You are on page 1of 12

What Is Transmitted in the Intergenerational Transmission of Violence?

Author(s): Pamela C. Alexander, Sharon Moore, Elmore R. Alexander and III


Source: Journal of Marriage and Family, Vol. 53, No. 3 (Aug., 1991), pp. 657-667
Published by: National Council on Family Relations
Stable URL: http://www.jstor.org/stable/352741
Accessed: 27-06-2016 09:40 UTC

REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/352741?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.

National Council on Family Relations, Wiley are collaborating with JSTOR to digitize, preserve and
extend access to Journal of Marriage and Family

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 09:40:48 UTC
All use subject to http://about.jstor.org/terms
PAMELA C. ALEXANDER, SHARON MOORE,* AND ELMORE R. ALEXANDER III**
University of Maryland

What Is Transmitted in the

Intergenerational Transmission of Violence?

with the intention of causing physical pain or in-


This study investigates the intergenerational trans-
jury to another person" (Straus, 1986: 141).
mission of violence among dating partners. Hier-
Therefore, figures referring to dating violence do
archical regression analysis was used to answer
not even include verbal abuse or threats, even
three questions: What is (are) (a) the effect of a
though the threat of assault could be more coer-
history of violence on dating violence, (b) the ef-
cive than a mild shove (Makepeace, 1981). When
fects of witnessing marital violence on one's own
the threat of violence or verbal abuse is also as-
attitudes toward women, and (c) the effect of atti-
sessed, the incidence increases dramatically (Lane
tudes toward women on dating violence? Severe
abuse by his father predicted a man's violent be-
and Gwartney-Gibbs, 1985; Thompson, 1987).
Dating violence is not only important as a phe-
havior. Witnessing marital violence predicted a
woman's liberal attitudes and a man's conserva- nomenon in itself but also because it precedes
marital violence and thus may provide a link in
tive attitudes. Finally, a discrepancy in attitudes
the intergenerational transmission of violence. In
toward women and particularly a woman's liberal
fact, Roscoe and Benaske (1985) interviewed
attitudes toward women predicted her violent be-
havior. women who had been physically abused by their
husbands and found many parallels between their
experiences of courtship violence and marital vio-
In recent years, attention has begun to focus on lence. Therefore, any attempt to understand and
the high prevalence of violence in dating relation- to intervene in the transmission of values that
ships. Makepeace (1981), for example, found that tolerate the use of violence in intimate relation-
more than one-fifth of his respondents reported ships must include a consideration of dating vio-
having been the victim or the perpetrator of pre- lence.
marital violence. Cate, Henton, Koval, Christo- Evidence on the intergenerational transmission
pher, and Lloyd (1982) reported a similar inci- of violence has focused on two main antecedents
dence of violence in their sample and also dis- of dating and marital abuse-the experience of
covered that half of this group had been involved growing up in an abusive home and the encultura-
in a reciprocally violent dating relationship. "Vio- tion of patriarchal values. Social learning theory
lence" is typically defined as "an act carried out has been used as a model to examine the effects of
either experiencing abuse oneself as a child or wit-
nessing abusiveness between one's parents. Re-
Department of Psychology, University of Maryland, College
Park, MD 20742-4411. searchers of marital and dating violence have
found that a history of abuse is related to a later
*Department of Psychology, Memphis State University, Mem- involvement in an abusive relationship for both
phis, TN 38152.
males and females (Bernard and Bernard, 1983;
**Department of Management, Kogod College of Business Ad- Kalmuss, 1984; Telch and Lindquist, 1984). How-
ministration, American University, Washington, DC 20016. ever, other researchers have noted that this history

Journal of Marriage and the Family 53 (August 1991): 657-668 657

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 09:40:48 UTC
All use subject to http://about.jstor.org/terms
658 Journal of Marriage and the Family

of abuse has differential effects upon men and quires evidence of an association between violence
women. Sigelman, Berry, and Wiles (1984) found and patriarchal values both within an abusive cou-
that having been abused as a child predicted later ple and between generations.
involvement in abuse for women but not for men, There are a number of reasons why the re-
whereas Telch and Lindquist (1984) found exactly search on intergenerational transmission of dating
the opposite results. With respect to the observa- and marital violence has produced such mixed re-
tion of marital violence, O'Leary and Curley sults. First, although most researchers acknowl-
(1986) found that abusive men were more likely to edge that violence is a continuous variable ranging
have witnessed violence between their parents from verbal abuse to varying degrees of physical
than were women in an abusive marriage. Similar- abuse, many studies make comparisons between
ly, Telch and Lindquist (1984) found that al- dichotomous groups only-couples engaging in
though more abused wives had witnessed marital physical violence and those who are not (e.g., Ber-
violence than had nonabused wives, they still were nard et al., 1985; Sigelman et al., 1984). Second,
less likely to have witnessed it than were their the research on violent couples' attitudes toward
abusive husbands. Therefore, although a history women does not attempt to address whether these
of either witnessing or experiencing abuse seems attitudes (either extremely traditional or extremely
to be prevalent among men and women in abusive liberal) were established prior to the initiation of
relationships, the exact nature of that influence abuse or were the result of the violence. For exam-
on men and women remains unclear. ple, while some women may become more mili-
Feminist theory has also been used as a con- tant about the subject of violence as a function of
ceptual framework for explaining the presence of having been abused, other women in abusive rela-
violence between males and females. Feminist tionships may come to evince a "learned helpless-
theory views violence as a manifestation of the pa- ness" response to the violence (Walker, 1983) and
triarchal structure in our culture, which is re- may actually espouse more traditional views.
flected in the patterns of behaviors and attitudes While the only certain way to determine causality
of individuals (Gentemann, 1984; Kalmuss, 1984). with respect to this question lies in the use of
Researchers have indeed discovered that abusive longitudinal research, a consideration of histori-
men frequently espouse more traditional views cal events relative to current attitudes can provide
about women than do nonabusive men (Bernard, some basis for making inferences about causa-
Bernard, and Bernard, 1985; Sigelman et al., tion. Finally, any study on the intergenerational
1984; Telch and Lindquist, 1984) or at least are transmission of violence needs to determine what
seen by their wives as having more traditional is actually transmitted or modeled from parents to
views (Rosenbaum and O'Leary, 1981). Converse- their children--abusive behaviors or the attitude
ly, more recent research has found no clear rela- that men are justified in maintaining the subordi-
tionship between men's attitudes toward women nation of women.
and violent behavior toward their partners (John- To these ends, the following study was de-
ston, 1988; Rouse, 1988). The evidence regarding signed to answer three sets of questions about dat-
women's attitudes is similarly mixed. Rosenbaum ing violence: (a) What is the effect of experiencing
and O'Leary (1981) found no difference between or witnessing violence in one's family of origin on
abused wives and those in discordant or satisfied subsequent verbal and physical abuse in a dating
marriages in their attitudes toward women. Simi- relationship? Moreover, does this experience of
larly, Korman and Leslie (1982) found no rela- family violence have differential effects on men
tionship between adherence to feminist ideology and women? (b) What is the effect of witnessing
and the experience of sexual aggression in a dating violence between one's parents on one's attitudes
relationship. Other researchers have concluded toward women? Is there a differential effect on
that the discrepancy between a couple's attitudes men and women? (c) What is the effect of one's
toward women's sex roles may be a more signifi- attitudes toward women on subsequent verbal and
cant predictor of violence than the actual values physical abuse in a dating relationship? Further-
of either the man or the woman (Rosenbaum and more, is there a differential effect on men and
O'Leary, 1981; Sigelman et al., 1984). In sum- women and is there an interaction between one's
mary, support for a feminist conceptualization of own attitudes and those of one's partner on subse-
the intergenerational transmission of violence re- quent dating violence?

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 09:40:48 UTC
All use subject to http://about.jstor.org/terms
Intergenerational Transmission of Violence 659

Figure 1 presents the hypothesized relation- of .77 to .88 for the Verbal Aggression scale and
ships among abuse in the family of origin, atti- .62 to .88 for the Violence scale.
tudes toward women, and dating violence. Respondents completed separate question-
naires to describe (a) their father's violent be-
havior toward their mother (CTSVIFM), (b) their
FIGURE 1. HYPOTHESIZED PREDICTORS OF ATTITUDES
TOWARD WOMEN AND DATING VIOLENCE
mother's violent behavior toward their father
(CTSVIMF), (c) their father's violent behavior
Abuse in Family of Origin toward themselves (CTSVIFS), (d) their mother's
(Experiencing & Witnessing) h violent behavior toward themselves (CTSVIMS),
(e) their own verbally aggressive and violent be-
havior toward their partner (CTSVBSP,
CTSVISP), and (f) their partner's verbally aggres-
Dating sive and violent behavior toward themselves
Violence (CTSVBPS, CTSVIPS). The present research re-
quired subjects to describe their parents' behavior
during the time that they were living at home and
Witnessing of Own t to describe their own behavior with their dating
Marital Violence Attitudes ? partner during the course of the dating rela-
tionship (cf. Johnston, 1988). The questionnaire
Partner's required respondents to report on the frequency
of occurrence of each conflict tactic (0 = never; 1
Presumed Attitudes
= once; 2 = twice; 3 = 3-5 times; 4 = 6-10
times; 5 = 11-20 times; and 6 = more than 20
METHOD times). Items were then summed to yield measures
of verbal aggression and violent behavior.
Subjects
As noted by Straus and Gelles (1986), several
Subjects in this study were selected from a group behaviors such as slapping, spanking, and hitting
of 459 men and women attending an urban South- with an object are commonly used as a means of
ern university who completed a series of question- discipline with children and yet would constitute
naires for extra credit. Only those individuals who abusiveness if used with a spouse. Therefore,
were not presently married but were involved in a these items were omitted from the measures of
dating relationship of at least six months duration parental violence toward the respondent
(n = 380) were included in the following analyses. (CTSVIFS, CTSVIMS), and these measures in-
The length of the dating relationships ranged stead reflect only severe violence. Finally, because
from six months to more than ten years (mean = the rates of observed physical violence by the
1.9 years, SD = .57). The final sample consisted father to the mother and by the mother to the
of 152 males and 228 females with a mean age of father were so highly correlated (r = .89, p <
20 years (SD = 3.2). .001), these two scales (CTSVIFM and
CTSVIMF) were averaged to yield one scale re-
flecting marital violence (CTSVIALL).
Measures

Conflict Tactics Scales (CTS). Developed by Attitudes toward Women Scale (A WS). This scale
Straus (1979), this is the most widely used measure was developed by Spence and Helmreich (1972) to
of family violence. Respondents are asked to indi- measure the views of respondents regarding the
cate whether and how often certain behaviors roles and rights of women in contemporary soci-
have ever been used to resolve conflict. The items ety. The present study used the short version
range from "discussed the issue calmly" to "used (Spence, Helmreich, and Stapp, 1973), which is
a knife or gun" and comprise three subscales-- composed of 25 items and which, like the longer
Reasoning, Verbal Aggression, and Violence. For version, is a Likert-type scale with four response
the purpose of this study, the Verbal Aggression options: agree strongly, agree mildly, disagree
and Violence subscales were used. Straus (1979) mildly, disagree strongly. Correlations between
has reported internal consistency reliability scores the full-length AWS and the short-form AWS

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 09:40:48 UTC
All use subject to http://about.jstor.org/terms
660 Journal of Marriage and the Family

TABLE 1. MEANS AND STANDARD DEVIATIONS FOR CONFLICT TACTICS SCALES AND
ATTITUDES TOWARD WOMEN

Males (n = 152) Females (n = 228)


Variable Mean SD Mean SD

Conflict Tactics Scales


CTSVIFS .934 3.195 .692 2.358
CTSVIMS .550 1.795 .387 1.636
CTSVIALL 3.976 8.844 4.436 8.290
CTSVBSP 9.942 7.292 9.473 7.072
CTSVBPS 9.934 7.386 7.507 7.310
CTSVISP 2.115 4.517 3.009 5.910
CTSVIPS 3.213 6.982 2.392 5.294
Attitudes toward women
AWSS 53.224 11.503 45.881 8.605
AWSP 48.710 10.701 60.477 11.797

Note: CTSVIFS = physical violence from father to self; CTSVIMS = physical violence from mother to self;
CTSVIALL = witnessing of physical violence between parents; CTSVBSP = verbal aggression from self to partner;
CTSVBPS = verbal aggression from partner to self; CTSVISP = physical violence from self to partner; CTSVIPS =
physical violence from partner to self; AWSS = attitudes toward women of self; AWSP = attitudes toward women
of partner.

were .968 for college men and .969 for college General Data Analysis Techniques
women (Spence et al., 1973).
Hierarchical regression analysis as described by
Respondents completed separate question-
Pedhazur (1982) was utilized to evaluate the three
naires to describe (a) their partner's presumed at-
research questions. For each dependent variable,
titudes toward women (AWSP) and (b) their own
the first step was to estimate an equation that con-
attitudes toward women (AWSS). Item scores
tained all of the main-effect independent vari-
were summed to yield a total score ranging from
ables. The second step was to test various interac-
25 (extremely liberal) to 100 (extremely conserva-
tion terms (created by multiplying main-effect
tive).
variables by each other) for significance. The in-
teractions were judged to be significant if their ad-
RESULTS dition to the equation containing all of the main
effects resulted in a significant increase in R2 (p
Means and standard deviations for the responses
< .05). For the significant interactions, separate
of males and females on each of the Conflict Tac-
equations were then calculated for (a) males and
tics Scales and on the Attitudes toward Women
females in the case of interactions involving sex
Scale are summarized in Table 1. Zero-order cor-
and for (b) high and low AWSS (one's attitudes
relations among family abuse variables, attitudes
toward women) for interactions involving that
toward women, and measures of dating violence
variable. By comparing the slopes of males vs. fe-
are summarized in Table 2.
males or individuals with liberal vs. conservative
attitudes, one is able to observe the change in the
nature of the relationship between the dependent

TABLE 2. ZERO-ORDER CORRELATIONS AMONG FAMILY ABUSE, ATTITUDES TOWARD


WOMEN, AND DATING VIOLENCE
Variables 1 2 3 4 5 6 7 8 9
1. CTSVIFS - .251 .642 -.056 -.076 -.017 -.055 .000 -.024
2. CTSVIMS .087 - .345 .018 -.075 .126 -.010 .078 .016
3. CTSVIALL .180 .472 - -.106 -.083 .053 -.007 .045 .039
4. AWSS -.044 -.026 .031 - .164 -.023 -.028 .031 .045
5. AWSP .103 -.040 .039 .096 - .365 .411 .544 .581
6. CTSVBSP .106 .007 .106 .066 .616 - .825 .725 .679
7. CTSVBPS .143 .054 .064 .086 .748 .821 - .697 .718
8. CTSVISP .117 -.071 .064 -.480 .780 .841 .766 - .938
9. CTSVIPS .105 -.012 -.003 .053 .867 .755 .886 .897 -

Note: Females are above the diagonal.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 09:40:48 UTC
All use subject to http://about.jstor.org/terms
Intergenerational Transmission of Violence 661

variable and the other variable involved in the in- (CTSVBPS), suggesting that males were more
teraction as either sex or attitudes vary. likely to report verbal abuse and that verbal abuse
was positively related to the length of the relation-
ship. There were no significant main effects to
Research Question 1:
predict the probability of being physically abusive
Abusiveness in the Family of Origin
to one's partner (CTSVISP). However, men were
The first research question examined the direct more likely to report being physically abused by
impact of witnessing abuse and being abused on their partner (CTSVIPS).
one's subsequent involvement in abusive relation- One interaction variable was significant for
ships. Four dependent variables were examined each of the dependent variables-the interaction
here: (a) CTSVBPS (verbally abused by partner); of sex and CTSVIFS (a history of severe abuse by
(b) CTSVBSP (verbally abused partner); (c) one's father). The derivation of the equations for
CTSVIPS (physically abused by partner); and (d) this interaction required calculating separate
CTSVISP (physically abused partner). Indepen- equations for males and females. This was ac-
dent variables consisted of the individual's sex, complished by substituting mean values of 0 for
the length of the dating relationship (RELTIME), males and 1 for females into the equation contain-
experience of physical abuse by one's father ing the interaction terms. Table 6 contains sum-
(CTSVIFS), experience of physical abuse by one's marized results for these four significant interac-
mother (CTSVIMS), and witnessing of abusive- tions. The slopes for these equations suggest that
ness between one's parents (CTSVIALL). The in- males alone are more likely to report being ver-
teraction of sex with each of the last three bally abusive (CTSVBSP), being verbally abused
variables was considered in the hierarchical analy- (CTSVBPS), being physically abusive
sis. Table 3 contains the results of the regression (CTSVISP), and being physically abused
analyses for the main-effect variables for Re- (CTSVIPS) by their partner if they have been
search Question 1. severely abused by their father.
Sex and the length of the relationship
(RELTIME) were significantly related to the
Research Question 2:
probability of being verbally abusive to one's
Attitudes toward Women
partner (CTSVBSP) as well as to the probability
of being verbally abused by one's partner Research Question 2 examined the direct impact

TABLE 3. RESULTS OF REGRESSION ANALYSES FOR THE EFFECT OF VIOLENCE IN THE FAMILY OF
ORIGIN ON DATING VIOLENCE (RESEARCH QUESTION 1)
Dependent Variable Sex CTSVIMS RELTIME CTSVIFS CTSVIALL R2 F
CTSVBSP
0 -.1014 .0588 .1477 -.0113 .0552 .0385 2.99*
t -1.9920 1.0610 2.8980 -.1930 .8820
p .0471 .2895 .0040 .8473 .3783
CTSVBPS
0 -.1724 .0281 .1558 .0152 -.0233 .0530 4.18***
t -3.4150 .5120 3.0810 .2610 -.0370
p .0007 .6093 .0022 .7939 .9701
CTSVISP
0 -.0833 -.0168 .0752 .0350 .0470 .0165 1.26
t -1.6200 -.2990 1.4600 .5910 .7420
p .1061 .7650 .1452 .5549 .4584
CTSVIPS
0 -.1292 .0056 .0725 .0373 -.0107 .0223 1.71
t -2.5180 .0990 1.4110 .6330 -.1700
p .0122 .9209 .1592 .5273 .8655
Note: CTSVBSP = verbal aggression from self to partner; CTSVBPS = verbal aggression from partner to self;
CTSVISP = physical violence from self to partner; CTSVIPS = physical violence from partner to self; CTSVIMS =
physical violence from mother to self; RELTIME = length of dating relationship; CTSVIFS = physical violence
from father to self; CTSVIALL = witnessing of physical violence between parents.
*p < .02. ***p < .001.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 09:40:48 UTC
All use subject to http://about.jstor.org/terms
662 Journal of Marriage and the Family

TABLE 4. RESULTS OF REGRESSION ANALYSES FOR THE EFFECT OF WITNESSING VIOLENCE BETWEEN
PARENTS ON ATTITUDES TOWARD WOMEN (RESEARCH QUESTION 2)
Dependent Variable Sex CTSVIALL R2 F
AWSS
0 -.2059 -.0409 .0450 8.89***
t -4.0830 -.8120
p .0001 .4174
Note: AWSS = attitudes toward women of self; CTSVIALL = witnessing of physical violence between parents.
***p < .001.

of witnessing violence between one's parents on impact of one's own attitudes toward women and
one's own attitudes toward women. The effects of one's perceptions of one's partner's attitudes to-
the independent variables of sex, CTSVIALL ward women on subsequent involvement in abu-
(witnessing violence between one's parents), and sive relationships. The dependent variables were
the interaction of sex and CTSVIALL on the de- the same as they were for Research Question 1
pendent variable of AWSS (one's own attitudes (CTSVBPS, CTIVBSP, CTSVIPS, and CTSVISP
toward women) were analyzed in a hierarchical -receiving and extending verbal and physical
analysis. Table 4 contains the results for the main- abuse). Independent variables were sex, one's
effect variables. own attitudes toward women (AWSS), and one's
Men described their attitudes toward women partner's presumed attitudes toward women
as significantly more conservative than did (AWSP). Both the interaction of sex with AWSS
women. Furthermore, the interaction of sex and and AWSP as well as the interaction of AWSS
CTSVIALL was significant. Table 6 contains with AWSP were considered in the hierarchical
these summarized results. The slopes for these analysis. Table 5 summarizes the results for the
equations suggest that females were significantly main-effects analyses for the four dependent vari-
less likely to be conservative in their attitudes ables.
toward women and males were significantly more Both sex and the presumed attitudes of one's
likely to be conservative if they had witnessed vio- partner (AWSP) were found to be significantly re-
lence between their parents (CTSVIALL). lated to being verbally abusive (CTSVBSP), being
verbally abused (CTSVBPS), being physically
abusive (CTSVISP), and being physically abused
Research Question 3:
(CTSVIPS) by one's partner. More specifically,
Attitudes toward Women and Dating Violence
men reported higher rates of verbal and physical
The third research question examined the direct abuse, both extended and received. Furthermore,

TABLE 5. RESULTS OF REGRESSION ANALYSES OF THE EFFECT OF ATTITUDES TOWARD WOMEN ON


DATING VIOLENCE (RESEARCH QUESTION 3)
Dependent Variable Sex AWSP AWSS R2 F
CTSVBSP
0 -.1029 .5113 -.0140 .2690 46.13****
t -2.2820 11.5160 -.3080
p .0230 .0001 .7582
CTSVBPS
/3 -.1730 .5953 .0009 .3826 77.67****
t -4.1730 14.5910 .0220
p .0001 .0001 .9822
CTSVISP
P-.0860 .6893 -.0156 .4787 115.08****
t -2.2580 18.3840 -.4070
p .0245 .0001 .6843
CTSVIPS
3 -.1345 .7577 -.0069 .5888 179.48****
t -3.9750 22.7560 -.2010
p .0001 .0001 .8404
****p < .0001.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 09:40:48 UTC
All use subject to http://about.jstor.org/terms
Intergenerational Transmission of Violence 663

TABLE 6. RESULTS OF REGRESSION ANALYSES OF INTERACTIONS INVOLVING SEX

Dependent Independent Variable Change Slope for Slope for


Variable Interacting with Sex Fa in R2 Males Females
CTSVBSP CTSVIFS 3.91* .01 .28 -.16
CTSVBPS CTSVIFS 7.48** .02 .46 -.18
CTSVISP CTSVIFS 4.68* .01 .68 -.15
CTSVIPS CTSVIFS 4.07* .01 .68 -.12
AWSS CTSVIALL 8.07** .02 1.74 -1.10
CTSVBSP AWSS 4.63* .01 -.01 -.07
CTSVBPS AWSS 14.63*** .03 .00 -.10
CTSVBPS AWSP 6.69** .01 .16 .11
CTSVISP AWSS 7.28** .01 -.03 -.14
CTSVISP AWSP 14.15*** .02 .30 .19
CTSVIPS AWSS 15.37*** .02 -.01 -.15
CTSVIPS AWSP 21.59*** .02 .34 .21

aF associated with the interaction. (df = 1, 376).


*p < .05. **p < .01. ***p < .001.

the reported conservatism of one's partner's atti- with one's partner's presumed attitudes toward
tudes toward women was found to predict verbal women (AWSP) were derived by calculating sep-
and physical abuse, both extended and received. arate equations for those with liberal attitudes to-
Table 6 provides summaries for the several sig- ward women (low AWSS) and for those with con-
nificant interactions of sex with AWSS and servative attitudes toward women (high AWSS).
AWSP, while Table 7 provides summaries for the This was accomplished by substituting values at
significant interactions of AWSS with AWSP. the 25th and 75th percentiles of the range of
The slopes of the equations for the interaction values for this sample for low AWSS and for high
of sex with one's own attitudes toward women AWSS, respectively. The equations suggest that,
(AWSS) suggest that women with conservative relative to those with liberal attitudes toward
views are significantly less likely than women with women (low AWSS), individuals with conserva-
liberal views to report verbal and physical abuse, tive attitudes toward women (high AWSS) are
both extended and received. Conversely, men's more likely to report being verbally and physically
reports of verbal and physical abuse, both ex- abusive if their partner's attitudes are seen as lib-
tended and received, do not appear to be related eral. On the other hand, they are less likely to re-
to their own attitudes toward women (AWSS). port being verbally and physically abused than are
The slopes of the equations for the interaction of those with liberal attitudes toward women if their
sex with one's partner's presumed attitudes to- partner's attitudes are seen as conservative.
ward women (AWSP) suggest that men who de-
scribe their partner's attitudes as conservative are
significantly more likely than women to report be- DISCUSSION
ing verbally abused, being physically abusive, and In considering these results, it is important to re-
being physically abused by their partner. member that they are based on the reports of the
The slopes of the equations for the interaction respondents. Therefore, they reflect the respon-
of one's own attitudes toward women (AWSS) dents' perceptions of themselves and their dating

TABLE 7. RESULTS OF REGRESSION ANALYSES OF INTERACTIONS OF AWSS AND AWSP

Dependent Change Slope of AWSP x Slope of AWSP x


Variable Fa in R2 Low AWSS High AWSS
CTSVBSP 12.69*** .03 .02 -.03
CTSVBPS 7.33** .01 .07 .04
CTSVISP 50.20*** .06 .01 -.12
CTSVIPS 52.72*** .05 .20 .09

Note: Low and high AWSS are defined as the 25th and 75th percentiles, respectively, of the range of values for
this sample.
aF associated with the interaction (df = 1, 375).
**p < .01. ***p < .001.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 09:40:48 UTC
All use subject to http://about.jstor.org/terms
664 Journal ofMarriage and the Family

partners and, as such, are subject to bias. By way Research Question 1:


of illustration, although it would not be unreason- Abusiveness in the Family of Origin
able to assume that the respondents and their
Witnessing violence between one's parents was
partners are from the same population, they re-
not predictive of either extending or receiving ver-
portedly differ on certain dimensions. This discus-
sion will attempt to reconcile some of the more bal and physical abuse in the dating relationship.
overt discrepancies. Similarly, being physically abused by one's
mother did not predict dating violence. Further-
more, physical abuse by one's father was not pre-
Incidence of Verbal and Physical Abuse dictive of a woman's involvement in dating vio-
Males reported being more verbally abusive to lence. However, physical abuse by one's father
their partners than did females as well as receiving was highly predictive of a man's extending and re-
substantially more verbal abuse from their part- ceiving both verbal abuse and physical abuse in a
ners than they themselves extended. On the other dating relationship (suggesting that the abuse he
hand, females did not report as much verbal abuse received was probably in response to his initiation
of it).
in their dating relationships. However, both males
and females reported that their own level of verbal The findings from this study were consistent
aggression was similar to that of their partners. with those of Rosenbaum and O'Leary (1981)
Similarly, males described themselves as being sig- and O'Leary and Curley (1986) that males are
nificantly more physically violent than did fe- more likely to model the abusive behaviors of
males and as being more abused themselves by their parents than are females. In this sample, the
their partners. Both males and females were very strongest direct determinant of later dating vio-
likely to characterize any physical violence in their lence was physical abuse of the male by his father.
relationship as being reciprocal (r = .897 and .938 This finding is consistent with the observation by
for males and females, respectively). Whether Comins (1984) that family environments charac-
males and females see themselves as equally re- terized by harsh physical punishment are associ-
sponsible for the violence or as merely fighting ated with the perpetration of dating violence. The
back is less clear. lack of influence of witnessing marital violence
There is no ready explanation for why males may be attributed to investigating a nonclinical
described more abusiveness overall. The idea that sample (as opposed to the clinical samples used by
the men might be more willing to acknowledge Rosenbaum and O'Leary, 1981, and O'Leary and
abuse in their dating relationships would be con- Curley, 1986) in which the roles of husband and
sistent with other findings. For example, Laner wife had not yet been adopted. The primacy of
(1983) suggested that men are more likely to re- the effect of physical abuse of the male by his
port a higher level of violence simply because they father suggests the importance of emphasizing
are more socialized to accept or approve of it. In this issue in the treatment of abusive couples.
any case, the levels of physical violence described
by this sample (32.5% described themselves as be- Research Question 2:
ing physically abusive and 31.8% as being physi- Attitudes toward Women
cally abused) were comparable to or even greater
The second regression analysis was used to deter-
than those found by other researchers (Cate et al.,
mine the influence of the family of origin in estab-
1982; Lane and Gwartney-Gibbs, 1985; Make-
peace, 1981). Therefore, these figures, combined lishing one's attitudes toward women. In particu-
lar, it was hypothesized that one's attitudes to-
with the high reciprocity in violence described by
the respondents, argue against an effect of social ward women would be influenced by having wit-
desirability in this study. nessed violence between one's parents. The signi-
The length of the dating relationship was ficant interaction between sex and witnessing
found to be significantly related to the amount of marital violence suggests that the latter may in-
deed have a powerful effect in the enculturation
verbal abuse extended and received. However, it
was not found to predict physical violence in the of values in children who see their parents fight.
The finding that males who witnessed marital vio-
relationship.
lence were more conservative in their current atti-

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 09:40:48 UTC
All use subject to http://about.jstor.org/terms
Intergenerational Transmission of Violence 665

tudes toward women is consistent with other re- to be relatively unrelated to the partner's pre-
search suggesting that the experience of witnessing sumed attitudes.
abuse between parents appears to have a strong Gender also interacted with attitudes toward
impact on boys (O'Leary and Curley, 1986; Telch women in determining the level of violent beha-
and Lindquist, 1984). The present research sug- vior extended and received. Men's reported in-
gests that, whether or not abusive behavior is sub- volvement in violence (either extended or re-
sequently modeled by the son in his own intimate ceived) was unrelated to their own attitudes to-
relationships, he is at least more likely to espouse ward women. On the other hand, women de-
a more conservative point of view. scribed themselves as most abusive and as being
Although past research has been equivocal most abused when their attitudes toward women
about the effects on a daughter of witnessing were very liberal. This reported level of violence
marital violence, the present research suggests dropped off quickly as women described them-
that she is more likely to espouse liberal attitudes selves as more conservative. With respect to the
toward women. Whether, however, these atti- partner's presumed attitudes toward women, both
tudes protect her from an abusive relationship men and women described themselves as more
herself is less clear. As will be demonstrated in the abusive and as receiving more abuse when they
subsequent section, any violence depends in part perceived their partner's attitudes to be more con-
on the interactions of the woman's attitudes with servative; however, this was significantly more
those of her partner. true for men than for women. It is noteworthy
that both men and women reported that it was the
woman's attitudes that were primarily associated
Research Question 3:
with the occurrence of violence. These findings
Attitudes toward Women and Dating Violence
are consistent with those of Sigelman et al. (1984),
One of the significant interactions to emerge from who also found that a discrepancy in attitudes
the third set of regression analyses was that of toward women predicted a woman's abusive be-
one's own attitudes toward women with one's havior but not a man's. They similarly found a
partner's presumed attitudes toward women in curvilinear relationship between dominance and
determining the level of violent behavior. Specifi- violence for women (suggesting that both men
cally, the level of violent behavior among individ- and women may be correct in associating violence
uals with liberal attitudes toward women was un- with either an extremely dominant or an extremely
likely to be related to the liberal or conservative passive stance on the part of the woman).
nature of their partner's views. However, those In spite of the consistency of report that the
liberal individuals who were being abused did de- woman's attitudes were most relevant to the oc-
scribe their partner as having more conservative currence of violence (perhaps because the violence
attitudes. On the other hand, individuals with affected a woman's attitudes about herself and
conservative attitudes toward women described other women significantly more than it affected a
themselves as most violent when they saw their man's attitudes about her and women in general),
partner as having liberal attitudes and not violent men and women did not describe in similar terms
at all when they saw their partner's attitudes as what the woman's attitudes toward women (and
also conservative. Furthermore, the report of vio- presumably herself) actually were. While women
lence received by conservative individuals did not in violent relationships described their attitudes as
appear to be related to their perceptions of their highly liberal, men in violent relationships de-
partner's attitudes as liberal or conservative. scribed their partner's attitudes as highly conser-
Therefore, self-reports from those with liberal at- vative. These findings suggest that the woman in a
titudes, from those with conservative attitudes, violent dating relationship may see herself as
from those who themselves were violent, and highly assertive in reaction to an oppressive part-
from those whose partner was violent toward ner. This liberal view of herself may paradoxically
them were all consistent in suggesting that a dis- help her justify the maintenance of the relation-
crepancy in attitudes toward women would result ship even though it is violent. On the other hand,
in the abuse of the liberal partner by the conserva- the man in a violent dating relationship appar-
tive one. Furthermore, abuse by an individual ently sees his partner as highly traditional and
with liberal attitudes toward women was reported conservative, possibly because the relationship is

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 09:40:48 UTC
All use subject to http://about.jstor.org/terms
666 Journal of Marriage and the Family

maintained in spite of its abusiveness. The impact ing marital violence, this study suggests that the
of these different perceptions of the woman in the observed discrepancy in attitudes toward women
violent relationship needs to be assessed. between violent partners exists prior to the occur-
Results from this study suggest that both social rence of the dating violence. Whether this dis-
learning theory and feminist theory are relevant to crepancy in attitudes is exacerbated by the vio-
the intergenerational transmission of violence. lence is not clear.
For example, from a social learning theory per- Our argument here is that people are influ-
spective, a man's involvement in dating violence is enced by and act on the basis of their perceptions
best predicted by his having been severely abused of others' attitudes. Whether the perceptions are
by his father. On the other hand, although his be- consistent with the intent or belief of the other is
havior is not directly predicted by having wit- irrelevant in the prediction of violence. However,
nessed violence between his parents, his attitudes the accuracy of the perceptions is indeed impor-
toward women apparently are. Subsequently, his tant with regard to treatment. Attitudes of violent
own attitudes are relevant to the perpetration of partners need to be addressed within the context
violence only as they interact with those of his of the relationship because that is the only way to
partner. assess each partner's attitudes and the accuracy of
A woman's involvement in dating violence is their perceptions of the other and to then resolve
not predicted by either having witnessed or having these differences. This study would therefore
experienced abuse in her family of origin. This is question whether consciousness-raising of abused
not to say, however, that a woman's early experi- women in and of itself is effective in precluding
ence is irrelevant to later involvement in violence. further violence. Liberal attitudes do not neces-
This study suggests that the experience of witness- sarily protect women from violent relationships.
ing marital violence directly contributes to the de- They appear neither to encourage individuals to
velopment of more liberal attitudes in women. Al- seek out partners with like-minded views nor to
though these liberal attitudes do not seem to pre- preclude or protect them from becoming involved
dict the type of man with whom she may have a in violence if the views of their partner are indeed
relationship, they are associated, from the different from their own. Therefore, it is essential
woman's self-report, with an increased probabil- to work closely with both members of a violent
ity of violence (extended and received) if her part- couple in order to promote a more egalitarian
ner's views should happen to be perceived as con- relationship and to help them achieve a consensus
servative-in other words, if her dating relation- about their roles. O'Leary, Curley, Rosenbaum,
ship should happen to mirror the prototype she and Clarke's (1985) discovery that assertiveness
developed in childhood. Therefore, it appears training is hazardous for abused wives may be
that the attitudes of both men and women are af- precisely because it involves a change in the be-
fected (in converse ways) by having witnessed vio- havior and attitudes of only one party and thereby
lence between their parents, but these attitudes may actually increase role discrepancy. Finally,
only have an effect on actual behavior when they any change in attitudes toward women must be di-
are perceived as discrepant between the partners. rected toward men as much as toward women if
Therefore, the enculturation of patriarchal atti- women are not to become even more vulnerable to
tudes described by feminist theory is particularly violence in a dating or marital relationship.
salient when one examines the interaction between In conclusion, this study provides evidence on
the two dating partners. two distinct mechanisms of intergenerational
It is important to clarify what this study can transmission of violence. Men seem to model vio-
and cannot justifiably suggest about the intergen- lent behaviors directly, as a function of experienc-
erational transmission of violence. On one hand, ing abuse by their father, which lends support to
the present research is correlational and cannot the role of social learning theory. In addition, the
guarantee causality either with respect to the role attitudes of both men and women are influenced
of the family of origin or with respect to whether a (although in opposite ways) by having witnessed
discrepancy in attitudes precedes dating violence violence between their parents. To the extent that
or results from it. On the other hand, to the extent the attitudes of dating partners are discrepant, the
that the current attitudes of both men and women result is an increased likelihood of violence.
are associated with the past experience of witness- Therefore, the enculturation of attitudes toward

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 09:40:48 UTC
All use subject to http://about.jstor.org/terms
Intergenerational Transmission of Violence 667

women is affected by having witnessed marital O'Leary, K. Daniel, Alison Curley, Alan Rosenbaum
and C. Clarke. 1985. "Assertion training for abused
violence (as would be suggested by feminist
wives: A potentially hazardous treatment." Journal
theory), but has a direct impact on actual be- of Marital and Family Therapy 11: 319-322.
havior only through the interaction of attitudes of Pedhazur, Elazar J. 1982. Multiple Regression in Be-
the partners. havioral Research: Explanation and Prediction (2nd
ed.). New York: Holt, Rinehart and Winston.
Roscoe, Bruce, and Nancy Benaske. 1985. "Courtship
violence experienced by abused wives: Similarities in
REFERENCES patterns of abuse." Family Relations 34: 419-424.
Rosenbaum, Alan, and K. Daniel O'Leary. 1981.
Bernard, J. Larry, S. L. Bernard, and Maxine L. Ber- "Marital violence: Characteristics of abusive cou-
nard. 1985. "Courtship violence and sex-typing." ples." Journal of Consulting and Clinical Psychol-
Family Relations 34: 573-576. ogy 49: 63-71.
Bernard, Maxine L., and J. Larry Bernard. 1983. Rouse, Linda P. 1988. "Conflict tactics used by men
"Violent intimacy: The family as a model for love in marital disputes." Pp. 176-191 in Gerald T.
relationships." Family Relations 32: 283-286. Hotaling, David Finkelhor, John T. Kirkpatrick,
Cate, Rodney M., June M. Henton, James Koval, and Murray A. Straus (eds.), Family Abuse and Its
F. Scott Christopher, and Sally Lloyd. 1982. "Pre- Consequences: New Directions in Research. Beverly
marital abuse: A social psychological perspective." Hills, CA: Sage.
Journal of Family Issues 3: 79-90. Sigelman, Carol K., Carol J. Berry, and Katherine A.
Comins, Catherine A. 1984. "Courtship violence: A re- Wiles. 1984. "Violence in college students' dating
cent study and its implications for future research." relationships." Journal of Applied Social Psychol-
Paper delivered at the Second National Conference ogy 5: 530-548.
for Family Violence Researchers, Durham, NH (Au- Spence, Janet T., and Robert Helmreich. 1972. "The
gust 7-10). Attitudes toward Women Scale: An objective instru-
Gentemann, K. M. 1984. "Wife beating: Attitudes of a ment to measure attitudes toward the rights and
non-clinical population." Victimology: An Interna- roles of women in contemporary society." Catalog
tional Journal 9: 109-119. of Selected Documents in Psychology 2: 66.
Johnston, Mildred E. 1988. "Correlates of early vio- Spence, Janet T. Robert Helmreich, and Joy Stapp.
lence experience among men who are abusive toward 1973. "A short version of the Attitudes toward
female mates." Pp. 192-202 in Gerald T. Hotaling, Women Scale (AWS)." Bulletin of Psychonomic So-
David Finkelhor, John T. Kirkpatrick, and Murray ciety 2: 219-220.
A. Straus (eds.), Family Abuse and Its Conse- Straus, Murray A. 1979. "Measuring intrafamily con-
quences: New Directions in Research. Beverly Hills, flict and violence: The Conflict Tactics Scales."
CA: Sage. Journal of Marriage and the Family 41: 75-88.
Kalmuss, Debra K. 1984. "The intergenerational trans- Straus, Murray A. 1986. "Prevention of family vio-
mission of marital aggression." Journal of Marriage lence." In The Prevention of Mental-Emotional
and the Family 42: 11-19. Disabilities. Resource papers to the Report of the
Korman, Sheila K., and Gerald R. Leslie. 1982. "The National Mental Health Association Commission on
relationship of feminist ideology and date expense the Prevention of Mental-Emotional Disabilities.
sharing to perceptions of sexual aggression in dat- Straus, Murray A., and Richard J. Gelles. 1986. "Socie-
ing." Journal of Sex Research 18: 114-129. tal change and change in family violence from 1975
Lane, Katherine E., and Patricia A. Gwartney-Gibbs. to 1985 as revealed by two national surveys." Jour-
1985. "Violence in the context of dating and sex." nal of Marriage and the Family 48: 465-479.
Journal of Family Issues 6: 45-59. Telch, Christy F., and Carol U. Lindquist. 1984. "Vio-
Laner, Mary R. 1983. "Courtship abuse and aggres- lent versus nonviolent couples: A comparison of pat-
sion : Contextual aspects." Sociological Spectrum 3: terns." Psychotherapy 21: 242-248.
69-83. Thompson, William E. 1987. "Courtship violence:
Makepeace, James M. 1981. "Courtship violence Toward a conceptual understanding." Youth and
among college students." Family Relations 30: 97- Society 18: 162-176.
102. Walker, Lenore E. 1983. "The battered woman syn-
O'Leary, K. Daniel, and Alison D. Curley. 1986. "As- drome study." Pp. 31-48 in David Finkelhor,
sertion and family violence: Correlates of spouse Richard J. Gelles, Gerald T. Hotaling, and Murray
abuse." Journal of Marital and Family Therapy 12: A. Straus (eds.), The Dark Side of Families: Current
281-289. Family Violence Research. Beverly Hills, CA: Sage.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 09:40:48 UTC
All use subject to http://about.jstor.org/terms

You might also like