30
This article was downloaded by: [University of South Florida] On: 07 February 2014, At: 06:07 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK Deviant Behavior Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/udbh20 Repetitive Intimate Partner Victimization: An Exploratory Application of Social Learning Theory John K. Cochran a , Christine S. Sellers a , Valerie Wiesbrock b & Wilson R. Palacios a a University of South Florida , Tampa, Florida, USA b University Hospitals Case Medical Center , Cleveland, Ohio, USA Published online: 09 Sep 2011. To cite this article: John K. Cochran , Christine S. Sellers , Valerie Wiesbrock & Wilson R. Palacios (2011) Repetitive Intimate Partner Victimization: An Exploratory Application of Social Learning Theory, Deviant Behavior, 32:9, 790-817, DOI: 10.1080/01639625.2010.538342 To link to this article: http://dx.doi.org/10.1080/01639625.2010.538342 PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or

Repetitive Intimate Partner Victimization: An Exploratory Application of Social Learning Theory

  • Upload
    uml

  • View
    1

  • Download
    0

Embed Size (px)

Citation preview

This article was downloaded by: [University of South Florida]On: 07 February 2014, At: 06:07Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,UK

Deviant BehaviorPublication details, including instructions forauthors and subscription information:http://www.tandfonline.com/loi/udbh20

Repetitive Intimate PartnerVictimization: An ExploratoryApplication of Social LearningTheoryJohn K. Cochran a , Christine S. Sellers a , ValerieWiesbrock b & Wilson R. Palacios aa University of South Florida , Tampa, Florida, USAb University Hospitals Case Medical Center ,Cleveland, Ohio, USAPublished online: 09 Sep 2011.

To cite this article: John K. Cochran , Christine S. Sellers , Valerie Wiesbrock &Wilson R. Palacios (2011) Repetitive Intimate Partner Victimization: An ExploratoryApplication of Social Learning Theory, Deviant Behavior, 32:9, 790-817, DOI:10.1080/01639625.2010.538342

To link to this article: http://dx.doi.org/10.1080/01639625.2010.538342

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all theinformation (the “Content”) contained in the publications on our platform.However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness,or suitability for any purpose of the Content. Any opinions and viewsexpressed in this publication are the opinions and views of the authors, andare not the views of or endorsed by Taylor & Francis. The accuracy of theContent should not be relied upon and should be independently verified withprimary sources of information. Taylor and Francis shall not be liable for anylosses, actions, claims, proceedings, demands, costs, expenses, damages,and other liabilities whatsoever or howsoever caused arising directly or

indirectly in connection with, in relation to or arising out of the use of theContent.

This article may be used for research, teaching, and private study purposes.Any substantial or systematic reproduction, redistribution, reselling, loan,sub-licensing, systematic supply, or distribution in any form to anyone isexpressly forbidden. Terms & Conditions of access and use can be found athttp://www.tandfonline.com/page/terms-and-conditions

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

repetitive intimate partnervictimization: an exploratoryapplication of sociallearning theory

John K. Cochran andChristine S. SellersUniversity of South Florida, Tampa,Florida, USA

Valerie WiesbrockUniversity Hospitals Case Medical Center,Cleveland, Ohio, USA

Wilson R. PalaciosUniversity of South Florida, Tampa,Florida, USA

The research literature on intimate partner violence(IPV) has documented a number of poignant factsthat serve as the foundation for this study. First, IPVis prevalent, frequent, and often repetitive.Moreover, repetitive violence within an intimaterelationship tends to escalate over time, both in itsfrequency of occurrence and in its severity. Wealso know that decisions to leave the relationshipdo not guarantee that the violence will end. Inaddition, the phenomenon of ‘‘mutualcombatancy,’’ prevalent in many intimate partnerrelationships, suggests that both parties in this

Received 20 November 2009; accepted 8 May 2010.Address correspondence to John K. Cochran, Ph.D., Department of Criminology,

University of South Florida, 4202 E. Fowler Ave., SOC #107, Tampa, FL 33620-8100.E-mail: [email protected]

Deviant Behavior, 32: 790–817, 2011

Copyright # Taylor & Francis Group, LLC

ISSN: 0163-9625 print=1521-0456 online

DOI: 10.1080/01639625.2010.538342

790

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

dyadic process co-share the roles of offender andvictim. Finally, we know that targets of IPV, liketheir abusers, tend to disproportionately come fromfamilies-of-origin in which violence and aggressionwere directly and=or vicariously experienced.These facts suggest that one possible starting pointfor the exploration of repetitive intimate partnervictimization (R-IPV) may derive from aninter-generational transmission, or cycle ofviolence theory, suggested more formally in sociallearning approaches to criminal and deviantbehaviors. The present study examines the extentto which measures of Akers’ social learningconstructs are able to predict repetitive intimatepartner victimization. Self-report data on intimatepartner violence among a sample of collegestudents reveal the social learning theory variables,differential association and differentialreinforcement in particular, are able to predict boththe prevalence and frequency of predict repetitiveintimate partner victimization.

It is now disingenuous to claim that the scholarly communityneglects to study the problem of intimate partner violence(IPV). The recent coordinated efforts by the Centers forDisease Control and Prevention at standardizing IPV termin-ology, definitions, measurement, and enhancing comparabil-ity across levels of data sources for the purposes of linkingglobal intervention and prevention efforts (Saltzman 2004)stands as a testament concerning the research community’slevel of commitment. Yet, despite such efforts, certain dimen-sions concerning intimate partner violence remain elusive. Assuch, in this article, we explore the issue of repetitive intimatepartner victimization (R-IPV). We follow the lead of Daigleand her colleagues (Daigle et al. 2008; see also Fisher et al.2010) who distinguish ‘‘repeat victimization’’ from ‘‘recur-rent victimization.’’ According to Daigle and her colleagues(2008), repetitive victimizations are limited to victims whoexperienced the same type of victimization more than oncewhile recurrent victimization subsumes repetitive victimiza-tions and it, thus, a broader construct. We limit our analysisto physical intimate partner victimization and hence examine

Social Learning and Intimate Partner Victimization 791

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

repetitive victimizations subsumed within the broaderconstructs of both intimate partner victimization (both physi-cal and emotional) and physical assaults across all combina-tions of victim–offender relationship.

From initial studies of spousal abuse to contemporaryscientific investigations concerning intimate partner violenceseveral poignant facts have emerged: it is prevalent, frequent,and often repetitive (Morse 1995; O’Leary et al. 1989; Straus1990; Tjaden and Thoennes 2000). Moreover repetitiveviolence within an intimate relationship tends to escalateover time, both in its frequency of occurrence within a fixedtemporal period of observation and in its severity (Andersonet al. 2003; Kimmel 2002). We also know that decisions toleave the relationship do not guarantee that the violence willend, as many victims who decide to leave often also return totheir abusers, many other continue to be abused by theirformer partner even if they do not return, and=or others mayinitiate a new although similarly abusive relationship(Anderson et al. 2003; Fleury et al. 2000; Griffing et al. 2003).These disturbing facts serve as a starting point for this article.

Additional facts, which direct this inquiry include thecommonality between intimate partner victimization andother more general forms of interpersonal violence (Duganand Apel 2005; Farrell and Pease 1993; Forrestal et al.2004; Martin and Fracchia 2004; Rich and Grey 2005;Romito et al. 2005; Taft et al. 2005; Tseloni and Pease2003). In addition, the phenomenon of bi-directionality or‘‘mutual combatancy’’ prevalent in many intimate partnerrelationships (Goodyear-Smith and Laidlaw 1999) suggeststhat both parties in this dyadic process may also co-sharethe roles of offender and victim, a finding also observed inthe research on interpersonal violence in general (Lauritsenet al. 1991). Finally, we know that targets of IPV, like theirabusers, tend disproportionately to come from families-of-origin in which violence and aggression were directlyand=or vicariously experienced (Heyman and Slep 2002;Kalmuss 1984; Walker 1984). Therefore, these facts suggestto us that one possible starting point for the exploration ofrepetitive intimate partner victimization (R-IPV) may derivefrom an intergenerational transmission, or cycle of violencetheory, suggested more formally in social learningapproaches to criminal and deviant behaviors (Sellers et al.

792 J. K. Cochran et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

2005). A quick perusal of the research literature on intimatepartner violence indicates that investigators have been usingthe intergenerational transmission of violence approach forseveral decades and that nearly all of them reference thetheoretical contributions of social learning theorists suchand Bandura, Sutherland, and Akers. However, these studiesdo not provide direct tests of social learning theory (see, forinstance, Corvo 2006; Corvo and Carpenter 2000; Kalmuss1984; Mihalic and Elliott 1997; Kerley et al. 2010; Kernsmith2006; Bevan and Higgins 2002; Swinford et al. 2000;Bassuck et al. 2006). Moreover, it is our contention thatthe intergenerational transmission of violence theory is nota fully articulated social learning theory. Its theoretical pro-positions are incomplete and facile relative to Akers sociallearning theory.

A SOCIAL LEARNING MODEL OF INTIMATE PARTNERVICTIMIZATION

One of the dominant theoretical perspectives on intimacyviolence is the Intergenerational Transmission Theory (IGT;Neugebauer 2000; Straus et al. 1980). IGT argues that witnes-sing or experiencing household violence during childhood islikely to lead to future violence or victimization in adulthoodas children learn to imitate such aggression in their laterintimate relationships (see Alexander et al. 1991; Bevan andHiggins 2002; Frias-Armenta 2002; Sugarman and Hotaling1989). IGT shares common theoretical elements consistentwith those of social learning processes. IGT stresses imitationwithin the family across generations and the transmissionof beliefs, values, and norms. While IGT contains theseelements of social learning processes, it is a discursive theor-etical rubric and lacks a systematic and logically consistenttheoretical framework delineating the specific mechanismsby which intimate partner victimization is learned. It is ourcontention that Akers’ (1985) social learning theory accom-modates and integrates the key theoretical elements of IGTinto a general theoretical framework that effectively addressesthese limitations. It is the purpose of this study to develop asocial learning model of repetitive intimate partner victimiza-tion and to provide an exploratory test of its efficacy withself-report data that can accommodate this purpose.

Social Learning and Intimate Partner Victimization 793

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

But why victimization? Both IGT and social learning theoryare theories of behavior; victimization is a status. We argue,IGT should also be predictive of intimate partner victimiza-tion. After all, witnessing violence and aggression in one’sfamily of origination should not only transmit messages con-sonant with the perpetration of violence but also its victimiza-tion. Witnesses of family violence, experiencing it, seeing itmodeled, reinforced, and justified, can become so inured toit that they can become readied and accustomed targetssuitable for their own victimization, especially repetitivevictimization. Such witnesses internalize family norms whichmay serve to neutralize the stigma of intimate partnerviolence, to accept it as normal, and perhaps even to approveof it under certain circumstances. These witnesses may alsobecome more inclined to understand the costs of leaving anabuser, to be willing to endure their own abuse so as to keepit from escalating and=or to prevent it from spreading to theirchildren, and so on. For these and other reasons, we feel thereis no basis for believing that IGT is an inappropriate theoreticalrubric for explaining intimate partner victimization as well asperpetration. Again, however, IGT is also a rather informallydeveloped, discursive learning theory that can be easily andfully subsumed (integrated) into Akers’ social learning theory.

Akers’ social learning theory (SLT) is comprised of four keyelements: imitation, definitions, differential associations, anddifferential reinforcement. Imitation refers to the extent towhich one emulates the behavior of role models. These rolemodels are significant others whom one admires, whomone has a perceived personal relationship, and whom onehas directly observed behaving. In the present context, thetheory, consistent with the tenets of IGT, predicts a greaterprobability of intimate partner victimization for those indivi-duals who have personally observed admired others engagingin acts of violence against their partners or tolerating their orother’s victimization.

Definitions, the second element of social learning theory,refers to the attitudes and values individuals hold regardingthe morality of the law in general and the wrongfulnessof specific deviant=criminal behaviors. These attitudes mayapprove, disapprove, or be morally neutral toward a specificdeviant behavior. Moreover, attitudes may vary in strengthand=or salience and they may be rendered situationally

794 J. K. Cochran et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

inoperative. That is, while deviant behavior is likely amongthose who approve of it, such approval is not required fordeviant behavior to occur. Instead, weakly held conven-tional morals and values or situationally neutralized moralsand values are sufficient to generate deviant behavior. Con-versely, the more strongly individuals endorse norms andvalues against deviant behavior, the less likely they are toengage in it. Thus, intimate partner victimization is predictedto be most likely among those who accept it, tolerate it,weakly oppose it, and=or have their opposition situationallyneutralized.

Differential association is the third element of social learn-ing theory. It refers to the influence of the definitions(attitudes) and behaviors of significant others. According tosocial learning theory, exposure to the definitions and beha-viors of others with whom one interacts has a powerful effecton one’s own definitions and behaviors. The impact of thisexposure varies according to the frequency, duration, inten-sity, and priority of the different associations individuals havewith others. In terms of intimate partner victimization, sociallearning theory predicts that the probability of repetitivephysical aggression by one’s partner is greater among thosewhose close associates (family, friends, and significant others)endorse and=or engage in such conduct themselves.

Lastly, differential reinforcement refers to the net balanceof anticipated costs and rewards associated with a givenbehavior. According to social learning theory, an act that isexpected to yield a greater balance of rewards than costs ismore likely to be repeated. Persons in intimate relationshipsmost inclined to experience repetitive intimate partner victi-mization are those who view tolerating such victimizationas more rewarding than costly, or who may view their toler-ance of such violence as less costly than their alternatives.The rewards of tolerating their partner’s violence may includemaintaining the relationship, place to live, and the continuedfinancial support for one’s self and children and feeling lovedby, wanted, and important to one’s partner. The rewards ofsuch tolerance may also derive from the advice and encour-agement one receives from significant others to stay in therelationship; also rewarding is the avoidance of their criti-cism. Conversely, repetitive intimate partner victimizationis less likely to occur among those for whom the anticipated

Social Learning and Intimate Partner Victimization 795

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

costs exceed any anticipated rewards. These costs are thevarious social and non-social losses a person suffers—orcould suffer—as a direct result of intimate partner violence.These could include the fear of the loss of the relationship,diminished self-worth, shame and embarrassment, socialapprobation, physical injury, and so on.

In summation, as applied to repetitive intimate partnervictimization, social learning theory predicts that the preva-lence and frequency of repetitive or repetitive victimizationis greater among those who (1) have witnessed others theyadmire using aggression against a partner or tolerating theirpartner’s use aggression against them; (2) hold definitions thatapprove, tolerate, only weakly disapprove, or are situation-ally neutralized with regard to the use of partner violence;(3) associate with significant others who hold definitionsconsistent with the use of partner violence and=or engagein partner violence themselves; and (4) anticipate a greaterbalance of social and non-social rewards than costs fromtolerating partner violence.

Akers’ social learning theory easily integrates and accom-modates the key theoretical elements of IGT while alsoaddressing the various limitations associatedwith it. However,Akers’ social learning theory has not, to date, been tested as anexplanation of intimate partner victimization. In fact, ourreview of the available research literature yields only two setsof direct tests of social learning theory against data on partnerviolence. The first of these, Boeringer and Akers (see Boeringeret al. 1991 and Akers 1998) found support for a social learningmodel of sexual aggression and rape in dating and acquaint-ance relationships in two studies of university males. In thesecond study, Sellers and her associates (Sellers et al. 2003;see also Sellers et al. 2005) examined the efficacy of a sociallearning model to explain dating=courtship violence amonga sample of college students; they also tested its ability tomediate the effects of gender. While their social learningmodel accounted for a significant amount of the variation incourtship violence and substantially reduced the effects ofgender, it did not completely mediate these gender effects.Beyond these two sets of studies, we are unaware of anyadditional tests of Akers’ social learning theory as a moregeneral explanation of intimate partner aggression and no testsof its efficacy with regard to intimate partner victimization.

796 J. K. Cochran et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

The current study provides an exploratory test of the scope ofthe social learning model by testing each of its four keyelements against self-report data on the prevalence andfrequency of intimate partner among a sample of currentlymarried or currently dating college students.

METHODS

The data for this study were gathered through a self-administered survey of students attending a large urbanuniversity in Florida. The students were surveyed in graduateand undergraduate classes randomly selected from the courseofferings of five colleges (Arts and Sciences, Business Admin-istration, Education, Engineering, and Fine Arts) during thefirst four weeks of the Spring 1995 semester. Courses weresampled from each college in proportion to the enrollmentseach college contributed to the university’s total enrollment.This sampling strategy targeted a total of 2,500 students; how-ever, absenteeism on the day of the survey and enrollments ofstudents in more than one sampled course produced an over-all response rate of 73%. The current study is based thosestudents who completed the questionnaire, who report beingcurrently involved in an intimate relationship (i.e., married ordating), and who also report having had at least one previousserious relationship (n¼ 1,124). The sociodemographicprofile of the sample was very similar to that of the totalenrollment at the university. While these data are quite dated,the research question examine herein is not be time depen-dent. Moreover, these data, unlike most other self-reporteddata collections, were specifically designed to examine theto test the predictive efficacy of Akers’ social learning theoryon intimate partner violence. Finally, while these self-reportdata are derived from a sample of college students, it is note-worthy that a substantial number of the respondents weremarried or co-habiting and as we report below the prevalenceand frequency of intimate partner violence among the studentssampled was quite substantial.

Dependent Variables

The dependent variables consist of three measures of self-reported victimization by their current partner. All are drawnfrom the physical aggression items in Straus’ (1979) Conflict

Social Learning and Intimate Partner Victimization 797

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

Tactics Scale. Readers should note that these data werecollected one year prior to the development of the revisedconflict tactics scale—CTS2 (see Straus et al. 1996). Specifi-cally, respondents were asked during their current marital ordating relationship how many times their partner had doneany of the following: (1) threw something at you, (2) pushed,grabbed, or shoved you, (3) slapped you, (4) kicked, bit, orhit you with a fist, (5) hit you with something, (6) beat youup, (7) threatened you with a knife or gun, and (8) used aknife or gun against you. Responses to these items werenever, once or twice, 3 to 5 times, 6 to 10 times, 11 to 20times, and 21 or more times, coded from 0 to 6.

A principal components factor analysis of these itemsproduced two factors with eigenvalues greater than 1.00;however, the scree discontinuity test revealed that a singlefactor solution best fits these data. Seven of the eight itemsproduced factor loadings on this factor greater than .55.Despite one weak loading, all eight items were retainedto maintain consistency with the Conflict Tactics Scale.Responses to these eight items were summed to create aphysical aggression victimization scale with possible valuesranging from 0 to 48 (Cronbach’s alpha¼ .82).

Because the distribution on this variable is so stronglyskewed, we have dichotomized it into a prevalence of physi-cal aggression indicator, where 0¼ no report of any form ofphysical aggression by one’s current partner (76.7%) and1¼ some form of physical aggression at least once by one’scurrent partner (23.3%). Given that social learning theorystresses the influence of social learning processes on themaintenance or continuation of behavior, we also employtwo additional measures of repetitive victimization derivedfrom the initial 8-item additive scale. The first of these, preva-lence of repeated victimization, is a trichotomous variablewhere 0¼ no victimization by current partner (76.7%),1¼ victimized once or twice by one’s current partner(8.1%), and 2¼ victimized more than twice by one’s currentpartner (15.2%). Finally, we also include a frequency ofvictimization variable in which values range from 0 to 6 ormore. This variable is naturally left-censored at zero and,because it is so positively skewed, we have also artificiallyright-censored it at the 95th percentile of its distribution(mean¼ 0.630, standard deviation¼1.388).

798 J. K. Cochran et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

Independent Variables

The independent variables in this study are drawn from Akers’four social learning constructs: imitation, differentialreinforcement, definitions, and differential associations. Weendeavored to measure the items comprising these constructsexactly as they were measured by Akers and his colleagues(1979), although modified to reflect intimate partner violencerather than adolescent substance use. Imitationwas measuredas an additive index of the total number of admired modelsthe respondent knows to have been physically victimizedby an intimate partner. Admired models included actors ontelevision or in movies, parents, siblings, other relatives,friends, and other people. Scores on the imitation index rangefrom 0, for no models observed, to 7, for all models observed(Cronbach’s alpha¼ .55).

Several items were used to measure differential reinforce-ment. First, respondents were asked to report the actual oranticipated reaction of three different sets of significantothers (i.e., friends, parents, and other significant persons)to respondent’s victimization by their partner. Respondentsindicated that these significant others would either 1¼approve and encourage it, 2¼ neither approve nor disapprove,3¼ disapprove but do nothing, 4¼disapprove and try to stopit, or 5¼ disapprove and report to the authorities. These itemswere transformed into standardized variables, entered into aprinciple components factor analysis from which a single-factor solution emerged (eigenvalue¼2.51) yielding factorloadings of .89 to .93. These items were then summed to forman additive scale (Cronbach’s¼ .90).

Second, a single three-point, ordinal measure of overallbalance of reinforcement for intimate partner victimizationwas included. This item ascertained the respondent’s percep-tion of the usual or anticipated net outcome from beingphysically victimized by an intimate partner (1¼mostlybad, 2¼ about as much good as bad, or 3¼mostly good).

Third, the net rewards-to-costs of intimate partner victimi-zation was measured by asking respondents to indicatewhich, if any, of five rewards and seven costs they associatedwith being physically victimized by their partner. The fiverewards include: ‘‘It showed me my partner really lovedme,’’ ‘‘It let my partner vent his=her frustration,’’ ‘‘I felt

Social Learning and Intimate Partner Victimization 799

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

wanted and important to my partner,’’ and ‘‘My partner wasmore affectionate afterwards.’’ The seven costs include: ‘‘I feltunwanted and=or worthless,’’ ‘‘My friends criticized me,’’ ‘‘Ifelt ashamed,’’ ‘‘I was in physical pain,’’ ‘‘I felt angry,’’ ‘‘Myfamily criticized me,’’ and ‘‘It ended the relationship.’’ Tocompute the net rewards-to-costs, the sum of identified costswas subtracted from the sum of identified rewards; thisproduced a measure with values ranging from �7 (all costswith no rewards) to þ5 (all rewards with no costs).

Finally, these three measures were entered into a principalcomponents factor analysis from which a single-factorsolution emerged (eigenvalue¼ 1.60) with loadings of .72to .73. These three measures were then summed to producethe differential reinforcement scale used in our analyses(Cronbach’s alpha¼ .56).

Definitions is an additive scale of three sub-scales each ofwhich addresses respondent’s attitudes toward the use ofpartner violence and the law in general. For each of thesesub-scales respondents were asked to indicate the extent towhich they agreed=disagreed with a series of Likert-typestatements (1¼ strongly agree to 5¼ strongly disagree), allof which have been transformed into standardized variablesprior to scale construction. The first of these sub-scales mea-sures is a three-item additive scale (Cronbach’s alpha¼ .54)comprised of the following items: ‘‘We all have a moral dutyto abide by the law’’ (reverse coded), ‘‘It is okay to break thelaw if we do not agree with it,’’ and ‘‘Laws against the use ofphysical violence, even in intimate relationships, should beobeyed’’ (reverse coded). A principle components factoranalysis of these three items produced a single-factor solution(eigenvalue¼ 1.57) with factor loading ranging from .65 to.78. The second sub-scale, approval of intimate partnerviolence, is comprised of two items: ‘‘It is against the lawfor a man to use violence against a woman even if they arein an intimate relationship’’ (reverse coded) and ‘‘It is againstthe law for a woman to use violence against a man even ifthey are in an intimate relationship’’ (reverse coded). Thissub-scale has an alpha reliability of .80. The third and finalsub-scale addresses respondent’s use of neutralizing defini-tions (Cronbach’s alpha¼ .48). It is comprised of the follow-ing three items: ‘‘Physical violence is a part of a normaldating=marital relationship,’’ ‘‘I believe victims provoke

800 J. K. Cochran et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

physical violence,’’ and ‘‘In dating=marital relationships,physical abuse is never justified’’ (reverse coded). A principalcomponents factor analysis of these three items produceda single-factor solution (eigenvalue¼ 1.50) with factorsloadings between .58 and .77.

These three sub-scales were then entered into a principlecomponents factor analysis which, in turn, produced a single-factor solution (eigenvalue¼ 1.85) with loadings of .76 to .81.These sub-scales were then summed to produce the measureof definitions used in our analyses (Cronbach’s alpha¼ .69).

Lastly, several items=sub-scales were employed to createour differential association scale. First, Mother’s, father’s,partner’s, and respondent’s best friend’s attitudes towardpartner violence were measured by asking respondents toindicate to what degree each of these significant others wouldapprove=disapprove of the use of physical violence against apartner (1¼ strongly disapprove; 4¼ strongly approve). Theseitems were transformed into standardized variables andentered into a principle components factor analysis. Theresults of the factor analysis reveal a single-factor solution(eigenvalue¼ 1.71) with factor loadings of .44 to .74. Thesestandardized items were then summed to form an additivemeasure of significant other’s definitions (Cronbach’salpha¼ .54).

Respondents were also asked to indicate the proportion oftheir best friends who had been physically victimized by apartner (1¼ none or almost none, 2¼ less than half, 3¼moremore than half, and 4¼ all or almost all). Lastly, respondentswere asked to report how often their parents, friends, andother relatives had been physically victimized by an intimatepartner (1¼never, 2¼ seldom, 3¼ usually, and 4¼ always).These three items were transformed into standardized vari-ables and entered into a principal components factor analysisfrom which two factors with eigenvalues greater than 1.00emerged; however, the scree discontinuity test suggests thata single-factor solution best fit these data. Factor loadingson this first factor range from .47 to .71. These three itemswere then summed to create a significant other victimizationsub-scale (Cronbach’s alpha¼ .64).

The proportion of significant other’s victimized, the signifi-cant other victimization sub-scale, and the significant others’definitions sub-scale were, in turn, entered into a principle

Social Learning and Intimate Partner Victimization 801

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

components factor analysis from which a single-factorsolution emerged (eigenvalue¼ 1.58) with loading of .68 to.79. These measures were then combined into a single com-posite differential associations scale (Cronbach’s alpha¼ .55).

Methods of Analysis

Because of the various ways in which we have operationa-lized respondents’ physical victimization by their currentpartner, we employ a variety of analytical techniques to testthe effects of the four elements of Akers’ social learningtheory. For the binary measure of the prevalence of physicalaggression we use logistic regression; for the trichotomousmeasure of repeated victimization we use logistic regressionfor polytomous=ordinal dependent variables—the non-significant Chi-squares reported in the score tests for the totalsample and both the male and female sub-samples suggeststhat the assumption of equal=parallel effects was sustainedwith these data (Menard 1995); and for the measure of thefrequency of victimization we employ negative binomialregression analyses (Ordinary Least Squares [OLS], Poisson,and Tobit analyses of these data generated substantivelysimilar findings to those reported herein). Akers’ social learn-ing theory is relatively silent as to whether or not the effects ofthese four constructs should be more or less intense as analy-ses move from one form of intimate partner victimization toanother; hence another reason for the exploratory nature ofthe present study. However, the relative influence of imitationshould, perhaps, be strongest for the onset of victimizationand, thus, most efficacious at predicting the prevalence ofvictimization. Conversely, Akers’s theory stresses the role ofdifferential reinforcement in the continuation of behavior (inour case victimization) and, hence, could be expected to haveits greatest efficacy at predicting repetitive victimization andthe frequency of victimization. We examine models for thetotal sample of respondents as well as separate models formales and females; we examine gender-specific modelsbecause we are concerned that the efficacy of Akers’ sociallearning model and=or its constituent elements to predictrepetitive intimate partner victimization may vary by gender(see Mihalic and Elliott 1997 or Sellers et al. 2003). Sellersand her colleagues (2003) found the effects of social learningconstructs on intimate partner violence to be gender-specific.

802 J. K. Cochran et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

RESULTS

Table 1 presents the results of the Logistic and negativebinomial regression models for the effects of the four sociallearning theory constructs on the prevalence=frequency ofphysical violence by one’s current partner. The first modelexamines the relative effects of these four measures on theprevalence of physical aggression by an intimate partnerutilizing logistic regression for a binary dependent variable;the secondmodel employs logistic regression for polytomous=ordered dependent variables to examine the relative effects ofthe four social learning theory variables with regard to thelikelihood of repeated victimization, and the third modelutilizes negative binomial regression techniques to test theefficacy of these four measures against a measure of the self-reported frequency of victimization by one’s current partner,also a measure of repetitive victimization.

The results of these three different analyses are remarkablysimilar: both the differential associations and differentialreinforcement measures are significantly and positivelyassociated with intimate partner victimization; conversely,

TABLE 1 Logistic and Negative Binomial Regression Models for theEffects of Social Learning Variables on the Prevalence=Frequency ofPhysical Violence by One’s Current Partner

Intimate partner victimization

Binarylogistic

regressionPolytomous=Orderedlogistic regression

Negativebinomialregression

(0 vs. 1þ) (0 vs. 1 vs. 2þ) (0,1,2,3,4,5,6þ)

Variable b se(b) p b se(b) p b se(b) p

Definitions .021 .016 .199 .022 .016 .165 .017 .017 .305Differentialassociations

.063 .016 .001 .065 .015 .001 .067 .014 .001

Imitation .011 .058 .851 .002 .057 .976 �.002 .060 .981Differentialreinforcement

.102 .018 .001 .112 .018 .001 .112 .016 .001

Intercept(s) �1.031 0.984 0.038�1.538

Social Learning and Intimate Partner Victimization 803

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

neither the definitions nor the imitation measure attainstatistically significant effects. According to these results thelikelihood of intimate partner victimization, the likelihoodof repeated victimization, and the frequency of victimizationare each greatest among those who anticipate more rewardsand fewer costs associated with intimate violence and amongthose who associate with significant others who have beenvictimized by their partners. More specifically, each one-unitincrease in differential association is associated with a 5–7%increase in the odds of being physically victimized andrepeatedly victimized by one’s current partner. Likewise,each one-unit increase in differential reinforcement is asso-ciated with a 9–12% increase in the odds of victimization=repeated victimization. While these three models attempt topredict three very different manifestations of intimate partnervictimization, the efficacy of each of the four social learningtheory constructs is nearly identical across form of intimatepartner victimization. Inconsistent with our expectations,the predictive efficacy of the imitation variable was invariantacross the three models as too was that of the differentialreinforcement measure. Importantly, as an exploratory testof Akers’ social learning theory, only two of the four constructattained statistically significant effects, differential associa-tions and differential reinforcement. While it is not uncom-mon for the effects of modeling=imitation to be rathermodest in previous tests of Akers’ social learning theory, itis atypical for the definitions construct to have such lowpredictive efficacy (see Pratt et al. 2010). Moreover, each ofthese models accounts for only 10–11% of the variance inintimate partner victimization, unusually low when com-pared to other tests of social learning theory.

Table 2 presents the results of separate analyses for malesand females of the models presented in Table 1. As with thefindings reported in Table 1, the results presented in Table 2are quite consistent within gender across the three differentmeasures of victimization experience and their associatedanalytic techniques. Conversely, and more importantly, theresults in Table 2 report gendered differences in the effectsof the four social learning variables on victimization experi-ence. For males, the likelihood of intimate partner victimiza-tion, the likelihood of repeated victimization, and thefrequency of victimization are each greatest among those

804 J. K. Cochran et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

TABLE

2Lo

gistic

andNegativeBinomialRegressionModelsfortheEffectsofSo

cial

LearningVariablesonthe

Prevalence

=Freq

uen

cyofPhysical

Violence

byOne’sCurren

tPartner

forMales

andFemales

Intimatepartner

victim

ization

Binarylogistic

regression

Polytomous=Ordered

logistic

regression

Neg

ativebinomialregression

(0vs.1)

(0vs.1vs.2þ)

(0,1,2,3,4,5,6þ)

Males

Females

Males

Females

Males

Females

Variable

bse(b)

pb

se(b)

pb

se(b)

pb

se(b)

pb

se(b)

pb

se(b)

p

Definitions

.004

.024

.860

.031

.022

.167

.007

.024

.775

.031

.022

.153

�.015

.027

.577

.043

.023

.055

Differential

associations

.045

.031

.153

.073

.019

.001

.047

.030

.125

.076

.018

.001

.057

.026

.032

.080

.018

.001

Imitation

.061

.098

.537

�.014

.073

.852

.066

.096

.493

�.032

.072

.654

�.004.097

.968

�.010

.078

.903

Differential

reinforcem

ent.077

.031

.012

.111

.024

.001

.095

.030

.002

.117

.024

.001

.112

.028

.001

.116

.024

.001

Intercep

t(s)

�1.055

�1.011

�1.607

�1.500

0.232

�0.059

�1.044

.951

805

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

who anticipate more rewards and fewer costs associatedwith intimate violence. For females, these odds are greatestamong both those who anticipate more rewards and fewercosts associated with intimate violence and among thosewho associate with significant others who have been victi-mized by their partners. However, tests for the equality ofthese maximum likelihood coefficients reveal that these gen-dered effects are not significantly different from one another(Brame et al. 1998).

If victimization is learned, then the learning should alsoinclude the adoption of approving and=or neutralizing defi-nitions. Table 3 presents the results of three OLS regressionmodels that examine this notion. The first models employsthe complete definitions scale as the dependent variable;the second and third models employ the approval of intimatepartner violence sub-scale and the neutralization sub-scaleas dependent variables. Exogenous variables in these modelsare respondent’s self-reported use of physical aggression in aprevious relationship and respondent’s report of victimiza-tion in a prior intimate relationship, each are coded dichot-omously (0¼ no, 1¼ yes). In addition, these model includethe three remaining social learning measures (i.e., the differ-ential associations scale, imitation index, and differentialreinforcement scale). Each model is limited to only thoserespondents who report at least one experience with physicalaggression by their current partner (n¼ 252).

TABLE 3 OLS Regression Models for the Effects of Prior Experiences withIntimate Partner Violence and Social Learning Variables on Approving=Neutralizing Definitions among Victims

Overall Approving Neutralizing

Variable b se(b) p b se(b) p b se(b) p

Offender �.726 .663 .1376 �.323 .257 .1048 �.212 .285 .2293Victim .380 .669 .2851 .466 .259 .0368 .013 .288 .4820Differentialassociations

.095 .064 .0680 .013 .025 .3024 .051 .027 .0312

Imitation .244 .231 .1456 .107 .089 .1159 .085 .099 .1951Differentialreinforcement

.203 .067 .0014 .050 .026 .0268 .047 .029 .0512

Intercept �.258 �.283 �.094

806 J. K. Cochran et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

TABLE

4OLS

RegressionModelsfortheEffectsofPriorExperienceswithIntimatePartner

Violence

andSo

cial

LearningVariablesonApproving=Neu

tralizingDefinitionsam

ongVictimsforMales

andFemales

Overall

Approving

Neu

tralizing

Males

Females

Males

Females

Males

Females

Variable

bse(b)

pb

se(b)

pb

se(b)

pb

se(b)

pb

se(b)

pb

se(b)

p

Offen

der

�.335

.697

.316

�.294

.331

.188

�.419

.258

.053

�.132

.129

.154

.008

.290

.489

�.010

.153

.258

Victim

.201

.352

.285

.360

.219

.050

.017

.130

.448

.199

.085

.010

.134

.147

.181

.155

.101

.063

Differential

associations

.184

.070

.004

.081

.034

.009

.057

.026

.014

.011

.013

.206

.089

.029

.001

.030

.016

.028

Imitation

.175

.215

.209

.129

.123

.148

.021

.080

.395

.026

.048

.293

.063

.090

.242

.053

.057

.175

Differential

reinforcem

ent

.065

.069

.175

.041

.042

.168

.057

.026

.014

.012

.016

.241

�.022

.029

.221

�.014

.020

.243

Intercep

t0.533

�1.302

0.237

�0.304

0.068

�0.593

807

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

Several interesting findings emerge from these models.First, these models do not perform especially well, account-ing for only 4–6% of the variance in the dependent variables.Relatedly, few of the exogenous variables attain statisticallysignificant effects. However, several effects are significantand worthy of our attention. For instance, victimization bya previous partner is positively and significantly associatedwith the adoption of definitions approving of intimate partnerviolence while the adoption of neutralizing definitions isgreater among those in association with others who haveexperienced intimate partner violence. Finally, the adoptionof both approving and neutralizing definitions is greatestamong those victims who anticipate greater rewards andfewer costs associated with their victimization.

The models presented in Table 3 were also analyzedseparately for males and females (see Table 4). Some notablegendered effects are evident. First, for males the adoption ofdefinitions regarding intimate partner violence, whetherapproving, neutralizing, or overall, appears to be a functionof their associations with significant others. The adoption ofapproving definitions, for males, is also greater among thosewho anticipate greater rewards and fewer costs associatedwith their victimization. For females, the adoption of neutra-lizing definitions is greater among those who associate withother who have experienced intimate partner violence, whilethe adoption of approving definitions is greatest among thosewho have themselves experienced higher levels of intimatepartner violence. Both victimization experiences and associ-ating with others who have been the victims of intimate part-ner influence the adoption of definitions regarding intimatepartner violence. Despite these differences in effects acrossmales and females, tests for the equality of these regressioncoefficients reveal that none are significantly different inmagnitude (Paternoster et al. 1998).

DISCUSSION

Perhaps the leading theoretical perspective accounting forintimate partner violence is the Intergenerational TransmissionTheory (IGT; Neugebauer 2000; Straus et al. 1980). Thistheory, at its core, is a form of social learning theory directedat predicting interpersonal violence and aggression among or

808 J. K. Cochran et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

between intimates. It stresses that intimate partner violence islearned=transmitted via both direct and indirect=vicariouswitnessing of violence within the family of origination(see Alexander et al. 1991; Bevan and Higgins 2002;Frias-Armenta 2002; Sugarman and Hotaling 1989). Childrenwitness the use of violence among family members, they see itmodeled, they see it reinforced, and they internalize moralmessages signifying its acceptance. Such children are at great-er risk of perpetrating intimate partner violence themselveswhen they enter into adulthood.

IGT has garnered tremendous empirical support over thelast several decades (see, for instance, Corvo 2006; Corvoand Carpenter 2000; Kalmuss 1984; Mihalic and Elliott1997; Kerley et al. 2009; Kernsmith 2006; Bevan andHiggins 2002; Swinford et al. 2000; Bassuck et al. 2006); itclearly is a viable theoretical explanation for the perpetrationof intimate partner violence. But, we argue, IGT should alsobe predictive of intimate partner victimization. After all,witnessing violence and aggression in one’s family oforigination should not only transmit messages consonantwith the perpetration of violence but also its victimization.Witnesses of family violence, experiencing it, seeing it mod-eled, reinforced, and justified, can become so inured to itthat they can become readied and accustomed targetssuitable for their own victimization, especially repetitivevictimization. However, IGT is essentially a discursivetheoretical paradigm without a formal set of theoretical pro-positions. Conversely, Akers’ social learning theory, whicheasily subsumes IGT, is a more formally stated theoreticalperspective amenable to more direct testing and falsification(Sellers et al. 2003; 2005).

Hence, our objective was to develop a social learningmodel of repetitive intimate partner victimization. We claimthat social learning theory can easily and logically subsume(integrate) the key theoretical elements found in intergenera-tional transmission theory and can do so within a parsimoni-ous, logically consistent, systematic, and general theoreticalframework. Furthermore, we have shown with self-reportdata on urban college students with a history of maritaland courtship violent victimization that this social learningmodel may be viable. However, our support for Akers’ sociallearning theory was also quite mixed.

Social Learning and Intimate Partner Victimization 809

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

Only two of the four social learning constructs significantlypredicted intimate partner victimization: differential associa-tions and differential reinforcement. However, the predictiveefficacy of these constructs was largely invariant across boththe various manifestations of intimate partner victimizationand gender. We expected the effects of imitation to be stron-gest for the initiation=onset of intimate partner victimization(i.e., prevalence), but that was not the case. Likewise, weexpected the efficacy of differential reinforcement to be great-est for repetitive=frequent victimization; it was not. Finally,our models could account, at best, for only very modestamounts of the variance in our dependent variables.

While our models did not account for a large percentageof the explained variance, we can conclude those whoanticipate a greater net balance of rewards to costs fromintimate partner victimization or who associate with otherswho have experienced intimate partner victimization aresignificantly more likely to initially experience, bere-victimized, and=or be frequently victimized. In addition,we found victimization by a previous partner to be positivelyand significantly associated with the adoption of definitionsapproving of intimacy violence while the adoption of neutra-lizing definitions is greater among those in association withothers who have experienced intimate partner violence.Finally, the adoption of both approving and neutralizingdefinitions is greatest among those victims who anticipategreater rewards and fewer costs associated with their victimi-zation. In sum, while we found some compelling evidencefrom this exploratory test in support of Akers’ social learningtheory, it is clear that additional theoretical developmenttoward the explanation if repetitive intimate partner victimi-zation is needed.

LIMITATIONS

Several key limitations may account for the overall perfor-mance of our models. First, our models can be construedas under-specified for it did not include control for the stan-dard sociodemographic variables. This under-specificationwas intentional on our part as our purpose was not to gener-ate a model that best predicts repetitive intimate partner vic-timization but to provide an exploratory test of the predictive

810 J. K. Cochran et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

efficacy of Akers’ social learning theory on repetitive intimatepartner victimization.

Second, ourmeasure of self-reported violence against a part-ner was derived from the physical aggression items in Straus’(1979) Conflict Tactic Scale. While the CTS was once the stan-dard in intimate partner violent research, its theoretical andexternal validity have been questioned (see Dawkins 2001;Dobash and Dobash 2004). The traditional CTS is viewed bysome (Dobash and Dobash 2004) as an ‘‘action-based’’ assess-ment that minimizes important nuances found in relationshipprocesses that may be of relevance to our theoretical under-standing of R-IPV.Moreover, the CTS does not minimize socialdesirability bias; whichmay account for the nature of our mod-els presented herein. As an alternative, a revised conflict tacticscale (CTS2) is now available and considered the most viableassessment tool for capturing such complex social and beha-vioral constellations as psychological trauma, for example,Post-Traumatic Stress Disorder (PTSD), substance use andabuse, psychological aggression, dysphoria, and conflictnegotiation skills, to name a few (Jones et al. 2002; Marcus2004; Najavitis et al. 2004; Newton et al. 2001; Reichenheimand Moraes 2004; Straus 2004, Straus et al. 1996). Given thenuances of repetitive intimate partner victimization, future stu-dies of repetitive intimate partner violence should considerusing the CTS2 in future data collections, but these future stu-dies should also strive, as we have, to include items that wouldalso allow measurement of key theoretical constructs such thatfull test of one or more theories is possible.

A third limitation with the present study is the fact thatseveral of the scales operationalizing key constructs fromAkers’ social learning theory evidenced rather modest alphareliability coefficients (imitation at .55 and definitions at.69); the limited internal consistencies in these measuresindicates a considerable amount of measurement error andinflated standard errors that could lead to their underachieve-ment. Despite this legitimate concern, our measures weredirectly taken from those employed by Akers and his collea-gues (1979) in the first full test of his social learning theory;our only modification was to make them applicable to inti-mate partner violence rather than adolescent substance use.

Fourth, our data are dated, cross-sectional, and limited to asample of college students. Nonetheless, we feel these data

Social Learning and Intimate Partner Victimization 811

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

are still sufficient for our research purposes. The age of one’sdata, in our view, is not a relevant concern when one’s pur-pose is to provide an exploratory test of a general theoryunless there is a reason to believe that the theory itself isnot time=period invariant. Likewise, the use of age-limitedrespondents in survey research, in our case college-students,is also not especially problematic when the purpose of theresearch is to conduct an exploratory test of a general theory,unless there is reason to believe that the reach of the theory islimited to a specific age-group or cohort. Akers’ social learn-ing theory is a general theory of deviance; it’s predictive effi-cacy is not limited to a certain age group or to a specific birthcohort, nor is it limited to a specific time period. However, amajor limitation with these data is their cross-sectional nat-ure. Akers’ social learning theory is a processual theory thatclearly specifies the temporal-ordering of effects; hence a fulltest of it requires longitudinal, panel data. Our use ofcross-sectional data proscribes us from being able to makeany assertions about the causal ordering of the observedassociations. Because of this serious limitation, we presentour study a simply an exploratory test from which we wereonly able to establish the directionality of effects.

The reality that a substantial number of persons experiencerepetitive intimate partner victimization motivated the presentstudy and should motivate others as well. Future research onthis issue would be well advised to employ panel designs,multiple waves of data collection, on representative samplesof the general population. The questionnaires employed insuch studies should include measures of intimate partner viol-ence (both offending and victimization experiences) and ahost of measures of key theoretical constructs of several gen-eral theories, including social learning theories, bonding=control theories, and strain=anomie theories.

REFERENCES

Akers, Ronald L. 1985. Deviant Behavior: A Social Learning Approach.Belmont, CA: Wadsworth.

Akers, Ronald L. 1998. Social Learning and Social Structure: A GeneralTheory of Crime and Deviance. Boston: Northeastern University Press.

Akers, Ronald L., Marvin D. Krohn, Lonn Lanza-Kaduce, and MarciaRadosevich. 1979. ‘‘Social Learning and Deviant Behavior: A Specific

812 J. K. Cochran et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

Test of a General Theory.’’ American Sociological Review44:636–655.

Alexander, P. C., S. Moore, and E. R. Alexander, III. 1991. ‘‘What isTransmitted in the Intergenerational Transmission of Violence?’’Journal of marriage and the Family 53:657–668.

Anderson, M. A., P. M. Gillig, M. Sitaker, K. McCloskey, K. Malloy, andN. Grigsby. 2003. ‘‘ ‘Why Doesn’t She Just Leave?:’ A DescriptiveStudy of Victim Reported Impediments to Her Safety.’’ Journal ofFamily Violence 18:151–155.

Bassuck, E., R. Dawson, and N. Huntington. 2006. ‘‘Intimate PartnerViolence in Extremely Poor Women: Longitudinal Patterns and RiskMarkers.’’ Journal of Family Violence 21:387–399.

Bevan, E. and D. J. Higgins. 2002. ‘‘Is Domestic Violence Learned? TheContribution of Five Forms of Child Maltreatment to Men’s Violenceand Adjustment.’’ Journal of Family Violence 17:223–245.

Boeringer, Scott, Constance L. Shehan, and Ronald L. Akers. 1991. ‘‘SocialContext and Social Learning in Sexual Coercion and Aggression: Asses-sing the Contribution of Fraternity Membership.’’ Family Relations40:558–564.

Brame, Robert, Raymond Paternoster, Paul Mazerolle, and Alex Piquero.1998. ‘‘Testing for the Equality of Maximum Likelihood RegressionCoefficients between Two Independent Equations.’’ Journal ofQuantitative Criminology 14:245–261.

Corvo, K. 2006. ‘‘Violence, Separation, and Loss in Families of Origin ofDomestically Violent Men.’’ Journal of Family Violence 21:117–125.

Corvo, K. and E. H. Carpenter. 2000. ‘‘Effects of Parental SubstanceAbuse on Current Levels of Domestic Violence: A Possible Elaborationof Intergenerational Transmission Processes.’’ Journal of FamilyViolence 15:123–135.

Daigle, Leah E., Bonnie S. Fisher, and Francis T. Cullen. 2008. ‘‘The Viol-ent and Sexual Victimization of College Women: Is Repeat Victimiza-tion a Problem?’’ Journal of Interpersonal Violence 23:1296–1313.

Dawkins, N. 2001. ‘‘Responses to Partner Abuse: Modeling Paths toSubstance Abuse, Suicide Attempts, and Retaliative PartnerViolence.’’ Unpublished doctoral dissertation. Emory University.

Dobash, R. P. and R. E. Dobash. 2004. ‘‘Women’s Violence to Men inIntimate Relationships.’’ British Journal of Criminology 44:324–349.

Dugan, Laura and R. Apel. 2005. ‘‘The Differential Risk of Retaliation byRelational Distance: A More General Model of Violent Victimiza-tion.’’ Criminology 43:697–729.

Farrell, G. and Pease, K. 1993. ‘‘Once Bitten, Twice Bitten: RepeatVictimisation and Its Implications for Crime Prevention.’’ CrimePrevention Unit Paper 46. London: Home Office.

Fisher, Bonnie S., Leah E. Daigle, and Francis T. Cullen. 2010. ‘‘WhatDistinguishes Single from Recurrent Sexual Victims? The Role of

Social Learning and Intimate Partner Victimization 813

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

Lifestyle-Routine Activities and First-incident Characteristics.’’ JusticeQuarterly 27:102–129.

Fleury, R. E., C. M. Sullivan, and D. I. Bybee. 2000. ‘‘When Ending theRelationship Does Not End the Violence: Women’s Experiences ofViolence by Former Partners.’’ Violence Against Women 6:1363–1383.

Forrestal, K. A., R. Riviello, K. Martin, and J. Fracchia. 2004. ‘‘Survey onInterpersonal Violence: An Economic Impact of Note and ChildhoodExposure to Domestic Violence in Continuing the Cycle.’’ Annals ofEmergency Medicine 44:S95.

Frias-Armenta, M. 2002. ‘‘Long-Term Effects of Child Punishment onMexican Women: A Structural Model.’’ Child Abuse and Neglect26:371–386.

Goodyear-Smith, F. A. and T. M. Laidlaw. 1999. ‘‘Aggressive Acts andAssaults in Intimate Relationships: Towards an Understanding of theLiterature.’’ Behavioral Sciences and the Law 7:285–304.

Griffing, S., D. F. Ragin, R. E. Sage, L. Madry, L. E. Bingham, and B. J.Primm. 2003. ‘‘Domestic Violence Survivors’ Self-identified Reasonsfor Returning to Abusive Relationships.’’ Journal of InterpersonalViolence 17:306–319.

Heyman, R. E. and A. M. S. Slep. 2002. ‘‘Do Child Abuse and Interper-sonal Violence Lead to Adulthood Family Violence?’’ Journal ofMarriage and the Family 64:864–870.

Jones, N. T., P. Ji, M. Beck, and N. Beck. 2002. ‘‘The Reliability andValidity of the Revised Conflict Tactics Scale (CTS2) in a FemaleIncarcerated Population.’’ Journal of Family Issues 23:441–457.

Kalmuss, D. 1984. ‘‘The Intergenerational Transmission of MaritalAggression.’’ Journal of Marriage and the Family 46:11–19.

Kerley, K. R., X. Xu, B. Sirisunyaluck, and J. M. Alley. 2010. ‘‘Exposure toFamily Violence in Childhood and Intimate Partner Perpetration orVictimization in Adulthood: Exploring Intergenerational Transmissionin Urban Thailand.’’ Journal of Family Violence 25:337–347.

Kernsmith, P. 2006. ‘‘Gender Differences in the Impact of Family ofOrigin Violence on Perpetrators of Domestic Violence.’’ Journal ofFamily Violence 21:163–171.

Kimmel, M. 2002. ‘‘Gender Symmetry in Domestic Violence.’’ ViolenceAgainst Women 8:1332–1363.

Lauritsen, Janet L., Robert J. Sampson, and John H. Laub. 1991. ‘‘Assessingthe Link between Offending and Victimization among Adolescents.’’Criminology 29:265–291.

Marcus, R. F. 2004. ‘‘Dating Partners’ Responses to Simulated DatingConflict: Violence Chronicity, Expectations, and Emotional Quality ofRelationship.’’ Genetic Social and General Psychology 130:163–188.

Menard, Scott. 1995. Applied Logistic Regression Analysis. Sage Univer-sity Paper series on Quantitative Applications in the Social Sciences,07–106. Thousand Oaks, CA: Sage.

814 J. K. Cochran et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

Mihalic, S. W. and Delbert Elliott. 1997. ‘‘A Social Learning TheoryModel of Marital Violence.’’ Journal of Family Violence 12:21–47.

Morse, B. 1995. ‘‘Beyond the Conflict Tactics Scale: Assessing GenderDifferences in Partner Violence.’’ Violence and Victims 10:251–272.

Najivitis, L. M., J. Sonn, M. Walsh, and R. D. Weiss. 2004. ‘‘DomesticViolence in Women with PTSD and Substance Abuse.’’ AddictiveBehaviors 29:707–715.

Neugebauer, R. 2000. ‘‘Research on Intergenerational Transmission ofViolence: The Next Generation.’’ The Lancet 355:1116–1117.

Newton, R. R., C. D. Connelly, and J. A. Landsverk. 2001. ‘‘An Examin-ation of Measurement Characteristics and Factorial Validity of theRevised Conflict Tactics Scale.’’ Educational and PsychologicalMeasurement 61:317–335.

O’Leary, K. D., J. Barling, I. Arias, A. Rosenbaum, J. Malone, and A. Tyree.1989. ‘‘Prevalence and Stability of Physical Aggression betweenSpouses: A Longitudinal Analysis.’’ Journal of Consulting and ClinicalPsychology 57:263–268.

Paternoster, Raymond, Robert Brame, Paul Mazerolle, and Alex Piquero.1998. ‘‘Using the Correct Statistical Test for the Equality of RegressionCoefficients.’’ Criminology 36:859–866.

Pratt, Travis C., Francis T. Cullen, Christine S. Sellers, L. Thomas Winfree,Jr., T. D. Madensen, Leah E. Daigle, N. E. Fearn, and J. M. Gau. 2010.‘‘The Empirical Status of Social Learning Theory: A Meta-analysis.’’Justice Quarterly 27:765–802.

Reichenheim, M. E. and C. L. Moraes. 2004. ‘‘Comparison between theAbuse Assessment Screen and the Revised Conflict Tactics Scale forMeasuring Physical Violence during Pregnancy.’’ Journal of Epidemi-ology and Community Health 58:523–527.

Rich, J. A. and C. M. Grey. 2005. ‘‘Pathways to Recurrent Trauma amongYoung Black Men: Traumatic Stress, Substance Use, and the ‘Code ofthe Street.’ ’’ American Journal of Public Health 95:816–824.

Romito, P., J. M. Turan, and M. DeMarchi. 2005. ‘‘The Impact of Currentand Past Interpersonal Violence on Women’s Mental Health.’’ SocialSciences and Medicine 60:1717–1727.

Saltzman, L. E. 2004. ‘‘Definitional and Methodological Issues Related toTransnational Research on Intimate Partner Violence.’’ ViolenceAgainst Women 10:812–830.

Sellers, Christine S., John K. Cochran, and Kathryn A. Branch. 2005.‘‘Social Learning and Partner Violence: A Research Note.’’ DeviantBehavior 26:379–395.

Sellers, Christine S., John K. Cochran, and L. Thomas Winfree, Jr. 2003.‘‘Social Learning Theory and Courtship Violence: An Empirical Test.’’Pp. 109–127 in Social Learning Theory and the Explanation of Crime,Advances in Criminological Theory, edited by Ronald L. Akers andGary F. Jensen. vol. 11. New Brunswick, NJ: Transaction.

Social Learning and Intimate Partner Victimization 815

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

Straus, Murray A. 1979. ‘‘Measuring Intrafamily Conflict and Violence: TheConflict Tactics (ct) Scales.’’ Journal of Marriage and the Family 41:75–86.

_______. 1990. ‘‘Injury and Frequency of Assault and the RepresentativeSample Fallacy in Measuring Wife Beating and Child Abuse.’’Pp. 75–91 in Physical Violence in American Families: Risk Factorsand Adaptations to Violence in 8,145 Families, edited by M. A. Strausand R. J. Gelles. New Brunswick, NJ: Transaction.

_______. 2004. ‘‘Cross-Cultural Reliability and Validity of the RevisedConflict Tactics Scales (CTS2): A Study of University Student DatingCouples in 17 Nations.’’ Cross-Cultural Research 38:407–432.

Straus, M. A., R. J. Gelles, and S. Steinmetz. 1980. Behind Closed Doors:Violence in the American Family. New York: Anchor=Doubleday.

Straus, M. A., S. L. Hamby, S. Boney-McCoy, and D. B. Sugarman. 1996.‘‘The Revised Conflict Tactics Scales (CTS2)—Development and Pre-liminary Psychometric Data.’’ Journal of Family Issues 17:283–316.

Sugarman, D. B. and G. T. Hotaling. 1989. ‘‘Dating Violence: Preva-lence, Context, and Risk Markers.’’ Pp. 3–32 in Violence in DatingRelationships: Emerging Social Issues, edited by M. A. Pirog-Goodand J. E. Stets. New York: Praeger.

Swinford, S. P., A. DeMaris, S. A. Cernkovich, and P. C. Giordano. 2000.‘‘Harsh Physical Discipline in Childhood and Violence in LaterRomantic Involvements: The Mediating Role of Problem Behaviors.’’Journal of Marriage and Family 62:508–519.

Taft, C. T., A. P. Piess, L. J. Stalans, K. C. Koenen, L. A. King, and D. W.King. 2005. ‘‘Risk Factors for Partner Violence among a NationalSample of Combat Veterans.’’ Journal of Consulting and ClinicalPsychology 73:1512–1529.

Tjaden, J. and N. Thoennes. 2000. ‘‘Prevalence and Consequences ofMale-to-Female and Female-to-Male Intimate Partner Violence asMeasured by the National Violence Against Women Survey.’’Violence Against Women 6:142–161.

Tseloni, A. and K. Pease. 2003. ‘‘Repeat Victimization: ‘Boots’ or ‘Flags.’ ’’British Journal of Criminology 43:196–212.

Walker, L. 1984. The Battered Woman Syndrome. New York: Springer.

JOHN K. COCHRAN is Professor of Criminology and Associate Dean for Faculty Affairsfor the College of Arts & Sciences at the University of South Florida. His scholarly interestsinclude the testing of both micro- and macro-social theories of criminal=deviant behaviorand investigating issues associated with capital punishment. He has published over ninetyarticles in peer-reviewed journals.

CHRISTINE S. SELLERS is Associate Professor and Associate Chair of Criminology at theUniversity of South Florida. Her scholarly interests include the development and testing oftheories of criminal behavior and the study of intimate partner violence and aggression. Sheis co-author with Ronald L. Akers of Criminological Theories: Introduction, Evaluation, andApplication (4th Edition), Roxbury.

816 J. K. Cochran et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4

VALERIE WIESBROCK is manager of the IRB Administration Office for the UniversityHospitals Case Medical Center in Cleveland, Ohio. Her research interests includeassessments of the prevalence of alcohol and substance use as well as dating violence incollege populations, the prevalence of mental illness in jails and prisons, the developmentof a shortened mental health screening tool for use in correctional systems, and theimplementation of a modified cognitive behavioral therapy program for incarceratedpersons with personality disorders.

WILSON R. PALACIOS is Associate Professor of Criminology at the University of SouthFlorida. His primary research interests are in the social epidemiology of illicit drug use=abuse, illicit drug markets (social networks), and qualitative research methods and analysis.

Social Learning and Intimate Partner Victimization 817

Dow

nloa

ded

by [

Uni

vers

ity o

f So

uth

Flor

ida]

at 0

6:07

07

Febr

uary

201

4