Arrested development: The effects of incarceration on the … · 2017-06-17 · Arrested development: The effects of incarceration on the development of psychosocial maturity JULIA

Arrested development: The effects of incarcerationon the development of psychosocial maturity

JULIA DMITRIEVA,a KATHRYN C. MONAHAN,b ELIZABETH CAUFFMAN,c AND LAURENCE STEINBERGd

aUniversity of Denver; bUniversity of Pittsburgh; cUniversity of California, Irvine; and dTemple University

Abstract

Improvements in temperance, perspective, and responsibility are a part of typical development of psychosocial maturity during adolescence. The existingliterature suggests that the developmental course of psychosocial maturity is influenced by normative variations in social context, but little is known about howatypical contexts, such as incarceration, influence its development. The study investigates how the development of psychosocial maturity is affected byincarceration, using data from a 7-year longitudinal study of 1,171 adolescent males. We compared the effects of confinement in juvenile facilities with varyingdegrees of focus on incarceration versus rehabilitation (i.e., secure vs. residential treatment facilities) and tested whether facility quality and age atincarceration moderate the effect of incarceration on psychosocial maturity. The results indicate that incarceration in a secure setting, but not a residentialtreatment facility, is associated with a short-term decline in temperance and responsibility. The total amount of time incarcerated in a residentialtreatment facility, but not a secure setting, had a negative effect on the developmental trajectory of psychosocial maturity. Age at incarceration, but not thefacility quality, moderated the effect of recent incarceration: older youths were more susceptible to short-term negative effects of recent incarceration in asecure setting, but they also benefited more than younger participants from short-term positive effects of incarceration in a residential treatment setting.Furthermore, youths who perceived their incarceration setting as unsafe evinced a decline in temperance. Future research and policy implications are discussed.

Adolescence is marked by increases in psychosocial maturitythat are reflected in improvements in temperance (the abilityto curb impulsive and aggressive behavior), perspective (theability to see things from multiple temporal and social van-tage points), and responsibility (the ability to function auton-omously; Steinberg & Cauffman, 1996). On average, temper-ance, perspective, and responsibility develop steadily acrossthe adolescent years, with growth slowing as youth transitioninto early adulthood (Monahan, Steinberg, Cauffman, &Mulvey, 2009, 2011). Evidence also suggests that there is sig-nificant variability in the degree and rate of development ofpsychosocial maturity across adolescence and young adult-hood, with some youth showing greater or faster gains andothers exhibiting little or very slow growth in maturity overtime (Monahan et al., 2009, 2011).

The extant literature on the development of psychosocialmaturity suggests that normative variations in social contextmay contribute to individual differences in the developmentof temperance, perspective, and responsibility. For example,exposure to positive parent–child relationships, prosocialpeers, and nurturing school and extracurricular contexts areeach linked to greater psychosocial maturity during the adoles-cent years (Steinberg, Chung, & Little, 2004). Less is knownabout how normative development proceeds in atypicalcontexts, a central question within the study of developmental

psychopathology. The current study investigates how the de-velopment of psychosocial maturity is affected by incarcera-tion, a highly atypical context. Studies of normative psychoso-cial development among incarcerated young people are almostnonexistent, even though this is not an uncommon experienceamong poor, ethnic minority male adolescents in America (Na-tional Council on Crime and Delinquency, 2007).

The juvenile justice system handles more than 1.6 milliondelinquency cases each year (Snyder & Sickmund, 2006).Annually, approximately 72,000 youth are incarcerated inthe United States (Sickmund, Sladky, Kang, & Puzzanchera,2008). Given the existing research on contextual contributorsto positive psychosocial development, there is good reason tobelieve that incarceration has a lasting, adverse impact on thedeveloping adolescent. Although the challenges and stressesof incarceration would tax the faculties of even those mostadept at coping with adversity, adolescent offenders facethis situation with the added disadvantage of immaturity. In-carceration separates youth from their parents at a time in de-velopment when youths’ well-being (Helsen, Vollebergh, &Meeus, 2000) and acquisition of coping skills (Patterson &McCubbin, 1987) are still influenced heavily by the family.Simultaneously, adolescents perceive their relationshipswith friends to be increasingly important (Helsen et al.,2000), and incarcerated adolescents are placed in an environ-ment where their only peer group is composed of other anti-social youth. Compounding the social isolation from typicalcontexts is the high rate of psychological disorder among ado-lescent offenders (Teplin, Abram, McClelland, Dulcan, &

Address correspondence and reprint requests to: Julia Dmitrieva, Departmentof Psychology, Frontier Hall, 2155 South Race Street, University of Denver,Denver, CO 80208; E-mail: [email protected].

Development and Psychopathology 24 (2012), 1073–1090# Cambridge University Press 2012doi:10.1017/S0954579412000545

1073

Mericle, 2002), which may render them particularly vulnera-ble to the stresses of incarceration.

Although all adolescents who are incarcerated share cer-tain basic experiences (e.g., separation from parents andfriends, restrictions on autonomy), incarceration itself is aheterogeneous category that includes a wide range of place-ments. Juvenile facilities vary considerably in their degreeof security and, at least in principle, in the degree to whichthey offer rehabilitative services. Although rehabilitation isone of the stated purposes of the juvenile justice system, to-day’s punitive “adult time for adult crime” climate has trans-formed many juvenile justice facilities into places designedmainly (or exclusively) for punishment and incapacitation(Scott & Steinberg, 2008).

The goals of many juvenile correctional institutions (se-cure or not) often run counter to the successful developmentof psychosocial maturity. That is, strategies that are adaptivefor successful social functioning in life outside of facilitiesthat promote the development of psychosocial maturity maybe maladaptive for life on the inside. For example, incarcer-ated individuals “do well” if they obey without questioningthe rules, even when many rules do not make sense or arenot fairly enforced (Kummerlowe, 1995). Incarcerated indi-viduals undergo repetitive and restrictive routines that areregulated by extensive and rigid rules. Consequently, theymay lose the ability to cope with the demands of an outsideworld that lacks rigid structure (Irwin & Owen, 2005). More-over, if incarcerated youths are discouraged from expressingconcerns, this may interfere with coping efforts, such as seek-ing social support, which in turn could impede adjustment(Shulman & Cauffman, in press). Because adolescence is atime when youth develop increasing temperance, perspective,and responsibility, it stands to reason that in contexts wherethese skills are not practiced and may be actively discouraged,adolescents may have limited ability to attain these psychoso-cial capacities. Spending a great deal of time in an incarcer-ated context may have an especially detrimental impact.

Whereas most research on adolescent development wouldlead one to hypothesize that incarceration would have adverseconsequences for the development of psychosocial maturity,there is an alternative view that needs to be acknowledged.Some proponents of the incarceration of juvenile offendersview incarceration as a form of rehabilitation, one that may pos-itively impact the growth of maturity. Although these writersacknowledge that a primary purpose of this sort of sanctionis punishment, they also believe that the regimentation andharsh discipline characteristic of incarceration environmentsis precisely what juvenile offenders need in order to developself-control and responsibility. For instance, one website pro-moting the use of boot camps notes that, “if you yell andyell at these troubled teens and discipline [them], they will gen-erally straighten out. Angry, aggressive, defiant and opposi-tional teens might improve under the strict, intense, military-style environment of a boot camp” (Teen Boot Camps, 2011).

Although there are multiple types of secure confinement, thetwo most common settings for juvenile offenders are secure

facilities and residential treatment facilities. There are notabledifferences between these two settings. Secure settings tendto be oriented more toward punishment and less toward reha-bilitation. These facilities operate on strict schedules and tendto provide few or no services to youth in their care. In contrast,residential treatment facilities tend to be more rehabilitationoriented, providing youth with a secure, but more residential,life style with greater flexibility of scheduling and indepen-dence. In general, residential treatment facilities tend to offeryouth access to, and require participation in, more rehabilitativeservices, such as schooling, counseling, and drug and alcoholtreatment (Mulvey, Schubert, & Chung, 2007). Undoubtedly,within each of these categories, there are variations in qualityand access to services, but generally secure facilities are charac-terized by a more extreme departure from the typical experi-ences of adolescence than are residential treatment facilities.As such, it may be the case that the type of setting a youthspends time in, not only how long they spend time lockedup, has important implications for how incarceration shapesthe development of psychosocial maturity.

Moreover, although it is possible that the type of incarcera-tion setting and length of time incarcerated may impact thedevelopment of psychosocial maturity, the age of the adoles-cent may moderate whether or how a youth is affected by in-carceration and the quality of the facility. Given that psycho-social maturity develops rapidly throughout adolescence,with growth slowing as youth transition into early adulthood(Monahan et al., 2009, 2011), it could be that earlier confine-ment may be more detrimental, because this is precisely thetime when gains in psychosocial maturity are most likely tooccur. Conversely, it could be that as psychosocial capacitiesdevelop throughout adolescence and early adulthood, incar-ceration at any point in this developmental period impactssubsequent psychosocial maturity. Understanding develop-mental variability in vulnerability to the effects of incarcera-tion on psychosocial maturity is an important question, withimplications for both juvenile justice policy and practice.

Because correctional facilities vary considerably in theirinstitutional climate, as well as the availability and qualityof the rehabilitative treatments they offer, it is likely that thecharacteristics of a given facility will also moderate whetherthere are detrimental effects of incarceration. Programs thatutilize a familylike setting (i.e., with a parental unit in chargeof a small number of youths), have small staff to youth ratios,do not sever inmates’ ties to family and community, and pro-vide professional counseling are more successful at keepingrecidivism low (Caldwell, Skeem, Solekin, & Van Rybroek,2006; Rosner, 1990). Educational and vocational servicesare also important, because for some youth these facilitiesare the only sources of such training (Chung, Little, & Stein-berg, 2005). Facilities that offer higher quality environments,regardless of whether they are secure or nonsecure, may havedifferential effects on youth development than those withlower quality environments.

In order to investigate how the experience of incarcerationaffects the development of psychosocial maturity, it is neces-

J. Dmitrieva et al.1074

sary to follow longitudinally a sample of adolescents andyoung adults in which a large proportion are likely to becomeincarcerated. The present study utilizes a sample of serious ju-venile offenders, who are precisely those youth most likely toexperience incarceration, and asks how incarceration experi-ences shape patterns of psychosocial development from mid-adolescence into early adulthood. In addition to examininghow incarceration generally impacts psychosocial develop-ment, we also investigate how incarceration in different typesof settings (i.e., secure facilities vs. residential treatment facil-ities) may differentially affect adolescent psychosocial devel-opment. Furthermore, we investigate how the age of an ado-lescent and the characteristics of a given facility, such assubjective feelings of safety, moderate the association be-tween time spent incarcerated and the development of psy-chosocial maturity.

Given the absence of prior research on this topic, there islittle guidance with respect to the formulation of specific hy-potheses. It is reasonable to expect that, because of theirharsher environment and relative absence of rehabilitativeservices, secure settings are relatively more detrimental tothe development of psychosocial maturity than residentialtreatment facilities. In contrast, it could be that any disruptionof the normative socialization experiences of adolescence (athome, with peers, or in school) is harmful to psychosocial de-velopment. Similarly, although it could be that the effects ofincarceration on psychosocial development are cumulative,such that greater total time incarcerated during adolescenceis linked with suppressed development of psychosocial devel-opment, it could also be that incarceration experiences sub-stantially impact psychosocial maturity only in the shortterm. Finally, it could be that the age of an adolescent orthe characteristics of a facility affect psychosocial develop-ment more than the type of facility or length of time incarcer-ated per se.

The present analyses are based on a 7-year longitudinalstudy of serious adolescent offenders, and the effects of incar-ceration on the development of psychosocial maturity fromages 14 to 25 are examined. We address four specific researchquestions. First, we examine how total time incarcerated dur-ing this developmental period affects psychosocial maturity.Second, we examine the short-term impact of incarcerationon psychosocial functioning. Third, we investigate whetherthe effects of incarceration are moderated by the age of theadolescent at the time of incarceration. Fourth and finally,we test whether the impact of incarceration on subsequentpsychosocial development varies based on the type and qual-ity of the facility. We examine each of these research ques-tions disaggregating secure facilities from residential treatmentfacilities. The longitudinal nature of this study allowed formultilevel testing of within- and between-individual differ-ences in the development of psychosocial maturity. Fur-thermore, the scope of data detailing youth facility experiencesprovided a rich source of information for investigation ofthe effects of institutional characteristics on the individualtrajectories of psychosocial maturity.

Method

Participants

The participants were 1,171 male adolescents in a larger pro-spective study of male and female serious juvenile offendersin two major metropolitan areas (for study details, see Schu-bert et al., 2004; there were not enough females in the sampleto examine the questions of interest in the present study). Thisstudy enrolled 14- to 17-year-old adolescents who had beenadjudicated of a felony offense, a serious property crime, amisdemeanor weapons offense, or a misdemeanor sexual as-sault. The proportion of male participants with drug law vio-lations was capped at 15%. Of 2,008 youths eligible for thestudy, 1,692 were located and contacted with recruitment in-formation. Parental consent and youth assent were obtainedfor 80% (N ¼ 1,355) of the youths contacted, 1,171 ofwhom were males and therefore were included in the presentinvestigation.

The majority of participants were from lower socioeco-nomic status backgrounds, with fewer than 3% of the adoles-cents having parents who graduated from college. Only 14%of the participants lived with both biological parents, 33%lived with a single mother (never married, divorced, or wi-dowed), 17% lived with a biological mother and a stepfather,20% lived with another adult relative, 5% lived with a singlebiological father, 4% had no adult in the home, and the re-maining 7% lived in other arrangements. Forty-two percentof the participants were African American, 34% were Hispa-nic, 19% White, and 5% other or biracial.

Procedure

Following a referral of eligible adolescents by the juvenilecourt, a trained interviewer contacted adolescents and theirguardians to ascertain their interest in participation. Afterthe appropriate consents and assents had been obtained, an in-terviewer met with an adolescent in a facility, the juvenile’shome, or another location in the community. Adolescentswere paid $50 for their participation in the baseline interview(when allowed by facility rules) and, in order to minimize at-trition, compensation for the follow-up interviews increasedgradually over time to a maximum of $150. All recruitmentand assessment procedures were approved by the institutionalreview boards of the participating universities.

The first (baseline) interview was conducted over 2 days intwo 2-hr sessions. The interviewer and participant sat side byside at a computer while the interviewer read the questionsaloud to avoid any problems due to difficulties with reading.Participants were reinterviewed every 6 months for the 3years following the baseline interview; after 36 months, par-ticipants were interviewed annually for the remaining 4 yearsof the study (a total of 11 interviews prospectively over 7years). Retention in the sample was excellent: 60% of indi-viduals completed all 11 interviews, 18% completed 10 inter-views, 8% of individuals completed 9 interviews, 5% of indi-

Arrested development 1075

viduals completed 8 interviews, 3% of individuals completed7 interviews, 2% of individuals completed 6 interviews, 2%of individuals completed 5 interviews, and 4% of individualscompleted fewer than 4 interviews.

Measures

Of interest in the present report are measures of psychosocialmaturity, incarceration in a secure facility or a residentialtreatment facility, characteristics of incarceration settings,and covariates that may impact whether youths were likelyto be incarcerated or not.

Psychosocial maturity. Psychosocial maturity is conceptual-ized as consisting of three separate components: temperance,perspective, and responsibility, each of which has two ele-ments (Steinberg & Cauffman, 1996). In addition to the threecomponents, we also create a global psychosocial maturitymeasure that combines the dimensions of temperance, per-spective, and responsibility. Temperance is assessed by im-pulse control and suppression of aggression; perspective in-cludes consideration of others and future orientation; andresponsibility includes personal responsibility and resistanceto peer influence. Four measures, described subsequently,were used to create these six indicators of psychosocial ma-turity, which are combined into the three conceptual dimen-sions of psychosocial maturity: the Weinberger AdjustmentInventory (Weinberger & Schwartz, 1990), which includessubscales that assess impulse control, suppression of aggres-sion, and consideration of others; the Future Outlook Inven-tory (Cauffman & Woolard, 1999), which was used to derivea measure of future orientation; the Psychosocial Maturity In-ventory (Greenberger, Josselson, Knerr, & Knerr, 1975),which includes a scale that assesses personal responsibility;and the Resistance to Peer Influence measure (Steinberg &Monahan, 2007).

Three subscales of the Weinberger Adjustment Inventorywere used: impulse control (e.g., “I say the first thing thatcomes into my mind without thinking enough about it”), sup-pression of aggression (e.g., “People who get me angry betterwatch out”), and consideration of others (e.g., “Doing thingsto help other people is more important to me than almost any-thing else”). Participants assessed how accurately a series ofstatements matched their own behavior (on a 5-point scale,from false to true). Each subscale was found to have adequatereliability (Cronbach a) and good fit to the baseline data (con-firmatory factor analysis): impulse control (eight items; a ¼0.76, normed fit index [NFI] ¼ 0.95, comparative fit index[CFI] ¼ 0.95, root mean square error of approximation[RMSEA] ¼ 0.07), suppression of aggression (seven items;a ¼ 0.78, NFI ¼ 0.96, CFI ¼ 0.97, RMSEA ¼ 0.06), andconsideration of others (seven items; a ¼ 0.73, NFI ¼0.98, CFI ¼ 0.99, RMSEA ¼ 0.04).

The Future Outlook Inventory is an eight-item measurethat includes items from the Life Orientation Task (Scheier& Carver, 1985), the Zimbardo Time Perspective Scale (Zim-

bardo, 1990), and the Consideration of Future ConsequencesScale (Strathman, Gleicher, Boninger, & Edwards, 1994).The inventory asks participants to rank the degree to whicheach statement reflects how they usually act, on a scale of 1(never true) to 4 (always true). The future orientation scorewas calculated as the mean of responses from each item(e.g., “I will keep working at difficult, boring tasks if Iknow they will help me get ahead later”). The scale showedgood reliability and an excellent fit to the baseline data(a ¼ 0.68, NFI ¼ 0.96, CFI ¼ 0.97, RMSEA ¼ 0.03).

The Psychosocial Maturity Inventory (Greenberger et al.,1975) includes a 30-item, reverse-scored subscale that as-sesses personal responsibility (e.g., “If something more inter-esting comes along, I will usually stop any work I’m doing”).Individuals respond on a 4-point scale from strongly disagreeto strongly agree. Personal responsibility was calculated asthe mean across all 30 items. The measure showed excellentreliability and an adequate fit to the baseline data (a ¼ 0.89,NFI ¼ 0.82, CFI ¼ 0.87, RMSEA ¼ 0.04).

Finally, the measure of Resistance to Peer Influence (Stein-berg & Monahan, 2007) assesses the degree to which adoles-cents act autonomously in interactions with their peers. Partic-ipants are read two conflicting statements (e.g., “Some peoplego along with their friends just to keep their friends happy” and“Other people refuse to go along with what their friends want todo, even though they know it will make their friends unhappy”)and are asked to choose the statement that most closely reflectstheir behavior. Next, participants are asked to rate the accuracyof the statement (i.e., “sort of true” or “really true”). Each itemis scored on a 4-point scale, ranging from 1 (really true) for thecharacterization indicating less resistance to influence to 4(really true) for the characterization indicating more resistanceto influence; answers of sort of true are assigned a score of 2 (ifassociated with the less resistant option) or 3 (if associated withthe more resistant option). Ten such items are presented to par-ticipants. Each item explores a different dimension of peer in-fluence (e.g., going along with friends, saying things one doesnot really believe), and one resistance to peer influence scorewas computed for this measure by averaging scores on the10 items. The measure showed good internal consistency andfit to the baseline data (a ¼ 0.73, NFI ¼ 0.92, CFI ¼ 0.94,RMSEA ¼ 0.04).

Confirmatory factor analysis (with baseline data) evalu-ated the structural validity of the psychosocial maturity con-struct, with the three first-order factors (i.e., temperance, re-sponsibility, and perspective) and the second-order globalpsychosocial maturity factor. The confirmatory factor modelhad a good fit: x2 (6) ¼ 26.47, p , .001; NFI ¼ 0.97, CFI ¼0.98, RMSEA ¼ 0.06 (0.04, 0.08). Consequently, the tem-perance, perspective, responsibility, and total psychosocialmaturity scores were created using the factor scores fromthe confirmatory factor analysis model.

Incarceration. Two types of incarceration settings were of in-terest: confinement in a secure facility and confinement in aresidential treatment facility. At each interview, individuals


completed a life calendar where they reported the number ofdays in each month since the prior interview that they hadbeen incarcerated and gave the name of each facility theyhad spent time in since the last interview. Interviewers beganthe life calendar by asking the youth to identify major lifeevents, such as the birth of a child, that had occurred sincethe prior interview. Participants were then asked if they hadbeen involved in an activity since the last interview (e.g.,“Were you incarcerated at all?”). If the respondent’s answerwas positive, the interviewer went through the recall periodmonth by month to determine in what month(s) the activityoccurred. If a youth had a problem remembering the exactmonth of an occurrence, the interviewer would use major lifeevents or information from other activity domains to help nar-row the date (e.g., “Did this happen around your birthday? Orwas it closer to when you moved to Penn Avenue?”). If a youthendorsed being in multiple facilities in the recall period, thenumber of days was recorded for each facility. Previous re-search suggests that retrospective data gathered using life cal-endar methods is accurate (Freedman et al., 1988) and thatthe data structure of the life calendar fits the structure of respon-dents’ autobiographical memories well (Belli, 1998). For ex-ample, Caspi and Amell (1994) found that life events reported3 years earlier matched the retrospective life history calendardata with over 90% accuracy. As such, the life calendar datacollection method can provide a more continuous and com-plete representation of life events than is possible with other in-terview or questionnaire measures.

The variables that assessed incarceration in secure and resi-dential treatment facilities were created from the life calendardata. The research staff classified each facility that was namedin the interview as a secure setting or a residential treatment fa-cility. We distinguished between the secure and residentialtreatment facilities based on the orientation of the facility. Con-finement settings where incarceration, not rehabilitation, is thestated goal were classified as secure facilities. These includeddetention centers (i.e., juvenile correctional facilities that areused both as short-term preadjudication placements or asmore permanent postadjudication and disposition placements),jails (i.e., typically local juvenile facilities that hold youth untiltrial or for relatively short sentences after trial), and prisons(i.e., usually state-run facilities that hold people for a longerterm after conviction). A very small proportion of secure set-tings included federal adult detention centers and Immigrationand Naturalization Services detention centers.

Confinement settings that have a stronger focus on rehabili-tation and providing services to residents were classified as res-idential treatment facilities. These facilities typically provide arange of services that are often centered on a specific modelof intervention (e.g., peer culture, physical challenge). Some fa-cilities focus more specifically on mental health treatment andtarget mentally ill adolescents. These settings usually providea more intensive assignment to therapy and other treatment ac-tivities. Although all residential treatment facilities provide a re-sidential program within a structured environment, they varywith respect to their security and access to the community.

Based on youth responses on the life calendar data, we cal-culated four incarceration variables: (a) the total proportion oftime incarcerated in secure settings across all years of datacollection; (b) the total proportion of time incarcerated in res-idential treatment facilities across all years of data collection;(c) recent secure incarceration, that is, the proportion oftime since the last interview spent in a secure facility (as op-posed to being in a different setting or on the streets; a time-varying covariate); and (d) recent residential treatment facilityincarceration, that is, the proportion of time since the last in-terview spent in a residential treatment facility (as opposed tobeing in a different setting or on the streets; also a time-vary-ing covariate).

The first two variables, total proportion of time in a securesetting and total proportion of time in a residential treatmentfacility are grand-mean centered in each analytic model.Thus, these variables test the effects of relatively more or rel-atively less incarceration over the course of the 7-year study(secure or residential treatment facility) compared to otheryouth in the sample. Specifically, these variables allow atest of how being incarcerated more or less in a given settingimpacts the development of psychosocial maturity over thecourse of the study.

The second two variables, the time-varying covariates thatreflect the proportion of time since the last interview spent ina facility, indicate recent incarceration in a secure setting andrecent incarceration in a residential treatment facility. Thesevariables represent the fluctuations in incarceration that occurthroughout the course of the study and allow a test of how rel-atively more or less incarceration in the previous recall periodfor an individual is linked with immediately subsequent de-velopmental changes in psychosocial maturity.

Over the course of the study, youths spent on average 23months (SD ¼ 23) in a secure facility and 4 months (SD ¼ 7)in a residential treatment facility. (Note that these stays arenot necessarily consecutive and can follow any possiblepattern of incarceration in secure settings, residential facil-ities, or street time across the 7 years of the study.) Youthscould report any possible combination of street time, incar-ceration in a secure facility, and incarceration in a residentialtreatment facility. Figure 1 illustrates the proportion of eachrecall period that youth, on average, reported being in a securefacility, a residential treatment facility, or in no facility. Fig-ure 2 illustrates how much of the recall period was spent ineither a secure facility (Figure 2a) or a residential treatmentfacility (Figure 2b). At each time point, approximately halfof the sample reported incarceration for some amount oftime in a secure setting during the relevant recall period.Across time, 20% to 27% of youth reported being incarcer-ated in a secure setting for 75% of the recall period ormore; 21% to 32% of youth reported being incarcerated ina secure facility for less than 75% of the recall period. Fewyouths were placed in residential treatment facilities. Overthe 7-year period, only 2% to 24% of the sample reportedtime spent in a residential treatment facility during a given re-call period; and there is a notable decline in placement in


residential treatment facilities as youth aged, reflecting thegreater use of treatment facilities in the juvenile justice systemthan the criminal justice system.

Characteristics of incarceration setting. If a youth endorsedconfinement in a facility, either a secure facility or a residen-tial treatment facility, he reported on six characteristics ofeach facility (adapted from Mulvey, Schubert, & Odgers,2010): (a) the number of mental health services received inthe facility, (b) the number of training services received inthe facility, (c) the presence of a caring adult, (d) the presenceof reentry planning, (e) perceived facility danger, and (f) thedegree of contact with antisocial peers. These variables wereassessed with respect to each facility a youth reported stayingin since the prior interview.

The Number of Mental Health Services is a 7-item checklistthat assessed youth attendance of treatment services such asdrug and alcohol treatment, psychological therapy, anger man-agement, or group therapy. The Training Services Checklist isthe 2-item checklist that assessed whether job and school train-ing was received by the youths. Reentry planning is a single-item variable that assessed whether youths have had a personassigned to them to help make arrangements for their returnto the community. The Presence of the Caring Adult is a 7-item checklist that assessed the presence of an adult figure inthe facility that youths admire, can get advice from, or coulddepend on for help. The Perceived Facility Danger is a 2-item scale that assessed how safe youths felt (on the scalefrom 1 ¼ in danger all of the time to 5 ¼ in danger none of

the time; reverse coded) from other youths and staff in the fa-cility. Finally, the Peer Delinquency Scale (adapted fromThornberry, Lizotte, Krohn, Farnworth, & Jang, 1994) is a12-item scale that assessed how many of youths’ peers (on ascale from 1 ¼ none of them to 5 ¼ all of them) engaged ina variety of antisocial activities such as fighting and theft.

In order to ascertain the accuracy of youths’ reports on thefacility characteristics, we computed intraclass correlationcoefficients (ICCs) for each of the six facility characteristics.For these analyses, we analyzed youths’ responses for 421 se-cure and residential treatment facilities. Each facility, on aver-age, received 12 (SD¼ 33) participants over the course of thestudy. Given the wide range of the number of participants perfacility, the ICC model that analyzed the complete data in-cluded many empty cells (e.g., a model for one facility thathas 100 participants/coders and five other facilities thathave 10 participants/coders contains 450 empty cells out of600 total cells), making it impossible to compute the reliabil-ity coefficients. Thus, we had to constrain the ICC analyses toa maximum of 30 participants/coders for each facility. The re-sulting ICCs were 0.92 ( p , .001) for the mental healthservices variable, 0.96 ( p , .001) for the training servicesvariable, 0.90 ( p , .001) for the caring adult variable, 0.88( p , .001) for presence of the reentry planning variable,0.34 ( p ¼ .15) for the perceived facility danger variable,and 0.32 ( p¼.19) for contact with antisocial peers. These re-sults indicate excellent agreement for objective markers of fa-cility characteristics (i.e., presence of mental health, training,reentry services, and caring adults) and lower agreement for

Figure 1. The average proportion of time spent in each facility type.


variables that assess youth perceptions of facility danger andextent of contact with antisocial peers.

Based on the results of an exploratory factor analysis, thesix characteristics were divided into either negative facilityfeatures (consisting of the perceived facility danger andpeer delinquency scales) or positive facility features (consist-ing of the number of mental health services, the number oftraining services, reentry planning, and the presence of thecaring adult checklists). If youth reported stays in multiple fa-cilities (i.e., multiple types of secure facilities), the responseson scales were aggregated across facilities and weighted bythe proportion of time spent in each facility. Responseswere calculated separately for secure facilities and residential

facilities. Thus, four total variables were calculated: (a)negative facility features in secure facility(ies), (b) negativefacility features in residential treatment facility(ies), (c) posi-tive facility features in secure facility(ies), and (d) positive fa-cility features in residential treatment facility(ies).

Covariates. The number of arrests prior to the baseline inter-view (coded from official records), age at first arrest, parentaleducation (ranging from 1 ¼ some grade school to 9 ¼ pro-fessional or graduate school), parental arrest history (codedas 0 ¼ neither parent arrested, 1 ¼ one parent arrested,and 2 ¼ both parents arrested), mental health symptoms,and youth intelligence were utilized as controls in all models.

Figure 2. The incarceration in (a) secure facilities or (b) residential treatment facilities over time.


This accounted for how prior offending, familial offending,and demographic characteristics may impact whether a youthis incarcerated and, if so, sentenced to time in a secure facilityor a residential treatment facility. Mental health symptomswere assessed using the Brief Symptom Inventory (Derogatis& Melisara, 1983), a 53-item self-report inventory assessingthe degree to which participants have been bothered (0 ¼ notat all to 4¼ extremely) in the past week by various symptoms.Intelligence was assessed using the Wechsler AbbreviatedScale of Intelligence (Wechsler, 1999). This measure pro-duces an estimate of general intellectual ability based ontwo subtests; the vocabulary test consists of 42 items that re-quire the participant to orally define four images and 37words presented both orally and visually and the matrix rea-soning test consists of 35 incomplete grid patterns that requirethe participant to select the correct response from five possi-ble solution choices. This test has been normed for indi-viduals 6 to 89 years of age.

Plan of analyses

Multilevel models were used to identify the average trajectoriesof the development of each dimension of psychosocial matur-ity (temperance, perspective, and responsibility) as well asglobal psychosocial maturity. For all analyses, observationswere modeled as a function of participant age centered at age14, the age of the youngest participant in the study. All modelscontrolled for the number of prior arrests at baseline, age at firstarrest, ethnicity, mental health symptoms, intelligence, paren-tal education, and parental arrest history.

Next, multilevel modeling was used to examine the effectsof the total proportion of incarceration time across the entirestudy period in either a secure facility or a residential treat-ment facility on psychosocial maturity. This model testedhow relatively more or less time in a secure facility or a resi-dential treatment facility impacted (a) the level of psychoso-cial maturity and (b) the development of psychosocial matur-ity over time. Because both the amount of time spent in secureand residential treatment facilities were entered into eachmodel simultaneously, the estimates for secure incarcerationcan be interpreted as the effects of time spent in a secure con-finement when incarceration in a residential treatment facilityequals zero and, conversely, the estimates for the residentialtreatment incarceration can be interpreted as the effects oftime spent in a residential treatment facility when incarcera-tion in a secure facility equals zero. In other words, the resultsestimate the effects of time in confinement in each respectivefacility type relative to no incarceration.

Subsequently, we tested how recent incarceration in theprior recall period (i.e., the time interval since the last inter-view) impacted the level of psychosocial maturity at theend of the recall period (i.e., the current interview date).This modeling strategy allowed us to separate the effects ofhaving high versus low rates of incarceration at any point dur-ing the study from the effects of having high versus low ratesof incarceration in the given assessment period.

Next, we investigated if the effects of recent incarcerationin a secure facility or a residential treatment facility weremoderated by the age of the adolescent (i.e., if recent incar-ceration was more detrimental to psychosocial maturity foryounger or older individuals). Finally, we tested if the effectsof recent incarceration in a secure facility or a residential treat-ment facility were moderated by the quality of the facilitieswhere a youth spent time in that period (i.e., the negative orpositive features of those facilities). Appendices A throughC provide the statistical formulas for the analyses that werecompleted.

Results

Growth of psychosocial maturity from adolescenceto early adulthood

We first identified the average trajectories of the developmentof temperance, perspective, responsibility, and global psy-chosocial maturity from adolescence to early adulthood(see Table 1). The pattern of growth was identical for globalpsychosocial maturity and each dimension of psychosocialmaturity. Growth models indicated a significant intercept,significant linear growth term, and significant quadraticgrowth term. On average, individuals increased in psychoso-cial maturity across adolescence, with gains in psychosocialmaturity slowing as youth transitioned into early adulthood.It is important that there was significant individual variabilityboth in the level of maturity (i.e., intercept) and the develop-mental rate of change (linear and quadratic rates of growth)over age (for more details, see Monahan et al., 2011).

The impact of total time incarcerated in secure facilitiesand residential treatment facilities on psychosocialmaturity

We next evaluated if the total proportion of time incarceratedin a secure facility or a residential treatment center over the 7-year study period explained individual variation in the (a)level or (b) rate of change of global psychosocial maturity,temperance, perspective, and responsibility (see Table 2).As a reminder, these and subsequent models estimated the ef-fects of confinement in each respective facility type relative tono incarceration. Figure 3 illustrates the impact of the totalproportion of time incarcerated in secure and residential treat-ment facilities on global psychosocial maturity. (The patternof results is nearly identical for temperance and responsibil-ity, but not perspective. For parsimony, we only illustratethis effect with global psychosocial maturity.) Figure 4 illus-trates the impact of the total time incarcerated in secure andresidential treatment facilities on perspective.

Compared to youths who were not incarcerated, those whospent more time in secure facilities over the course of thestudy evinced lower levels of psychosocial maturity at age14. Greater time spent incarcerated in secure settings overthe course of the study, however, was unrelated to the devel-


opment of global psychosocial maturity, temperance, or re-sponsibility from adolescence to early adulthood. It was un-expected, however, that youths who spent more time incarcer-ated in secure facilities had faster increases in perspectiveduring the adolescent years than youths who were not incar-cerated. As a result, the initial differences in perspective foryouths who spent longer versus shorter time in a secure facil-ity disappear by age 25 (see Figure 4).

In contrast to the effects of total confinement in a securefacility, youths who spent more time in residential treatmentcenters evinced higher levels of global psychosocial maturity,perspective, and responsibility at age 14. However, youthswho spent more time incarcerated in a residential treatmentfacility over the 7-year study also demonstrated slower devel-opmental gains in these measures of psychosocial maturity.As can be seen in Figures 3 and 4, the relative standing of

Table 2. Impact of total time incarcerated in secure facility or RTF on global psychosocial maturity, temperance,perspective, and responsibility

Global Psychosocial Maturity Temperance Perspective ResponsibilityB B B B

Fixed Effects

Intercept 0.84 2.93 2.42 1.90Total time in secure facility 20.07*** 20.26*** 20.21*** 20.10**Total time in RTF 0.08* 0.29 0.27* 0.23**

Linear slope 0.03*** 0.09*** 0.12*** 0.07***Total time in secure facility 0.00 0.01 0.02* 0.00Total time in RTF 20.01* 20.05 20.06*** 20.02*

Quadratic slope 20.001*** 20.004*** 20.01*** 20.003***

Random Effects

Intercept 0.021*** 0.38*** 0.20*** 0.07***Linear slope 0.002*** 0.04*** 0.02*** 0.01***Quadratic slope ,0.001*** ,0.001*** ,0.001*** ,0.001***Level 1 error 0.008*** 0.17*** 0.07*** 0.05***Deviance 216,063 15,111 5,742 642No. of parameters 25 25 25 25

Note: RTF, residential treatment facility. Total time in secure facility and total time in RTF variables are grand mean centered.*p , .05. **p , .01. ***p , .001.

Table 1. Unconditional multilevel models of global psychosocial maturity, temperance, perspective, andresponsibility


Fixed Effects

Intercept 0.85*** 3.01*** 2.45*** 1.95***Linear slope 0.03*** 0.09*** 0.11*** 0.07***Quadratic slope 20.001*** 20.004*** 20.01*** 20.003***

Random Effects


***p , .001.


youth incarcerated in a residential treatment setting for longversus short time changes over the age span of the study.Longer residential treatment facility incarceration at age14 is associated with higher psychosocial maturity. This in-carceration-related gain in psychosocial maturity disappearsbetween the ages 18 and 20. After age 20, however, a longertotal time in a residential treatment facility is associatedwith lower psychosocial maturity. Total time in residential

treatment facilities was not associated with changes intemperance.

The impact of recent incarceration in secure facilities andresidential treatment facilities on psychosocial maturity

Next we investigated how incarceration in either a securefacility or a residential treatment facility during the time

Figure 3. The impact of the total incarceration in a secure setting and residential treatment facility on global psychosocial maturity.

Figure 4. The impact of the total incarceration in a secure setting and residential treatment facility on perspective.


period immediately preceding the interview was associatedwith global psychosocial maturity, temperance, perspective,and responsibility (see Table 3). As in prior analyses, theeffects of recent incarceration are tested compared to noincarceration.

When youths had recently spent a greater amount of timein a secure facility, they evinced lower levels of global psy-chosocial maturity, temperance, and responsibility (but notperspective). That is, greater time spent in a secure setting ap-

peared to arrest the development of global psychosocial ma-turity, temperance, and responsibility at the subsequent age.To illustrate this effect, Figure 5 shows the effect of being in-carcerated in a secure facility 0% of the time since the priorinterview, 50% of the time since the prior interview, and100% of the time since the prior interview between ages 17and 18 on global psychosocial maturity. The pattern of resultsis identical for other ages in the sample and for temperanceand responsibility (in addition to global psychosocial matur-

Table 3. Impact of recent time incarcerated in secure facility or RTF on global psychosocial maturity, temperance,perspective, and responsibility


Fixed Effects

Intercept 0.84 2.94 2.41 1.93Total time in secure facility 20.06** 20.22** 20.19*** 20.10**Total time in RTF 0.06 0.17 0.25* 0.17Recent time in secure facility 20.01** 20.04** 20.02 20.02*Recent time in RTF 0.01 0.03 0.02 0.03*

Linear slope 0.02*** 0.07*** 0.11*** 0.07***Total time in secure facility 0.00 0.01 0.02* 0.00Total time in RTF 20.01 20.03 20.05** 20.02

Quadratic slope 20.001*** 20.002 20.01*** 20.003***

Random Effects


Note: RTF, residential treatment facility.*p , .05. **p , .01. ***p , .001.

Figure 5. The impact of recent incarceration in a secure facility on global psychosocial maturity.


ity). For parsimony, we only show the effect at age 18 forglobal psychosocial maturity.

In contrast, recent time spent in a residential treatment fa-cility (i.e., the amount of time spent in a residential treatmentfacility during the time period immediately prior to the inter-view) was not associated with subsequent maturity one wayor the other, except in the case of responsibility. Longer incar-ceration in a residential treatment facility appeared to have apositive effect on subsequent responsibility, with longertime spent in a residential treatment facility associated withsubsequently higher levels of responsibility. Figure 6 illus-trates the effect of being incarcerated in a residential treatmentfacility 0% of the time since the prior interview, 50% of thetime since the prior interview, and 100% of the time sincethe prior interview between ages 17 and 18 on responsibility.Note that this same pattern is found at all other ages in thesample.

We also tested if the effects of recent incarceration in a se-cure setting or in a residential treatment facility were sus-tained over time (i.e., if the effects of recent incarcerationwere sustained at later time points). The results indicatedthat the detrimental effects of incarceration in a secure settingon global psychosocial maturity, temperance, and responsi-bility and the positive effect of recent incarceration in a resi-dential treatment facility on responsibility are only found im-mediately after that period of incarceration.1

Age as a moderator of the impact of recent incarcerationin secure facilities and residential treatment facilities onpsychosocial maturity

Next we investigated if the impact of recent incarceration insecure facilities or residential treatment facilities was mod-erated by the age of the individual (see Table 4). Specifi-

cally, we tested if incarceration in secure settings or residen-tial treatment facilities might have stronger effects onpsychosocial maturity if they occurred relatively earlier orlater in development. The results indicated that age moder-ated the short-term effects of incarceration in a secure facil-ity on responsibility and the short-term effects of incarcera-tion in a residential treatment facility on temperance.Figure 7 demonstrates age-related changes in the short-term effect of incarceration in a secure facility on responsi-bility. Longer secure facility incarceration at age 14 is asso-ciated with higher subsequent responsibility. However, thisincarceration-related boost in responsibility disappears byage 18. After age 18, incarceration in a secure facility leadsto short-term declines in responsibility, however.

A different pattern of results emerged with respect to theshort-term effects of residential treatment incarceration ontemperance (Figure 8). In this case, longer incarceration ina residential treatment facility is associated with lower sub-sequent temperance among younger youths. This incarcera-tion-related decline in temperance disappears by at 18, afterwhich incarceration in a residential treatment facility is as-sociated with increases in temperance. There were no othersignificant age-related interactions, indicating that theshort-term negative effects of incarceration in a secure set-ting on global psychosocial maturity and temperance andthe positive short-term effects of incarceration in a residen-tial treatment facility on responsibility noted previously areconsistent across age.

Facility quality as a moderator of the impact of recentincarceration in secure facilities and residential treatmentfacilities on psychosocial maturity

Finally, we tested if the impact of recent incarceration in a se-cure facility or a residential treatment center on subsequentlevel of global psychosocial maturity, temperance, perspec-

Figure 6. The impact of recent incarceration in a residential treatment facility on responsibility.

1. The detailed results of these analyses are available from the authors.


Table 4. Age as a moderator of recent time incarcerated in secure facility or RTF on global psychosocial maturity,temperance, perspective, and responsibility

Global Psychosocial Maturitya Temperancea Perspectivea Responsibilitya

B B B B

Fixed Effects

Intercept 0.85 2.98 2.41 1.89Total time in secure facility 20.05* 20.23* 20.20** 20.13**Total time in RTF 0.07 0.23* 0.26* 0.17Recent time in secure facility 20.02 20.12** 20.01 0.05Recent time in RTF 20.02 20.16* 20.01 0.07Recent time in secure facility×Age 0.00 0.01 0.00 20.01**Recent time in RTF×Age 0.01 0.04* 0.01 20.01

Linear slope 0.02*** 0.06*** 0.11*** 0.07***Total time in secure facility 0.00 0.00 0.02 0.01*Total time in RTF 20.01 20.05 20.05** 20.01

Quadratic slope 20.001** 20.001 20.01*** 20.003***

Random Effects

Intercept 0.03*** 0.47*** 0.24*** 0.09***Linear Slope 0.003*** 0.04*** 0.02*** 0.01***Quadratic Slope ,0.001*** ,0.001*** ,0.001*** ,0.001***Level 1 Error 0.008*** 0.17*** 0.07*** 0.05***Deviance 214,547 14,776 5,068 579No. of parameters 29 29 29 29

Note: RTF, residential treatment facility.aIn order to aid model convergence, the interaction terms for the secure facility and RTF quality variables were tested in two separate models.*p ,. 05. **p , .01. ***p , .001.

Figure 7. The effect of recent incarceration in a secure facility on responsibility at each age.


tive, and responsibility were moderated by the quality of thefacility (i.e., if youths stayed in secure settings or residentialtreatment facilities with more negative features or more pos-itive features).

Although negative facility features were related to psycho-social maturity (main effects), neither positive nor negative fa-cility features moderated the link between incarceration timeand subsequent maturity (see Table 5).2 Confinement in securesettings with more negative features was associated with lowersubsequent levels of global psychosocial maturity, temperance,and responsibility compared to no incarceration. Similarly, in-carceration in residential treatment centers with more negativefeatures was associated with lower subsequent levels of globalpsychosocial maturity and temperance (relative to no incarcera-tion), but not perspective or responsibility. Positive features ofsecure facilities and residential treatment facilities were unre-lated to any of the subsequent measures of psychosocial matur-ity. However, the aspects of the facility context (either positiveor negative) did not moderate the short-term effects of incar-ceration on the development of psychosocial maturity, suggest-ing that the length of time spent in a certain type of incarceratedsetting and the characteristics of a given facility independentlyimpact psychosocial maturity.

Discussion

Little is known about the impact of incarceration on norma-tive adolescent development. Given the existing literatureon what facilitates positive development of psychosocial ma-turity, we would expect that the incarceration environmentwould impede the development of psychosocial capacities.Incarceration is an extreme departure from the normative ado-lescent socialization experience and has contextual featuresthat may be reasonably hypothesized to lead to stunted devel-opment of psychosocial maturity. Specifically, because incar-ceration limits youths’ opportunities to practice responsiblejudgment and places them exclusively among deviant peers,who are unlikely to be models of mature temperance, respon-sibility, or perspective, it is reasonable to hypothesize that theextreme socialization experience of incarceration may impactthe development of psychosocial maturity.

Our results provide evidence that, in the short term, secureincarceration impacts the development of psychosocial ma-turity in ways consistent with such hypotheses. Incarcerationin a secure facility was associated in the short term with aslowing of gains in global psychosocial maturity, temper-ance, and responsibility. It is surprising that the negative ef-fect of incarceration in a secure setting increased with age.It is interesting, however, that we generally failed to observemeaningful sustained differences in maturity as a function ofthe total amount of time spent in secure confinement through-out the 7-year study period.

Taken together, the short- and long-term effects suggestthat although secure incarceration impedes psychosocial de-

Figure 8. The effects of recent incarceration in a residential treatment facility on temperance at each age.

2. In a separate set of analyses not presented here, we separately tested eachof the six facility quality indicators as moderators of the link between re-cent incarceration in a secure facility or a residential treatment facility andsubsequent psychosocial maturity. Consistent with results for the com-bined negative and positive facility features scales, no facility quality mea-sures functioned as significant moderators.


velopment, the net effect is not cumulative and is short lived.It is notable that “total time incarcerated” reflects the numberof months confined, but not the specific timing (so amongyouths confined for 2 years, some may have been confinedduring Years 1 and 2, whereas others may have been confinedduring Years 6 and 7, and still others may have been confinedfor multiple shorter periods). Thus, although the lack of anobserved relation between the total number of months incar-cerated in a secure setting and the net change in psychosocialmaturity over 7 years indicates that the negative effects of in-carceration observed in the short term may not be sustained inthe long term, the average length of secure confinement (23months) was relatively short compared to the overall study pe-riod, indicating that most subjects spent the majority of thetime in nonconfined settings during which they (apparently)recovered from the negative impact of incarceration on theirpsychosocial maturity. It is also important to note that wefound no support for the notion that incarceration in harsh set-

tings promotes the development of self-control or responsibil-ity, as has been asserted by proponents of boot camps andother similar institutions.

In contrast to our findings for secure incarceration, we findlong-term negative effects of incarceration in residential treat-ment facilities on psychosocial maturity. Youths who spent agreater proportion of time incarcerated in a residential treat-ment facility over the course of the study showed slower netgrowth in psychosocial maturity, perspective, and responsi-bility. Thus, although youths placed in residential treatmentfacilities evinced higher age-14 scores on these three mea-sures of psychosocial maturity, they reached young adulthoodwith lower global psychosocial maturity, perspective, and re-sponsibility scores. The short-term effects of incarceration ina residential treatment facility were contrary to our expecta-tions; that is, incarceration in a residential treatment facilitytemporarily bolsters the development of responsibility andtemperance (among older youths).

Table 5. Facility context as a moderator of recent time incarcerated in secure facility or RTF on global psychosocialmaturity, temperance, perspective, and responsibility

Global Psychosocial Maturitya Temperancea Perspectivea Responsibilitya

B B B B

Fixed Effects

Intercept 0.78 2.79 2.10 1.90Total time in secure facility 20.09** 20.32** 20.22** 20.07Total time in RTF 0.03 0.00 0.21 0.36*Recent time in secure facility 0.00 0.02 20.04* 20.01Recent time in RTF 0.02 0.10 20.01 0.01Positive features secure facility 0.00 0.00 0.01 0.00Negative features secure facility 20.02*** 20.10*** 0.01 20.02*Positive features RTF 0.00 0.00 0.01 0.00Negative features RTF 20.02* 20.11* 20.01 20.04Recent time in Secure Facility×Positive

Features Secure Facility 0.00 0.00 0.01 0.00Recent Time in Secure Facility×Negative

Features Secure Facility 0.01 0.06 20.02 0.02Recent Time in RTF×Positive

Features RTF 0.00 0.00 20.01 0.00Recent Time in RTF×Negative

Features RTF 0.00 0.02 20.03 20.01Linear slope 0.03*** 0.08*** 0.12*** 0.06***

Total time in Secure Facility 0.01** 0.05** 0.03** 0.01Total time in RTF 0.00 0.02 20.04 20.01

Quadratic Slope 20.001*** 20.004** 20.01*** 20.003***

Random Effects

Intercept 0.019*** 0.37*** 0.15*** 0.07***Linear slope ,0.001*** 0.005*** 0.002*** 0.001***Quadratic slope —b —b —b —b

Level 1 error 0.008*** 0.16*** 0.07*** 0.05***Deviance 26,488 6,668 2,856 805No. of parameters 28 28 28 28

Note: RTF, residential treatment facility.aIn order to aid model convergence, the interaction terms for the secure facility and RTF quality variables were tested in two separate models.bThe variance component was omitted to aid model convergence.*p , .05. **p , .01. ***p , .001.


We also find that youths who spend greater time in securesettings evince lower psychosocial maturity to begin with,whereas those who spend more time in residential treatment fa-cilities report higher initial psychosocial maturity. It is likelythat this reflects sorting at the time of adjudication, where judgesmake decisions about assignment to more or less punitive incar-ceration settings based on their (or their probation officers’) per-ceptions of youths’ temperance, perspective, and responsibility.Although we control for the history of prior offenses, our infor-mation is limited to the type and number of prior offenses, notthe details of the offense. It may be that juveniles who are moreimpulsive, short-sighted, and susceptible to peer influence of-fend in ways that the courts perceive as especially dangerous,requiring longer incarceration in a secure setting.

Contrary to our original expectations, youths who spentmore time in a secure setting showed faster overall growthof their perspective-taking abilities. However, this effect ap-pears to be a function of longer-incarcerated youths catchingup to nonincarcerated youths over time, rather than a true fa-cilitative effect, because, as noted earlier, incarcerated youthsbegin with lower initial perspective scores. This pattern offindings may reflect regression to the mean, rather than a gen-uine positive effect of the secure confinement (especiallygiven the observed short-term negative effects). Thus, weare hesitant to draw firm conclusions from this observation.Further research is necessary to understand this effect.

It is important to note that our failure to find a long-termstunting effect of incarceration in a secure setting does notdefinitively prove the absence of such an effect. Our findingsmay have been influenced by several factors. First, partici-pants in this study were between the ages of 14 and 17 atthe onset of the study. Given that two-thirds of the partici-pants in this sample had previous arrest records prior to thebeginning of the study, many of them had experienced incar-ceration prior to their baseline interview. It is possible that ayouth’s earliest incarceration experiences have the most pro-found effect on his trajectory of psychosocial maturity andthat subsequent incarcerations have more limited effects.This may explain our ability to find a long-term negative ef-fect for incarceration in the residential treatment facilities butnot the secure facilities (where youths are more likely to havehad previous incarceration experiences). Studies that includeyounger participants or are restricted to first-time offenderswould be needed to confirm this. Second, because youthswho receive the harshest placements are the least mature tobegin with, they may have a poor psychosocial prognosis, re-gardless of their incarceration experiences. Finally, becauseyouths who receive longer sentences in secure facilities aremore immature to begin with, their regression to the meanmight have masked any underlying effect of incarcerationon the psychosocial maturity slope.

The effects of secure incarceration on the development ofpsychosocial maturity are apparent, however, when we con-sider the short-term impact of being locked up. Incarcerationin secure confinement appears to temporarily disrupt the devel-opment of temperance and responsibility. Even though these

effects are short term, it does not mean that they are not impor-tant. That youths who are the most immature to begin with re-ceive the longest sentences in secure confinement, coupledwith the finding that secure confinement disrupts their psycho-logical functioning even further, has implications for publicsafety. If youths leave secure confinement even more impulsiveand susceptible to peer influence, they have an elevated risk ofrearrest, consistent with previous research associating harsherpunishment with increased recidivism (Loughran, Mulvey,Schubert, Fagan, Piquero, & Losoya, 2009). Because repeat of-fenders are subject to harsher sentences, the short-term effectsof incarceration may initiate a deleterious spiral, wherein themost immature youths are treated in ways that further dampenthe development of psychosocial maturity.

Although the present study is strengthened by its longitu-dinal design, focus on a unique population, and highly detaileddata on incarceration experiences, it is nevertheless limited inseveral respects. First, the present analyses were designed spe-cifically to examine the effects of incarceration on psychosocialmaturity. Whether incarceration affects juveniles’ mentalhealth, legal socialization, educational attainment, and otheroutcomes are important questions that are beyond the scopeof this article. Second, because of the naturalistic nature ofthe study, we cannot completely disentangle the effects of in-carceration from preexisting individual differences in psycho-social maturity. However, our finding that recent incarcerationin secure settings is associated with a short-term decline in psy-chosocial maturity was tested over and beyond the youths’ in-dividual trajectories of psychosocial maturity, adding greaterconfidence to these findings. Third, our study relied on self-re-port measures both of facility experiences and psychosocialmaturity. We are confident that our assessment of incarcerationis as accurate as possible because of careful use of the life cal-endar method as well as our familiarity with the facilities (re-call that confined youths were interviewed in the facility, sowe had ample opportunity to corroborate youths’ descriptions),but we have no similar validation of our self-report measures ofpsychosocial maturity (i.e., correlation with behavioral mea-sures of similar constructs). However, we have no reason to ex-pect that these reports are biased in ways that would create theparticular patterns of findings observed here. Fourth and fi-nally, we do not know if, given our sample, we observed a re-stricted range of psychosocial maturity or a different rate ofchange in psychosocial maturity that might have been observedamong nonantisocial youth. However, given that the focus ofthe present study was to examine how incarceration impactsthe development of psychosocial maturity, it is difficult to imag-ine finding a sample that would be more appropriate to study.

Despite these limitations, the results of our analyses haveimportant implications for policy and practice. Given socie-ty’s reliance on incarceration for its putative impact onpublic safety, incarceration will likely remain a reality fora considerable proportion of American youths, and dispro-portionately so for youths of color. However, our study in-dicates that steps can be taken to reduce the iatrogenic ef-fects of incarceration on psychosocial maturity and to


improve public safety by reducing recidivism. Such effortsshould include enhanced efforts to ensure safe incarcerationcontexts. Future study is needed to test whether such efforts

cannot only protect youths from the short-term negative ef-fects of incarceration but also have an additional benefit ofreducing juvenile offending.

References

Belli, R. F. (1998). The structure of autobiographical memory and the eventhistory calendar: Potential improvements in the quality of retrospectivereports in surveys. Memory, 6, 383–406.

Caldwell, M., Skeem, J., Solekin, R., & Van Rybroek, G. (2006). Treatmentresponse of adolescent offenders with psychopathy features—A 2-yearfollow-up. Criminal Justice and Behavior, 33, 571–596.

Caspi, A., & Amell, J. (1994). The reliability of life history calendardata (DPPP Tech. Report 94–01). Madison, WI: University ofWisconsin.

Cauffman, E., & Woolard, J. (1999). The Future Outlook Inventory. Unpub-lished manuscript, MacArthur Network on Adolescent Development andJuvenile Justice.

Chung, H., Little, M., & Steinberg, L. (2005). The transition to adulthood foradolescents in the juvenile justice system: A developmental perspective.In W. Osgood, M. Foster, C. Flanagan & G. Ruth (Eds.), On your ownwithout a net: The transition to adulthood for vulnerable populations(pp. 68–91). Chicago: University of Chicago Press.

Derogatis, L., & Melisaratos, N. (1983). The Brief Symptom Inventory: Anintroductory report. Psychological Medicine, 13, 595–605.

Freedman, D., Thornton, A., Camburn, D., Alwin, D., & Young-Demarco, L.(1988). The life history calendar: A technique for collecting retrospectivedata. Sociological Methodology, 18, 37–68.

Greenberger, E., Josselson, R., Knerr, C., & Knerr, B. (1975). The measure-ment and structure of psychosocial maturity. Journal of Youth and Ado-lescence, 4, 127–143.

Helsen, M., Vollebergh, W., & Meeus, W. (2000). Social support from par-ents and friends and emotional problems in adolescence. Journal ofYouth and Adolescence, 29, 319–335.

Irwin, J., & Owen, B. (2005). Harm and the contemporary prison. In A. Lie-bling & S. Maruna (Eds.), The effects of imprisonment (pp. 94–117).London: Willan.

Kummerlowe, C. (1995). Coping with imprisonment: A long-termer’s view.In T. Flanagan (Ed.), Long-term imprisonment: Policy science and cor-rectional practice (pp. 41–50). Thousand Oaks, CA: Sage.

Loughran, T. A., Mulvey, E. P., Schubert, C. A., Fagan, J., Piquero, A. R., &Losoya, S. H. (2009). Estimating a dose–response relationship betweenlength of stay and future recidivism in serious juvenile offenders. Crim-inology, 47, 699–740.

Monahan, K. C., Steinberg, L., Cauffman, E., & Mulvey, E. (2009). Trajec-tories of antisocial behavior and psychosocial maturity from adolescenceto young adulthood. Developmental Psychology 45, 1654–1668.

Monahan, K. C., Steinberg, L., Cauffman, E., & Mulvey, E. (2011). Psycho-social (im)maturity from adolescence to early adulthood: Distinguishingbetween adolescence-limited and persistent antisocial behavior. Manu-script submitted for publication.

Mulvey, E. P., Schubert, C. A., & Chung, H. L. (2007). Service use after courtinvolvement in a sample of serious adolescent offenders. Child and YouthServices Review, 29, 518–544.

Mulvey, E. P., Schubert, C. A., & Odgers, C. A. (2010). A method for mea-suring organizational functioning in juvenile justice facilities using resi-dent ratings. Criminal Justice and Behavior, 37, 1255–1277.

NationalCouncilonCrimeandDelinquency. (2007).Andjustice forsome:Differ-ential treatment of youth of color in the justice system. Oakland, CA: Author.

Patterson, J. M., & McCubbin, H. I. (1987). Adolescent coping style and be-haviors—Conceptualization and measurement. Journal of Adolescence,10, 163–186.

Rosner, L. (1990). Documenting the destruction of a successful juvenile re-habilitation program. Journal of Offender Rehabilitation, 16, 175–185.

Scheier, M. F., & Carver, C. S. (1985). Optimism, coping and health: Assess-ment and implications of generalized outcome expectations. Health Psy-chology, 4, 219–247.

Schubert, C. A., Mulvey, E. P., Steinberg, L., Cauffman, E., Losoya, S. H.,Hecker, T., et al. (2004). Operational lessons from the pathways to desis-tance project. Youth Violence and Juvenile Justice, 2, 237–255.

Scott, E., & Steinberg, L. (2008). Rethinking juvenile justice. Cambridge,MA: Harvard University Press.

Shulman, E., & Cauffman, E. (in press). Coping with incarceration: How dojuvenile offenders adjust? Journal of Research on Adolescence.

Sickmund, M., Sladky, T. J., Kang, W., & Puzzanchera, C. (2008). Easy ac-cess to the census of juveniles in residential placement. Office of JuvenileJustice and Delinquency Prevention. Retrieved August 6, 2011, fromhttp://www.ojjdp.gov/ojstatbb/ezacjrp/

Snyder, H. N., & Sickmund, M. (2006). Juvenile offenders and victims: 2006national report. Washington, DC: US Department of Justice, Office ofJustice Programs, Office of Juvenile Justice and Delinquency Prevention.

Steinberg, L., & Cauffman, E. (1996). Maturity of judgment in adolescence:Psychosocial factors in adolescent decision making. Law and Human Be-havior, 20, 249–272.

Steinberg, L., Chung, H. L., & Little, M. (2004). Reentry of young offendersfrom the justice system: A developmental perspective. Youth Violenceand Juvenile Justice, 2, 21–38.

Steinberg, L., & Monahan, K. C. (2007). Age differences in resistance to peerinfluence. Developmental Psychology, 43, 1531–1543.

Strathman, A., Gleicher, F., Boninger, D., & Edwards, C. S. (1994). The con-sideration of future consequences: Weighing immediate and distant out-comes of behavior. Journal of Personality and Social Psychology, 66,742–752.

Teen Boot Camps. (2011). What problems do boot camps fix? Retrieved August18, 2011, from http://www.teenbootcamps.com/index/boot+camps/FAQ

Teplin, L. A., Abram, K. M., McClelland, G. M., Dulcan, M. K., & Mericle,A. A. (2002). Psychiatric disorders in youth in juvenile detention. Ar-chives of General Psychiatry, 59, 1133–1143.

Thornberry, T. P., Lizotte, A. J., Krohn, M. D., Farnworth, M., & Jang, S. J.(1994). Delinquent peers, beliefs, and delinquent behavior: A longitu-dinal test of interactional theory. Criminology, 32, 47–83.

Wechsler, D. (1999). Wechsler Abbreviated Scale of Intelligence. New York:Psychological Corporation.

Weinberger, D. A., & Schwartz, G. E. (1990). Distress and restraint as super-ordinate dimensions of self-reported adjustment: A typological perspec-tive. Journal of Personality, 58, 381–417.

Zimbardo, P. G. (1990). The Stanford Time Perspective Inventory. Stanford,CA: Stanford University.

Appendix A

Formulas of unconditional growth models

Level 1:

Yti ¼ p0i þ p1i(ageti � 14)þ p2i(ageti � 14)2 þ eti:

Level 2:

p0i ¼ b00 þ u0i,

p1i ¼ b10 þ u1i,

p2i ¼ b20 þ u2i,

where Yti is the outcome variable (i.e., global psychosocial maturity,responsibility, perspective, or temperance), for youth i at time t; b00

is the intercept, which is the value of the outcome at age 14; b10 isthe linear rate of growth over age; and b20 is the quadratic rate ofgrowth over age. Coefficients u0i through u2i (where significant) al-lowed for the estimation of variability in youth developmental trajec-tories of the outcome variables.


Appendix B

Formula of conditional growth model: Testing the impactof recent incarceration on psychosocial maturity

Level 1:

Yti ¼ p0i þ p1i(ageti � 14)þ p2i(ageti � 14)2

þ p3i(time-varying secure facility incarcerationti)

þ p4i(time-varying residential

treatment facility incarcerationti)þ eti:

Level 2:

p0i ¼ b00 þ b01(average secure facility incarcerationi)

þ b02(average residential treatment facility

incarcerationi)þ u0i,




p2i ¼ b20 þ u2i,

p3i ¼ b30,

p4i ¼ b40,

where b30 and b40 are the effects of incarceration in a secure facilityand/or residential treatment facility during the preceding assessmentinterval on subsequent psychosocial maturity score, and b01 throughb12 are the effects of the average amount of incarceration in a securefacility and a residential treatment facility over the course of thestudy on the psychosocial maturity intercept and slope. Thus, b30

and b40 estimate the immediate effects of incarceration on psychoso-cial maturity, b01 and b11 test whether less mature youths also tendto be incarcerated for longer times over the course of the study, andb02 and b12 examine whether youths with longer cumulative incar-ceration experiences tend to experience stunted development of psy-chosocial maturity.

Appendix C

Formula of conditional growth model: Testing the facilitycontext as a moderator of the impact of recentincarceration on psychosocial maturity

Level 1:

Yti ¼ p0i þ p1i(ageti � 14)þ p2i(ageti � 14)2

þ p3i(time-varying secure facility incarcerationti)

þ p4i(time-varying residential treatment facility

incarcerationti)þ p5i( positive secure facility featuresti)þ � � � þ p8i(negative residential treatment facility featuresti)

þ p9i( positive secure facility featuresti)

� (time-varying secure facility incarcerationti)þ � � �þ p12i(negative residential treatment facility featuresti)

� (time-varying residential treatment facility

incarcerationti)þ eti:

Level 2:


þ b02(average residential treatment

facility incarcerationi)þ u0i,

p1i ¼ b10 þ b11(average secure incarcerationi)



p2i ¼ b20 þ u2i,

p3i ¼ b30,

� � �p12i ¼ b120,

where b50–b80 represent the direct effect of the positive and negativefeatures of secure and residential treatment facilities, and b90–b120

represent their interaction with the effect of stay for each of thosetwo respective facility types.


Documents

Arrested development: The effects of incarceration on the … · 2017-06-17 · Arrested development: The effects of incarceration on the development of psychosocial maturity JULIA