Family Violence

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=udbh20

Download by: [The University of Texas at El Paso] Date: 21 July 2016, At: 13:27

Deviant Behavior

ISSN: 0163-9625 (Print) 1521-0456 (Online) Journal homepage: http://www.tandfonline.com/loi/udbh20

Social Learning Theory and Partner Violence: A Research Note

Christine S. Sellers , John K. Cochran & Kathryn A. Branch

To cite this article: Christine S. Sellers , John K. Cochran & Kathryn A. Branch (2005) Social Learning Theory and Partner Violence: A Research Note, Deviant Behavior, 26:4, 379-395, DOI: 10.1080/016396290931669

To link to this article: http://dx.doi.org/10.1080/016396290931669

Published online: 01 Sep 2006.

Submit your article to this journal

Article views: 2038

View related articles

Citing articles: 18 View citing articles

social learning theory and partner violence: a research note

Christine S. Sellers, John K. Cochran, and Kathryn A. Branch University of South Florida, Tampa, Florida, USA

Despite being a well-established area of inquiry, the study of intimate partner violence and aggression is limited by a lack of theoretical development. In this study we show how the key theoretical elements of two of the more prominent explanations for partner violence, intergenerational transmission theory and male peer support theory, are easily subsumed by Akers’ social learning theory. We then test the explanatory power and scope of social learning theory against self-report data on the prevalence of physical aggression in marital and dating relationships among a large sample of university students. The results show that our models account for a modest amount of the variation in intimate partner violence with several indicators of two of the four social learning theory constructs (differential association and differential reinforcement) attaining statistical significance; the effects of both imitation and definitions fail to attain significance.

Received 5 March 2003; accepted 7 February 2005. Address correspondence to Christine S. Sellers, Department of Criminology, University

of South Florida, 4202 E. Fowler Ave., SOC 107, Tampa, FL 33620-8100. E-mail: csellers@ chumal.cas.usf.edu

Deviant Behavior, 26: 379�395, 2005 Copyright # Taylor & Francis Inc.

ISSN: 0163-9625 print/1521-0456 online

DOI: 10.1080/016396290931669

379

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

The study of violence and aggression against partners in inti- mate relationships began in the 1970s with studies of spouse abuse, particularly physical aggression. This area of inquiry rapidly expanded to include non-physical forms of aggression, child abuse and neglect, elder abuse, abuse of parents, abuse among siblings, aggression within same-sex couples, and dating=courtship violence. Research in partner violence now enjoys several academic journals devoted to this area of inquiry, has produced hundreds, if not thousands, of scholarly works, receives considerable extramural sup- port, has prominence at professional meetings for a variety of disciplines, and has played a very influential role inform- ing public opinion and affecting legislation and public policy. In short, the study of violence among intimates is no longer a new and emergent field of study; it has arrived.

Yet this area of research reasonably can be criticized for the paucity of theoretical scholarship into the etiology of partner violence. Two efforts have made significant progress toward this end: intergenerational transmission theory (Strauss, Gelles, and Steinmentz 1980) and male peer support theory (DeKeseredy and Schwartz 1993). Intergenerational transmission theory argues that witnessing or experiencing household violence during childhood is likely to lead to future violence or victimization in adulthood as children learn to imitate such aggression in their later intimate rela- tionships. Though an attractive theory due to its policy impli- cations, empirical tests of intergenerational transmission theory have been mixed and inconclusive (Sugarman and Hotaling 1989).

Alternatively, male peer support theory stresses the influ- ence of patriarchy within intimate male-female relationships, especially when the male partner perceives that his authority has been challenged. A challenge to the male’s authority (i.e., when his partner refuses him, argues with him, nags or berates him, threatens to end the relationship, etc.) leads him to experience stress and to seek support from his male peers, often from within all-male social groups such as frater- nities, athletic teams, or gangs. Within these groups, patriar- chal values and norms are promulgated that legitimate, justify, and even encourage physical aggression against the female partner. While male peer support theory has received a small amount of empirical support, its theoretical scope is

380 C. S. Sellers et al.

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

quite limited. That is, male peer support theory is specifically designed to explain male-to-female partner violence; it can- not account for female-to-male or same-sex partner violence.

Despite their limitations, each of these theories of partner violence shares common theoretical elements consistent with those of social learning processes. Intergenerational transmission theory stresses imitation within the family across generations while male peer support theory stresses the transmission of group values. While both theories share these elements of learning, both lack a systematic and logi- cally consistent theoretical framework that delineates the specific mechanisms by which partner violence is learned. Moreover, each lacks a general level of scope to account for all forms of partner=intimate violence. It is our contention that Akers’ (1973, 1977, 1985) social learning theory accom- modates and integrates the key theoretical elements of both intergenerational transmission theory and male peer support theory into a general theoretical framework that effectively addresses these limitations. It is the purpose of this study to develop a social learning model of intimate partner violence and to test its efficacy with self-report data specifically designed for this purpose.

A SOCIAL LEARNING MODEL OF INTIMATE PARTNER VIOLENCE

Social learning theory (SLT) is comprised of four key ele- ments: imitation, definitions, differential associations, and differential reinforcement. Imitation refers to the extent to which one emulates the behavior of role models. These role models are significant others whom one admires, whom one has a perceived personal relationship, and whom one has directly observed behaving. In the present context the theory, consistent with the tenets of intergenerational transmission theory, predicts a greater probability of the use of physical aggression against one’s partner for those individuals who have personally observed admired others engaging in acts of violence against their partners.

Definitions, the second element of social learning theory, refers to the attitudes and values individuals hold regarding the morality of the law in general and the wrongfulness of specific deviant=criminal behaviors. These attitudes may

Social Learning Theory and Partner Violence 381

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

approve, disapprove, or be morally neutral toward a specific deviant behavior. Moreover, attitudes may vary in strength or salience and they may be rendered situationally inoperative. That is, while deviant behavior is likely among those who approve of it, such approval is not required for deviant beha- vior to occur. Instead, weakly held conventional morals and values or situationally neutralized morals and values are suf- ficient to generate deviant behavior. Conversely, the more individuals that endorse norms and values against deviant behavior, the less likely they are to engage in it. Thus, partner violence is predicted to be most likely among those who accept it, weakly oppose it, or have their opposition situa- tionally neutralized. All of these attitudinal contexts are con- sistent with the patriarchal norms and values that are the foundation of male-to-female partner violence asserted by male peer support theory. However, social learning theory allows for a greater range of definitions that, in turn, broad- ens its scope into a more general theory of crime and deviant behavior.

Differential association is the third element of social learn- ing theory. Because it refers to the influence of the definitions (attitudes) and behaviors of significant others on individuals’ conduct, it too is logically consistent with similar influences implied by both intergenerational transmission theory and male peer support theory. According to social learning theory, exposure to the definitions and behaviors of others with whom one interacts has a powerful effect on one’s own definitions and behaviors. The impact of this exposure varies according to the frequency, duration, intensity, and priority of the different associations individuals have with others. In terms of partner violence, social learning theory predicts that the probability of physical aggression against one’s partner is greater among those whose close associates (family, friends, and significant others) endorse or engage in such conduct themselves.

Lastly, differential reinforcement refers to the net balance of anticipated costs and rewards associated with a given behavior. While not an element of either intergenerational transmission theory or male peer support theory, the concept of differential reinforcement is not inconsistent with their theoretical argumentation. According to social learning theory, an act that is expected to yield a greater balance of

382 C. S. Sellers et al.

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

rewards than costs is more likely to be engaged. Persons in intimate relationships most inclined to employ physical aggression against their partner are those who view partner violence as more rewarding than costly. Such rewards may derive from the act of partner violence itself, such as the domination, power, and control one has over another. The rewards of partner violence also may derive from the support and encouragement one receives from others (such as the all- male peer groups identified by male peer support theory). Conversely, partner violence is less likely to occur among those for whom the anticipated costs exceed the anticipated rewards. These costs are the various social and non-social losses a person suffers—or could suffer—as a direct result of engaging in partner violence. These could include the fear of arrest, the loss of the relationship, guilt=remorse, shame and embarrassment, social approbation, physical injury, and so forth.

As applied to partner aggression, social learning theory predicts that the prevalence of such violence is greater among those who have witnessed others they admire using aggression against a partner; who hold definitions that approve, only weakly disapprove, or are situationally neutra- lized with regard to the use of partner violence; who associ- ate with significant others who hold definitions consistent with the use of partner violence and=or engage in partner violence themselves; and who anticipate a greater balance of social and non-social rewards from partner violence than costs.

Akers’ social learning theory easily integrates and accom- modates the key theoretical elements of the leading theories of intimate partner violence while also addressing the various limitations associated with these other theories. However, it has not, to date, been extensively tested as an explanation of intimate partner violence. Our review of the available research literature yields only two sets of direct tests of social learning theory against data on partner viol- ence. The first of these, Boeringer and Akers (see Boeringer, Shehan, and Akers 1991 and Akers 1998) found support for a social learning model of sexual aggression and rape in dating and acquaintance relationships in two studies of university males. In the second study, Sellers and her associates (Sellers, Cochran, and Winfree 2003) examined the efficacy

Social Learning Theory and Partner Violence 383

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

of a social learning model to explain dating=courtship viol- ence among a sample of college students; they also tested its ability to mediate the effects of gender. While their social learning model accounted for a significant amount of the variation in courtship violence and substantially reduced the effects of gender, it did not completely mediate these gender effects. Beyond these two sets of studies, we are unaware of any additional tests of social learning theory as a more general explanation of intimate partner aggression. The current study provides a further test of the scope of the social learning model by testing each of its four key elements against self-report data on the prevalence of physical aggression against an intimate partner among a sample of currently married or currently dating college students.

METHODS

The data for this study were gathered through a self-adminis- tered survey of students attending a large urban university in Florida. The students were surveyed in graduate and under- graduate classes randomly selected from the course offerings of five colleges (Arts and Sciences, Business Administration, Education, Engineering, and Fine Arts) during the first four weeks of the Spring 1995 semester. Courses were sampled from each college in proportion to the enrollments each college contributed to the university’s total enrollment. This sampling strategy targeted a total of 2,500 students; however, absenteeism on the day of the survey and enrollments of stu- dents in more than one sampled course produced an overall response rate of 73%. The current study is based on two sub- sets of 1,641 students who completed the questionnaire and who report having had at least one serious relationship: (1) for analyses examining the use of violence in current dating relationships we use 945 students who indicated that they were currently dating, going steady, engaged, and=or cohabi- ting with an intimate partner, and (2) for analyses examining violence in marital relationships we use 255 students who indicated that they were currently married. Married stu- dents were on average older than the total sample and the dating subsample (30.4 years of age vs. 24.2 and 22.5, respectively). Married students also were more likely to be graduate students, seniors, and juniors and less likely to be

384 C. S. Sellers et al.

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

freshmen or sophomores. Other than these two differences, the socio-demographic profiles of both subsamples were nearly identical to that for the sample as a whole. Moreover, the socio-demographic profile of the overall sample was very similar to that of the total enrollment at the university.

The dependent variable consists of a measure of self- reported violence against a current partner. The measure is drawn from the physical aggression items in Strauss’ (1979) Conflict Tactics Scale. Specifically, respondents were asked how many times they had done any of the following during their current marital or dating relationship: (1) threw some- thing at a partner, (2) pushed, grabbed, or shoved a partner, (3) slapped a partner, (4) kicked, bit, or hit a partner with a fist, (5) hit their partner with something, (6) beat up their part- ner, (7) threatened their partner with a knife or gun, and (8) used a knife or gun against their partner. Responses to these items were never, once or twice, 3 to 5 times, 6 to 10 times, 11 to 20 times, and 21 or more times, coded from 0 to 6.

A principal components factor analysis of these items pro- duced two factors with eigenvalues greater than 1.00; how- ever, the scree discontinuity test revealed that a single factor solution best fits these data. Five of the eight items produced factor loadings on this factor greater than .70; the remaining three items yielded loadings of .28 to .52. Despite weaker loadings, these three items were retained to maintain consist- ency with the Conflict Tactics Scale. Responses to these eight items were summed to create a physical aggression scale with possible values ranging from 0 to 48 (Cronbach’s alpha ¼ .79). Because the distribution on this variable is so strongly skewed, we have dichotomized it into a ‘‘prevalence of physical aggression’’ indicator, where 0 ¼ never used any form of physical aggression against a partner and 1 ¼ used some form of physical aggression at least once against a partner.

The independent variables in this study are drawn from Akers’ four social learning constructs: imitation, differential reinforcement, definitions, and differential associations. Imi- tation was measured as an additive index of the total number of admired models which the respondent had actually seen using physical actions, such as hitting, slapping, kicking, or punching a partner during a disagreement. Admired models included actors on television or in movies, parents, siblings,

Social Learning Theory and Partner Violence 385

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

other relatives, friends, and other people. Scores on the imitation index range from 0, for no models observed, to 7, for all models observed (Chronbach’s alpha ¼ .53).

Several items were used to measure differential reinforce- ment. First, respondents were asked to report the actual or anticipated reaction of four different sets of significant others (i.e., friends, parents, partners, and other significant persons) to respondent’s use of violence against a partner. Respon- dents indicated that these significant others would either 1 ¼ approve and encourage it, 2 ¼ neither approve nor disapprove, 3 ¼ disapprove but do nothing, 4 ¼ disapprove and try to stop it, or 5 ¼ disapprove and report to the autho- rities. Second, a single three-point, ordinal measure of over- all balance of reinforcement for partner violence was included. This item ascertained the respondent’s perception of the usual or anticipated net outcome of using violence against a partner (1 ¼ mostly bad, 2 ¼ about as much good as bad, or 3 ¼ mostly good). Third, the net rewards-to-costs of using violence against a partner was measured by asking respondents to indicate which, if any, of eight rewards and eight costs they associated with using aggression against a partner. The eight rewards were: ‘‘It gave me a satisfying and rewarding feeling,’’ ‘‘It made me feel more masculine and tough,’’ ‘‘It ended the argument,’’ ‘‘It got my partner off my back,’’ ‘‘I felt powerful,’’ ‘‘My friends respected me more,’’ ‘‘My partner respected me more,’’ and ‘‘I felt more in control.’’ The eight costs were: ‘‘It made my relationship more stressful,’’ ‘‘My friends criticized me,’’ ‘‘I got arrested,’’ ‘‘It made me feel out of control,’’ ‘‘I felt ashamed,’’ ‘‘It made the argument worse,’’ ‘‘My family criticized me,’’ and ’’ I felt guilty.’’ To compute the net rewards-to-costs, the sum of identified costs was subtracted from the sum of identified rewards; this produced a measure with values ranging from �8 (all costs with no rewards) to þ8 (all rewards with no costs). Finally, two items were combined to assess respon- dent’s perceived certainty and severity of formal criminal justice responses to intimacy violence.

Definitions is measured by three additive scales: a two- item additive scale of respondent’s attitudes toward the law in general, a three-item additive scale of approving defini- tions of partner violence, and a three-item additive scale of partner violence neutralizing definitions. For each item

386 C. S. Sellers et al.

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

comprising these scales, respondents were asked to indicate the extent to which they agreed=disagreed with the following Likert-type statements (1 ¼ strongly agree to 5 ¼ strongly disagree): ‘‘We all have a moral duty to abide by the law’’ (reverse coded), ‘‘It is okay to break the law if we do not agree with it,’’ ‘‘It is against the law for a man to use violence against a woman even if they are in an intimate relationship’’ (reverse coded), ‘‘It is against the law for a woman to use violence against a man even if they are in an intimate relationship’’ (reverse coded), ‘‘Laws against the use of phy- sical violence, even in intimate relationships, should be obeyed’’ (reverse coded), ‘‘Physical violence is a part of a normal dating=marital relationship,’’ ‘‘I believe victims provoke physical violence,’’ and ‘‘In dating=marital relation- ships, physical abuse is never justified’’ (reverse coded).

Lastly, several items were employed to create our three measures of differential association. First, a three-item addi- tive scale of mother’s, father’s, and respondent’s best friend’s attitudes toward partner violence were measured by asking respondents to indicate to what degree each of these signifi- cant others would approve=disapprove of the use of physical violence against a partner (1 ¼ strongly disapprove; 4 ¼ strongly approve). Respondents also were asked to indi- cate the proportion of their best friends who had used viol- ence against a partner (1 ¼ none or almost none, 2 ¼ less than half, 3 ¼ more than half, and 4 ¼ all or almost all). Lastly, respondents were asked to report how often their mother, father, and best friend had used physical actions against a partner (1 ¼ never, 2 ¼ seldom, 3 ¼ usually, and 4 ¼ always); these items comprise the third composite measure of differential association.

RESULTS

We employ logistic regression models to test the combined and relative influence of these four social learning theory variables on the prevalence of intimate violence. We test these variables across three models: currently married stu- dents, currently dating students, and the total sample of stu- dents who report having had at least one serious relationship. The results of these analyses are reported in Table 1.

Social Learning Theory and Partner Violence 387

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

T A B L E 1

L o g is ti c R e g re ss io n M o d e ls

fo r th e E ff e c ts

o f S o c ia l L e a rn in g V a ri a b le s o n th e P re v a le n c e o f P h y si c a l

V io le n c e A g a in st

O n e ’s

P a rt n e r

T o ta l sa m p le

(n ¼

1 ,1 0 3 )

M a rr ie d sa m p le

(n ¼

2 2 5 )

D a ti n g sa m p le

(n ¼

8 4 8 )

b se (b )

O R

b se (b )

O R

b se (b )

O R

Im it a ti o n (I n d e x )

� .0 1 9

.0 6 7

0 .9 8

.1 6 8

.1 3 9

1 .1 8

� .0 6 6

.0 8 1

0 .9 4

D if f. R e in fo rc e m e n t

(R e w a rd s- c o st s)

.0 9 6 �

.0 3 0

1 .1 0

.1 6 1 �

.0 6 9

1 .1 8

.0 8 7 �

.0 3 5

1 .0 9

D if f. R e in fo rc e m e n t

(N e t o u tc o m e )

.5 3 3 �

.1 9 3

1 .7 0

1 .5 9 0 �

.8 1 0

4 .9 0

.5 5 7 �

.2 0 7

1 .7 5

D if f. R e in fo rc e m e n t

(R e a c ti o n s o f o th e rs )

.0 6 7 �

.0 2 2

1 .0 7

.1 0 0

.0 5 3

1 .1 1

.0 6 8 �

.0 2 5

1 .0 7

D if f. R e in fo rc e m e n t

(P e rc e iv e d ri sk )

� .0 1 8

.0 1 2

0 .9 8

� .0 0 2

.0 2 8

0 .9 9

� .0 2 7

.0 1 4

0 .9 7

D e fi n it io n s (g e n e ra l)

� .0 4 1

.0 4 0

0 .9 6

.0 5 4

.0 9 2

1 .0 6

�. 0 6 8

.0 4 7

0 .9 3

D e fi n it io n s (s p e c if ic )

.0 2 2

.0 4 6

1 .0 2

� .1 6 1

.1 0 6

0 .8 5

.0 6 1

.0 5 2

1 .0 6

D e fi n it io n s (n e u tr a li z in g )

.0 1 1

.0 4 1

1 .0 1

.0 2 8

.0 9 6

1 .0 3

� .0 0 4

.0 4 8

0 .9 9

D if f. A ss o c ia ti o n s

(O th e rs

d e fi n it io n s)

.0 6 9 �

.0 2 9

1 .0 7

.0 6 1

.0 6 0

1 .0 6

.0 6 8 �

.0 3 5

1 .0 7

D if f. A ss o c ia ti o n s

(P re v . o f v io le n c e )

.2 4 7 �

.1 1 9

1 .2 8

.2 8 0

.3 0 9

1 .3 2

.2 8 1 �

.1 3 4

1 .3 3

D if f. A ss o c ia ti o n s

(F re q . o f V io le n c e )

.0 2 4

.0 2 7

1 .0 3

.0 5 1

.0 6 4

1 .0 5

.0 3 9

.0 3 2

1 .0 4

In te rc e p t

� 1 .2 1 5

� 2 .0 3 2

� 1 .2 3 6

� 2 L o g L (i n te rc e p t)

1 2 9 1 .0

2 8 5 .0

9 7 2 .2

� 2 L o g L (m

o d e l)

1 1 9 2 .1

2 4 1 .6

8 9 8 .9

M o d e l X 2

9 8 .9

4 3 .5

8 0 .3

C o rr e c te d P se u d o -R

2 .1 3 9

.2 9 0

.1 6 2

� p <

.0 5 .

388

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

The first model reported in Table 1 presents a test of social learning theory’s efficacy at explaining partner violence within the total sample of students who report that they have had at least one serious relationship. The model accounts for 13.9% of the variation in the prevalence of physical aggression against a partner with 5 of the 11 parameter esti- mates for social learning theory constructs attaining statisti- cal significance. Three of the four measures of differential reinforcement measures and two of the three measures of differential association are each positively and significantly associated with partner violence. Neither the imitation index nor any of the three definitions variables are significantly associated with intimate partner violence.

The odds of partner violence increase with the anticipation of social or non-social reinforcement. For each one-unit increase in the balance of rewards to costs associated with partner violence, the odds of such violence increase 10%. Likewise, for each one-unit increase in the level of approval for intimacy violence from one’s significant others, the like- lihood of partner violence increases by 7%. And, for each one-unit increase in the overall outcome associated with partner violence, the odds of such violence increase by 70%. Of course, the opposite also is true; for each one-unit decrease in anticipated reinforcement, social approval, and overall outcome is associated, respectively, with 10%, 7%, and 70% decreases in the odds of partner violence.

Associating with others who use or approve of the use of violence against one’s partner also is associated with signi- ficantly increased odds of such violence. Each one-unit increase in the prevalence of partner violence among one’s significant others is associated with a 33% increase in the odds that they themselves have used such violence. Like- wise, each one-unit increase in the level of attitudinal support for intimacy violence among one’s significant others increases the odds of partner violence by 7%.

Based upon these data, the prevalence of intimacy viol- ence is associated with embeddedness in a social network in which intimacy violence is employed, condoned, and rewarded. While such findings appear, at first glance, to be consistent with DeKesseredy and Schwartz’s (1993) male peer support theory, it is very important to note that these findings are not exclusive to males, but, instead, are equally

Social Learning Theory and Partner Violence 389

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

relevant to female acts of partner violence (see Sellers, Cochran, and Winfree 2003). Given this, a gender-neutral and general nature of Akers social learning theory offers a more satisfying theoretical account. In addition, while the lack of significant effects for measures of imitation and defi- nitions suggest that the explanatory role of social learning processes may be limited, the failure of these measures to attain significant effects is especially invalidating of the inter- generational transmission theory proffered by Strauss, Gelles, and Steinmentz (1980).

The results for the currently dating subsample are strikingly similar to those for the total sample. Again, neither the imi- tation index nor any of the three measures of definitions is significantly associated with intimacy violence while the same three measures of differential reinforcement and the same two measures of differential association do attain stat- istically significant, positive associations. This model can account for only 16.2% of the variation in the prevalence of intimacy violence.

Conversely, the results for the currently married subsample are somewhat different from those for the total sample of stu- dents. Among married students, intimate partner violence is significantly and positively associated with only two the four measures of differential reinforcement (balance of rewards to costs and overall or net outcome), yet this model accounts for 29% of the variation in the prevalence of intimacy violence.

Finally, tests for the equality of logistic regression coeffi- cients for the effects of significant social learning variables on the prevalence of physical aggression against one’s part- ner between the those currently married and those currently dating revealed no significant differences in these effects (Brame et al. 1998).

In sum, these results suggest that the influence of social learning processes on the prevalence of physical aggression against one’s partner are consistent across samples and that these influences primarily take the form of anticipating social and non-social reinforcement from the use of aggression against one’s partner and associating with others who use or approve of the use of violence against one’s partner.

390 C. S. Sellers et al.

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

DISCUSSION

Despite being a fairly well-established field of inquiry, inti- mate partner violence still suffers from its rather atheoretical nature. To date, very few theoretical accounts of intimate partner violence have been generated. Of these, two of the more prominent approaches are especially noteworthy. These are the intergenerational transmission theory originally developed by Strauss and his associates (1980) and DeKesseredy and Schwartz’ (1993) male peer support theory. Intergenerational transmission theory has enjoyed great popularity despite the rather inconclusive nature of empirical tests of its validity. Conversely, male peer support theory is severely limited in its theoretical scope even though it has received a small but consistent amount of empirical support.

It is our contention that the study of intimate partner viol- ence will best advance through the development and testing of systematic and logically consistent general theories of its cause (Sellers 1999). Toward this end we have introduced Akers’ social learning theory. We have established that social learning theory can easily and logically subsume (integrate) the key theoretical elements of both intergenerational trans- mission theory and male peer support theory and can do so within a parsimonious, logically consistent, systematic, and general theoretical framework. Furthermore, we have shown with self-report data on marital and courtship violence that this social learning model is viable. Those who anticipate a greater net balance of rewards to costs from intimate partner violence or who associate with others who use violence against their partners, who hold attitudes supportive of such violence, and who do not react negatively to others use of intimacy violence are significantly more likely to use viol- ence against their partners.

Moreover, our models further establish the generality of Akers’ social learning theory. That is, while Akers’ theory has successfully withstood more than 100 attempts at falsifi- cation (Sellers et al. 2000) it has been criticized as a theory of minor and=or youthful misbehavior (Curran and Renzetti 1994; Rojek and Jensen 1996; Morash 1999) because most tests of its empirical validity have been limited to adolescent substance use (Akers et al. 1979; Akers and Cochran 1983;

Social Learning Theory and Partner Violence 391

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

Krohn et al. 1985; Sellers and Winfree 1990; Winfree and Bernat 1998; Winfree and Griffiths 1983; Winfree, Griffiths, and Sellers 1989) and other minor forms of youthful deviance such as academic dishonesty (Lanza-Kaduce and Klug 1986), computer crimes, (Skinner and Fream 1997), or premarital sex (DiBlasio and Benda 1990). This study shows that social learning theory also can explain a signifi- cant amount of variation in serious violent and adult crime= deviance (see also Boeringer et al. 1991; Akers 1998; Sellers et al. 2000; Akers et al. 1989).

We note, however, that not each of the four social learning constructs were significant predictors of intimate partner violence; in fact, only two of the four produced measures that attained statistically significant effects: differential association and differential reinforcement. Measures of both definitions and imitation consistently failed to attain statisti- cally significant associations with intimate partner violence in these data. The absence of significant effects for the three measures of definitions is quite unusual when compared to the findings of most tests of Akers’ social learning theory; in fact a recent meta-analysis of such studies revealed that measures of definitions are typically among the strongest social learning elements, second only to measures of differ- ential association (Sellers et al. 2000).

While early research established a relationship between attitudinal acceptance of violence and its usage (Cate et al. 1982; Stets and Priog-Good 1987), we believe that university students have subsequently become heavily indoctrinated against the endorsement of violence. Thus, responses to the items that comprise our definitions scales may reflect a social desirability bias.

The failure of the imitation index to significantly predict intimate partner violence also is understandable. The role models from which imitation effects derive are often the same persons with whom one is in differential association. As such our measures of imitation and differential asso- ciation may tend to covary such that the effects of one (imitation) are subsumed by those for the other (differential association).

As uniquely predicted by social learning theory, the influ- ence of the social and non-social rewards relative to costs associated with partner violence is a powerful and important

392 C. S. Sellers et al.

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

correlate that is rarely included in studies of intimacy viol- ence. This is unfortunate because such neglect may have stunted the development of this field of research. Moreover, because differential reinforcement integrates the idea of ‘‘rational actors’’ into the social learning process, the absence of inquiry into the rewards and costs associated with partner violence has denied policy-makers key insights into potentially effective forms for intervention.

Finally, social learning theory’s ability to ‘‘explain’’ inti- mate partner violence varies across the different forms of intimate partners sampled. That is, the amount of variation in the prevalence of intimate partner violence among mar- ried students accounted for by social learning theory (29%) was twice that accounted for among dating students (16.2%). While we are not able to fully address why this is the case, we suspect that the reason may be due, at least in part, to the more enduring and embedded context of mar- riage, on average, compared to that of dating relationships, especially among college students. Clearly additional work is badly needed both in the development and testing of alter- native theories of intimate partner violence and in the contin- ued development and testing of Akers’ social learning theory.

REFERENCES

Akers, Ronald L. 1973. Deviant Behavior: A Social Learning Approach. Belmont, CA: Wadsworth.

———. 1977. Deviant Behavior: A Social Learning Approach. Second ed. Belmont, CA: Wadsworth.

———. 1985. Deviant Behavior: A Social Learning Approach. Third ed. Belmont, CA: Wadsworth.

———. 1998. Social Learning and Social Structure: A General Theory of Crime and Deviance. Boston, MA: Northeastern University Press.

Akers, Ronald L. and John K. Cochran. 1983. ‘‘Adolescent Marijuana Use: A Test of Three Theories of Deviant Behavior.’’ Deviant Behavior 6:323�346.

Akers, Ronald L., Marvin D. Krohn, Lonn Lanza-Kaduce, and Marcia Radosevich. 1979. ‘‘Social Learning and Deviant Behavior: A Specific Test of a General Theory.’’ American Sociological Review 44:635�55.

Akers, Ronald L., Anthony J. LaGreca, John K. Cochran, and Christine S. Sellers. 1989. ‘‘Social Learning Theory and Alcohol Behavior among the Elderly.’’ Sociological Quarterly 30:625�38.

Social Learning Theory and Partner Violence 393

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

Boeringer, Scot, Constance L. Shehan, and Ronald L. Akers. 1991. ‘‘Social Context and Social Learning in Sexual Coercion and Aggression: Assessing the Contribution of Fraternity Membership.’’ Family Relations 40:558�64.

Cate, R. M., J. M. Heston, J. Koval, F. S. Christopher, and S. Lloyd. 1982. ‘‘Premarital Abuse: A Social Psychological Perspective.’’ Journal of Family Issues 3:79�90.

Curran, Daniel and Claire M. Renzetti. 1994. Theories of Crime. Boston, MA: Allyn and Bacon.

DeKeseredy, Walter S. and Martin D. Schwartz. 1993. ‘‘Male Peer Support and Woman Abuse: An Expansion of DeKeseredy’s Model.’’ Sociological Spectrum 13:393�413.

DiBlasio, F. and Bruce B. Benda. 1990. ‘‘Adolescent Sexual Behavior: A Multivariate Analysis of a Social Learning Model.’’ Journal of Youth and Adolescence 5:449�466.

Lanza-Kaduce, Lonn and Mary Klug. 1986. ‘‘Learning to Cheat: The Interaction of Moral Development and Social Learning Theories.’’ Deviant Behavior 7:243�59.

Morash, Mary. 1999. ‘‘A Consideration of Gender in Relation to Social Learning and Social Structure: A General Theory of Crime and Deviance.’’ Theoretical Criminology 3:451�62.

Rojek, Dean G. and Gary F. Jensen. 1996. Exploring Delinquency: Causes and Control. Los Angeles, CA: Roxbury.

Sellers, Christine S. 1999. ‘‘Self-Control and Intimate Violence: An Exam- ination of the Scope and Specification of the General Theory of Crime.’’ Criminology 37:375�404.

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, vol. 11, edited by Ronald L. Akers and Gary F. Jensen. New Brunswick, NJ: Transaction.

Sellers, Christine S., Travis C. Pratt, L. Thomas Winfree, Jr., and Francis T. Cullen. 2000. ‘‘The Empirical Status of Social Learning Theory: A Meta-Analysis.’’ Paper presented at the annual meetings of the Amer- ican Society of Criminology, November, San Francisco, CA.

Sellers, Christine S. and L. Thomas Winfree, Jr. 1990. ‘‘Differential Asso- ciations and Definitions: A Panel Study of Youthful Drinking Beha- vior.’’ International Journal of the Addictions 25:755�71.

Skinner, William F. and A. M. Fream. 1997. ‘‘A Social Learning Theory Analysis of Computer Crime among College Students.’’ Journal of Research in Crime and Delinquency 34:495�518.

Stets, Jan E. and Maureen A. Priog-Good. 1987. ‘‘Violence in Dating Relationships.’’ Social Psychology Quarterly 50:237�46.

394 C. S. Sellers et al.

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16

Straus, Murray A. 1979. ‘‘Measuring Intrafamily Conflict and Violence: The Conflict Tactics (CT) Scales.’’ Journal of Marriage and the Family 41:75�88.

Straus, Murray A., Richard J. Gelles, and Suzanne Steinmetz. 1980. Behind Closed Doors: Violence in the American Family. New York: Anchor=Doubleday.

Sugarman, David B. and Gerald T. Hotaling. 1989. ‘‘Dating Violence: Prevalence, Context, and Risk Markers.’’ Pp. 3�32. In Violence in Dating Relationships: Emerging Social Issues, edited by Maureen A. Pirog-Good and Jan E. Stets. New York: Praeger.

Winfree, L. Thomas, Jr. and Frances P. Bernat. 1998. ‘‘Social Learning, Self Control and Substance Abuse by Eighth Grade Students: A Tale of Two Cities.’’ Journal of Drug Issues 28:539�58.

Winfree, L. Thomas, Jr. and Curt Taylor Griffiths. 1983. ’’Social Learning and Adolescent Marijuana Use: A Trend Study of Deviant Behavior in a Rural Middle School.’’ Rural Sociology 48:219�39.

Winfree, L. Thomas, Jr., Curt T. Griffiths, and Christine S. Sellers. 1989. ‘‘Social Learning Theory, Drug Use, and American Indian Youths: A Cross-Cultural Test.’’ Justice Quarterly 6:395�417.

Social Learning Theory and Partner Violence 395

D ow

nl oa

de d

by [

T he

U ni

ve rs

it y

of T

ex as

a t

E l

P as

o] a

t 13

:2 7

21 J

ul y

20 16