Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2011 Aug 17.
Published in final edited form as: J Interpers Violence. 2008 Apr 22;24(4):551–568. doi: 10.1177/0886260508317189

Agreement on Intimate Partner Violence among a Sample of Blue-Collar Couples

Carol B Cunradi 1, Melina Bersamin 1, Genevieve Ames 1
PMCID: PMC3157480  NIHMSID: NIHMS304887  PMID: 18430971

Abstract

This study assessed level of agreement about the occurrence of specific acts of past-year male-to-female partner violence (MFPV) and female-to-male partner violence (FMPV) among a sample of unionized construction industry workers (95% male) and their spouses or cohabiting partners who completed separate and confidential telephone interviews on IPV, work, job stress, and drinking (n=897 couples). Past-year IPV was measured with the Physical Assault subscale of the revised Conflict Tactics Scales (CTS2). Level of agreement was assessed with Cohen’s kappa statistic. Lower-bound estimates (based on male-female agreement that an IPV event occurred) and upper-bound estimates (based on reports that an IPV event had occurred, whether corroborated or not) were calculated for MFPV, FMPV, and any past-year IPV. Results indicated that item-specific agreement between partners for most IPV behaviors was low (kappa < .40). The following lower- and upper-bound rates were estimated: 6.7% and 21.2% for MFPV; 7.1% and 24.2 for FMPV; and 10.1% and 30.2% for any partner-to-partner violence. The findings suggest that single-point estimates of IPV prevalence are biased, and that lower- and upper-bound estimates of IPV using collateral reports should be calculated when possible. Additionally, these estimates underscore the importance of conducting IPV research among understudied populations, such as working-class couples, that appear to be at elevated risk for IPV.

INTRODUCTION

A considerable body of research provides evidence that both men and women engage in intimate partner violence (IPV) within their relationships (see review in Archer, 2000). Straus (2005), for example, found that of the 495 couples in the 1985 National Family Violence Survey in which one or more assaults was reported by a woman respondent, the man was the only violent partner in 25.9% of the cases; the woman was the only violent partner in 25.5% of cases; and both were violent in nearly half (48.6%) of cases. Caetano et al. (2005), in an analysis of a national household sample of white, black, and Hispanic couples (n=1440), found that most couples who reported partner violence reported bidirectional IPV (i.e., both MFPV and FMPV) in the past year. Male partner violence, however, has more negative consequences for its victims than female violence; women experience more injury and other psychological sequelae than men (Archer, 2000). Despite the greater public health burden posed by male IPV, Straus (2005) argues that women’s acts of IPV are a serious social problem as well. First, even “minor” assaults by women put them in danger of more severe retaliation by men. Second, women’s acts of IPV help perpetuate the cultural norms that make “the marriage license a hitting license” (Stets & Straus, 1990). Third, acts of marital aggression between parents, regardless of gender perpetrator, are likely to be harmful for children who witness such events (Straus, 1999).

Johnson (1995, 2006) and Johnson and Ferraro (2000) propose that in order to meaningfully study IPV, distinctions must first be made about the type of IPV being studied. For example, they hypothesize that the large differences in gender symmetry seen between IPV based on samples obtained from agencies, clinics and shelters versus samples based on community or national household surveys are due to qualitative differences in the forms of IPV, such as situational couple violence and intimate terrorism. The former is theorized to characterize the type of situational outbursts that may occur between couples, typically in the course of conflict. Situational couple violence can also be bi-directional, and usually involves ‘moderate’ acts (e.g., pushing, shoving, grabbing, slapping), although escalation to severe episodes (e.g., hitting with fist, kicking) is possible. Intimate terrorism is characterized by a pattern of more severe violence typically associated with terms such as ‘wife beating’ and ‘battered women.’ This type of violence is a form of terroristic control on the part of one partner, typically of women by their male partners. It involves the systematic use of violence, as well as other control tactics, such as threats, emotional abuse, isolation, and economic dependency (Johnson & Leone, 2005). The current study is based on a cohort of unionized construction industry workers and their cohabiting spouses or partners, and is thus akin to a community-based sample. We therefore expect that the cases of IPV obtained in the context of a family conflict survey interview will overwhelmingly represent situational couple violence (Johnson & Leone, 2005). This is important for several reasons. For example, situational couple violence has deleterious health consequences, although they are not as severe as those for intimate terrorism. Johnson & Leone (2005) found that women who experienced situational couple violence experienced significantly more depressive symptoms, and were significantly more likely to use antidepressants, compared to women who did not experience any couple violence. In addition, moderate couple violence can potentially progress over time to more severe levels (Feld & Straus, 1990).

Obtaining reports from the male and female partners in the same relationship is an essential first step for research addressing IPV gender symmetry (Straus, 2006) and obtaining accurate prevalence estimates (Szinovacz & Egley, 1995). The available evidence, however, suggests that agreement between partners on the occurrence of IPV is low. Armstrong and colleagues (2002), for example, conducted a review of fifteen studies in which agreement between male and female partners was assessed. They concluded that (1) couples tend to agree on the nonoccurrence of IPV; (2) in most of the studies, there is considerable disagreement as to the occurrence of violence in the relationship; and (3) women tend to report higher levels of violence than men.

The purpose of the current study is to assess the level of agreement about past-year IPV among a sample of blue-collar couples who participated in a cross-sectional study of IPV, work stress, and drinking. Conducting IPV research among blue-collar couples is important for a number of reasons. First, there is considerable evidence, based on general population samples, that IPV prevalence is higher among certain sociodemographic groups, including younger couples, members of racial/ethnic minorities, and couples with household indicators of lower socioeconomic status, such as unemployment, and lower education and income levels (Field & Caetano, 2004; Fox, Benson, Demaris, & Van Wyk, 2002; Sorenson, Upchurch, & Shen, 1996). Despite this evidence, there is a dearth of IPV research among working-class populations, including families in which one or both partners are employed in blue-collar or service occupations. The current study seeks to address this gap. Second, the results of a pilot study conducted among a sample of construction industry workers prior to the current study showed elevated rates (26%) of self-reported past-year IPV perpetration (Cunradi, Ames, & Moore, in press). Although collateral reports were not available in the pilot study, the current study included independent reports from each member of the couple, and allowed for event-specific analysis of male-female agreement on past-year acts of partner violence perpetration and victimization. Because previous research found that higher levels of education and income are associated with lower levels of disagreement about the occurrence of IPV (Medina et al., 2004), we expected couples within this blue-collar sample to exhibit low levels of agreement on the occurrence of past-year IPV.

METHODS

Sample and Data Collection

The study was carried out with the cooperation of a large union representing construction industry workers in Northern California. The goal of the survey was to obtain separate, confidential telephone interviews on work, job stress, IPV and drinking with 1,000 union workers and their spouses or cohabiting partners (i.e., 1,000 couples). The results of an ethnographic component of the study will be reported elsewhere. The study protocol was approved by the Institutional Review Board of Pacific Institute for Research and Evaluation. Study eligibility requirements for workers were (1) membership in the construction industry union; (2) currently married or cohabiting with the same partner for at least 12 months; and (3) physically and mentally able to complete a telephone interview in English or Spanish. Spouse/partners were required to be physically and mentally able to complete a telephone interview in English or Spanish. Population Research Systems (PRS), an experienced San Francisco-based survey research firm, conducted the survey data collection from August 2006 through January 2007. The research protocol required that initial contact be made with the union member, and that the worker’s permission be obtained to contact their spouse/partner by telephone. Informed consent was obtained from each participant, and the voluntary, confidential nature of the study was emphasized. To help safeguard confidentiality, interviewers tried to ensure that each respondent was in a quiet, private setting, and that they felt comfortable answering questions about their relationship with their partner. Additionally, each respondent was offered the opportunity to reschedule the interview at another time. These procedures were repeated before the interviewer began the section covering partner violence. At the conclusion of the interview, each respondent was asked if “…they would like to talk further or get help with some of the relationship issues that we asked you about…” If requested, male respondents were given the telephone number of their county’s mental health services; female respondents were given the telephone number for the National Domestic Violence Hotline. On average, telephone interviews lasted 30 minutes, and each respondent was given $25.00 for their participation.

Worker Recruitment and Participation

The union provided the study researchers with a database containing its current membership of approximately 35,000 workers, excluding retirees. Of these, information in English and Spanish regarding the study was mailed to 10,884 people randomly selected from the database. Fully trained professional bilingual survey interviewers attempted to contact by telephone every union member who had been sent the study mailing. It was determined that a total of 3,960 were not eligible to participate in the study (e.g., not part of study population; did not speak English or Spanish). Study eligibility could not be determined (answering machine, busy signal, etc.) for 3,842 workers. The remaining 3,082 met the study eligibility requirements. Of these, 1,088 workers completed the telephone survey, and 1119 (36.3%) refused to participate. An additional 875 workers met eligibility requirements, but asked for a follow-up call before consenting to complete the interview. Ultimately, this group was not re-contacted because the targeted sample quota (i.e., interviews with 1,000 workers) was achieved. For purposes of calculating a response rate, we assumed the same refusal rate (36.3%) for these 875 workers, which would have resulted in 557 of the 875 consenting to complete the interview had they been re-contacted. Based on this assumption, the effective response rate for workers was 53.4% (1645/3082).

Spouse/Partner Recruitment and Participation

Of 1,088 workers who completed the interview, 95.6% gave their consent for their spouse/partners to be contacted. Of 1,040 potential spouse/partner respondents, 35 (3.4%) were unable to be contacted (answering machine, busy signal, etc.), 56 eligible spouse/partners (5.4%) refused participation, and 927 spouse/partners completed the interview. An additional 21 spouse/partners met eligibility requirements and asked to be called back for the interview. Due to the budgetary and time limitations, these spouse/partners were not re-contacted prior to the completion of data collection. The final sample for the study consisted of 927 married/cohabiting couples, and an additional 161 workers for whom we were unable to collect collateral reports from their spouse/partner. Because the focus of this paper is on male-female agreement, 30 same-sex couples were excluded from the current analysis. Results are therefore presented for 897 couples.

Measurement of Intimate Partner Violence

Past-12 month IPV was measured with the physical assault subscale of the revised Conflict Tactics Scales (CTS2). Straus and colleagues (1996) reported the internal consistency reliability (alpha) for this subscale was .86. The subscale asks about the occurrence of 12 behaviors that the respondent may have perpetrated against their spouse/partner, and that their spouse/partner may have perpetrated against them. Of these 12 behaviors, 5 are classified as minor violence: (1) threw something at my partner that could hurt; (2) twisted my partner’s arm or hair; (3) pushed or shoved my partner; (4) grabbed my partner; and (5) slapped my partner. The severe violence behaviors are (1) used a knife or gun on my partner; (2) punched or hit my partner with something that could hurt; (3) choked my partner; (4) slammed my partner against a wall; (5) beat up my partner; (6) burned or scalded my partner on purpose; and (7) kicked my partner.

Statistical Analysis

Paralleling methodology proposed by Schafer et al. (1998) we assessed partner concordance for reports of partner violence by examining partner agreement and disagreement for 12 violent acts from the physical assault subscale of the CTS2. That is, based on both female and male reports for past year behavior, we determined the percentage of males and females that concurred that a violent act occurred (WYMY; woman yes, man yes), or did not occur (WNMN; woman no, man no), and the percentage of males and females that disagreed (WYMN or WNMY) about the occurrence of a specific violent act. In addition, we calculated Cohen’s kappa for each behavior in order to statistically gauge partner agreement. Cohen’s kappa is believed to be a more robust measure of reliability than percentage agreement as it accounts for agreement that may occur by chance (Cohen, 1960). Fleiss (1981) suggests the following rule of thumb for determining a qualitative interpretation of kappa: < .40, poor; .40-.75; fair to good; > .75, excellent. As recommended by Schafer and colleagues (1998), we calculated lower- and upper-bound estimates of MFPV, FMPV, and any partner-to-partner violence. Lower-bound estimates are based on couple agreement on the occurrence of specific events. Upper-bound estimates are based on reports that specific events had occurred, whether corroborated or not.

RESULTS

Sample Characteristics

Sample characteristics are shown in Table 1. At the individual level, the sample comprised 5.5% female union workers and 94.5% male union workers. Male workers reported having been married or cohabiting with their spouse or partner for an average of 12.2 years whereas female workers indicated that the average length of their married or cohabiting relationship was 8.6 years. Approximately 67% of male workers and 69% of female workers indicted that they were raising any children. Note, however that these couple characteristics were only reported by the primary respondent, the union member. The average age for females in the sample (independent of union status) was 38.8 years as compared to 40.7 years for males. Among males, 18.5% of the sample was under 30 years of age and 19.5% were 50 or older. Among females, 22.8% of the sample was under 30 years of age and 17% were 50 or older. A greater percentage of females (14.8%) had completed four or more years of college compared to males (5.6%). In 78% of the couples, partners were married or cohabiting with someone of the same race. For example, in 4.4% of the couples both partners were African American, in 24.6% both partners were Hispanic, and in 47.3% both partners were Caucasian.

Table 1.

Participant Characteristics for 897 Couples1

Characteristic Female
% (n)		 Male
% (n)
Mean age (SD) 38.8 (11.1) 40.6 (11.2)
Spouse/partner of Union Member 94.5% (n = 848) 5.5% (n = 49)
Race/ethnicity
 Native American Indian or Alaska Native 1.5% (n = 13) .9% (n = 8)
 Filipino .16% (n = 14) .6% (n = 5)
 Asian American 1.8% (n = 16) 1.1% (n = 9)
 Black, African American 4.9% (n = 43) 5.8% (n = 49)
 Latino, Hispanic 32.1% (n = 281) 30.8% (n = 261)
 Native Hawaiian or Other Pacific Islander .6% (n = 5) .4% (n = 3)
 White, Caucasian 53.9% (n = 472) 55.9% (n = 474)
 Multiple ethnicity 3.5% (n = 31) 4.6% (n = 39)
Raising a child 2 69.4% (n = 34) 67.3% (n = 571)
4-year college degree (BA, BS, or equivalent) or higher 14.8% (n = 133) 5.6% (n = 50)
Years of cohabitation or marriage, mean (SD) 2 8.6 (7.4) 12.2 (10.1)
Open in a new tab
1

Respondents did not answer every question so the sample size varies from item to item

2

Question only asked of Union member, regardless of gender

Male-to-Female Partner Violence

Table 2 shows the rates of occurrence for each type of male-to-female partner violence (MFPV). The percentages in the first column (WYMY) and fourth column (WNMN) represent agreement between couple members that an event had or had not occurred, respectively. The second column (WYMN) and third column (WNMY) represent disagreement about the occurrence of an event, and the fifth column contains the kappa statistic that assesses level of agreement. The rates indicate that the most common MFPV events reported by at least one person in the couple were pushing or shoving (13.5%) and grabbing (13.5%), followed by threw something that could hurt (8.5%). Rates of severe MFPV events were generally much lower than the rates of minor MFPV events. For example, less than 1% of couples reported male-to-female acts of burning or scalding, or using a knife or gun. When any type of MFPV is considered based on collaborated and uncollaborated reports (i.e., upper-bound estimate), 21.1% of couples had at least one type of event.

Table 2.

Rates of Male-to-Female Partner Violence among Couples

Violent Act WYMY,
%		 WYMN,
%		 WYMY,
%		 WYMN,
%		 κ
Threw something that could hurt 1.3% 4.1% 3.0% 91.5% 0.24
Pushed or shoved 3.0% 4.6% 5.9% 86.5% 0.31
Slapped .8% 1.2% 1.5% 96.5% 0.36
Beat up .2% 1.0% .8% 98.0% 0.19
Choked .1% .9% .6% 98.4% 0.13
Burned or scalded 0.00% .1% .3% 99.6% NC
Used a knife or gun .2% .1% .2% 99.4% NC
Twisted arm or hair .4% 1.9% 3.0% 94.6% 0.13
Punched or hit with something that could hurt 0.00% 1.6% 1.0% 97.4% NC
Slammed against a wall .1% 1.0% 1.9% 97.0% 0.06
Kicked .1% .9% .6% 98.4% 0.13
Grabbed 2.8% 3.6% 7.2% 86.5% 0.29
Any violence 6.7% 5.1% 9.3% 78.9% 0.40
Open in a new tab

Note. WYMY = both partners report that an act occurred (woman yes, man yes). WYMN = woman reports that an act occurred man does not (woman yes, man no). WNMY = woman reports that an act did not occur, male reports act did occur (woman no, man yes). WNMN = both partners report that an act did not occur (woman no, man no). NC = not calculated.

Agreement between partners, however, was low as indicated by the kappa statistics. For example, 3% of couples agreed that there was at least one episode of male-to-female pushing or shoving, and 86.5% agreed that this had not occurred. Among another 4.6% of couples, the woman reported that this had occurred, but the man did not, and among another 5.9% of couples, the man reported that this had occurred, but the woman did not. Despite being one of the most prevalent types of MFPV, level of agreement for pushing or shoving was low (kappa=.31). Level of agreement was lowest for the severe CTS2 items, and could not be calculated for some of the items due to missing or low cell counts (burned or scalded, used a knife or gun, punched or hit with something that could hurt). Regarding rates of any past-year MFPV, 6.7% of couples agreed that at least one event had occurred. In 5.1% of couples, the woman reported at least one MFPV event but the man did not, and in 9.3% of couples, the man reported at least one MFPV event but the woman did not. Thus, the lower-bound estimate for any MFPV was 6.7%, and the upper-bound estimate was 21.1%.

Female-to-Male Partner Violence

Table 3 shows the rates of occurrence for each type of female-to-male partner violence (FMPV). Similar to the results for MFPV, the rates indicate that the most common FMPV events reported by at least one person in the couple were pushing or shoving (16%), throwing something that could hurt (12.9%), and grabbing (10.7%). Rates of severe FMPV events were generally much lower than the rates of minor FMPV events. For example, less than 1% of couples reported female-male choking, burning or scalding, or using a knife or gun. When any type of FMPV is considered based on corroborated and uncorroborated reports (i.e., upper-bound estimate), 24.2% of couples had at least one type of event.

Table 3.

Rates of Female-to-Male Partner Violence among Couples.

Violent Act WYMY,
%		 WYMN,
%		 WYMY,
%		 WYMN,
%		 κ
Threw something that could hurt 2.1% 3.9% 6.8% 87.1% 0.23
Pushed or shoved 3.5% 5.0% 7.5% 84.0% 0.29
Slapped 1.5% 2.3% 3.1% 93.1% 0.32
Choked 0.00% .2% .6% 99.2% NC
Burned or scalded 0.00% .1% .2% 99.7% NC
Used a knife or gun 0.00% .3% .6% 99.1% NC
Twisted arm or hair .1% .8% 1.9% 97.2% 0.07
Punched or hit with something that could hurt .4% 1.6% 4.7% 93.3% 0.10
Slammed against a wall 0.00% .4% 1.0% 98.5% NC
Beat up .1% .7% .4% 98.8% 0.16
Kicked .4% 1.1% 1.8% 96.6% 0.22
Grabbed 1.1% 2.8% 6.8% 89.3% 0.14
Any violence 7.2% 6.0% 11.1% 75.8% 0.36
Open in a new tab

Note. WYMY = both partners report that an act occurred (woman yes, man yes). WYMN = woman reports that an act occurred, man does not (woman yes, man no). WNMY = woman reports that an act did not occur, man reports act did occur (woman no, man yes). WNMN = both partners report that an act did not occur (woman no, man no). NC = not calculated.

Agreement between partners regarding FMPV was low as indicated by the kappa statistics. For example, 3.5% of couples agreed that there was at least one episode of female-to-male pushing or shoving, and 84% agree that this had not occurred.

Among another 5.0% of couples, the woman reported that this had occurred, but the man did not, and among another 7.5% of couples, the man reported that this had occurred, but the woman did not. Although this was the most prevalent type of FMPV, level of agreement for pushing or shoving was low (kappa=.29). Level of agreement was lowest for the severe CTS2 items, and could not be calculated for some of the items due to missing or low cell counts (choked, burned or scalded, used a knife or gun, punched or hit with something that could hurt). Regarding rates of any past-year FMPV, 7.2% of couples agreed that at least one event had occurred. In 6% of couples, the woman reported at least one FMPV event but the man did not, and in 11.1% of couples, the man reported at least one FMPV event but the woman did not. Thus, the lower-bound estimate for any FMPV was 7.2%, and the upper-bound estimate was 24.2%.

Finally, we calculated lower- and upper-bound estimates for the occurrence of any partner-to-partner violence. The lower bound estimate was based on agreement between partners for any male-to-female partner violence or any female-to-male partner violence, resulting in an estimate of 10.1%. The upper-bound estimate for any partner violence was based on either partner indicating an occurrence of interpersonal violence as either the victim or perpetrator, resulting in an estimate of 30.2%.

Factors Associated with Disagreement

A multinomial regression was conducted to assess whether a set of demographic variables predicted group membership among those couples where at least one or both members reported any violence. Couples in which males, but not females, reported any MFPV, and couples in which females, but not males, reported any MFPV, were compared to couples in which both members reported any MFPV. A similar analysis was conducted to examine group membership for those who reported FMPV. Predictor variables included age of partner as well as age difference, a dummy variable for Hispanic or Caucasian ethnicity for each spouse as well as a variable for whether both members in the couple were Hispanic, White, or other. In addition, we included an indicator of college education for each partner. Results, however, indicated that none of the variables predicted whether partners agreed or disagreed about whether any IPV had occurred in the past year (data not shown).

DISCUSSION

Consistent with findings reported by Armstrong et al. (2002) in their literature review regarding couple agreement about the occurrence of IPV, agreement between partners on IPV in the current study was low. This was the case for assessments of specific acts of MFPV and FMPV, and for the occurrence of any violence. The results thus lend support to the idea of obtaining collateral reports from both partners on the occurrence of IPV whenever possible. This permits the assessment of agreement between partners so that when agreement is low, lower- and upper-bound estimates of IPV can be calculated. Although this interval approach makes the modeling of IPV more complex, it is likely to result in less biased prevalence estimates than if single-point estimates are relied upon. Alternatively, another approach to modeling IPV is to merge across specific items. In this study, this yielded the best (though still only fair) level of agreement. It may be that couples do not agree as much on the specifics of what happened in an incident as much as they agree that an incident of IPV occurred.

In contrast to findings reported by Schafer et al. (1998), men in this sample generally reported a higher incidence of male-to-female and female-to-male violence than the women. For example, 16% of men reported at least one type of MFPV, while 11.8% of women reported at least one type of MFPV. Similarly, 18.3% of the men reported at least one type of FMPV, but only 13.2% of women reported at least one type of FMPV. In an analysis of the 1995 National Study of Couples (n=1599), Schafer et al. (1998) found that women were significantly more likely to report that IPV had occurred, regardless of the gender of the perpetrator. One possible explanation for the gender differences between these two studies is that different methodologies were used. For example, the current study is based on a sample of blue-collar workers (and their spouse/partners) from an industrial union in a particular geographic area of the U.S. (e.g., Northern California). The sample therefore is relatively homogeneous from an occupational and social class perspective. In contrast, the National Study of Couples employed a multistage, multicluster sampling frame in order to obtain a representative sample of married/cohabiting couples from the general household population across the 48 contiguous states. As such, the sample was quite diverse regarding employment and overall socioeconomic status. Additionally, couples in the current study were interviewed by telephone; couples in the National Study of Couples were interviewed face-to-face. These methodological differences may partially explain the apparent gender differences in IPV reporting that is evident between the two studies. Moreover, our research protocol required that initial contact be made with the union member (95% male), and that the worker’s permission be obtained to contact their spouse/partner by telephone. The impact of interviewing the man in a couple first, and then the woman, with both of their knowledge, is unknown, but it is plausible that this produced a “chilling” effect on the woman’s report of IPV. Each respondent was assured of study confidentiality, and that their responses would not be shared with their spouse/partners. Nevertheless, the effect of such constraint would result in underreporting, and thereafter underestimation of IPV in the sample.

The upper-bound estimates for MFPV (21.1%) and FMPV (24.2%) found in this study suggest that IPV rates are elevated in this blue-collar sample compared to prevalence estimates based on national surveys that represent a cross-section of occupational and social class groups. Schafer et al. (1998), for example, found an upper-bound estimate of 13.6% for MFPV, and 18.2% for FMPV, among a national sample of couples. The upper-bound IPV estimates in the current study are consistent with the 26% prevalence rate of IPV perpetration reported among a sample of construction workers who participated in a pilot study of IPV (Cunradi et al., in press). Blue-collar couples may be at elevated risk for IPV compared to white-collar or professional couples. Previous research has linked IPV risk with work-related factors, such as lower income and unemployment (Bachman & Saltzman, 1995; Cunradi, Caetano, & Schafer, 2002; Hotaling & Sugarman, 1986; Sorenson et al., 1996), male occupational status (i.e., blue-collar or low-status jobs) (Fox et al., 2002; Kaufman Kantor, 1990), and occupational stress factors (Fox et al., 2002; Jasinski, Asdigian, & Kantor, 1997). Despite these findings, relatively few studies (e.g., Cunradi et al., in press; Leonard, Bromet, Parkinson, Day, & Ryan, 1985) have been conducted among blue-collar populations in order to more fully investigate the factors that may contribute to elevated risk for IPV. Using a mixed-methods approach, including telephone survey interviews with couples and face-to-face semi-structured interviews with workers, future analyses based on the current study will be able to explore many of the hypothesized mechanisms that potentially underlie the occurrence of IPV among blue-collar couples. Identifying such mechanisms can provide the basis for formulating IPV prevention strategies that target the environmental factors that may shape the context in which IPV occurs.

This study contains a number of limitations. First, the cross-sectional study design does not provide the type of information that would allow for temporal inference. Another limitation concerns possible non-response bias among the study population. A typical concern when conducting IPV survey research, particularly through telephone interviews, is that women who are experiencing abuse will be less likely to participate, resulting in biased estimates of IPV. Waltermaurer et al. (2003), for example, estimated an 8-13% deflated IPV prevalence of moderate to highly abused women, and an 8 to 11% inflated prevalence of non-abused women among a sub-sample of 135 women who would willingly participate in a household survey. McNutt and Lee (2000) found that physically abused women were more likely than other women to say they would participate in telephone surveys, but among severely victimized women, those living with their partner were less willing to participate than those not cohabiting. Black and colleagues (2006), in an analysis of two national telephone surveys, found that the overwhelming majority of respondents (92%) favored questions on interpersonal violence being asked, and there was no difference based on victimization history in willingness to answer questions about violence. Most of the research on IPV non-response bias is based on female participants; little is known, however, about non-response bias among male IPV victims and perpetrators. Caetano et al. (2003) found that male and female non-respondents in a longitudinal study of IPV were not more likely to have reported IPV (either as victims or perpetrators) at baseline. Given the sampling strategy we employed, and the higher non-response rates among males than females in this study, it is difficult to gauge what effect non-response may have had on the study’s IPV prevalence estimates. For example, 4.4% of workers would not give permission for their spouse/partner to be interviewed for the study. Because we lack dyadic data for these 44 workers, they were excluded from the present analysis; future analysis will explore if these workers reported higher rates of IPV compared to workers who gave consent for the spousal interview.

Another limitation is that, due to time constraints, participants were only asked questions from the Physical Assault subscale of the revised CTS2 (Straus et al., 1996); no information was gathered on psychological aggression, injury, or sexual assault. Finally, our subgroup analysis on factors that may underlie couple disagreement on the occurrence of IPV produced null results. In a qualitative study of disagreement on MFPV, Armstrong et al. (2001) found that study participants think women and men remember differently; women remember more than men; both choose what they want to remember; and both remember that they were right in the conflict. In a study based on a sample of couples in which the male partners were recently abstinent polysubstance abusers, Medina et al. (2004) found that men and women’s memory ability, problem solving, disinhibition, and verbal ability were significantly related to disagreement about the most recent IPV episode. In addition, their analysis found that the men’s and women’s ethnicity was significantly related to disagreement about IPV, with African Americans and other ethnic minorities more likely to disagree with their partners than white men and women. Men’s and women’s higher education were also associated with lower levels of disagreement. These factors, however, did not predict agreement or disagreement in our sample. It is possible that because couples in this study were recruited from the same occupational group and social strata, there may not be enough variability in educational level among the sample to replicate the findings of Medina et al.

Despite these limitations, a number of study strengths should be noted. For example, this study is based on a sample of unionized construction industry workers and their spouse/partners. As noted above, there is a dearth of IPV studies that are based on blue-collar samples; the current study helps address this gap. Additionally, this study was able to obtain separate, confidential reports on IPV from both partners. Each partner was asked to report on their own behaviors, and their spouses’ or partners’ behaviors, regarding past-year IPV. Inclusion of both partners in IPV research is critically important for evaluating gender symmetry in physical assaults (Straus, 2006) and accurate assessment of IPV prevalence (Szinovacz & Egley, 1995). The use of collateral reports in the current study also allows for an interval approach to IPV prevalence estimation, since low agreement between partners on the occurrence of IPV suggests that single-point prevalence estimates are biased (Schafer et al., 1998).

In summary, the event-specific analysis provides evidence that most IPV within the sample of blue-collar couples represents situational couple violence (e.g., pushing, shoving, grabbing). Our data suggest that obtaining couple data when possible is essential for estimating prevalence rates of IPV perpetration and victimization. Furthermore, the IPV prevalence estimates reported here underscore the importance of conducting IPV research among understudied populations, such as working-class couples, that appear to be at elevated risk for IPV. Future research should investigate

Acknowledgments

The project described was supported by Grant Number 5 R01 AA015444 from the National Institute on Alcohol Abuse and Alcoholism; Genevieve Ames, Ph.D., Principal Investigator; Carol Cunradi, M.P.H., Ph.D., Co-Principal Investigator. The content is solely the responsibility of the authors and does not necessarily represent the National Institute on Alcohol Abuse and Alcoholism or the National Institutes of Health. The authors are grateful to the union members and their spouses/partners for their participation in the study.

REFERENCES

  1. Archer J. Sex differences in aggression between heterosexual partners: A meta-analytic review. Psychological Bulletin. 2000;126(5):651–680. doi: 10.1037/0033-2909.126.5.651. [DOI] [PubMed] [Google Scholar]
  2. Armstrong TG, Heideman G, Corcoran KJ, Fisher B, Medina KL, Schafer J. Disagreement about the occurrence of male-to-female intimate partner violence: a qualitative study. Family and Community Health. 2001;24(1):55–75. doi: 10.1097/00003727-200104000-00008. [DOI] [PubMed] [Google Scholar]
  3. Armstrong TG, Wernke JY, Medina KL, Schafer J. Do partners agree about the occurrence of intimate partner violence? A review of the current literature. Trauma, Violence, & Abuse. 2002;3(3):181–193. [Google Scholar]
  4. Bachman R, Saltzman LE. Violence against women: Estimates from the redesigned survey (No. NCJ-154348) U.S. Dept. of Justice, Office of Justice Programs; Washington, DC: 1995. [Google Scholar]
  5. Black MC, Kresnow M-J, Simon TR, Arias I, Shelley G. Telephone survey respondents’ reactions to questions regarding interpersonal violence. Violence and Victims. 2006;21(4):445–459. [PubMed] [Google Scholar]
  6. Caetano R, Ramisetty-Mikler S, Field CA. Unidirectional and bidirectional intimate partner violence among white, black, and Hispanic couples in the United States. Violence and Victims. 2005;20(4):393–405. [PubMed] [Google Scholar]
  7. Caetano R, Ramisetty-Mikler S, McGrath C. Characteristics of non-respondents in a US national longitudinal survey on drinking and intimate partner violence. Addiction. 2003;98(6):791–797. doi: 10.1046/j.1360-0443.2003.00407.x. [DOI] [PubMed] [Google Scholar]
  8. Cohen J. A coefficient of agreement for nominal scales. Educational & Psychological Measurement. 1960;20:37–46. [Google Scholar]
  9. Cunradi CB, Ames G, Moore RS. Prevalence and correlates of intimate partner violence among a sample of construction industry workers. Journal of Family Violence. doi: 10.1007/s10896-008-9209-0. in press. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cunradi CB, Caetano R, Schafer J. Alcohol-related problems, drug use, and male intimate partner violence severity among US couples. Alcoholism: Clinical and Experimental Research. 2002;26(4):493–500. [PubMed] [Google Scholar]
  11. Field CA, Caetano R. Ethnic differences in intimate partner violence in the U.S. general population: The role of alcohol use and socioeconomic status. Trauma, Violence, & Abuse. 2004;5(4):303–317. doi: 10.1177/1524838004269488. [DOI] [PubMed] [Google Scholar]
  12. Fleiss JL. Statistical methods for rates and proportions. 2d ed. Wiley; New York: 1981. [Google Scholar]
  13. Fox GL, Benson ML, Demaris AA, Van Wyk J. Economic distress and intimate violence: Testing family stress and resources theory. Journal of Marriage and Family. 2002;64:793–807. [Google Scholar]
  14. Hotaling G, Sugarman D. An analysis of risk markers in husband to wife violence: The current state of knowledge. Violence and Victims. 1986;1(2):101–124. [PubMed] [Google Scholar]
  15. Jasinski JL, Asdigian NL, Kantor GK. Ethnic adaptations to occupational strain: Work-related stress, drinking, and wife assault among Anglo and Hispanic husbands. Journal of Interpersonal Violence. 1997;12(6):814–831. [Google Scholar]
  16. Johnson MP. Patriarchal terrorism and common couple violence: Two forms of violence against women. Journal of Marriage and the Family. 1995;57:283–294. [Google Scholar]
  17. Johnson MP. Conflict and control: Gender symmetry and asymmetry in domestic violence. Violence Against Women. 2006;12(11):1003–1018. doi: 10.1177/1077801206293328. [DOI] [PubMed] [Google Scholar]
  18. Johnson MP, Ferraro KJ. Research on domestic violence in the 1990s: Making distinctions. Journal of Marriage and the Family. 2000;62:948–963. [Google Scholar]
  19. Johnson MP, Leone JM. The differential effects of intimate terrorism and situational couple violence: Findings from the National Violence Against Women Survey. Journal of Family Issues. 2005;26(3):322–349. [Google Scholar]
  20. Kaufman Kantor G. The ‘drunken bum’ theory of wife beating. In: Straus M, Gelles R, editors. Physical violence in American families. Transaction; New Brunswick: 1990. [Google Scholar]
  21. Leonard KE, Bromet EJ, Parkinson DK, Day NL, Ryan CM. Patterns of alcohol use and physically aggressive behavior in men. Journal of Studies on Alcohol. 1985;46(4):279–282. doi: 10.15288/jsa.1985.46.279. [DOI] [PubMed] [Google Scholar]
  22. McNutt L-A, Lee R. Intimate partner violence prevalence estimation using telephone surveys: Understanding the effect of nonresponse bias. American Journal of Epidemiology. 2000;152(5):438–441. doi: 10.1093/aje/152.5.438. [DOI] [PubMed] [Google Scholar]
  23. Medina LK, Schafer J, Shear PK, Armstrong TG. Memory ability is associated with disagreement about the most recent conflict in polysubstance abusing couples. Journal of Family Violence. 2004;19(6):387–398. [Google Scholar]
  24. Schafer J, Caetano R, Clark CL. Rates of intimate partner violence in the United States. American Journal of Public Health. 1998;88(11):1702–1704. doi: 10.2105/ajph.88.11.1702. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Sorenson S, Upchurch D, Shen H. Violence and injury in marital arguments: risk patterns and gender differences. American Journal of Public Health. 1996;86(1):35–40. doi: 10.2105/ajph.86.1.35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Stets J, Straus M. The marriage license as a hitting license: A comparison of assaults in dating, cohabiting, and married couples. In: Straus M, Gelles R, editors. Physical violence in American families. Transaction; New Brunswick: 1990. [Google Scholar]
  27. Straus MA. The controversy over domestic violence by women: A methodological, theoretical, and sociology of science analysis. In: Arriaga X, Oskamp S, editors. Violence in intimate relationships. Sage Publications, Inc.; Thousand Oaks, CA: 1999. pp. 17–44. [Google Scholar]
  28. Straus MA. Women’s violence toward men is a serious social problem. In: Loseke DR, Gelles RJ, Cavanaugh MM, editors. Current controversies on family violence. 2nd ed. Sage Publications; Newbury Park, NJ: 2005. [Google Scholar]
  29. Straus MA. Future research on gender symmetry in physical assaults on partners. Violence Against Women. 2006;12(11):1086–1097. doi: 10.1177/1077801206293335. [DOI] [PubMed] [Google Scholar]
  30. Straus MA, Hamby SL, Boney-McCoy S, Sugarman DB. The Revised Conflict Tactics Scales (CTS2): Development and Preliminary Psychometric Data. Journal of Family Issues. 1996;17(3):283–316. [Google Scholar]
  31. Szinovacz ME, Egley LC. Comparing one-partner and couple data on sensitive marital behaviors: the case of marital violence. Journal of Marriage and Family. 1995;57:995–1010. [Google Scholar]
  32. Waltermaurer E, Ortega CA, McNutt L-A. Issues in estimating the prevalence of intimate partner violence: assessing the impact of abuse status on participation bias. Journal of Interpersonal Violence. 2003;18(9):959–974. doi: 10.1177/0886260503255283. [DOI] [PubMed] [Google Scholar]

RESOURCES