[acb-hsp] IPV: College Students

[J.Rayl]

Domestic Violence Beliefs and Perceptions among College Students

by Erin L. Nabors , Tracy L. Dietz , Jana L. Jasinski

With more than three decades of research on intimate partner violence, domestic violence

is now generally recognized as a serious social problem. However, despite existing

research evidence (Sugarman & Hotaling, 1989 ), it is not as widely known that dating

couples are significantly more likely to be violent in their relationships than married

couples. Specifically, college students experience extremely high rates of dating

violence that range between 20% and 50% (Bryant & Spencer, 2003; Lloyd, 1991; Makepeace,

1981, 1986; Shook, Gerrity, Jurich, & Segrist, 2000; Straus, 2004). Because dating

violence among college students is such a widespread problem, it is important to

understand what lies at the foundation of this type of abuse. One possible factor

is a belief system supporting the use of violence against intimate partners.

Researchers have found a strong correlation between beliefs supportive of domestic

violence and acts of intimate partner violence (Archer & Graham-Kevan, 2003; Archer

& Haigh, 1999; Bryant & Spencer, 2003; Riggs & O'Leary, 1996). However, studies investigating

the question of who is most likely to hold beliefs accepting domestic violence are

limited in both number and scope and often arrive at inconsistent findings. The current

study provides an analysis of the relationships between sociodemographic variables

and beliefs supportive of domestic violence among college students using recently

available data from the Relationship Characteristics Study (Dietz & Jasinski, 2003).

In addition, the results from this study also contribute to the understanding of

a newly developed scale to measure beliefs about domestic violence (Carlson & Worden,

2005; Worden & Carlson, 2005). With the knowledge provided by this research, prevention

programs can be more effectively directed toward college students who are most likely

to hold beliefs supportive of domestic violence.

LITERATURE REVIEW

Dating Violence Among College Students

Undeniably, college students experience exceptionally high rates of dating violence.

Worldwide figures from the International Dating Violence Survey (Straus, 2004 ) suggest

that anywhere from 17% to 45% of university students had physically assaulted their

intimate partner in the year prior to the survey, and up to 20% had assaulted their

partner so severely as to cause injury. In fact, rates of severe violence among university

students are considerably higher than among the general population (see, e.g., Tjaden

& Thoennes, 2000). Results from the International Dating Violence Survey (Straus,

2004) indicate that between 4% and 20% of students had used severe forms of violence,

including using a gun or knife on their partner, punching or hitting with a solid

object, choking, slamming their partner against a wall repeatedly, beating up their

partner, purposefully burning or scalding their partner, or kicking their partner

(Straus, 2004). The types of severely violent acts used by dating partners are consistent

across studies (Makepeace, 1981; Straus, 2004) and often result in injury. Straus

(2004), for example, found that 2% of college students seek medical attention because

of injuries inflicted by their partners.

These high rates of victimization do not go unnoticed by students. One study, for

example, found that 79% of college students think dating violence is a major problem

(Knickrehm & Teske, 2000). It is likely that this awareness is a result of some level

of personal involvement, because students often know someone involved in a physically

abusive dating relationship (Makepeace, 1981). It is certainly possible that knowledge

and beliefs may influence the types of decisions college students make when faced

with violence in their own or their friends' relationships. Moreover, developing

successful prevention and intervention programs may depend on a more complete understanding

of both the relationship between beliefs and behaviors and the factors that are associated

with beliefs supportive of violence.

Beliefs and Domestic Violence

Among the factors identified as risk markers for dating violence, there is evidence

of a strong correlation between holding beliefs supportive of domestic violence and

committing violent acts against partners (Archer & Graham-Kevan, 2003; Archer & Haigh,

1997; Dibble & Straus, 1980; Kaufman Kantor, Jasinski, & Aldarondo, 1994; Price et

al., 1999). For example, Dibble and Straus (1980) found that 28% of respondents believe

that slapping their partners is either necessary, normal, or good. Of that 28%, one-third

reported physically abusing their partners, while only 8% of respondents not supportive

of slapping their partners reported perpetrating domestic violence. Dibble and Straus

(1980) also found that 5% of participants believe that slapping, pushing, grabbing,

shoving, and throwing something at their partner is either necessary, normal, or

good. Similar findings have been observed among college students (Archer & Graham-Kevan,

2003; Archer & Haigh, 1999; Bryant & Spencer, 2003; Knickrehm & Teske, 2000; Riggs

& O'Leary, 1996). In fact, Archer and Graham-Kevan (2003) found that beliefs supportive

of domestic violence are more predictive of abuse in intimate relationships among

college students than among either women in domestic violence shelters or men in

prison convicted of physically abusing their partners. These findings are particularly

troublesome when combined with the high rates of perpetration of partner violence

in this population. Furthermore, Straus, Kaufman Kantor, and Moore's (1997) analysis

indicates that youth is one factor associated with a greater likelihood of believing

in the acceptability of violence against an intimate partner. Consequently, an analysis

of the factors associated with beliefs endorsing the use of violence among a sample

of college students is particularly important.

Sociodemographic Characteristics and Beliefs Supporting Intimate Partner Violence

Although there is evidence that domestic violence beliefs are associated with perpetrating

abusive acts against partners, research probing the question of who is most likely

to hold beliefs supportive of domestic violence is lacking. Studies dealing with

this topic have demonstrated only tentative correlations between beliefs accepting

violent acts toward partners and gender. Moreover, only a limited amount of research

examines connections between domestic violence beliefs and other demographic characteristics

such as socioeconomic status, age, and relationship status.

The results of research investigating the relationship between beliefs supportive

of domestic violence and gender are ambiguous. Some researchers, for example, have

found no gender differences in the beliefs accepting violence in intimate relationships

(Archer & Haigh, 1997; Arias & Johnson, 1989; Dibble & Straus, 1980; Mwamwenda, 1999).

For instance, Mwamwenda (1999) found that approximately 25% of both male and female

college students approve of domestic violence. However, other researchers have found

that men are more likely than women to accept physical aggression toward partners

(Archer & Graham-Kevan, 2003; Bryant & Spencer, 2003; Carlson, 1999; Locke & Richman,

1999; Riggs & O'Leary, 1996; Simon et al., 2001).

The relationships among beliefs supportive of violence and other sociodemographic

characteristics such as race and ethnicity and socioeconomic variables such as education

and income have also been considered, often with inconsistent results (Dibble & Straus,

1980; Locke & Richman, 1999; Simon et al., 2001). Although several studies have found

that college students of color are more likely than White students to hold beliefs

accepting violence toward partners (Locke & Richman, 1999; Simon et al., 2001), other

researchers have found that White respondents were more likely to approve of the

use of violence toward a marital partner (Straus et al., 1997). In an effort to understand

racial and ethnic differences in beliefs about intimate partner violence, Klein,

Campbell, Soler, and Ghez (1997) argued that these differences may be a function

of deeply ingrained beliefs about gender roles and responsibilities. Evidence for

a relationship between socioeconomic variables and belief systems is also not definitive.

Although at least one study has found that participants without a high school diploma

are most likely to accept physical abuse in intimate relationships (Simon et al.,

2001), other researchers have found no significant association with education (Straus

et al., 1997). Similarly, contradictory results have been found with respect to the

association between income and domestic violence beliefs (Dibble & Straus, 1980;

Simon et al., 2001; Straus et al., 1997).

In addition to gender, race and ethnicity, and socioeconomic status, researchers

have examined the correlation between age and beliefs accepting physically abusive

acts toward partners. In this area, the research findings are more consistent. Most

studies have shown that beliefs supportive of domestic violence decrease with age

(Archer & Haigh, 1997; Carlson, 1999; Simon et al., 2001; Straus et al., 1997). More

precisely, one study found that acceptance of hitting a partner is higher among participants

younger than 35 years old than among any other age group (Simon et al., 2001). The

majority of college students would fall into this age group, and there is evidence

that college students in dating relationships are more supportive of abusive acts

toward partners (Carlson, 1999 ) compared to other age groups.

College students experience violence within intimate relationships at a rate that

is much higher than the general population, and there is some evidence to suggest

that beliefs about domestic violence, among other factors, contribute to this risk.

At the same time, we know little about the factors that are associated with these

beliefs. Moreover, comparisons with existing research are often difficult because

of different samples and a lack of standardized measures. This article builds on

existing research by comparing the rates of attitudes and beliefs reported by college

students to those of an adult, community-dwelling sample in the state of New York,

using a newly developed measure of attitudes toward and beliefs about domestic violence.

In addition, psychometric analysis of the measure using this college student sample

is performed, which extends the work of previous investigators using these items

(Carlson & Worden, 2005; Worden & Carlson, 2005). Finally, the research adds to the

literature by investigating the relationships among sociodemographic characteristics

and attitudes and beliefs regarding the acceptance of domestic violence and beliefs

about the causes of it among college students using scales created from the items.

DATA AND METHODS

Data for this investigation are taken from the cross-sectional Relationship Characteristics

Study conducted by two of the authors (Dietz & Jasinski, 2003). Data were collected

by questionnaires completed during class in large undergraduate introductory sociology

and anthropology courses at a university in the southeastern United States during

the fall of 2001. The convenience sample consisted of 1,938 student participants.

These courses fulfill general education requirements for the university. Thus, this

selection allowed for the inclusion of students from a variety of disciplines. The

students were informed of their right to accept or decline participation in the project

as well as of the procedures used to ensure anonymity and confidentiality. Participation

in the study was not a course requirement, and completion of the questionnaire took

less than 75 minutes. Following the administration of the questionnaires, students

were provided with information sheets detailing the purpose of the study and information

on where they could obtain help in the area with relationship problems.

The Relationship Characteristics Study includes sociodemographic questions about

respondents' and partners' gender, respondents' age, race and ethnicity, educational

level, family income, parents' education, parents' marital status, respondents' relationship

status, cohabitation, length of relationship, and sexual activity within their relationship.

Participants also answer questions regarding video game use, family commitment, social

desirability, alcohol and controlled substance use, and pets. Additionally, the study

includes the BEM Sex-Role Inventory and the Revised Conflict Tactics Scale. Finally,

the Relationship Characteristics Study poses questions about participants witnessing

violence between parents or guardians, their victimization and help-seeking behaviors,

and their beliefs toward domestic violence.

MEASURES

Dependent Variables

The dependent variables are composed of two sets of variables created by Bonnie Carlson

and Alissa Politz Worden (Carlson & Worden, 2001, 2005; Worden & Carlson, 2005).

These variables include a set of 10 items that are designed to determine whether

the respondent endorses certain causes of domestic violence. Respondents are asked

to agree or disagree with a series of statements. Half of the items in the set are

designed to measure myths or misconceptions about the causes of violence and include

e specially those types of myths that result in victim blaming. The other half of

the items measure causation endorsements reflecting risk factors for domestic violence

that have been identified through empirical research. Many of these items reflect

the relationship between substance abuse and family violence and the escalation of

verbal abuse to physical violence as well as the socialization of aggressive behavior

within society.

A second set of five items is designed to measure whether respondents believe particular

behaviors constitute domestic violence. Each of the five items are presented first

with a male perpetrator and female victim and then re-presented with a female perpetrator

and male victim. Respondents are instructed to answer whether they would consider

each of the following domestic violence (1 = yes, 0 = no): punching with a fist,

slapping during an argument, using physical force for sex, following a former partner

all over town, and insulting a partner by calling him or her a "stupid slob." The

items in the current study differ from the original in that response options are

dichotomous (agree/disagree; yes/no) rather than scaled.

Independent Variables

Gender . Respondents were asked to indicate whether they were male or female.

Racial/Ethnic Category . Respondents were asked to indicate which category best described

them from the following options: Asian, African American (Black), Caucasian (White),

Native American (American Indian, Samoan, or Hawaiian), Hispanic (Latino/a), and

Other.

University Year . Respondents were asked to indicate whether they were a freshman,

sophomore, junior, senior, or other.

Parents' Education . Respondents were asked to indicate the highest level of education

achieved by their mother and their father. They were given the following options

on both questions: less than high school, high school graduate, some college, two-year

college graduate (for example, community college), four-year college graduate, some

graduate school, and graduate degree.

Family Income . Respondents were asked to indicate their family's yearly income using

the following categories: under $9,999, $10,000-$19,999, $20,000-$29,999, $30,000-$39,999,

$40,000-$49,999, $50,000-$59,999, $60,000-$69,999, $70,000-$79,999, and $80,000 or

more.

Parents' Current Marital Status . The following categories were used: married to

each other, separated, divorced, never married to each other, or one or both parents

have died.

Relationship Status . Respondents were asked whether they were currently, previously,

or had never been in a relationship.

ANALYTIC STRATEGY

The purpose of this study is threefold. First, we attempted to replicate the studies

conducted by Carlson and Worden (Carlson & Worden, 2001, 2005; Worden & Carlson,

2005). Thus, we began by conducting an analysis of the frequency distributions of

the independent and dependent variables. The second focus of this research is to

extend the work of Carlson and Worden by identifying latent factors. Thus, we conducted

a series of factor analyses of the sets of items created by Carlson and Worden, followed

by alpha reliability testing of the resulting scales. Finally, to identify the factors

associated with attitudes and beliefs about domestic violence, a series of regression

analysis models using the independent variables identified above and the scales created

from the factor analysis were examined.

RESULTS

Table 1 displays the sample's sociodemographic characteristics. The sample is relatively

evenly split with 59% women and 41% men. Approximately 71% of the participants are

White. The remaining participants are relatively evenly distributed, with roughly

10% Black and 11% Hispanic. The majority of participants are freshmen (66%), 14%

are sophomores, and 11% are juniors. Respondents' parents' education varies, but

the median educational level for both is a two-year college degree. Median family

income for the sample was between $60,000 and $69,999 per year. About two-thirds

of respondents' parents are currently married to each other. Approximately 45% of

participants are currently in a relationship, 40% have previously been in a relationship,

and 15% have never been in a relationship.

Overall, as presented in Table 2, participants demonstrate a high level of domestic

violence causation endorsement. More than half of participants agree with domestic

violence myths, with two exceptions. Only 9% of participants agree that "a lot of

what is called 'domestic violence' is really just a normal reaction to day-to-day

stress and frustration," and 26% agree that "some women who are abused secretly want

to be treated that way." Generally, the respondents in this sample were very likely

to agree with the statements that related to empirically derived causes of domestic

violence. In fact, more than two-thirds of all respondents in this study endorsed

each of the empirically based statements pertaining to causes of domestic violence.

We did , however, find some differences in the pattern of responses when compared

to the original study (Carlson & Worden, 2005; Worden & Carlson, 2005). Students

in our sample were more likely to endorse victim-blaming statements than were the

respondents in the original study. In addition, the students in our sample also were

more likely to endorse statements about the inevitability of violence.

Frequencies and percentages for the domestic violence definition items are displayed

in Table 3. Respondents are more likely to define the action as domestic violence

if the perpetrator is male than if the perpetrator is female. Nearly all of the participants

consider a man or woman physically forcing a partner to have sex (97% and 89%, respectively)

or punching a partner (97% and 96%, respectively) to be domestic violence. Most participants

define a husband slapping his wife as domestic violence (87%), but less than two-thirds

of participants (61%) define a wife slapping her husband as domestic violence. About

45% of respondents believe that a man or woman following a former partner all over

town to try to get back together with that partner is domestic violence. However,

less than one-third of participants think that a man or woman insulting a partner

is domestic violence. The last column of the table shows that, with few exceptions,

many of the rates for individual items are remarkably consistent with rates reported

by the scale developers (Carlson & Worden, 2005; Worden & Carlson, 2005).

Domestic Violence Beliefs Scales

Using the data from this sample, we analyzed the internal consistency of the subsets

defined by Carlson and Worden. An internal consistency estimate of reliability was

computed for each of the subsets described by these researchers, including the five

items that represent beliefs about causation that are myths, the five items that

represent beliefs about causation that are empirically founded, the entire set of

beliefs about what constitutes domestic violence, and then the last set divided according

to gender of perpetrator and victim. The internal consistency estimates generated

were generally low and are presented in Table 4.

Because the internal consistency estimates generated with this sample were low, we

elected to extend the original work by determining whether the two sets of items-domestic

violence causation and definitions-might have latent factor structures with this

sample that differed from those theoretically proposed by the developers (Carlson

& Worden, 2005; Worden & Carlson, 2005). Using factor analysis, we examined the possible

latent factor structures by conducting two separate analyses, one for each set of

items. The dimensionality of the 10 items from the domestic violence cause endorsement

measure was analyzed using maximum likelihood factor analysis. Four criteria were

used to determine the number of factors to rotate: the a priori hypothesis that the

measure was unidimensional, the eigenvalues, the scree test, and the interpretability

of the factor solution. The eigenvalues and scree plot indicate that the initial

hypothesis of unidimensionality is incorrect. Based on the eigenvalues and scree

plot, two factors were rotated using a varimax rotation procedure. However, the rotated

solution did not yield interpretable factors. Consequently, scales based on the theoretical

divisions proposed by Carlson and Worden (2005) were used for all further analyses.

Simple additive scales were created for the five myth-based and five empirically

based statements, respectively.

Maximum likelihood factor analysis was also used to analyze the dimensionality of

the 10 items from the measure of domestic violence definitions. Again, four criteria

were used to determine the number of factors to rotate: the a priori hypothesis that

the measure was unidimensional, the eigenvalues, the scree test, and the interpretability

of the factor solution. The eigenvalues and scree plot indicate that the initial

hypothesis of unidimensionality is incorrect. Based on the eigenvalues and scree

plot, three factors were rotated using a varimax rotation procedure. Table 5 shows

the rotated solution, which yields three interpretable factors: physical and sexual

abuse beliefs, stalking beliefs, and verbal abuse beliefs. The physical and sexual

abuse beliefs factor accounts for 20.74% of the item variance, the stalking beliefs

factor accounts for 17.93% of the item variance, and the verbal abuse beliefs factor

accounts for 17.45% of the item variance. Only one item, which asks participants

to specify whether they would consider it domestic violence for a wife to slap her

husband during an argument, did not load on any factor, indicating that participants

do not regard women slapping men to be an act of domestic violence. Because this

item did not load on any factor, it is excluded from subsequent analyses.

Additionally, an internal consistency estimate of reliability was computed for each

additive scale. The myth-based cause endorsement scale has a mean score of 2.55,

a standard deviation of 1.15, and a coefficient alpha of .44. Likewise, the empirically

based cause endorsement scale has a mean score of 3.82, a standard deviation of 1.20,

and a coefficient alpha of .48. Higher scores on both the myth endorsement scale

and the empirically founded scale indicate more agreement with domestic violence

cause endorsement. The physical and sexual abuse belief additive scale ranges from

zero to five and has a mean score of 4.64, a standard deviation of .85, and a coefficient

alpha of .68. The stalking belief and verbal abuse belief additive scales both range

from zero to two. The stalking belief scale has a mean score of .90, a standard deviation

of .97, and a coefficient alpha of .94; and the verbal abuse belief scale has a mean

score of .84, a standard deviation of .84, and a coefficient alpha of .89. Higher

scores on the physical and sexual abuse belief, stalking belief, and verbal abuse

belief scales indicate more agreement with domestic violence definitions. These coefficients

are presented in Table 4 with the scores for the theoretically based scales.

Using the results from these factor analyses, we created additive indices to be used

as the dependent variables in the subsequent multiple regression models. As mentioned

previously, because no reliable scales emerged during the psychometric analysis phase

of the analysis for the causation items, we elected to create two additive indices

based on the developers' delineation of the set of items. Thus, an index of causation

based on myth was calculated, and an index of causation based on empirically founded

causes was calculated. Indices measuring respondents' definitions of domestic violence

were calculated by summing participants' scores for physical and sexual abuse beliefs,

stalking beliefs, and verbal abuse beliefs, for a total of five scales.

Sociodemographic Characteristics Associated With Domestic Violence Beliefs

Five multiple linear regression models were analyzed to examine the relationships

between sociodemographic variables and attitudes and beliefs about the definitions

and causes of domestic violence among college students. Tolerances for the predictor

variables were examined to ensure against multicollinearity. Although age was identified

as a possible predictor of perpetrating violent acts against partners in earlier

studies, age was not included as a predictor variable in this study due to multicollinearity

with university year and the intrinsic lack of variance in age associated with college

students.

The physical and sexual abuse belief scale was the dependent variable in the first

analysis; the verbal abuse belief scale was the dependent variable in the second

analysis; the stalking belief scale was the dependent variable in the third analysis;

the domestic violence myth-based causation scale was the dependent variable in the

fourth analysis; and the empirically based causation scale was the dependent variable

in the fifth analysis.

The regression equation for the stalking belief scale was not significant, R 2 =

.005, F (11, 1,819) = .783, p = .658. These findings indicate that sociodemographic

characteristics are not associated with domestic violence stalking beliefs, differing

from prior research and our hypotheses. Conversely, the regression equation for the

physical and sexual abuse belief scale was significant, R 2 = .038, F (11, 1,824)

= 6.525, p < .000, as was the regression equation for the verbal abuse belief scale,

R 2 = .017, F (11, 1,822) = 2.894, p = .001. The regression equation for the myth-based

causation scale was significant, R 2 = .036, F (11, 1,801) = 6.103, p < .000. In

addition, the regression equation for the empirically based causation scale was also

significant, R 2 = .039, F (11, 1,807) = 6.663, p < .000.

Table 6 displays the multiple regression results for the physical and sexual abuse

belief scale. Findings demonstrate that men are more likely to hold beliefs supportive

of physical and sexual abuse than women, ß = -.168, p = .000. Results also indicate

that the participants who are farther along in their university education are less

likely than those who are not as far along to hold beliefs supportive of physical

and sexual abuse, ß = .064, p = .006. A comparison of mean physical and sexual abuse

belief scores of freshmen, sophomores, juniors, and seniors provides further insight

into this finding. Average scores for freshmen, sophomores, and juniors are between

4.60 and 4.66, while the average score for seniors is 4.83, indicating a significant

change in beliefs between junior and senior university years, F (3, 1,901) = 3.958,

p = .008.

Table 6 also presents the multiple regression results for the verbal abuse belief

scale. Only race and ethnicity is significantly correlated with verbal abuse beliefs.

Findings show that Hispanic participants and participants with races and ethnicities

other than White, Black, or Hispanic are least likely to hold beliefs supportive

of verbal abuse, ß = .072, p = .003 and ß = .056, p = .019, respectively. Furthermore,

having a Black racial and ethnic background is not significantly correlated with

verbal abuse beliefs.

Four variables were statistically related to scores on the myth-based causation scale.

Men scored statistically higher on this scale than women, ß = -.157, p < .000. In

addition, Black respondents were less likely to score higher on the scale, ß = -.046,

p = .05. No other ethnic group differences were found. Those respondents who were

currently in a relationship scored significantly higher on the myth-based causation

scale, ß = .059, p = .04. Finally, there was a negative relationship between year

in school and score on the myth-based causation scale, ß = -.072, p = .003.

Only three variables were statistically associated with scores on the empirically

based causation scale. Women scored statistically higher on this scale than men,

ß = .174, p < .000. In addition, Black respondents were less likely to score higher

on the scale than those with other racial or ethnic backgrounds, ß = -.050, p = .04.

No other racial or ethnic group differences emerged. Finally, there was a positive

relationship between year in school and score on the empirically based causation

scale, ß = .049, p = .04.

DISCUSSION AND CONCLUSION

This study demonstrates the importance of investigating domestic violence beliefs

to accurately pinpoint factors associated with beliefs regarding specific types of

domestic violence. Using data from a college student sample, the current study sought

to address several issues. First, we wanted to provide further evidence of the utility

of a newly created measure of attitudes and beliefs about what is defined as domestic

violence and the causes of domestic violence. Our results provide evidence that the

measures created by Carlson and Worden (2001) provide generally consistent results

across populations. For example, our findings are consistent with those of Carlson

and Worden (2005) in that respondents were more likely to endorse those items that

involved a male perpetrator and female victim as domestic violence than vice versa.

More than half of the respondents in Worden and Carlson's (2005) study of adults

in New York State endorsed the statement that some violence is caused by women starting

physical fights and that most women could find a way to get out of an abusive relationship

if they really wanted to. The results with this sample of college students were consistent

with those results, although a substantially larger percentage of our college student

sample endorsed the latter myth. This is not surprising, given that they are perhaps

more likely to be college educated once they reach the age of the respondents in

the Worden and Carlson study. In addition, more than half of this sample endorsed

the notion that some violence is caused by the way women treat men, while only 45%

of the Worden and Carlson sample endorsed this myth. Interestingly, the participants

in this student sample endorsed the empirically based causes for domestic violence

at a much greater rate than did the adult sample from New York used by Worden and

Carlson.

With a few exceptions, there was only moderate deviation from the rates for individual

items in the measure when comparing rates reported by the scale developers (Carlson

& Worden, 2005; Worden & Carlson, 2005) to those from this sample. In comparison

to the rates reported by the developers in their adult sample from New York, the

college students in our sample were more likely to endorse all but one of the empirically

founded causations and one of the myths of causation. The patterns for attitudes

and beliefs about what constitutes domestic violence were less consistent. Perhaps

most interesting was that the student sample was much less inclined than the adult

sample to identify most examples of female-perpetrated aggression as domestic violence.

The exceptions to this were that the respondents in the student sample were more

likely to define female-perpetrated stalking and a woman punching her male partner

as domestic violence. These noted differences may be related to age differences in

the samples or regional differences; the Carlson and Worden (2005) sample was geographically

isolated to the state of New York and the student sample was isolated to a large

metropolitan university in Florida. Clearly, additional research is needed to verify

the cross-cultural and cross-generational reliability and validity of the measures,

but these results suggest that the measures are promising in the efforts to develop

standardized measures of beliefs and attitudes regarding domestic violence.

The developers of the measures presented the items as subsets divided by the gender

of perpetrator for the definition of domestic violence items (Carlson & Worden, 2005).

Our psychometric analysis of these items reveals little reliability in using the

items in the manner proposed by the developers. Because our response choices were

dichotomous rather than the original scaled five category response choices, however,

it is possible that the reduction in variability could account somewhat for the lower

alpha coefficients compared to the original. Factor loadings for the variables illustrate

that, in this student sample, respondents were more likely to differentiate domestic

violence by type of abuse, including physical and sexual abuse, stalking, and verbal

abuse rather than based on gender. Furthermore, the developers suggested that the

beliefs about causation might be divided based on whether the beliefs had been substantiated

through empirical research or whether they were myths. In their explanation of the

set of items, they present factor analysis results, but acknowledge that the results

were not robust. Our results confirm the lack of robustness in their findings, because

obvious latent factors with high internal consistency likewise did not emerge among

our sample. It is worth further noting, however, that a woman slapping her husband

did not load onto any of the factors. This may result from a tendency in our society

to trivialize and normalize female-perpetrated acts of violence such as slapping

in intimate relationships.

Our analyses further demonstrate the importance of exploring relationships between

sociodemographic variables and beliefs supportive of specific types of domestic violence

among college students. Results show that some sociodemographic variables are correlated

with beliefs related to physical and sexual abuse beliefs and verbal abuse beliefs

as well as beliefs about domestic violence causation related to myth-based and empirically

based causes, but not to stalking beliefs. However, the results are not entirely

consistent with past research. Contrary to previous studies (e.g., Dibble & Straus,

1980; Locke & Richman, 1999; Simon et al., 2001) and expectations, race and ethnicity,

father's education, mother's education, family income, parents' marital status, and

relationship status are not significantly correlated with physical and sexual abuse

beliefs or verbal abuse beliefs. The relationships between sociodemographic variables

and physical and sexual abuse and verbal abuse also differ.

Gender was also associated with physical and sexual abuse beliefs and beliefs in

causation. Interestingly, men scored higher on the myth-based scale, while women

scored higher on the empirically based scale-indicating that women, at least at this

university, are much better educated about the causes of domestic violence. In addition,

women were more likely than men to define examples of physical and sexual abuse as

domestic violence. University year was also correlated with these same indices. Those

who had completed more education reported higher scores on the empirically based

causation scale and were more likely to define examples of physical and sexual abuse

as domestic violence. Meanwhile, those with less education were more likely to score

higher on the causation based on myths scale. This suggests that the students at

this university may be learning about what domestic violence is and what the real

risk factors are. At the same time, however, these differences are not large, and

we would assume any programming effect would have a greater impact than what the

scores indicate. It is also possible that these changes may simply be a result of

having more experiences in intimate relationships as well as getting older. Gender

and year at the university are not associated, however, with beliefs about verbal

abuse constituting abuse. Interestingly, individuals who reported being in a current

relationship were more likely than others not currently in a relationship to endorse

myths as causes of domestic violence. This finding is perplexing but may be related

to rates of domestic violence in the relationship. Future research might include

measures of domestic violence perpetration and victimization as predictors of beliefs

to further explore this result. Although the Relationship Characteristics Study (Dietz

& Jasinski, 2003) did contain measures of relationship violence, including them in

our analysis was beyond the scope and purpose of the current study.

Results regarding race and ethnicity and domestic violence beliefs differ from past

studies as well (e.g., Locke & Richman, 1999; Simon et al., 2001). Non-White respondents

were expected to be more likely than White respondents to hold beliefs supportive

of domestic violence. Likewise, being Black or African American was found to reduce

the score on both causation indices. However, race and ethnicity was not correlated

with physical and sexual abuse beliefs or stalking beliefs. Additionally, Hispanic

participants and participants with racial and ethnic backgrounds other than White,

Black, and Hispanic were significantly less likely to hold beliefs accepting verbal

abuse than White respondents. These findings might be attributed to variety of explanations.

First, this study examines correlations between sociodemographic variables and beliefs

related to each type of violence separately, while the majority of previous research

investigates domestic violence beliefs more generally (e.g., Dibble & Straus, 1980;

Straus et al., 1997). Also, in contrast to most past studies of associations between

race and ethnicity and domestic violence beliefs (Locke & Richman, 1999), this study

included racial and ethnic categories in addition to White and African American.

Finally, previous research has demonstrated that Hispanic ethnicity, particularly

Cuban ethnicity, is associated with lower rates of cultural approval of intimate

partner violence (Kaufman Kantor et al., 1994). Because the data for this study were

collected at a Florida university with a relatively large Hispanic population, it

is possible that there is a high proportion of Cuban respondents included in the

sample, which could affect results regarding Hispanic respondents' domestic violence

beliefs.

The diversity and size of the sample and a design that replicated a measure of domestic

violence beliefs are strengths of the current study. At the same time, several items

limit the generalizability of our findings to the overall population of college students

and to the adult population in general. Clearly, this sample is not reflective of

the population in terms of social class. Most participants recorded that their family

income is $60,000 or more. Moreover, future researchers should attempt to draw nongeographically

isolated samples that are more generalizable to the adult population in the United

States so that anomalies in the current findings can be further explored. A more

generalizable sample can also provide additional support for the refinement or adoption

of these measures for the study of domestic violence across populations. In addition,

the amount of variance in the dependent variables as explained by the independent

variables was relatively small as evidenced by the low R 2 values. Although this

is consistent with the results from the original study, it does suggest that additional

factors may be important to consider in any investigation of domestic violence beliefs.

Despite the limitations, however, the results of these analyses do provide important

contributions to the field. First, they provide empirical support for the potential

use of the newly created measures of attitudes and belief about the definitions and

causes of domestic violence. Such replication is crucial before any measure can be

fully adopted for use. In addition, these results demonstrate the need to further

explore the unique attitudes and beliefs held by college students. The literature

suggests that relationships among younger individuals are more likely to be violent,

and it is important to understand the beliefs and attitudes of young adults and adolescents

specifically if effective educational campaigns are to be designed. Further, these

analyses suggest that the educational campaigns at this university may be somewhat

effective given that those who have been in school longer are less likely to endorse

myths as causes of domestic violence and to define physical and sexual abuse as domestic

violence and more likely to report that the empirically based causes of domestic

violence are real causes. Unfortunately, the results also demonstrate that greater

efforts may be needed to educate college students about verbal abuse and that increased

attention should be directed toward educating men about domestic violence.

-1-

Questia, a part of Gale, Cengage Learning. www.questia.com

Publication Information:

Article Title: Domestic Violence Beliefs and Perceptions among College Students.

Contributors: Erin L. Nabors - author, Tracy L. Dietz - author, Jana L. Jasinski

- author. Journal Title: Violence and Victims. Volume: 21. Issue: 6. Publication

Year: 2006. Page Number: 779+. © 2006 Springer Publishing Company. Provided by ProQuest

LLC. All Rights Reserved.

Next Page

Jessie Rayl
EM: thedogmom63 at frontier.com
PH:304.671.9780
www.facebook.com/eaglewings10

"But they that wait upon the LORD shall renew their strength; they shall
mount up with wings as eagles. They shall run, and not be weary"--Isaiah 40.31
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.acb.org/pipermail/acb-hsp/attachments/20110725/fb10746f/attachment-0001.html>