Problem Drinking, Unemployment, and Intimate Partner Violence among a Sample of Construction Industry Workers and their Partners

Abstract

The purpose of this study is to determine the contribution of male unemployment and each partner’s problem drinking to risk for male-to-female partner violence (MFPV) and female-to-male partner violence (FMPV) among a sample of construction industry workers and their spouses/partners. Participants in the sample (n=848 couples) completed cross-sectional health behavior surveys. Multivariate logistic regression models of MFPV and FMPV, with adjustment for demographic and psychosocial variables, were developed. Approximately 20% of couples reported MFPV, and 24% reported FMPV. Results indicated that couples in which the male was a problem drinker, and in which the male worker reported being currently unemployed, were at risk for MFPV. Number of months unemployed by the male worker was significantly associated with FMPV, but problem drinking was not associated with this outcome. Male and female impulsivity were significantly associated with risk for MFPV and FMPV, and the male’s report of adverse childhood events was associated with increased likelihood of MFPV. There was no evidence for the effects of unemployment being moderated (exacerbated) by problem drinking. Workplace-based prevention efforts may be a feasible and important strategy to reduce problem drinking and partner violence among high-risk occupational groups.

Problem drinking and male unemployment have long been noted as risk factors for intimate partner violence (IPV). For example, in a survey of 100 battered wives, Gayford (1975) found that approximately half reported that their husbands engaged in frequent heavy drinking, and a similar proportion reported that their husbands experienced occasional, frequent, or chronic unemployment. The overall goal of this cross-sectional study is to estimate the contribution of each partner’s problem drinking and the male’s unemployment to risk for past-year male-to-female partner violence (MFPV) and female-to-male partner violence (FMPV) among a sample of married/cohabiting construction industry workers and their spouses/partners after taking into account a range of sociodemographic and psychosocial characteristics of the couples.

Analyzing the role of each partner’s problem drinking and male unemployment in relation to MFPV and FMPV among a sample from this population is significant for a number of reasons. First, evidence suggests that blue-collar couples experience elevated rates of IPV compared to couples in white-collar or professional occupations. For example, results of a pilot survey conducted among a sample of 100 construction industry workers prior to the current study showed elevated rates (26%) of self-reported IPV perpetration (Cunradi et al., 2008). Similarly, in an analysis of 897 blue-collar couples from the current study in which one spouse/partner is a construction worker, Cunradi et al. (in press) found that approximately 21% of the couples reported MFPV, and 24% reported FMPV. In comparison, Schafer et al. (1998) reported MFPV rates of 14%, and FMPV rates of 18%, among a national sample of married and cohabiting couples. Second, in a national sample of full-time workers aged 18 to 64, workers in the construction industry had the highest rate (16%) of past-month heavy alcohol use (drinking 5 or more drinks on the same occasion on 5 or more day in the past 30 days) among workers in all industry categories (Larson et al., 2007). Because of research linking indices of heavy drinking with increased IPV risk (see reviews in Leonard, 1993, 2001), it is imperative to explore the association of problem drinking to IPV among workers (and their partners) in this occupational group. Third, workers in the construction trades are subject to multiple work stressors, including seasonal fluctuations in work, stretches of lay off (i.e., temporary unemployment), and long commutes to construction job sites (Cunradi et al., 2008). Construction is also one of the most dangerous industries in the United States, and accounts for a disproportionate share of work-related injuries and deaths (Waehrer et al., 2007). For example, in 2004, construction workers accounted for 7.7% of the U.S. workforce, but suffered 22.2% of the nation’s 5,764 work-related deaths (Bureau of Labor Statistics, 2005). Since men’s occupational stressors have been associated with IPV (Cano & Vivian, 2003; Cascardi & Vivian, 1995), focusing on this population of workers and their spouses/partners will help to elucidate the prevalence and correlates of health-related disparities, reflected in the disproportionate prevalence of IPV and problem drinking among blue-collar families.

Although not a “necessary or sufficient cause” of IPV, problem drinking (e.g., heavy or binge drinking; intoxication) on the part of the male often precedes or accompanies acts of IPV (Leonard, 2005). In addition, some research suggests that problematic drinking patterns on the part of the male and female are associated with both MFPV and FMPV among couples in the general household population (Cunradi et al., 1999, 2002a). Context of drinking and other potential moderator variables may be of critical importance for understanding why alcohol contributes to IPV for some couples under some circumstances but not others (Testa, 2004). Several theoretical explanations of the alcohol-IPV relationship have been proposed. While a full discussion of these theories is beyond the scope of this article, Klostermann and Fals-Stewart (2006) recently reviewed the evidence for three proposed mechanisms underlying the alcohol-IPV association: the spurious cause model in which the alcohol-IPV relationship is the result of these variables being related to other factors that influence both drinking and IPV; the indirect effects model in which alcohol use has detrimental effects on relationship quality by increasing marital discord, which in turn increases the likelihood of IPV; and the proximal effects model in which alcohol intoxication is a proximal causal agent of IPV via the psychopharmacologic effects of alcohol on cognitive processing or through alcohol-related expectancies (Leonard & Quigley, 1999). The preponderance of evidence for the spurious cause model is weak in that the association between alcohol and IPV remains significant even when a range of psychosocial and sociodemographic variables related to both behaviors are controlled for (Leonard, 2005). Likewise, the indirect effects model is not well supported empirically because the alcohol-IPV association remains significant even when marital satisfaction, discord, and similar variables are statistically accounted for (e.g., Fals-Stewart et al., 2005). Klostermann and Fals-Stewart (2006) suggest that there is now considerable empirical support for the proximal effects model, including longitudinal studies that have found that the husband’s alcohol use predicted subsequent marital aggression (Heyman et al., 1995; Leonard & Senchack, 1996; Quigley & Leonard, 2000). Studies conducted among male alcoholics have shown that the occurrence of IPV was significantly reduced after the men completed treatment for alcohol dependence (O'Farrell et al., 2003; Stuart et al., 2003). Fals-Stewart (2003) found that likelihood of men engaging in IPV on days in which alcohol was consumed was 8 times higher among men entering treatment for domestic violence, and 11 times higher among men entering treatment for alcoholism.

Loss of employment has been recognized for decades as a stressful life event (e.g., Holmes & Rahe, 1967), and early family violence researchers theorized that elevated rates of family violence were linked to socially structured stress due to conditions such as unemployment and concomitant loss of the male breadwinner’s role (Gelles, 1980,1985). Especially for lower SES couples, male unemployment may precipitate economic distress, which in turn may trigger IPV (Fagan & Browne, 1994). For example, couples experiencing economic distress due to unemployment may have more arguments, either directly or indirectly due to money problems, which may manifest ultimately in IPV. Unemployed males may also be at home more than employed males, and may thus have more opportunities to interact negatively with their spouses/partners (Benson et al., 2003). Although research findings as to contribution of male unemployment to IPV risk have not been entirely consistent, the results generally point to male unemployment as a risk factor for MFPV in cross-sectional studies (Melzer, 2002) and longitudinal analyses (Benson et al., 2003; Caetano et al., 2005). In addition, some research suggests that the association between male unemployment and IPV may differ based on the race/ethnicity of the couple (e.g., Cunradi et al., 1999; Jasinksi et al., 1997).

Because women are more likely than men to sustain injuries as a result of IPV (Archer, 2000; Morse, 1995; Tjaden & Thoennes, 1998), MFPV has been regarded as the more urgent public health issue, and has received considerably more research attention than FMPV. Nevertheless, because rates of FMPV have been shown to equal or exceed MFPV rates among couples in the general household population (Caetano et al., 2000; Cunradi, 2007; Schafer et al., 1998; Straus, 1995), it is important to consider the contribution of factors such as problem drinking and unemployment to both types of IPV in order to further public health prevention efforts. The current study aims to increase our understanding of the role of male and female problem drinking to the occurrence of IPV among an at-risk, understudied population, and tests whether the association between male unemployment and IPV is moderated by problem drinking.

Methods

Sample and Data Collection

This mixed methods study (i.e., survey and ethnography) was carried out with the cooperation of a large union representing 35,000 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 married/cohabiting union workers and their spouses or cohabiting partners (i.e., 1,000 couples). Results of the ethnographic component will be reported elsewhere. The study protocol for the protection of human subjects was approved by the Institutional Review Board of Pacific Institute for Research and Evaluation. Survey data collection was conducted from August 2006 through January 2007. On average, telephone interviews lasted 30 minutes, and each respondent was mailed a $25.00 check for participating in the study. For a detailed description of the study data collection protocol and participant recruitment, see Cunradi et al. (in press).

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. Fully trained professional bilingual survey interviewers completed telephone interviews with 1,088 workers (53.4% response rate). 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.

Of 1,088 workers who completed the interview, 95.6% gave their consent for their spouse/partners to be contacted. A total of 927 spouses/partners completed the telephone survey interview. The final sample consisted of 927 married/cohabiting couples, and an additional 161 workers who lacked collateral reports from their spouse/partner. Because the current study focuses on the contribution of the male construction worker’s unemployment to the occurrence of MFPV and FMPV within the dyad, 30 same-sex couples, and 49 couples composed of female construction workers and male spouses/partners, were excluded from the study. The analyses herein are limited to 848 couples comprising male construction workers and female spouses/partners.

Measures

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: (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; (5) slapped my partner; (6) used a knife or gun on my partner; (7) punched or hit my partner with something that could hurt; (8) choked my partner; (9) slammed my partner against a wall; (10) beat up my partner; (11) burned or scalded my partner on purpose; and (12) kicked my partner. Separate variables were created for MFPV and FMPV. Violence was considered to have occurred if at least one partner reported a violent incident in the past year, regardless of whether the incident was corroborated by the other partner. Thus, if either partner reported occurrence of a violent incident, the partner violence variable (MFPV or FMPV, depending on the gender of the perpetrator) was coded “1;” if neither reported an incident, the variable was coded ‘0.’ This method allows for the correction of under-reporting of violence common in one partner data (Caetano et al., 2000). Previous analysis among the study’s couples showed that among those in which at least one partner reported any MFPV, or any FMPV, less than one third of the couples agreed about the occurrence of the event (Cunradi et al., in press).

Unemployment

Construction workers were asked two questions about unemployment. First, they were asked, “Are you currently on lay off?” A dichotomous variable was created for this measure. Next, they were asked, “Over the past 12 months, about how much time have you spent on lay off?” Response categories were (1) none; (2) less than one month; (3) between one and two months; (4) between two and three months; (5) between three and six months; and (6) more than six months. A continuous measure of months on layoff was created from this item using values corresponding to the midpoints of the intervals implied by the response options (i.e., 0 months, 0.5 months, 1.5 months, 2.5 months, 4.5 months, 9 months), so that the odds ratios obtained from our analyses would index the change in odds in terms of months rather than in terms of ordinal scale points.

Problem Drinking

Problem drinking was assessed with the Alcohol Use Disorders Identification Test (AUDIT), a screener developed by the World Health Organization to identify persons whose alcohol use may have become harmful to their health (Saunders et al., 1993). Males are classified as problem drinkers if their AUDIT score is 8 or higher; females are positively classified if their score is 5 or higher (Reinert & Allen, 2007). Gender-specific dichotomous variables were created to indicate the presence/absence of problem drinking.

Interpersonal Conflict at Work

Workers were asked, “Thinking back over the past 12 months, about how many times has each of the following things happened to you?” (a) Had a heated argument with your supervisor; (b) Had a heated argument with a co-worker; (c) Been in a physical fight with a co-worker. Response categories are: 10 or more times; 6–9 times; 2–5 times; once; never. Previous analysis found that reliability (Cronbach’s α) of this measure was 0.63 (Cunradi et al., 2008). This measure was included as a control variable in the analyses since it was expected that workers experiencing greater interpersonal conflict at work would be more likely to engage in IPV.

Impulsivity

Impulsivity was measured with a set of questions that have been used in previous national alcohol surveys (Caetano et al., 2000). Respondents were asked, “How well do the following statements describe you? WouId you say that this describes you quite a lot, some, a little, or not at all? (1) I often act on the spur-of-the-moment without stopping to think; (2) You might say I act impulsively; (3) Many of my actions seem to be hasty.” Reliability (Cronbach’s α) of this measure was 0.79.

Adverse Childhood Experiences

Childhood exposure to violence, alcoholism, and other adverse events was measured with a modified version of the Adverse Childhood Experiences (ACE) scale (Felitti et al., 1998). The modified ACE (Cabrera et al., 2007) asks about the following experiences while they were growing up as a child: (1) parent/caregiver-perpetrated physical abuse, (2) psychological abuse or (3) sexual abuse; (4) alcoholism or problem drinking by a household member; (5) depression or mental illness of a household member; and (6) domestic violence toward mother or caregiver. Respondents are defined as exposed to a category if they responded affirmatively to one or more of the questions in that category. A scale of exposure to adverse childhood experiences, ranging from 0 – 6, was created by summing the number of positive responses to each of the six categories. Reliability (Cronbach’s α) of this measure was 0.70.

Sociodemographic Characteristics

These factors included age, relationship length, and each partner’s race/ethnicity and highest level of education. Because the age of partners within couples is highly correlated, a variable was created representing couple mean age. Self-reported respondent race/ethnicity (Native American or Alaska Native; Filipino; Asian; Black or African American; Latino or Hispanic; Native Hawaiian or other Pacific Islander; White or Caucasian; or Other) was re-categorized as white or Caucasian, Latino or Hispanic, Black or African American, and other. Education was categorized as some high school, high school graduate, some college, and college graduate.

Analytic Strategy

The bivariate associations between sample characteristics and IPV were assessed with chi-square tests of independence for categorical variables (Table 1) and t-tests for continuous measures (Table 2). A series of multivariate logistic regression models was developed to assess the contribution of problem drinking, current layoff, and months on layoff to the risk of IPV as follows: The first set of models included variables representing couple characteristics, male characteristics, and female characteristics in relation to the risk of MFPV (Table 3). Model 1 included a measure of current layoff; Model 2 included a measure of months on layoff. Parallel analyses were conducted in order to separately test the impact of unemployment on IPV as an acute event (i.e., current layoff) as well as its chronic effect (i.e., months on layoff). The same model building procedures were carried out with FMPV as the dependent variable (Table 4). The models included an interaction term in order to test whether the association between unemployment and IPV was moderated by problem drinking. Because there was no evidence for the effects of unemployment being moderated (exacerbated) by problem drinking (data not shown), the results of the main effects models are presented.

Table 1.

Sample Characteristics of Male Workers and Female Spouse/Partners (n=848 couples)

	Number	MFPV (%)	χ2, df	FMPV (%)	χ2, df
Males
Education:
Some high school	161	26.1	5.37,	24.8	0.71,
High school graduate	407	20.4	3	23.8	3
Some college	232	17.4		23.0
College graduate	47	14.9		19.1
Race/ethnicity:
White	454	17.2	6.50,	21.9	11.36*,
Hispanic	250	22.1	3	20.9	3
Black	38	31.6		44.7
Other	83	22.9		25.3
Problem Drinking:
Yes	141	34.3	20.77***,	31.4	6.17*,
No	688	17.3	1	21.7	1
Current Layoff
Yes	178	30.5	13.79***,	29.4	3.97,
No	663	17.8	1	22.2	1
Females
Education:
Some high school	127	21.2	3.54,	21.2	2.76,
High school graduate	289	18.3	3	22.5	3
Some college	306	23.4		26.6
College graduate	126	16.7		20.6
Race/ethnicity:
White	453	17.0	7.62,	20.5	7.74,
Hispanic	272	25.5	3	24.4	3
Black	31	22.6		38.7
Other	80	20.3		29.1
Problem Drinking:
Yes	81	24.7	0.98,	28.4	1.16,
No	761	20.0	1	23.1	1

Note.

Numbers may not add up due to missing data

^*p < 0.05;

^***p < 0.001.

Table 2.

Mean differences among Male Workers and Female Spouses/Partners by Male-to-Female and Female-to-Male Partner Violence (n=848 couples)

	MFPV	N	Mean (SD)	t, df	FMPV	N	Mean (SD)	t, df
Couple characteristics
Relationship length (y)	No	670	12.96 (10.53)	5.350***	No	644	13.05 (10.37)	4.971***
	Yes	172	9.14 (7.69)	354.514	Yes	198	9.35 (8.75)	382.232
Mean age (y)	No	674	40.41 (11.09)	4.793***	No	647	40.67 (10.90)	5.294***
	Yes	172	36.32 (9.68)	296.295	Yes	199	36.05 (10.33)	844
Male characteristics
Impulsivity score	No	671	1.71 (0.76)	−5.997***	No	643	1.71 (0.76)	−4.953***
	Yes	169	2.14 (0.87)	237.859	Yes	197	2.06 (0.88)	292.330
Adverse childhood events	No	673	0.73 (1.12)	−3.539***	No	646	0.73 (1.11)	−2.893**
	Yes	172	1.10 (1.26)	243.904	Yes	199	1.03 (1.29)	293.857
Interpersonal work conflict	No	671	1.30 (2.89)	−2.346*	No	644	1.24 (2.98)	−3.278**
	Yes	171	2.15 (4.47)	207.445	Yes	198	2.25 (4.03)	266.664
Months on layoff	No	667	1.24 (2.16)	−2.831**	No	639	1.21 (2.09)	−2.964**
	Yes	171	1.77 (2.27)	836	Yes	199	1.79 (2.46)	291.886
Female characteristics
Impulsivity score	No	674	1.57 (0.66)	−3.482**	No	647	1.56 (0.66)	−3.913***
	Yes	172	1.79 (0.76)	240.174	Yes	199	1.80 (0.75)	297.008
Adverse childhood events	No	674	0.94 (1.34)	−2.929**	No	647	0.95 (1.33)	−2.371*
	Yes	172	1.30 (1.45)	250.003	Yes	199	1.22 (1.45)	308.178

Note.

2-tailed significance:

^*p < 0.05;

^**p < 0.01;

^***p < 0.001.

Table 3.

Multivariate Models of Male-to-Female Partner Violence, Male Workers and Female Spouse/Partners (n=848 couples)

	Model 1		Model 2
Couple characteristics	OR	95% CI	OR	95% CI
Relationship length (y)	0.98	0.95, 1.01	0.97	0.95, 1.00
Mean age (y)	0.99	0.96, 1.01	0.98	0.96, 1.01
Male characteristics
Education:
Some high school	1.15	0.41, 3.22	1.13	0.41, 3.16
High school graduate	0.95	0.37, 2.43	0.89	0.35, 2.27
Some college	0.87	0.33, 2.31	0.86	0.32, 2.27
College graduate (ref.)	1.00	--	1.00	--
Race/ethnicity:
White	0.95	0.48, 1.90	0.92	0.46, 1.84
Hispanic	1.13	0.53, 2.41	1.11	0.52, 2.36
Black	4.33	1.00, 18.78	4.39*	1.01, 19.02
Other (ref.)	1.00	--	1.00	--
Problem Drinking:
Yes	1.98**	1.24, 3.14	2.00**	1.26, 3.17
No (ref.)	1.00	--	1.00	--
Impulsivity score	1.68***	1.32, 2.14	1.64***	1.29, 2.09
Adverse childhood events	1.18*	1.01, 1.38	1.19*	1.01, 1.39
Interpersonal work conflicts	1.02	0.97, 1.07	1.02	0.97, 1.08
Current Layoff
Yes	1.59*	1.02, 2.48
No (ref.)	1.00	--
Months of Layoff			1.06	0.98, 1.16
Female characteristics
Education:
Some high school	0.85	0.37, 1.95	0.86	0.38, 1.95
High school graduate	0.81	0.42, 1.57	0.83	0.43, 1.60
Some college	1.48	0.79, 2.78	1.47	0.79, 2.76
College graduate (ref.)	1.00	--	1.00	---
Race/ethnicity:
White	0.79	0.39, 1.60	0.76	0.37, 1.55
Hispanic	1.37	0.63, 3.00	1.38	0.63, 3.02
Black	0.34	0.06, 1.81	0.31	0.06, 1.69
Other (ref.)	1.00	--	1.00	---
Problem Drinking:
Yes	0.91	0.48, 1.74	0.91	0.48, 1.75
No (ref.)	1.00	--	1.00	--
Impulsivity score	1.52**	1.16, 2.01	1.51**	1.15, 2.00
Adverse childhood events	1.12	0.98, 1.28	1.14	1.00, 1.31

Note.

^*p < 0.05;

^**p < 0.01;

^***p < 0.001.

Table 4.

Multivariate Models of Female-to-Male Partner Violence, Male Workers and Female Spouse/Partners (n=848 couples)

	Model 1		Model 2
Couple characteristics	OR	95% CI	OR	95% CI
Relationship length (y)	0.99	0.96, 1.02	0.99	0.97, 1.02
Mean age (y)	0.97**	0.94, 0.99	0.96**	0.94, 0.99
Male characteristics
Education:
Some high school	0.89	0.34, 2.35	0.85	0.32, 2.25
High school graduate	0.92	0.39, 2.20	0.89	0.37, 2.12
Some college	0.93	0.38, 2.29	0.94	0.38, 2.31
College graduate (ref.)	1.00	--	1.00	--
Race/ethnicity:
White	1.32	0.68, 2.54	1.29	0.67, 2.50
Hispanic	0.98	0.47, 2.04	0.99	0.47, 2.07
Black	5.04*	1.26, 20.23	4.80*	1.18, 19.46
Other (ref.)	1.00	--	1.00	--
Problem Drinking:
Yes	1.31	0.74, 1.80	1.29	0.81, 2.04
No (ref.)	1.00	--	1.00	--
Impulsivity score	1.42**	1.13, 1.79	1.41**	1.12, 1.77
Adverse childhood events	1.14	0.99, 1.33	1.13	0.97, 1.31
Interpersonal work conflicts	1.04	0.99, 1.10	1.05	1.00, 1.10
Current Layoff:
Yes	1.16	0.74, 1.80
No (ref.)	1.00	--
Months of Layoff			1.11*	1.02, 1.20
Female characteristics
Education:
Some high school	0.94	0.43, 2.05	0.96	0.44, 2.11
High school graduate	0.95	0.52, 1.73	0.98	0.54, 1.80
Some college	1.30	0.73, 2.31	1.38	0.77, 2.47
College graduate (ref.)	1.00	--	1.00	---
Race/ethnicity:
White	0.57	0.30, 1.06	0.59	0.31, 1.10
Hispanic	0.82	0.40, 1.67	0.82	0.40, 1.68
Black	0.46	0.10, 2.10	0.44	0.09, 2.03
Other (ref.)	1.00	--	1.00	---
Problem Drinking:
Yes	0.95	0.52, 1.74	0.91	0.54, 1.81
No (ref.)	1.00	--	1.00	--
Impulsivity score	1.56**	1.20, 2.02	1.58**	1.21, 2.05
Adverse childhood events	1.08	0.95, 1.23	1.08	0.95, 1.23

Note.

^*p < 0.05;

^**p < 0.01;

^***p < 0.001.

Results

Approximately 17% of male workers were classified as problem drinkers, as were 10% of female spouses/partners. Rates of MFPV (Table 1) were twice as high among couples in which the male was a problem drinker compared to couples in which the male was not a problem drinker (34% vs. 17%). Similarly, rates of FMPV were significantly higher among couples in which the male was a problem drinker compared to couples in which the male was not a problem drinker (31% vs. 22%). Current layoff was reported by 21% of male workers, and rates of MFPV were significantly higher among couples in which the male was currently on layoff compared to couples in which the male was not on current layoff (31% vs. 18%). There were no significant differences in rates of MFPV or FMPV based on the female’s education, race/ethnicity, or problem drinking status. FMPV rates varied significantly based on male race/ethnicity. While rates of MFPV and FMPV appear elevated among these couples, previous analyses of the data indicate that the overwhelming majority of aggressive acts between partners can be classified as “moderate” (e.g., pushing, shoving, grabbing) and thus likely represent situational or common couple violence (Cunradi et al., in press).

There were significant differences in mean couple, male, and female characteristics based on whether or not the couple reported past-year MFPV and FMPV (Table 2). Average relationship length was significantly shorter among those that reported MFPV and FMPV compared to couples not reporting MFPV or FMPV (9 years vs. 13 years). Similarly, the mean age of couples reporting MFPV and FMPV was younger than that of couples not reporting MFPV or FMPV (36 years vs. 40 years). Impulsivity scores and number of adverse childhood events were greater among males in couples reporting MFPV and FMPV compared to males in couples not reporting MFPV or FMPV. Males in couples reporting MFPV and FMPV had significantly greater interpersonal work conflict and months on layoff compared to males in couples not reporting MFPV or FMPV. The mean impulsivity scores and number of adverse childhood events were greater among females in couples reporting MFPV and FMPV compared to females in couples not reporting MFPV or FMPV.

Multivariate correlates of MFPV are shown in Table 3. In Model 1, couples in which the male was categorized as a problem drinker were approximately twice as likely to report MFPV compared to couples in which the male was not a problem drinker (Odds Ratio [OR] = 1.98; 95% Confidence Interval [CI] 1.24, 3.14; p < 0.01). Couples in which the male was on current layoff were at significantly elevated risk for MFPV compared to couples which the male was not on current layoff (OR=1.59; 95% CI 1.02, 2.48; p < 0.05). Male impulsivity (OR=1.68; 95% CI 1.32, 2.14; p< 0.001), male adverse childhood events (OR=1.18; 95% CI 1.01, 1.38; p < 0.05) and female impulsivity (OR=1.52; 95% CI 1.16, 2.01; p < 0.01) were also significantly associated with MFPV. In Model 2, the male worker’s months of layoff were not significantly associated with MFPV. Couples in which the male was categorized as a problem drinker were twice as likely to report MFPV compared to couples in which the male was not a problem drinker (OR=2.00; 95% CI 1.26, 3.17; p < 0.01), and couples in which the male reported his race/ethnicity as black were at elevated MFPV risk compared to couples in which the male’s race/ethnicity was categorized as ‘other’ (OR=4.39; 95% CI 1.01, 19.02). Male impulsivity, male adverse childhood events, and female impulsivity were significantly associated with MFPV in Model 2. Female problem drinking was not linked with MFPV risk in Models 1 or 2.

Multivariate correlates of FMPV are shown in Table 4. In Model 1, mean couple age was inversely associated with risk of FMPV (OR=0.97; 95% CI 0.94, 0.99; p <0.01). Couples in which the male was a problem drinker were not at elevated risk for FMPV, nor were couples in which the male was on current layoff. Male impulsivity (OR=1.42; 95% CI 1.13, 1.79; p <0.01) and female impulsivity (OR=1.56; 95% CI 1.20, 2.02; p <0.01) were significantly associated with FMPV risk. In Model 2, mean couple age was also inversely associated with FMPV risk (OR=0.96; 95% CI 0.94, 0.99; p < 0.01). Male problem drinking was not linked with elevated risk for FMPV. Months of layoff, however, were associated with increased risk of FMPV (OR=1.11; 95% CI 1.02, 1.20; p <0.05). Male impulsivity and female impulsivity were significantly associated with elevated FMPV risk. In both models, couples in which the female was a problem drinker were not at elevated risk for FMPV compared to couples in which the female was not a problem drinker. Couples in which the male’s race/ethnicity was categorized as black were at elevated FMPV risk in both models compared to couples in which the male’s race/ethnicity was categorized as ‘other.’

Discussion

Results of this cross-sectional study showed that approximately 17% of male construction industry workers, and 10% of the female spouses/partners, met gender-specific AUDIT criteria for problem drinking. Findings for male workers are consistent with data indicating that prevalence of past-month heavy drinking is highest among workers in the construction industry (Larson et al., 2007). While a direct comparison for females by occupational category cannot be made, data from the 2002–2005 National Surveys on Drug Use and Health indicate that between 2.2% and 2.7% of females aged 26 and older reported past-month heavy alcohol use (Substance Abuse and Mental Health Services Administration, 2006). Thus, the problem drinking rate among female spouses/partners (10%), although lower than that of their husbands/partners, appears quite elevated relative to women in the general population. Previous research suggests presence of spousal concordance for alcohol use (Demers et al., 1999; Leadley et al., 2000), which may be attributed to assortative mating (i.e., individuals choose to marry partners who are already similar to themselves; Leonard & Eiden, 1999; Leonard & Mudar, 2003).

In multivariate analyses, couples in which the male was a problem drinker were twice as likely to report past-year MFPV compared to couples in which the male was not a problem drinker. Couples in which the female was a problem drinker, however, were not at elevated risk for MFPV compared to couples in which the female was not a problem drinker. While the role of the male’s drinking has been consistently linked to MFPV in both cross-sectional (Pan et al., 1994) and longitudinal studies of couples (Leonard & Senchack, 1996), the role of the female’s drinking in the occurrence of MFPV is less clear. In an analysis of newlywed couples, Leonard and Quigley (1999) found that the husband’s drinking was associated with MFPV, but concluded that there was little evidence to support a contributory role for the wife’s drinking. Similarly, Testa et al. (2003), in a longitudinal examination of a sample of 18–30 year old women, found that women’s heavy drinking did not predict subsequent experiences of IPV in ongoing or new relationships. A longitudinal analysis by Quigley and Leonard (2000), however, indicated that the husband’s drinking patterns in early marriage predicted MFPV in the context of the wife’s alcohol use. In other words, husband and wife alcohol use in the first year of marriage interacted to predict MFPV in the second and third years of marriage. Cunradi et al. (2002a) found that couples in which the female reported alcohol-related problems were at risk for both moderate and severe MFPV compared to couples in which the female did not report alcohol-related problems. In the same analysis, male alcohol-related problems were also linked with both outcomes, but the measures of association were not as strong. These somewhat disparate findings concerning the role of the female partner’s drinking patterns and alcohol-related problems in relation to MFPV suggests that further research in this area is warranted.

Approximately 21% of the male workers in the sample reported being on current layoff at the time of the survey, and couples in which the worker was on current layoff were at elevated risk for MFPV compared to couples in which the worker was not on current layoff. Interestingly, the number of months on layoff in the past year reported by the worker was not associated with increased MFPV risk. One explanation for this finding is that current layoff may create a type of acute stress for the couple that can fuel conflict and physical aggression on the part of the male partner. Due to the cross-sectional study design, however, it is not possible to directly test this hypothesis. Another type of work stress, interpersonal work conflict, was not found to be associated with MFPV risk in this study. In contrast, Cunradi et al. (2008) found that interpersonal work conflict was significantly associated with IPV among a sample of 100 construction industry workers after adjustment for age, gender, and race/ethnicity. The low reliability of this measure may explain the absence of a significant association between interpersonal conflict at work and partner violence in the current study.

Male and female impulsivity were significantly associated with increased MFPV risk in both models. These findings are consistent with previous research linking these factors to the overall occurrence of MFPV (Schafer et al., 2004) and to risk for both moderate and severe male IPV (Cunradi et al., 2002a). Impulsivity is often characterized as an inability to regulate certain behaviors, such as aggression (Plutchik & van Praag, 1997); as such, it can contribute to risk for IPV. In addition, male adverse childhood events were positively associated with elevated MFPV risk among the couples in this study, a finding that is in accord with studies linking adult IPV perpetration and victimization with exposure to childhood violence (e.g., witnessing violence between one’s parents and experiencing childhood physical abuse) (Gover et al., in press; Heyman & Slep, 2002; Schafer et al., 2004; Whitfield et al., 2003) and other adverse experiences such as parental alcohol abuse. In fact, exposure to childhood adverse experiences and household dysfunction has been linked with numerous deleterious adult health and behavioral outcomes, including smoking (Anda et al., 1999), unintended pregnancy (Dietz et al., 1999), sexually transmitted disease (Hillis et al., 2000), alcohol abuse (Dube et al., 2002), and depression (Anda et al., 2002; Chapman et al., 2004). Furthermore, a dose-response relationship between adverse childhood experiences and comorbid conditions in adulthood has been demonstrated (Anda et al., 2006). From a public health standpoint, this suggests that significant resources should be devoted to primary, secondary, and tertiary prevention strategies (Felitti et al., 1998).

Neither male nor female problem drinking was associated with increased risk for FMPV among couples in this study. In an analysis of a national sample of couples, Cunradi et al. (2002b) found a significant cross-sectional relationship between female alcohol-related problems and past-year FMPV. In longitudinal follow-up with these couples, the female’s average weekly alcohol consumption was significantly associated with the recurrence of FMPV, and the male’s average weekly alcohol consumption was associated with incidence of FMPV (Caetano et al., 2005). Because FMPV has not received as much research attention as MFPV (Holtzworth-Munroe, 2005; Straus, 1999, 2005), less is known about the role of the male and female’s drinking patterns and alcohol problems in the occurrence of FMPV. A recent study among a sample of married/cohabiting female alcoholic patients and their non-substance abusing male partners found that at 1-year post-treatment follow-up, couples who had received behavioral couples therapy (BCT) plus individual-based treatment reported fewer drinking days, fewer drinking-related negative consequences, higher dyadic adjustment, and reduced partner violence compared to couples who underwent individual-based treatment only or psychoeducational attention control treatment (Fals-Stewart et al., 2006). Previous studies using data from male alcoholics have shown that BCT is effective in reducing drinking, improving couple functioning, and reducing IPV (Fals-Stewart et al., 2005). Couple-based interventions that address the substance use of both partners need to be developed and evaluated in order to reduce and prevent male- and female-perpetrated IPV.

Results showed that couples in which the male was on current layoff were not at elevated risk for FMPV compared to couples in which the male was not on current layoff. Months of male unemployment, however, was significantly associated with FMPV risk. Given the male’s traditional role as breadwinner, one explanation for these findings is that increasing months of the male’s unemployment produces a type of cumulative stress that results in conflict and aggression on the part of the female. Male unemployment was not found to be associated with FMPV among a national sample of couples (Cunradi et al., 1999). In longitudinal follow-up with the couples, however, male unemployment was found to be a significant predictor of FMPV recurrence and incidence (Caetano et al., 2005). Additional research is needed to investigate how and why male unemployment may contribute to risk for FMPV. Both models showed that risk for FMPV significantly decreases with couple age. This finding is consistent with other studies showing that younger couples are at greater risk for IPV compared to older couples (Cunradi et al., 2002b; DeMaris et al., 2003; Melzer, 2002; Sorenson et al., 1996). Male and female impulsivity were also significantly associated with FMPV risk in both models, a finding in accord with other research (e.g., Schafer et al., 2004). Finally, the finding that couples in which the male’s race/ethnicity was categorized as black were at elevated FMPV risk compared to couples in which the male’s race/ethnicity was categorized as ‘other’ is also consistent with prior research showing significant racial/ethnic differences in rates of FMPV among U.S. (Caetano et al., 2000).

A number of study limitations should be noted. First, the cross-sectional study design does not allow causal inferences to be made regarding the observed associations among study variables. A second 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. For example, 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. On the other hand, 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. In the current study, males had a higher non-response rate than females, and it is difficult to gauge what effect this pattern of non-response had on the study’s findings. A related concern is the potential reporting biases of survey participants in the context of an occupational health survey. For example, despite assurances of confidentiality, heavy drinkers may have underreported their alcohol consumption. This would result in an underestimation of consumption, and an observed distribution of consumption that is somewhat flatter that the actual distribution. Assuming a true association between alcohol and the outcome of interest, this underestimation would most likely lead to an attenuation of the association. Also, due to time constraints, participants were only asked questions from the Physical Assault subscale of the CTS2 (Straus et al., 1996); no information was gathered on psychological aggression, injury, or sexual assault. Lastly, respondent household income was not directly measured in the survey. Trends over the last two decades, however, indicate that inflation-adjusted wages of most workers in blue-collar and service industries are flat compared to those in white-collar and professional occupations (Norris, 2006). Based on data from the U.S. Bureau of Labor Statistics (2007), the median hourly wage of carpenters in California as of May 2007 was $23.77

Despite its limitations, this study helps illuminate how male unemployment and male and female problem drinking, as well as other psychosocial factors, contribute to the occurrence of both MFPV and FMPV among a sample of construction industry workers and their spouses/partners. While past research shows that lower SES couples are at higher risk for IPV, relatively few studies have investigated the correlates of IPV among specific blue-collar occupational groups (e.g., Cunradi et al., 2008; Leonard et al., 1985). Furthermore, to our knowledge, this is the first such study to separately model correlates of MFPV and FMPV among a sample of blue-collar couples in which the male workers are employed in an occupational category (construction) with high rates of heavy drinking compared to other occupational categories.