, J. H., & Fisch, R. (1974).Change: Principles of problem formation andproblem resolution. New York: Norton.

Williams, C. L., Ben-Porath, Y. S., Uchiyama, C ,Weed, N. C , & Archer, R. P. (1990). Externalvalidity of the new Devereux Adolescent BehaviorRating scales. Journal of Personality Assessment,55(1-2), 73-85.

Received June 30, 1994Revision received October 2, 1995

Accepted October 2, 1995 •

New Editor Appointed

The Publications and Communications Board of the American Psychological Association announcesthe appointment of Kevin R. Murphy, PhD, as editor of the Journal of Applied Psychology for a six-year term beginning in 1997.

As of March 1, 1996, submit manuscripts to Kevin R. Murphy, PhD, Department of Psychology,Colorado State University, Fort Collins, CO 80523-1876.

Changes in Parenting Practices and Adolescent Drug Abuse ... · Changes in Parenting Practices and Adolescent Drug Abuse During Multidimensional Family Therapy Susan E. Schmidt, Howard

Documents