A Research Review and AlternativeHypothesis Explaining the Link Between

Learning Disability and Delinquency

Katherine A. Larson

Current hypotheses explaining the link between learning disability and delinquencyare reviewed and evaluated. Research from diverse fields is integrated and an alterna-tive hypothesis is proposed to explain the link between learning disability and delin-quency. The alternative hypothesis postulates that ineffective social cognitiveproblem-solving skills increase risk for delinquency in learning disabled youth. Futureresearch is suggested.

D elinquency is of serious concernto contemporary society (Larson,

in press). Frequency and length ofcommitment (California Youth Au-thority, 1983) as well as cost for incar-ceration of delinquents has increasedin recent years. Moreover, adjudicatedjuveniles continue to be at great riskfor becoming adult criminals (Cali-fornia Youth Authority, 1983).

It is quite clear, despite disparity inprevalence rates, that incidence oflearning disability in the delinquentpopulation is much higher than in thegeneral population. The number ofstudents being served in LD programsis 4.3% ofthe K-12 school populationaccording to 1982-1983 child countdata (Gerber, 1984). Estimates of prev-alence of learning disability amongdelinquents range from 26% (Comp-troller General of the United States,1977) to 73% (Swanstrom et al., 1977).Learning disabled (LD) youth are ad-judicated at about twice the rate asnon-leaming-disabled (NLD) youth(Broder, Dunivant, Smith, & Sutton,1981), and LD youth have greater re-cidivism and parole failure (Califor-nia Youth Authority, 1982; Larson,1986).

Consequently, development of ef-fective delinquency prevention andtreatment methods for learning dis-abled youth is critically important. Inturn, development of effective preven-tion and treatment programs is related

to knowledge of why LD youth are atincreased risk for delinquency. Thus,there is a compelling and urgent needfor empirical investigations examin-ing potential causal links betweenlearning disability and delinquency.

Prior to such research efforts, thereis an immediate need to synthesizehistorical and current data and eval-uate empirical support of existingcausal hypotheses. This paper willreview and evaluate previously pro-posed hypotheses as well as interpretresearch that suggests an alternativehypothesis.

CAUSAL EXPLANATIONSOF THE LINK BETWEEN LDAND DELINQUENCY

Factors such as anomie, social dis-organization, poverty, and racismundoubtedly increase risk for delin-quency (Shoemaker, 1984). Yet dataindicate that LD youth and youth withattention deficits are at greater risk fordelinquency than NLD peers whenfactors of age, race, and socioecono-mic status are held constant (Broder etal, 1981; Keilitz & Dunivant, 1986;Satterfield, Hoppe, & Schell, 1982).Accordingly, professionals havelooked beyond demographic variablesto explain the link between learningdisability and delinquency. Currently,there are three hypotheses that at-

tempt to explain the link betweenlearning disability and delinquency.

The School Eailure Hypothesis

This vew postulates that learningdisability leads to school failure, whichleads to a negative self-image, which inturn results in school dropout anddelinquency. This hypothesis is anextrapolation of Cohen's (1955) socio-logically based "middle-class measur-ing rod" explanation of delinquency,proposing that need to achieve inschool is adopted and aspired to by allsocial classes. Accordingly, it is hy-pothesized that youth with learningdisabilities seek out delinquent-pronepeer groups to satisfy increased needsfor recognition and achievement. Theschool failure hypothesis implies thatschool suspension and dropout in-crease an LD youth's opportunity fordelinquent behavior.

Evaluation of the School FailureHypothesis. Support for the schoolfailure hypothesis comes from clinicalobservation, school records, and testsof basic academic skills. These studiesconsistently report school failure anddropout to be a strong and persistentcorrelate of official delinquency,school misbehavior, and self-reporteddelinquent behavior (e.g.. West, 1984).Further support for this hypothesiscomes from research by Levin, Zig-mond, and Birch (1985), who foundthat LD youth have greater rates ofschool dropout than NLD peers.

Nevertheless, evidence from aca-demic remediation studies does notsupport a direct causal relationshipbetween academic achievement anddelinquency in LD youth. Keilitz andDunivant (1986) evaluated an inten-sive academic intervention programfor LD delinquents and found thatdelinquency was reduced in some sub-jects; however, changes in delinquentbehavior were not related to improve-ment in academic achievement. Theauthors suggest that interpersonal rap-port between tutor and delinquentaccounted for reduction in delinquen-cy. Larson (1986) also found academicintervention unrelated to delinquentbehavior. From a population of youth

Volume 21, Number 6. June/July 1988 357

from three correctional institutions, 60youth categorized as either learningdisabled, low achieving, or averageachieving, were matched on demo-graphic, neighborhood, and criminalhistory variables. Larson found that23% of the learning disabled, 14% ofthe low achieving, and 19% of the av-erage achieving delinquents earned ahigh school diploma while incar-cerated. The high school diploma didnot reflect level of academic achieve-ment but it did represent the amountof academic learning time a youth hadbeen provided while incarcerated. Yet,recidivism rates for the three groupswere 52%, 42%, and 27%, respectively.Furthermore, during parole, 91% ofnew criminal convictions for the LDdelinquents were for very seriouscrimes (i.e., a 2-year confinement timefor the first offense) compared to 71%of the low-achieving delinquents'new convictions.

Finally, research does not support acausal relationship between droppingout of school and delinquency. Elliottand Voss (1974) examined the correla-tion between school dropout anddelinquency in 2,000 California stu-dents and found that students whodropped out of school did indeed havehigher rates of delinquency than thosewho graduated; however, the relation-ship between dropout and delinquen-cy was reversed. That is, delinquencyrates for school dropouts peakedbefore dropping out and declined afterquitting school. Evidence from thisstudy in conjunction with the Levin etal. (1985) data suggest that poor socialadjustment, including delinquency,may account for school dropout in LDyouth. If this is the case, school failurewould be an effect and not a cause ofsocial misbehavior and delinquencyfor learning disabled youth.

The DifferentialTreatment Hypothesis

According to the differential treat-ment position, youth with learningdisabilities and nonhandicappedpeers engage in the same rate and kindof delinquent behaviors; however,police, social workers, and other of-ficials treat LD youth differently so as

to increase incidence of arrest and/oradjudication. Support for this hypo-thesis comes primarily from researchby Broder and colleagues (Broder etal, 1981; Zimmerman, Rich, Keilitz, &Broder, 1981). Broder et al. found thatLD youth were adjudicated at abouttwice the rate as NLD peers while bothgroups self-reported similar delin-quent behavior. Consequently, theresearchers postulated differentialtreatment to account for the higherincidence of adjudication in LDyouth.

Evaluation of the Differential Treat-ment Hypothesis. There are severallimitations of Broder et al. data. First,the self-report instrument and scoringin the Broder research failed to mea-sure chronicity and seriousness ofdelinquent behavior. That is, Broder etal. (1981) asked LD and NLD youth toconfirm if they had ever engaged in aspecific delinquent act. These datashowed the two groups to be similar interms of self-reported delinquent be-havior. A problem with this method isthat the two groups could appearequally delinquent even though onegroup engaged in delinquent behaviorat a much greater rate than the othergroup. Additionally, as is the difficultywith most self-report instruments, theitem content of the scale primarilymeasured less serious criminal be-haviors. Such instruments are lesscapable of differentiating populationson the basis of incarcerable delin-quent behavior (e.g., Hindelang,Hirschi, & Weis, 1979), and whenitems are aggregated into a singleglobal score, as was the case in Broderet al., the trivial items further outweighthe more serious items as an assess-ment of delinquency. Thus, the twogroups were not differentiated onseriousness of delinquent behavior.Indeed, the definition of delinquencyin this study was confounded becauseno distinction was made betweenstatus offenders and criminal of-fenders.

Recognizing the need for measuringchronicity, Zimmerman et al. (1981)examined the same subjects by askingthem to estimate how many times intheir entire life they had engaged in a

specific delinquent act. Criticisms ofthe aforementioned Broder et al. studyregarding severity apply to the Zim-merman data. Additionally, Zimmer-man et al.'s self-report data are furthersuspect because thinking and memoryproblems in subjects with learningdisabilities are likely to interfere withability to make accurate estimates oflifetime-frequency behavior. Zimmer-man et al. argued that the self-reportinstrument had adequate reliabilityand validity; however, reliability fig-ures appear not to be derived from alearning disabled population. The factthat Zimmerman et al. found that LDyouth reported significantly/ewer de-linquent acts than NLD peers, tends toincrease skepticism in the reliability ofthis LD self-report data. It seemsunreasonable that a population ofyouth consistently found to be lesssocially effective, more impulsive, andfrequently more aggressive shouldcommit fewer delinquent acts thannonhandicapped peers.

Broder et al. (1981) and Zimmer-man et al. (1981) proposed the dif-ferential treatment hypothesis to ex-plain the apparent confiict betweenhigh LD official adjudication ratesand low LD self-report rates of delin-quent behavior. Such disparate find-ings may more appropriately beexplained as a function of the self-report instrument and scoring.

Other evidence suggests that LDyouth behave differently from non-handicapped peers in terms of delin-quent behavior. Keilitz and Dunivant(1986), in a study of 1,900 youths,found that LD youth self-reportedsignificantly greater rates of delin-quency and violent acts than matchedNLD peers. Official arrest and adju-dication rates were also higher for theLD adolescents. An 11-year follow-upstudy of 2,700 paroled delinquents(California Youth Authority, 1982)found that characteristics associatedwith a learning disability, such assevere low achievement in readingand math, and reduced mental ap-titude and evidence of neurologicalabnormality, significantly differen-tiated young parolees on chronicityand degree of violent behavior.

Furthermore, the differential treat-

358 Journal of Learning Disabilities

ment hypothesis is contradicted byZimmerman et al.'s (1981) own re-search, which found that, after adju-dication, LD and NLD delinquentshad equal probability of incarcera-tion. Similarly, Keilitz and Dunivant(1986) found that youth with learningdisability were no more likely thannonhandicapped peers to receive se-vere disposition from the court. Lar-son (1985) also found no significantdifferences between age of first con-finement and duration of confinementin LD and low-achieving delinquents.

Some have proposed that differen-tial treatment takes place at the arrestphase prior to adjudication or disposi-tion procedures (Keilitz & Dunivant,1986). This may be the case; however,other research finds that arrest issignificantly correlated with frequen-cy and severity of delinquent behavior(e.g., Blumstein & Cohen, 1976). Thus,higher rates of arrest and adjudicationfor youth with learning disabilitiesmay more directly reflect frequent andserious delinquent behavior. In sup-port of this conclusion, Keilitz andDunivant (1986) found a positive cor-relation between LD self-report, adju-dication, ana arrest; this corroboratesother juvenile criminology researchdemonstrating that self-report ratesand seriousness of delinquent be-havior strongly correlate with gettingcaught and being known by the police(e.g., Friday & Stewart, 1977). Refuta-tion of the differential treatment hy-pothesis is further presented by Keilitzand Dunivant (1986), who found in alongitudinal study that LD boys weremore likely than NLD peers to be-come delinquent as they grow older.

The Susceptihility Hypothesis

This position contends that learn-ing disabilities are frequently accom-panied by "a variety of socially trou-blesome personality characteristics"(Murray, 1976, p. 26). This hypothesiscan be interpreted from two distinctperspectives.

Evaluation of the Susceptibility Hy-pothesis. One interpretation takes theposition that social skill is a responsepredisposition or underlying per-

sonality trait. Thus, observable be-havior is only a refiection of "amount"of social skillfulness. Youth withlearning disabilities are seen as atincreased risk for delinquency be-cause they are "low" in social skill-fulness.

This approach is a tautological useofthe concept of social skill. That is, ifa learning disabled youth does notbecome delinquent it is inferred thatthe youth was in possession ofa satis-factory level of social skill; simultan-eously it is explained that the youth'snondelinquent behavior occurred be-cause the youth had a satisfactory levelof social skill. McFall (1982) has point-ed out that an individual's "inferredlevel of social skill cannot be used inthis way to account for the quality ofthe very behavior from which the skillswere inferred" (p. 3).

The alternative interpretation ofthesusceptibility hypothesis takes theposition that negative social-person-ality characteristics (e.g., impulsivity,poor reception to social cues) arespecific social characteristics that in-crease the likelihood of delinquentbehavior. This perspective avoids theproblems of an underlying trait con-cept of social skill; however, an impor-tant criticism of this hypothesis is thatit fails to address the issue of whatcaused the inappropriate social char-acteristics in the first place. That is,sources other than learning dis-abilities produce negative social char-acteristics that might increase suscep-tibility to delinquency. Additionally,and significantly, this interpretation ofthe susceptibility hypothesis lacks aunified framework for explaining howa variety of socially ineffective be-haviors can be accounted for by anunderlying learning disability.

An Alternative Hypothesis

As reviewed, a number of pro-fessionals have speculated that learn-ing disability contributes to increasedrisk for delinquent behavior, and whileit seems clear that a correlational asso-ciation between learning disability anddelinquency has been demonstrated,data supporting a causal relationshiphave not been generated by research.

The school failure and differentialtreatment hypotheses are not support-ed by empirical evidence. The suscep-tibility hypothesis lacks empiricaltesting, and the concept of suscep-tibility appears to be too global to besystematically testable. Clearly, analternative hypothesis is warranted toexplain the link between learning dis-ability and delinquency.

One approach to developing analternative causal hypothesis wouldbe to identify a specific skill related tosocial adjustment for which evidenceexists that (a) differences or deficits inthe skill are empirically associatedwith delinquent behavior, (b) sociallymaladjusted LD youth are highly like-ly to be ineffective in the skill, and (c)the skill identified must be shown tomediate various behavioral responsesacross different social situations (toaccount for a variety of socially inef-fective behaviors, any of which mightbe a delinquent act or related to adelinquent episode).

One skill that may fulfill theserequirements is social cognitive prob-lem solving. D'Zurilla and Goldfried(1969) conceptualized problem solv-ing as both a self-control procedureand a learning process involvingcognitive strategies. Social cognitiveproblem-solving skills are hypothe-sized as general skills applicable to avariety of situations and potentiallyuseful for increasing generalization ofsocially competent responses.

The following three sections willreview relevant problem-solving re-search suggesting that difficulties insocial cognitive problem-solving skillsare associated with social maladjust-ment, that they mediate social com-petence, and that learning disabledyouth and delinquent youth are morelikely to have difficulties with theseskills than are nonhandicapped andnondelinquent peers.

SOCIAL COGNITIVEINEFFECTIVENESS ANDSOCIAL MALADJUSTMENT

There has been considerable sup-port for the hypothesis that ineffectivesocial cognitive problem-solving skills

Volume 21, Number 6. June/July 1988 359

lead to emotional and behavioral dis-orders (e.g., D'Zurilla & Goldfried,1969). Individuals whose social prob-lem-solving processes are characteris-tically ineffective are more likely to beviewed as maladjusted, socially in-competent, or abnormal.

Spivack, Platt, and Shure (1976)found drug addict, psychiatric, and"poorly self-regulated" adolescentpopulations to be less effective than acomparison group of socially compe-tent peers on cognitive social problemsolving-skills such as weighing prosand cons, generating options, concep-tualizing a step-by-step process to agoal, and perceiving the situation fromanother's perspective. Furthermore,Spivack et al. found that ineffectivecognitive social problem-solving skillswere consistently manifested in social-ly incompetent populations of all ageswhen IQ and verbal fiuency wereheld constant.

NONSOCIAL COGNITIVEPROBLEM-SOLVINGDIFnCULTIES ANDLEARNING DISABILITIES

It has been consistently reportedthat, when faced with nonsocial tasks,youth with learning disabilities fre-quently exhibit profound difficultiesin many cognitive problem-solvingfunctions. Learning disabled indi-viduals are thought to fail many tasksnot only because of basic knowledgedeficits but also because they fail tospontaneously employ appropriateproblem-solving skills (e.g., Reid &Hresko, 1981; Torgesen, 1982). In-deed, some believe that ineffectiveproblem-solving abilities underlielearning disabilities. Difficulties insuch skills include identifying rele-vant variables, impulse control, eval-uation of possible strategies, per-sistence, and self-monitoring (e.g.,Forness & Esveldt, 1975; Hagen, Bar-clay, & Newman, 1982).

SOCIAL COGNITIVEPROBLEM-SOLVINGDIFFICULTIES ANDLEARNING DISABILITIES

Given that youth with learning dis-abilities manifest cognitive inefficien-

cies in nonsocial tasks, and con-sidering the frequently noted socialdifficulties of LD youth, it is reason-able to speculate that many LD youthwill manifest ineffective cognitiveproblem-solving skills when facingsocial tasks.

Indeed, investigators have pointedout that youth with learning dis-abilities have difficulties, compared tononhandicapped peers, in a variety ofcognitive social problem-solving skillssuch as interpreting the mood or com-munications of others (e.g., Lerner,1981), generating and choosing qualitysocial solutions (e.g.. Hazel, Schu-maker, Sherman, & Sheldon, 1982),predicting and evaluating social con-sequences (e.g., Bruno, 1981; Hazel etal, 1982), interpreting situations (e.g..Pearl & Cosden, 1982), taking theperspective of others (e.g., Bachara,1976), and controlling impulsive be-havior (e.g., Hallahan, Kauffman, &Lloyd, 1985). Schumaker and hercolleagues (Hazel et al., 1982; Schu-maker, Hazel, Sherman, & Sheldon,1982) found that NLD adolescentsscored significantly better on socialcognitive problem solving than LDand delinquent youth, who scoredsimilarly. Schumaker et al. did reportthat the LD group was heterogeneousin terms of social cognitive problem-solving ability, with some LD youthperforming as well as NLD peers.Unfortunately, because the LD youthwere not assessed on social adjust-ment, it is not possible to determinefrom this study how much of thevariance in social adjustment wasaccounted for by skill in social prob-lem solving.

SOCIAL COGNITIVEPROBLEM-SOLVINGDIFFICULTIES ANDDELINQUENCY

Delinquent youth, a heterogeneouspopulation including many youthwho fit clinical descriptions of learn-ing disability (Prout, 1981), are consis-tently found to be less skillful in avariety of social cognitive problem-solving skills when compared to non-delinquents. Poor perspective taking(e.g.. Chandler, 1973; Little, 1979), poorimpulse control (e.g.. White, 1965), and

inability to generate multiple andeffective solutions (e.g., Larson, 1985;Freedman, Rosenthal, Donahoe, &Schludt, 1978) are three specific prob-lem-solving difficulties historicallydemonstrated in delinquent popula-tions.

Researchers have documented othersocial cognitive problem-solving skilldifficulties in delinquent youth. Forexample, after testing 1,173 institu-tionalized delinquents, using the Cali-fornia I-level topology (interraterreliability was .88), Jesness (1971)reported that 54% of the delinquentshad difficulty in (a) understandingthat behavior has something to dowith getting what one wants (i.e.,general orientation to plan); (b) es-timating differences among othersand between self and others andunderstanding needs, feelings, andmotives of others (i.e., perspective tak-ing); (c) planning for the future beforeacting (i.e., impulse control and prob-lem formulation); and (d) predictingaccurately others' responses to theirown actions (i.e., predicting conse-quences and decision making). Schu-maker et al. (1982) found both LDyouth and delinquents less skilledthan normal peers on the socialproblem-solving skills of identifyingthe problem, generating solutions,evaluating consequences, and select-ing an adaptive solution. Larson(1985) found that nearly all socialsolutions generated by a randomly se-lected group of delinquents were inef-fective as judged by independentraters using a criterion-referencedrater's manual.

Additionally, self-assessment ofproblem-solving ability has been pos-tulated by Ravell (1979) as an essen-tial skill for effective problem solving.Larson (1985) found that incarcerateddelinquents who were learning dis-abled or low achieving were unable toaccurately assess or predict their com-petence in generating effective solu-tions to social problems.

MEDIATIONAL CAPACITYOF SOCIALCOGNITIVE SKILLS

As previously argued, to support aspecific skill deficit as a causal factor

360 Journal of Learning Disabilities

for increasing delinquent behavior, itis necessary to demonstrate the medi-ational capacity of that skill on overtsocial behavior and ultimately ondelinquent behavior. Currently, thereis a paucity of research testing themediational capacity of social prob-lem-solving skills on overt socialbehavior as directly measured in sam-ples of learning disabled or delinquentyouth.

Nevertheless, existing studies sug-gest a mediational function of socialcognitive problem-solving skill onovert social behavior as measureddirectly. For example, Feindler, Mar-riott, and Iwata (1984) trained 36acting-out adjudicated youth in self-regulation and cognitive problem solv-ing. Daily disruptive behavior andaggressive incidents in school weresignificantly reduced compared to notreatment control. Snyder and White(1979) found cognitive self-instructiontraining, compared to contingencyawareness and no treatment control,resulted in significant improvement indaily living social skills of incar-cerated delinquents. Larson and Ger-ber (1987) found that social cognitivetraining significantly enhanced so-cially relevant overt social behavior inLD and low-achieving incarceratedyouth compared to matched controls.Moreover, in this study, overt behaviorchanges were positively correlatedwith changes in cognitive problem-solving skill.

Fvaluation of the Alternative Hypo-thesis. Evidence indicates that socialcognitive problem-solving difficultiesare associated with social maladjust-ment and specifically with delinquen-cy. Youth with learning disabilitiesappear to be at greater risk than non-handicapped peers for cognitive dif-ficulties when faced with both socialand nonsocial problems. Moreover,delinquent youth and youth withlearning disabilities are found toexhibit similar difficulties with socialproblem-solving skills. Both groupsare reported to be ineffective in skillssuch as perspective taking, impulsecontrol, defining the problem, gener-ating multiple and effective solutions,predicting consequences, and under-

standing and using relevant socialcues. Lastly, and importantly, socialcognitive problem-solving skills ap-pear to mediate overt social behavior.

Given this evidence, an alternativeexplanation ofthe link between learn-ing disability and delinquency ap-pears reasonable although in need ofclear empirical support. Specifically,youth with learning disabilities maybe at increased risk for delinquencybecause they are more likely to be inef-fective in social cognitive problem-solving skills.

It is not clear from the currentresearch how deficits in social prob-lem solving might increase risk fordelinquent responses. One explana-tion is that deficiencies in social prob-lem solving impair ability to controlimpulsive responses, assess socialproblem variables, generate effectivesocial strategies, and regulate, throughself-monitoring, on-going social inter-action. Many researchers (e.g., Kol-ligian & Sternberg, 1987) would labeldeficits in these skills as a deficiencyor inefficiency in metacognition.

The proposal that ineffective meta-cognition impairs functioning in LDyouth is not new. Researchers havehypothesized that academic learningdifficulties of LD youth may be theresult of ineffective metacognitiveskills (e.g.. Baker, 1982; Loper, 1980;Wong, 1979). Several researchers (e.g..Brown, Bransford, Ferrara, & Cam-pione, 1983; Caranaugh & Perlmutter,1982) have pointed out the difficulty ofand need for clarifying the concept ofmetacognition. Nevertheless, recentresearch by Slife, Weiss, and Bell(1985) upholds the potential utility ofthe concept for describing learningdisabled populations in terms ofknowledge about and regulation ofcognition.

SOCIAL METACOGNITION

Metacognitive skills have been de-fined as superordinate skills requiringknowledge about how to know andknowledge about one's own cognitiveprocessing (Brown, 1975; Flavell &Wellman, 1977). Researchers such asFlavell, Wellman, and Brown haveonly alluded to metacognitive skills in

terms of social problem solving, al-though application of metacognitiveawareness and metacognitive controlconcepts to the social domain seemsreasonable and useful.

The awareness aspect of metacogni-tion has been defined to include anawareness of person variables (i.e.,self-appraisal of personal attributesthat affect the task) and awareness oftask variables (i.e., knowledge of pa-rameters, expectations, and situationattributes that affect the task). Accord-ingly, metacognitive social awarenessskill can be conceived as the ability toidentify and differentiate relevantsocial cues about oneself and others aswell as about the social situation ingeneral. Delinquent youth and youthwith learning disabilities appear tohave difficulties with this type of socialmetacognition. As previously re-viewed, both groups of youth havebeen found to be ineffective in com-prehending and using relevant socialvariables, in being aware of another'sperspective, and in accurately evaluat-ing their own social problem-solvingcompetence.

The control aspect of metacognitionhas been defined to include skills thatregulate efficient selection and on-going application of effective behav-iors. Accordingly, metacognitive so-cial control skill can be defined as theability to control impulsive responses,define the problem, generate appro-priate solutions, evaluate conse-quences, and monitor performance.As previously reviewed, researchershave consistently found delinquentyouth less effective in these kinds ofsocial cognitive skills compared tonondelinquents.

CURRENT AND FUTURERESEARCH TESTING THEALTERNATIVEHYPOTHESIS

Research is needed to systemati-cally define and specify social meta-awareness and social meta-controlskills. Thus far, there has been almostno effort to specifically identify anddescribe social knowledge and skillthat fulfills the meta-awareness func-tion. In a social situation this would

Volume 21. Number 6, June/July 1988 361

include knowledge of variables aboutself, about others, about social norms,or about the social environment. Re-search examining this hypothesis willhave to determine if there are socialvariables consistently relevant to avariety of social situations and ifknowledge of these variables en-hances general social performance.Preliminary evidence (Larson, 1983)indicates that there may be somesocial variables that socially compe-tent adolescents perceive as signifi-cant for solving a variety of socialproblems. Larson asked "sociallycompetent" high school students andmatched incarcerated delinquents toname factors (i.e., about themselves,about others, and about the situationin general) they felt were important tofind out about or know about whentrying to solve a social problem. Therewere statistically significant qualita-tive as well as quantitative differencesbetween delinquent and socially com-petent adolescent responses regardingsituational, self, and other variablesidentified as important to social prob-lem solving.

Not only will it be necessary to de-scribe social meta-awareness varia-bles; it will also be essential to under-stand how social meta-awarenessskills interact with social meta-controlskills. That is, effective problem solv-ing has been hypothesized (Hagen etal., 1982) as based upon a state ofacquired knowledge (i.e., awareness)as well as a process of using thatknowledge (i.e., control). Accordingly,it may be that skill in both meta-awareness and meta-control functionsis necessary for effective social func-tioning because of the possible recip-rocal nature of these knowledge areas.

In this regard, while working withLD delinquents, Larson found thatmeta-awareness of a social variablewas insufficient for enhancing themeta-control skill of solution genera-tion. The delinquents simply did notknow how to "use" the awarenessknowledge. For example, "emotionallevel" was one variable taught asimportant to recognize and interpretwhen facing a social problem situa-tion. In previous research this variablehad been identified by socially compe-

tent adolescents as important for solv-ing social problems. Delinquents withlearning disabilities learned to spon-taneously recall this as an importantproblem-solving variable. However,during simulation problem-solvingexperiences, determination of emo-tional level as "high" or "low" had nodifferential affect on cognitive controlskills in terms of what delinquentssuggested the solution or strategyshould be. Delinquents also failed tomodify their solutions on the basis ofmeta-awareness of other variablessuch as confidence level, expertise,number of people involved, relation-ship to the other, and so forth. Findingthat delinquents had difficulty in usingmeta-awareness knowledge suggeststhat researchers need to examinewhether delinquent youth and learn-ing disabled youth are merely una-ware of relevant social meta-variablesor whether they are unable to use theknowledge that awareness of the vari-able brings in regulating solutiongeneration.

Other lines of research must providea more direct test of the hypothesisthat social metacognitive ineffective-ness increases risk for delinquency inLD youth. Studies are needed todemonstrate that social metacognitivetraining enhances overt social be-havior in delinquent and LD youth. Inthese studies it will be important todocument that enhancement of socialbehavior is positively correlated withchanges in metacognitive skills.

Evidence suggests that cognitivetraining needs to train both meta-awareness and meta-control skills aswell as how to use these skills in areciprocal way. Previous social prob-lem-solving studies have failed to trainmeta-awareness skills; instead train-ing has focused exclusively uponmeta-control skills. Yet training effectshave been weak for changing overtsocial behavior or enhancing socialbehavior in novel contexts. Larsonand Gerber (1987) trained both meta-awareness and meta-control skills andfound significant positive behaviorchanges.

Demonstrating the mediational ca-pacity of social metacognition intraining studies is insufficient, how-

ever, for concluding that social me-tacognitive ineffectiveness increasesrisk for delinquency. That is, causalinfiuences cannot be inferred solelyfrom successful treatment. In order tofully support the proposed hypothesis,data will be needed that demonstratethat delinquents actually manifestinefficiencies in social metacognitioncompared to nondelinquents and thatsocially maladjusted and adjustedyouth with learning disabilities differon social metacognitive problem-solv-ing abilities.

Lastly, there is a need to examinethe relationship of specific socialmetacognitive skills to overt behavior.For example, a components analysisstudy assessing and comparing thepotential effectiveness of training im-pulse control skills, meta-awarenessskills, and meta-control skills wouldhelp to elucidate the relative impor-tance of these specific cognitive skillsto delinquent behavior.

This paper has attempted to inte-grate existing research data for thepurpose of developing a plausiblecausal hypothesis elucidating a func-tional relationship between learningdisability and delinquency. Testingrelationships within the hypothesiswill require extensive and systematicresearch. It is hoped that issues andresearch questions discussed in thispaper will encourage future researchefforts addressing this important issuein contemporary society.

ACKNOWLEDGMENTS

I wish to thank Melvyn Semmel. Michael Gerber.and Joan Lieber for their comments on a draft of thispaper.

ABOUT THE AUTHOR

Katherine A. Larson is a researcher in the SpecialEducation Research Laboratories, Graduate Schoolof Education at the University of California. SantaBarbara. Current research interests include delin-quency prevention, social skill training for be-haviorally disordered adolescents, and dropout pre-vention for inner city youth. Dr. Larson is currentlyprinciple investigator for PROJECT SELE-STARTtesting the efficacy of social problem-solving trainingfor enhancing parole adjustment in LD delinquents.Dr. Larson is also principal investigator for PRO-JECT MAIN STREET a 3-year intervention pro-gram aimed at reducing inappropriate referral tospecial education or school dropout in high-risk

362 Journal of Learning Disabilities

young adolescents. Address: Katherine A. Larson.University of California, Graduate School of Educa-tion. Special Education Program. Santa Barbara,CA 93106.

REEERENCES

Bachara. G.H. (1976). Empathy in learning dis-abled children. Perceptual and Motor Skills. 43,541-542.

Baker. L. (1982). An evaluation of the role ofmetacognitive deficits in learning disabilities.Topics in Learning and Learning Disabilities, 2(1), 27-35.

Blumstein. A.. & Cohen. A. W. (1976). Estimation ofindividual crime rates from arrest records. A Jour-nal of Criminal Law and Criminology, 70, 561-585.

Broder, P.K.. Dunivant. N. Smith. EC. & Sutton.P.L. (1981). Eurther observations on the link be-tween learning disabilities and juvenile delin-quency. Journat of Educational Psychology. 73.838-850.

Brown. A.L (1975). The development of memory inknowing about knowing, and knowing how toknow. In H.W. Reese (Ed.), Advances in childdevelopment and behavior (Vol. 10. pp. 103-152).New York: Academic Press.

Brown, A.L. Bransford, J.D.. Eerrara. ILA.. & Cam-pione, J.C (1983). Learning, remembering, andunderstanding. In J.H. Elavell & E.M. Markman(Eds.). Handbook of child psychology (4th ed.).Cognitive development (Vol. 3. pp. 420-494). NewYork: Wiley.

Bruno. R.M. (1981). Interpretation of pictorially pre-sented social strategies in learning disabled andnormal children. Journal of Learning Dis-abilities. 14. 350-352.

California Youth Authority. (1982). Early iden-tification ofthe chronic offender (Tech. Rep.). Sac-ramento: Author.

California Youth Authority. (1983). Annual report:Program description and statistical summary(Tech. Rep.). Sacramento: Author.

Caranaugh. J.C, & Perlmutter. M. (1982). Meta-memory: A critical examination. Child Develop-ment, 53. 11-28.

Chandler. M. (1973). Egocentrism and antisocialbehavior: Assessment and training of social per-spective-taking skills. Developmental Psychology.9. 326-332.

Cohen. A.K. (1955). Delinquent boys. New York:Eree Press.

Comptroller General of the United States. (1977).Learning disabilities: The link to delinquencyshould be determined. Washington. DC: GeneralAccounting Office.

D'Zurilla. TJ.. & Goldfried. M.R. (1969). A be-havioral-analytic model for assessing compe-tence. In CD. Spielberger (Ed.). Current topics inclinical and community psychology (Vol. I, pp.21-40). New York: Academic Press.

Elliott, D.S., & Voss. H.L (1974). Delinquency anddropout. Lexington. MA: D.C. Heath.

Eeindler, E.. Marriott. SA., & Iwata, M. (1984).Group anger control training for junior highschool delinquents. Cognitive Therapy & Re-search. 8(3), 299-311.

Elavell. J.H. (1979). Metacognition and cognitive

monitoring. American Psychologist, 34. 906-911.

Elavell. J.H.. & Wellman, H.M. (1977). Meta-memory. In R.V. Kail & J.W. Hagen (Eds.),Perspectives on the development of memory andcognition (pp. 3-33). Hillsdale. NJ: Erlbaum.

Eorness, S.. & Esveldt. K. (1975). Classroom obser-vation of children with learning and behaviorproblems. Journal of Learning Disabilities, 8,382-385.

Ereedman. B.J.. Rosenthal, L. Donahoe. CP.,Schludt. D.G.. & McEall, R.M. (1978). A socio-behavioral analysis of skill deficits in delinquentand nondelinquent adolescent boys. Journal ofConsulting and Clinical Psychology, 46. 1448-1462.

Eriday P.C, & Stewart. V.L. (Eds.). (1977). Youth,crime and juvenile justice: International perspec-tives. New York: Praeger.

Gerber. M.M. (1984). The department of education'ssixth annual report to congress on P.L 94-142: Iscongress getting a full story? Exceptional Chil-dren. 51(3), 209-224.

Hagen, J W., Barclay. CR. & Newman. R.S. (1982).Metacognition, self-knowledge, and learning dis-abilities: Some thoughts on knowing and doing.Topics in Learning and Learning Disabilities. 2,19-26.

Hallahan, D.P.. Kauffman, J.M., & Lloyd J.W.(1985). Introduction to learning disabilities (2nded.). Englewood Cliffs: Prentice-Hall

Hazel. J.S, Schumaker, J.B.. Sherman. JA.. &Sheldon, J. (1982). Application ofa group trainingprogram in social skills and problem solving tolearning disabled and non-leaming disabledyouth. Learning Disability Quarterly. 5. 398-408.

Hindelang. MJ.. Hirschi. T, & Weis, J.G. (1979).Correlates of delinquency: The illusion of dis-crepancy between self-report and official mea-sures. American Sociological Review, 44(6).995-1014.

Jesness. CE. (1971). The Preston typology study. Anexperiment with differential treatment in aninstitution. Journal of Research in Crime andDelinquency. 8. 38-52.

Keilitz. I. & Dunivant. N (1986). The relationshipbetween learning disability and juvenile delin-quency: Current state of knowledge. Remedialand Special Education, 7(3), 18-26.

KoUigian, J.. & Sternberg. RJ. (1987). Intelligence,information processing, and specific learning dis-abilities: A triarchic synthesis. Journal of Learn-ing Disabilities. 20(1). 8-17.

Larson, KA. (1983). Proximity matrix analysis ofsocial metacognition in delinquent and non-delinquent adolescents. Paper presented atCANHC-ACLD State Conference, Oakland.CA.

Larson. KA. (1985). Effectiveness of metacognitivetraining for social adjustment in learning dis-abled and non-learning-disabled delinquents.Dissertation Abstracts International. 46(04).948-A.

Larson. KA. (1986). Success on parole: A com-parison study ofLD and low-achieving youthfuloffenders. Paper presented at American Educa-tional Research Association National Conference.San Erancisco.

Larson, KA. (in press). Special education: Delin-quent and adjudicated youth. Internationalencyclopedia of education research and studiessupplement (Vol 1). Oxford. England: Per-gamon Books.

Larson. KA., & Gerber, M.M. (1987). Effects ofsocial metacognitive training for enhancing overtbehavior in learning disabled and low-achievingdelinquents. Exceptional Children, 54(3), 201-211.

Lerner. J.W. (1981). Learning disabilities: Theories,diagnosis, and teaching strategies (3rd ed.).Dallas: Houghton Miffiin.

Levin, £., Zigmond N, & Birch. J (1985). A follow-up study of 52 learning disabled adolescents.Journal of Learning Disabilities. 18(1), 2-7.

Little, V.L (1979). The relationship of role-takingability to self control in institutionalized juvenileoffenders. Dissertation Abstracts International,39, 2292B.

Loper. A.B. (1980). Metacognitive developmentimplications for cognitive training. ExceptionalEducation Quarterly. 1(1), 1-8.

McEall. R.M. (1982). A review and reformation ofthe concept of social skills. Behavioral Assess-ment, 4. 1-33.

Murray. CA. (1976). The link between leamingdis-abilities and juvenile delinquency: Current theoryand knowledge. Washington. DC: U.S. Govern-ment Printing Office.

Pearl. R.. & Cosden, M. (1982). Sizing up a situation:LD children's understanding of social interac-tions. Learning Disability Quarterly, 5. 371-373.

Prout. J. (1981). The incidence of suspected excep-tional educational needs among youth in juvenilecorrectional facilities. Journal of CorrectionalEducation, 32. 22-34.

Reid D.K. & Hresko. W.P. (1981). A cognitiveapproach to ieaming disabilities. New York:McGraw-Hill

Satterfield. J.H. Hoppe. CM.. & Scheil, A.M. (1982).A prospective study of delinquency in 110 adoles-cent boys with attention deficit disorders and 88normal boys. American Journal of Psychiatry. 6.795-798

Schumaker. J.B.. Hazel, J.S.. Sherman, JA.. &Sheldon. J. (1982). Social skill performance oflearning disabled, non-leaming disabled, anddelinquent adolescents. Learning DisabilityQuarterly. 5, 388-397.

Shoemaker. D.J.. (1984). Theories of delinquency.New York: Oxford University Press.

Slife. B.D.. Weiss, J. & Bell, T. (1985). Separabilityof metacognition and cognition: Problem solvingin learning disabled and regular students. Journalof Educational Psychology. 77(4). 437-445.

Snyder, J. & White, M. (1979). The use of cognitiveself-instruction in the treatment of behaviorallydisturbed adolescents. Behavior Therapy. 10,227-235.

Spivack. G. Platt, J, & Shure. N.B. (1976). Theproblem-solving approach to adjustment. SanFrancisco: Jossey-Bass.

Swanstrom, W.. Randle. C, Livingston. V, Mac-rqfic. K, Canfield M.. AAmold S. (1977). BlueEarth/Nicollet Counties prevalent study. Unpub-

(continued on p. 369)

Volume 21, Number 6, June/July 1988 363

Raskind, M.H. (1983). Non-verbal communicationsignals in behavior-disordered and non-disor-dered learning disabled boys and non-learningdisabled boys. Learning Disability Quarterly,6(1). 12-19.

Rutter. Af., Tizard. J.. & Whitmore. K. (1970).Education, health and behavior. London. Eng-land: Longman.

Saloner. M.. & Gettinger. M. (1975). Sociat inferenceskills in learning disabled and non-disabledchildren. Psychology in the Schools, 22(2). 201-207.

Sandberg, S.T., Rutter. A/., * Taylor. E. (1978).Hyperkinetic disorder in psychiatric clinic atten-ders. Developmental Medicine and Child Neurol-ogy, 20. 279-299.

Schachar. R.. Rutter. M.. & Smith, A. (1981). Thecharacteristics of situationally and pervasivelyhyperactive children: Implications for syndromedefinition. Journal of Child Psychology and Psy-chiatry. 22. 375-392.

Siperstein, G.N. Bopp. MJ. & Bak, J.J (1977). Thesocial status of learning disabled children. Cam-bridge. MA: Research Institute of Educational

Problems.Taylor. W.L (1953). Cloze procedure: A newtoolfor

measuring readability. Journalism Quarterly. 30.415-433.

Thompson. M. (1978. October). Stress reaction todyslexia. Paper presented at International Con-ference on the Child Under Stress. Monte Carlo.Monaco.

Wender. P.H. (1971). Minimal brain dysfunction inchildren. New York: Wiley.

Continued from p. 363

lished manuscript. Rochester, MN: CorrectionalLearning Disabilities Research Project

Torgesen. J.K. (1982). The leamingdisabled child asan inactive learner: Educational implications.Topics in learning and Learning Disabilities. 2,41-45.

West. W.G.(1984). Young offenders and the state: ACanadian perspective in delinquency. Toronto:

Butterworths Press.White. SH. (1965). Evidence for a hierarchical

arrangement of learning processes. In L Lisett &C. Spiber (Eds.), Advances in child developmentand behavior (Vol. 2). New York: AcademicPress.

Wong, B.L. (1979). The role of theory in learningdis-abilities research: Part II A selective review of

current theories of learning and reading dis-abilities. Journal of Learning Disabilities, 12(10).15-24

Zimmerman, J, Rich. W.D.. Keilitz. /., & Broder.P.K. (1981). Some observations on the link be-tween learning disabilities and juvenile delin-quency. Journal of Criminal Justice. 9, 1-17.

(of

How your students^ can learn to

controltheir own

behavior!Proven successful in research,

classroom and home environmenfs.Reduces disrupfive and sfressful behavior. Improvesacademic achievement, self-esfeem, physical healfhand concenfration abilify. Endorsed by fhePennsylania School Counselors Associafion. Basedupon nationally esfablished goals for educafion.

PLUS Order now andreceive FREE wifh yourorder an Adult Relaxaflonfape with complefeprogram overview. A$12.95 value.

• UmM^^ PA Residents• • ^ F a d d 6% Tax

plus Shipping and Handling.

inciuded:30 Minute Video Tape120 Minutes of AudioinstructionCompiete instructor'sManualAccess to individuaiTraining Programsand ConsultingServices.

[80

N T Minner spaces^RO. Box 81267, Pittsburgh, PA 15217 • (800) 835-2246 (ext. 404)

1988, Nancy Mromor M.Ed., Licensed Psychologist

Volume 21, Number 6, June/July 1988 369