Divorcing Mothers’ Use of Protective Strategies: Differences Over Time and by Violence Experience

Abstract

Objective

The current study considered protective strategy use at various points in time for divorcing mothers with a range of marital violence experiences (including no history of violence and different types of violence).

Method

Divorcing mothers (N = 170) from one Midwestern county participated in two in-person interviews that included structured assessments of intimate partner violence (IPV) during the last year of marriage and use of protective strategies at three time points – last year of marriage and at separation (both assessed at Time 1) and since the first interview (at a three month follow-up).

Results

Divorcing mothers, regardless of marital violence experience, used an array of private and public protective strategies during the last year of marriage, at separation, and at the three-month follow up interview. In general, mothers who experienced coercive controlling violence reported using more protective strategies than mothers who experienced situational couple violence or no violence. Strategy use peaked at separation with an increase in public strategy use.

Conclusions

All divorcing mothers use a range of active coping strategies over time, yet findings emphasize the importance of making distinctions between types of IPV so that divorcing mothers receive individualized resources and support that best meet their needs and minimize potentials risks and losses as they navigate the divorce process.

Introduction

The literature on intimate partner violence (IPV) has shifted in the last few decades from viewing abused women as helpless victims to seeing them as active agents in their efforts to achieve nonviolence within their romantic relationships (Gondolf & Fisher, 1988). Numerous studies (Goodman, Dutton, Weinfurt, & Cook, 2003; Riddell, Ford-Gilboe, and Leipert, 2009) and survivor theory (Gondolf & Fisher, 1988) indicate that women who experience IPV draw from a wide variety of strategies to resist violence and protect themselves and (if they are mothers) their children. This literature has generated a rich understanding of the ways that abused women use protective or help-seeking strategies. Yet, a number of unanswered questions remain. One question is how women’s use of protective strategies changes across the course of relationship dissolution. Despite the recognition that violence and abuse do not always stop upon leaving the relationship (Brownridge, 2006; Fleury, Sullivan, & Bybee, 2000), researchers have has largely prioritized examining strategies for leaving an abusive relationship – neglecting women’s use of post-separation strategies. Another question is what role motherhood plays in strategy use over time. Children can serve as both a barrier and catalyst to help-seeking and leaving an abusive relationship (Bennet, Goodman, & Dutton, 1999; Davies, Ford-Gilboe, & Hammerton, 2009; Hardesty & Ganong, 2006). Further, risk of victimization after separation may be greater for women with children. As mothers, these women often have ongoing contact with abusive former partners through parenting-related responsibilities (e.g., visitation exchanges), leaving them accessible to their former partners (Davies et al., 2009; Hardesty & Ganong, 2006; Wuest, Ford-Gilboe, Merritt-Gray, & Berman, 2003) and in need of ongoing protective strategies. To address these core questions, we contribute to the literature by examining victimized and nonvictimized mothers’ protective strategy use during marriage and after separation.

We situate our study within Gondolf and Fisher’s (1988) survivor theory, as well as Hamby’s (2014) definition of protective strategies – any step taken to minimize the likelihood of experiencing a risk or loss, including but not limited to physical violence or abuse, financial risk (e.g., loss of income), social and personal risk (e.g., loss of friendships), or institutional and legal risk (e.g., loss of child custody). The study examines how a community sample of divorcing mothers with different experiences of marital IPV use protective strategies throughout the divorce process. Specifically, we measure protective strategy use with Goodman et al.’s (2003) Intimate Partner Violence Strategies Index (IPVSI) and include mothers who did not experience IPV in their marital relationship but, nevertheless, may utilize protective strategies both before and after marital separation to protect themselves and their children from potential risks and losses specific to the divorce process (e.g., loss of child custody). The inclusion of divorcing mothers who did not report experiencing marital IPV is a novel application of the IPVSI, but is well situated within Hamby’s (2014) definitions of protective strategies, risks, and losses.

Protective Strategy Use

Scholars have identified a range of strategies women use to cope with IPV. The IPVSI is a comprehensive measure that identifies six types of protective strategies that fall into two categories – private and public. Private strategies are carried out in isolation and include placating (e.g., doing whatever the partner wants), resisting (e.g., trying to avoid the partner), and safety planning (e.g., creating an escape plan). Placating and resisting were once considered evidence of helplessness, but are now understood to be active and deliberate protective strategies (Goodman et al., 2003). Public strategies are carried out through disclosure and interaction with social support networks and public agencies and include informal (e.g., family, friends, neighbors), formal (e.g., therapists, shelters), and legal help (e.g., police, attorney).

In studies using the IPVSI, abused women have reported using multiple strategies (e.g., an average of 24 out of 36 measured strategies; Goodkind, Sullivan, & Bybee, 2004) and the same strategy on multiple occasions (Bennett Cattaneo, Stuewig, Goodman, Kaltman, & Dutton, 2007). Although placating and resistance strategies are the most commonly used, women report them to be the least effective in stopping violence or preventing reabuse (Goodkind et al., 2004; Riddell et al., 2009). Safety planning strategies are often the least commonly used (Riddell et al., 2009), despite reportedly being one of the most helpful and effective strategies in stopping violence (Goodman et al., 2003; Riddell et al., 2009). Although abused women often seek help from informal, formal, and legal sources, many prefer private over public strategies (Goodman et al., 2003). In terms of public strategy use, abused women reportedly prefer and more commonly use informal versus formal help-seeking. Nonetheless, between 40 – 53% of abused women seek help from formal support networks at some point in their relationships (Coker, Derrick, Lumpkin, Aldrich, & Oldendick, 2000; Goodman et al., 2003). Goodman and Fels Smyth (2011) hypothesized that women may be less likely to seek formal than informal help because formal help may be seen as more extreme or less reversible than confiding in friends and family.

The types of strategies that abused women use often shift over time. For example, women may begin using private strategies, but may include public strategies as the violence worsens or as private strategies prove ineffective in preventing or minimizing further risk (Goodman et al., 2003; Liang, Goodman, Tummala-Narra, & Weintraub, 2005). Indeed, Lempert (1996) found that abused women only sought public help when they believed there was nothing more they could do on their own to stop the violence. Decisions to end the relationship also relate to changes in strategies, as safety planning and disclosing to support networks are primarily used to increase options or resources as women prepare to leave (Goodman et al., 2003). Safety planning strategies can be preceded by or used in conjunction with seeking public support to safely exit the relationship. It also is likely that informal social support enables women to further pursue formal support (see review by Goodman, Dutton, Vankos, & Weinfurt, 2005). Because mothering places women at an increased risk for ongoing violence and abuse post-separation (Davies et al., 2009), it is likely that mothers would report using a greater number of strategies than non-mothers following separation.

Despite a greater understanding of women’s use of protective strategies, limitations remain. Although research has shown that women’s use of strategies shifts over time, few studies have examined both private and public strategy use over time. Additionally, a substantial percentage of women in the reviewed studies share children with their abusive partner (e.g., 41% Goodman et al., 2003); the presence of children has been identified as a key contextual factor in understanding abused women’s risk, especially after separation (Davies et al., 2009; Zanville & Bennett Cattaneo, 2012). Yet, little attention has been paid to how mothers use protective strategies both during marriage and separation. The protective strategy literature is further complicated by studies that document women’s differential use of protective strategies based on characteristics of the physical violence they experience (e.g., severity; Ansara & Hindin, 2010; Coker et al., 2000; Goodkind et al., 2004; Goodman et al., 2003) with no consideration of the type or context of the violence (e.g., coercive controlling versus situational couple violence; Goodman et al., 2003). The current literature also lacks the use of comparison groups of nonabused mothers. Thus, we are unsure how strategy use is a part of the broader experience of relationship dissolution versus unique to the experience of violence.

Different Types of IPV and Protective Strategy Use

Theory and empirical evidence identify two main types of IPV – coercive controlling violence (CCV) and situational couple violence (SCV) – differentiated by the context within which physical violence occurs (i.e., coercive control vs. violence that arises from escalating conflict; Johnson, 2008). To our knowledge, only one study has examined women’s (81% were mothers) differential use of protective strategies based on experiencing CCV versus SCV (Leone, Johnson, & Cohan, 2007). They found that CCV was associated with greater use of all types of protective strategies with the exception of informal help-seeking. Other studies (Ansara & Hindi, 2010; Goodkind et al., 2004; Goodman et al., 2003; Zanville & Bennett Cattaneo, 2012) have documented similar findings without explicitly classifying violence as CCV or SCV. These studies classified women as experiencing infrequent and minor IPV with no or low levels of injury as well as little psychological abuse (likely aligning with SCV) versus more frequent, moderate to severe, and injurious IPV as well as psychological abuse (likely aligning with CCV). The latter group of women in these studies reported greater use of all types of protective strategies except informal help-seeking. Each of these studies examined women’s use of protective strategies during the relationship or retrospectively (i.e., “in the past twelve months”) after leaving the relationship (e.g., filing for a protective order).

Whether these patterns hold true over time (e.g., during relationship, at separation, and post-separation) is unknown. Longitudinal studies that consider these relational conditions would shed light on when and why women use strategies over time. Additionally, most studies have utilized agency samples (e.g., domestic violence shelters, criminal courts) that largely capture the experiences of CCV, whereas SCV is more commonly identified in community or nationally representative samples (Johnson, 2008). Finally, none of these studies include nonabused women for comparison purposes. Thus, the existing literature does not reflect the full range of protective strategy use that is expected in community samples of separating mothers with different IPV experiences.

Present Study

The current study sought to address several methodological and theoretical gaps in the protective strategy literature. First, although many studies have documented the salient role that children play in whether or not mothers stay or leave abusive relationships, we are unaware of any studies that emphasize mothers’ use of protective strategies across time. Second, current understanding of strategy use draws primarily on studies that collected retrospective data at one time point (see Sylaska & Edwards, 2014 for review). Third, to our knowledge, no study to date has applied Johnson’s typology – coercive controlling and situational couple violence – or included nonabused women to examine both public and private protective strategy use among women in a community sample. Thus, our longitudinal, descriptive study of a community sample of divorcing mothers with a range of IPV experiences was designed to address these gaps with the following research questions and hypotheses:

Research Question 1: What protective strategies do divorcing mothers with different experiences of IPV use during the last year of marriage?

Consistent with survivor theory (Gondolf & Fisher, 1988) and Hamby’s (2014) definition of protective strategies, risks, and losses, we hypothesize that divorcing mothers, regardless of IPV victimization or type, will utilize an array of protective strategies to protect themselves and their children from physical violence, non-violent abuse tactics, and potential risks and losses inherent in the separation and divorce process. Survivor theory posits that women increase their resistance and strategy use as violence escalates (Gondolf & Fisher, 1988) and women with greatest risk (e.g., more physical and psychological abuse) engage in more strategy use (Goodkind et al., 2004; Goodman et al., 2003; Leone et al., 2007). Thus, based on the reviewed literature, we hypothesize that divorcing mothers who experience CCV will report a greater number of overall, private, and public protective strategies than mothers who experience SCV, and that mothers who experience SCV will report using more strategies than nonabused mothers.

Research Question 2: How does the use of protective strategies differ during the last year of marriage, at separation, and after separation (at a three month follow-up) among mothers with different experiences of IPV?

Consistent with the previous literature that has examined protective strategy use over time (Goodman et al., 2003; Liang et al., 2005), we hypothesize that all of the divorcing mothers will continue using private strategies following separation. Additionally, we hypothesize that mothers’ public strategy use will increase at the time of separation (increasing total strategy use) as these strategies are more typically used as women begin to leave relationships (Goodman et al., 2003). We expect that our hypotheses from research question one regarding mothers’ differential use of protective strategies depending on type of IPV will hold constant over time as CCV likely yields greater risks and potential losses following separation than SCV and no violence, and SCV likely yields greater risks and potential losses than no violence.

Methods

Sources of Data

Data were from two longitudinal projects on the experiences of mothers in the process of divorce: the Co-Parenting Project (CoP; conducted in 2009–2010) and the Mothers’ and Kids’ Experiences in Transition Project (MAKE IT; 2010–2014). Both studies involved the same inclusion criteria, recruitment methods, and assessments. Mothers completed in-person interviews (two in the first study, five in the second study) at three-month intervals. The current study utilizes data from the first two time points of both projects. At Time 1, mothers completed measures of their experiences during the last year of marriage and at separation; at the follow-up interview three months later they reported on their experiences since the first interview.

Participants

Women named in a divorce filing within the past 12 weeks were identified via public court records in a large Midwest county. Recruitment letters were mailed to 804 women named in a divorce filing between September 2009 and December 2012 inviting them to participate in a study of mothers’ experiences with divorce. When addresses were not available, the attorney listed in the file was asked to pass on a letter to the potential participant. Follow-up phone calls were made to potential participants with known phone numbers. Inclusion criteria for the study included: (a) had at least one child under age 18; (b) had custody of their child(ren) at least 25% of the time; (c) had been physically separated from their former partner for less than three years; and (d) could understand and speak English. Of the 804 women, 190 (23.6%) met eligibility criteria and completed the first interview; 118 (14.7%) were successfully contacted but did not take part (5 began the study but were subsequently determined to be ineligible, 50 were eligible and interested but did not show up for scheduled interviews or did not respond to scheduling efforts, 36 declined to participate, and 27 did not meet criteria); and 496 (61.7%) never responded to recruitment efforts. The analytic sample for the current study consists of 170 mothers who completed the first two interviews; 51 were from CoP and 119 from MAKE IT.

At the first interview, mothers in the analytic sample (N = 170) were 20.83 to 54.67 years old (M = 35.81, SD = 7.34). When asked to report their race/ethnicity, most indicated they were White (n = 139, 81.8%); 20 (11.8%) self-identified as Black or African American, 5 (2.9%) as Asian or Asian American, 5 (2.9%) as biracial, and 1 (0.6%) as Latino/Hispanic. Mothers had between one and four biological or adopted children with their former partner (M = 1.75, SD = 0.75) and had been married an average of 9.77 years (SD = 5.91; range = 0.17 to 27.42 years). Mothers had been physically separated from their former partner an average of 8.17 months (range = 1 – 34 months, SD = 6.90) and under one-fifth (17.6%) of divorces had been finalized. The majority of mothers were employed full time (n = 101, 59.4%); 32 (18.8%) were employed part time, 19 (11.2%) were involuntarily unemployed (e.g., due to disability, difficulty finding a job, or being laid off), and 18 (10.6%) were voluntarily unemployed (e.g., student). Twenty-one mothers had a high school degree or less (12.3%), 61 (35.9%) had some college, 60 (35.3%) had Bachelor’s degree, and 28 (16.6%) had a Master’s or Doctoral degree. Participants in the two studies were comparable on the above set of demographics with one exception: at the first interview, a higher percentage of mothers in CoP reported being involuntarily unemployed (vs. employed full-time) than those in MAKE IT, X² (3, N = 170) = 8.13, p = 0.043.

Procedures

Interviews consisted of a structured questionnaire administered by trained interviewers who read instructions and questions out loud, and wrote down responses. Interviews lasted between 60 – 90 minutes. IRB approval and Federal Certificates of Confidentiality were obtained for both studies, and numerous precautions were taken to ensure mothers’ safety and privacy. Participants’ rights and potential risks were discussed in the informed consent process, including limits of confidentiality (i.e., mandatory child abuse reporting laws). Study safety protocol and procedures were guided by Parker and Ulrich (1990). Most interviews took place in public locations (e.g., private room in library). Some interviews were conducted at the respondent’s home; in these cases, two researchers went to the interview. Mothers in CoP received a stipend of $30 for the first and $35 for the second interview, whereas mothers in MAKE IT received a stipend of $35 for the first and $40 for the second interview. All mothers received a list of resources related to divorce and violence.

Measures

Demographic variables

At Time 1, mothers reported their age, race/ethnicity, highest level of education, employment status, number of children with their former partner, length of their marriage, the amount of time since separation, and whether their divorce had been finalized.

Type of intimate partner violence during marriage

Mothers were classified into three mutually exclusive IPV groups based on responses to measures of physical violence and coercive control during marriage (assessed at Time 1). Physical violence was measured using the physical assault subscale of the Revised Conflict Tactics Scale (CTS2; Straus, Hamby, Boney-McCoy, & Sugarman, 1996) and Goodman et al.’s (2003) modified version of the sexual coercion subscale of the CTS2. The CTS2 is widely used in violence research and has been established as a valid and reliable measure (Straus et al., 1996). The physical assault subscale includes 11 items that address direct physical violence (e.g., “He pushed or shoved me”; “He slammed me against a wall”). One item that did not assess direct physical assault was administered but not used to identify women had experienced violence (i.e., “Threw something at you that could hurt you”). Two items from the sexual coercion subscale that assessed the use of force were also included in the analyses (e.g., “He used force to make me have sex”). Mothers reported whether each item occurred during their marriage. Ninety respondents (52.9%) reported at least one act of physical violence during their marriage; 80 (47.1%) reported no violence. (Frequency and severity of violence were also assessed, but were not used in the current analyses.)

Coercive control was measured using the Dominance-Isolation Subscale of the Psychological Maltreatment of Women Inventory (PMWI) – Short Form (Tolman, 1992). Respondents rated how often (1 = never to 5 = always) they experienced seven acts (e.g., “He monitored my time and made me account for my whereabouts”) during the last year of marriage. Responses were summed to calculate the frequency with which control tactics were used (M = 15.81, SD = 7.29, range 7–35). The PMWI is widely used to measure coercive control (Johnson, 2008) and has good discriminant validity and reliability (Bell, Goodman, & Dutton, 2007); alpha in the current study was 0.86.

Using a previously established classification method (Hardesty, Crossman, Haselschwerdt, Raffaelli, Ogolsky, & Johnson, 2015), mothers were assigned to one of three groups: 1) no violence (NV; no act of violence reported on the CTS2, any score on the PMWI); 2) situational couple violence (SCV: at least 1 act of violence as reported on the CTS2, a score of 18 or less on the PMWI); and 3) coercive controlling violence (CCV; at least 1 act of violence as reported on the CTS2, a score of 19 or greater on the PMWI).

Protective strategies

Protective strategies were measured using the IPV Strategies Index (IPVSI; Goodman et al., 2003). The original version of the IPVSI has 39 items designed to capture the range of protective strategies women use in response to their partner’s behavior. Although this scale was designed for women who experience IPV, the scale includes items that women who do not experience violence may use to cope with separation/divorce or conflict with their partner (e.g., “tried to avoid an argument with him”). The IPVSI has shown good ecological and convergent validity (Goodman et al., 2003). For the current study, eight items not relevant to physically separated women were removed (e.g., “tried to keep things quiet for him”) and one item was revised from “Ended (or tried to end) relationship” to “Ended (or tried to end) contact with him.” The 31 items administered in the current study are listed in Table 1. The IPVSI yields an overall count of protective strategies women used as well as the number of public and private strategies used in response to their (former) partner’s behavior. Strategy use in response to their (former) partners’ behavior was assessed during three time periods: the last year of marriage, at separation (both assessed at Time 1), and since the first interview (at follow-up). Private strategies (16 items) are placating (4 items), resisting (6 items) and safety planning (6 items); public strategies (15 items) are use of informal help (4 items), use of formal help (7 items), and use of legal resources (4 items; see Table 1 for the full list of items in each subcategory). Scores for overall number of protective strategies, and for private vs. public strategy use, were computed by summing. Summary scores for the six subtypes of strategies (placating, resisting, safety planning, informal, formal, and legal help) were also computed.

Table 1.

Percent Endorsement of Protective Strategies by IPV Group

Protective Strategies – Last Year Marriage	NV (80)	SCV (48)	CCV (42)	X2 / p value
Placating
Did whatever he wanted to stop the violence	7.5a	21.3b	54.8c	76.35 / .000
Tried not to cry during the violence	7.5a	31.9b	59.5c	73.11 / .000
Tried to avoid him	50.0a	62.5a,b	73.8b	6.74 / .03
Tried to avoid an argument with him	63.7a	95.8b	95.2b	27.39 / .000
Resisting
Fought back physically	2.5a	22.9b	50.0c	43.06 / .000
Refused to do what he said	35.0a	68.8b	71.4b	20.92 / .000
Used or threatened to use a weapon against him	1.3a	4.2a,b	9.5b	4.78 / .09
Left home to get away from him	42.5a	64.6b	64.3b	8.22 / .02
Ended or tried to end contact with him	13.8a	31.3b	40.5b	12.51 / .01
Fought back verbally	63.7a	93.6b	88.1b	18.71 / .000
Safety Planning
Developed a code so others would know you were in danger	0a	6.3b	14.3b	11.33 / .003
Removed/hid weapons	3.8	6.3	4.8	1.28 / .87
Kept important phone numbers you could use to get help	11.3a	18.8a,b	35.7 b	10.56 / .005
Hid important papers from him	20.0a	20.8a	50.0b	13.94 / .001
Put knife, gun, or other weapon where you could get to it	1.3a	0a	14.3b	14.73 / .001
Changed locks/somehow improved security	2.5	2.1	7.1	3.22 / .52
Use of Informal Help
Talked to family/friends about what to do to protect yourself/children	21.3a	39.6b	42.9b	7.82 / .02
Stayed with family or friends	16.3a	25.0a,b	42.9b	10.32 / .01
Sent kids to stay with family or friends	12.5a	14.6a	38.1b	16.09 / .003
Made sure there were always other people (adults) around	11.3a	14.6a	40.5b	16.04 / .000
Use of Formal Help
Tried to get help from clergy	18.8a	43.8b	31.0a,b	9.26 / .01
Tried to get help from employer/coworker	31.3	39.6	35.7	2.39 / .66
Talked to a doctor or nurse	30.0	39.6	35.7	1.29 / .53
Called mental health counselor for yourself	41.3	27.1	38.1	2.66 / .26
Stayed in a shelter	0	4.2	0	5.14 / .08
Talked to someone at a DV program, shelter or hotline	1.3a	10.4b	19.0b	11.97 / .003
Tried to get help for alcohol/substance abuse	1.3	0	0	3.68 / .45
Use of Legal Resources
Filed for an order of protection	2.5	0	4.8	2.23 / .33
Filed/tried to file criminal charges	0a	2.1a,b	7.1b	6.14 / .05
Sought help from legal aid	11.3	14.6	7.1	1.25 / .54
Called the police	3.8a	12.5a	33.3b	20.66 / .000

Note: NV = no intimate partner violence; SCV = situational couple violence; CCV = coercive controlling violence. Within each protective strategy, percentages with different superscripts differ significantly from each other (p < .05).

Analyses

Preliminary analyses were conducted to examine whether demographic factors potentially associated with marital violence differed across the three IPV groups (e.g., age, race, education, employment status, number of children, length of marriage, time since separation, and whether the divorce was final). Chi-square tests of association were used for categorical variables and ANOVAs for continuous variables. These tests revealed two significant differences (in age and education); thus, these variables were controlled in the main analyses.

Univariate and multivariate analyses were conducted to examine the research questions. For descriptive purposes, we examined whether endorsement of each specific protective strategy differed by IPV group (using chi-square tests of association). Following this, mean differences in protective strategies attributable to time and IPV group were examined by conducting two repeated-measures analyses of variance (ANOVAs), one for overall strategies and one for private versus public strategies. In each analysis, protective strategy use across time (last year of marriage, at separation, and at follow-up) was the repeated variable, IPV group (NV, SCV, CCV) was the between-groups measure, and age (in years) and highest level of education (dichotomized as 0 = less than a bachelor’s degree, 1 = 4 year degree and higher) were entered as statistical controls (covariates). Significant interactions were followed up by one-way ANOVAs with pairwise comparisons to pinpoint differences between IPV groups. To describe patterns of protective strategy use over time, one-way ANOVAs were conducted for the six protective strategy subscales. To aid in interpretation of results, exact p-levels are reported (except for pairwise comparisons presented in Table 2).

Table 2.

Means and Standard Deviations of Protective Strategies by Time Period and IPV Group

	During last year of marriage (assessed at T1)				At separation (assessed at T1)				Since first interview (assessed at follow-up)
	NV n = 80	SCV n = 48	CCV n = 42	Group Comparison	NV n = 80	SCV n = 48	CCV n = 42	Group Comparison	NV n = 80	SCV n = 48	CCV n = 42	Group Comparison
Private	3.26 a (2.51)	5.48 b (2.48)	7.33 c (3.15)	F = 33.50, p < .000	3.63 a (2.74)	5.65 ab (2.63)	6.86 b (2.79)	F = 21.34, p < .000	1.89 a (2.04)	3.50 b (3.14)	4.12 b (3.08)	F = 11.42, p < .001
Placating	1.29 a (1.09)	2.10 b (1.04)	2.83 c (1.08)	F = 29.72, p < .000	1.36 a (1.01)	1.96 b (.82)	2.19 b (0.94)	F = 12.45, p < .000	0.81a (0.83)	1.25 b (0.96)	1.21 b (0.87)	F = 4.90, p = .009
Resistance	1.59 a (1.31)	2.83 b (1.33)	3.24 b (1.61)	F = 23.44, p < .000	1.58 a (1.45)	2.65 b (1.36)	2.98 b (0.98)	F = 18.91, p < .000	0.71 a (0.98)	1.40 b (1.51)	1.62 b (1.25)	F = 9.23, p = .000
Safety planning	0.39 a (0.74)	0.54 a (0.85)	1.26 b (1.43)	F = 11.30, p < .000	0.69 a (0.85)	1.04 a (1.17)	1.69 b (1.44)	F = 11.28, p < .000	0.36 a (0.70)	0.85 ab (1.35)	1.29 b (1.71)	F = 8.35, p = .000
Public	2.03 a (2.07)	2.88 ab (2.35)	3.76 b (2.38)	F = 8.57, p < .000	2.81 a (2.44)	4.04 ab (2.74)	5.14 b (2.83)	F = 11.26, p < .000	1.01 a (1.40)	2.54 b (2.63)	2.64 b (2.56)	F = 11.77, p < .000
Informal	0.61 a (1.02)	0.94 a (1.14)	1.64 b (1.25)	F = 11.78, p < .000	0.93 a (1.25)	1.56 ab (1.38)	2.10 b (1.32)	F = 11.61, p < .000	0.25 a (0.67)	0.81 b (1.16)	1.10 b (1.27)	F = 11.32, p < .000
Formal	1.24 (1.17)	1.65 (1.34)	1.60 (1.42)	F = 1.92, p = .150	1.41 (1.32)	1.83 (1.39)	1.64 (1.28)	F = 1.55, p = .216	0.56 a (0.84)	1.17 b (1.24)	0.93 ab (1.22)	F = 5.11, p = .007
Legal	0.18 a (0.50)	0.29 ab (0.50)	0.52 b (0.74)	F = 5.19, p = .007	0.48 a (0.71)	0.65 a (0.79)	1.40 b (1.27)	F = 15.19, p < .000	0.20 a (0.46)	0.56 b (0.87)	0.62 b (0.88)	F = 6.40, p = .002
Total strategies	5.29 a (4.08)	8.35 b (4.22)	11.10 c (4.73)	F = 26.38, p < .001	6.44 a (4.69)	9.69 b (5.04)	12.00 b (4.87)	F = 19.55, p < .001	2.90 a (3.14)	6.04 b (5.42)	6.76 b (5.37)	F = 13.09, p < .001

Note. NV = no intimate partner violence; SCV = situational couple violence; CCV = coercive controlling violence. Values are Means (Standard Deviations). Subscale range is 0 – 4 for informal, legal, and placating; 0 – 6 for safety and resistance; 0 – 7 for formal. Total strategies (0–31) is the sum of the subscales.

^a,b,cWithin each measurement period, means with different superscripts differ significantly from each other (p < .05).

Results

Protective Strategy Use During Marriage

The first research question asked what protective strategies mothers with different experiences of IPV used during the last year of marriage. Table 1 displays percent endorsement of the 31 protective strategies for mothers with different experiences of IPV, together with results of group comparisons based on chi-square analyses. Group differences were seen on 21 of the 31 comparisons. The two violence groups (particularly CCV) had the highest endorsement of most protective strategies. Nevertheless, many mothers in the no violence group also reported using protective strategies during the last year of marriage. We noted that several mothers in the no violence group said they used placating strategies in response to their partner’s violence (first two items in Table 1). Post-hoc analyses showed that these mothers had either reported verbal abuse during the interview or had endorsed the CTS2 item that did not assess direct physical assault.

Table 2 displays descriptive statistics for summary indicators of protective strategies by group and time period. Consistent with the item-level analyses, the three IPV groups differed in number of total strategies used (bottom row) during the last year of marriage: mothers who experienced CCV reported the highest level of protective strategy use (M = 11.10), followed by those who experienced SCV (M = 8.35), then the NV group (M = 5.29). A similar pattern was seen for private strategy use. For public strategy use, the CCV group was significantly higher than the NV group, but SCV did not differ from the other groups. Differences were also observed on all subscales except formal help (see Table 2).

Protective Strategy Use Over Time

The second research question asked how the use of protective strategies differed over time among mothers with different experiences of violence.

Overall use of protective strategies

The multivariate analysis revealed a main effect for time (F = 6.59, df = 2, p = .002); within-subjects contrasts indicated both a linear (F = 6.07, df = 1, p = .015) and a quadratic (F = 7.70, df = 1, p = .006) effect for time. In the overall sample, the number of protective strategies was lower during the last year of marriage then at separation, then declined by the time of the follow-up interview (Ms = 7.59, 8.73, 4.74, respectively). There was also a between-subjects effect for violence type (F = 27.64, df = 2, p < .001). As reported above, the three IPV groups differed significantly in protective strategy use during the last year of marriage; the two violence groups also reported using more protective strategies than the NV group at separation and at follow-up (Table 2, last row). The two-way interaction of time by IPV group was not significant.

Private versus public protective strategy use

There was a main effect for strategy type (F = 10.16, df = 1, p = .002), with mothers reporting more private than public strategies at all three time periods. Means for private versus public protective strategies use were 4.89 versus 2.69 during last year of marriage, 4.99 versus 3.74 at separation, and 2.89 versus 1.85 at follow-up. The previously reported main effect for time also emerged (F = 6.69, df = 2, p = .001).

Analyses also revealed two-way interactions for time X strategy type (F = 4.56, df = 2, p = .012), strategy type X IPV group (F = 6.48, df = 2, p = .002), and a 3-way interaction for time X strategy type X IPV group (F = 3.04, df = 4, p = .017). These results indicated that use of protective strategies showed different patterns of change over time depending on strategy type (private vs. public) and IPV group. Means are displayed in Table 2, and the pattern of results depicted in Figures 1 and 2.

Figure 1.

Mean Number of Private Protective Strategies by Time and IPV Group

Figure 2.

Mean Number of Public Protective Strategies by Time and IPV Group

There was a general linear trend for private strategy use (F = 10.78, df = 1, p = .001), and a weaker quadratic trend (F = 4.48, df = 1, p = .036), indicating an overall drop in use of private strategies over time. As shown in Figure 1, this pattern was evident for all three IPV groups. There were group differences in use of private protective strategies during the last year of marriage, with mothers in the CCV group using significantly more private strategies than the other two groups, and the SCV group using more than the NV group. At separation, the CCV group reported using significantly more private protective strategies than the NV group, but the SCV group did not differ from the other two groups. At follow-up, the two violence groups did not differ from each other, and both were higher than the NV group.

A different pattern was seen for public strategies (Figure 2), which showed only a quadratic trend, peaking at separation and then declining (F = 7.25, df = 1, p = .008). Mothers in the CCV group reported significantly higher levels of public strategy use than the NV group at all-time points. The SCV group did not differ from the other two groups during the last year of marriage and at separation and was similar to the CCV group at follow-up.

Protective strategies subscales

Results of descriptive analyses to assess mothers’ use of specific protective strategies over time are also displayed in Table 2. The overall pattern of results was generally consistent with the analysis of aggregated private and public strategies reported earlier. Where significant group differences were observed, mothers in the CCV group used more protective strategies than those in the NV group with one exception (formal strategies at the follow-up interview). During the last year of marriage and at separation, mothers in the SCV group were sometimes similar to the NV mothers and sometimes to the CCV mothers. At the time of the follow-up interview, the SCV and CCV groups did not differ on any subscales (consistent with the overall patterns depicted in Figures 1 and 2).

Discussion

The present study builds upon the previous protective strategy literature in several important ways. First, to our knowledge, our study is the first to make distinctions between public and private strategy use among mothers with a variety of IPV experiences (including no violence). By applying Hamby’s (2014) holistic definition of protective strategies, our findings acknowledge the multiple ways that divorcing mothers manage potential losses and risks, give credence to the numerous ways in which victimized mothers actively resist IPV, and decenter physical violence as the main or only risk that mothers must navigate (Gondolf & Fisher, 1988; Goodman et al., 2003; Hamby, 2014). Finally, current understanding of strategy use draws primarily on studies that collected data at one time point and summed across the relationship (see Sylaska & Edwards, 2014 for review). We contribute to Liang et al.’s (2005) call for research that examines patterns of help-seeking over time.

Divorcing Mothers as Active Agents During Marriage

Our first research question asked what strategies divorcing mothers with different IPV experiences used during the last year of marriage. Consistent with our first hypothesis, all mothers actively utilized a range of protective strategies – both private and private – during the year prior to separation. This finding highlights the applicability of survivor theory (Gondolf & Fisher, 1988) and Hamby’s (2014) holistic definitions of protective strategies, risks and losses for all women, regardless of IPV type; however, it appears that mothers experiencing CCV are at greater risk and perceive greater potential losses, and, thus, use more overall, private, and public protective strategies during marriage. These findings also contribute to the relationship dissolution literature in general (i.e., not IPV specific) by identifying strategies nonabused, divorcing mothers use; this literature is mostly centered around support processes, such as perception of and satisfaction with available social support, often excluding the specific strategies individuals use during dissolution (Sarason & Sarason, 2006).

Our second hypothesis was mostly supported – CCV was associated with a greater number of overall, private, and public strategies during marriage than SCV or no violence, and SCV was associated with more overall and private strategies than no violence. This is consistent with previous literature that found greater protective strategy use among women who experienced CCV (Leone et al., 2007). Nonetheless, the mothers in our sample who experienced no violence utilized a substantial number of strategies, specifically informal and formal help-seeking during the last year of marriage, which is indicative of perceived potential risks and losses associated with their relationship. Strategy use during marriage became more complex when we examined use of specific protective strategies. Consistent with previous literature, mothers in the CCV group were significantly more likely to utilize three of the six subcategories of strategies in comparison to the SCV group. We did not identify any differences in formal or legal strategy use, which is in contrast to previous research that documented a significant, positive association between CCV or proxies of CCV (e.g., greater physical and psychological abuse) and greater formal and legal help-seeking (Ansara & Hindi, 2010; Goodkind et al., 2004; Leone et al., 2007; Zanville & Bennett Cattaneo, 2012). However, our findings become more consistent with previous literature when we examined specific formal help-seeking items. For example, across all violence groups, items specific to IPV shelters were less likely to be endorsed in comparison to items such as getting help from clergy, calling a mental health counselor, or talking to a doctor or nurse. Only a small percentage of victimized women report seeking help from IPV shelters (Coker et al., 2000; Hamby & Bible, 2009). Therefore, the relatively low percentage of mothers in our sample who sought out IPV shelter (only 4.2% endorsed in SCV group, none in CCV group) or agency help (19% for CCV, 10.4% for SCV) is not surprising. The mothers in our study reported seeking a moderate amount of professional support (e.g., mental health counselor, clergy), which is also consistent with previous literature (Ansara & Hindin, 2010; Coker et al., 2000), especially studies that emphasize mothers’ use of formal support (Letourneau, Young Morris, Stewart, Hughes, Critchley, & Secco, 2013).

Although mothers who experienced CCV during marriage utilized a greater number of private strategies than mothers who experienced SCV, there was no significant difference in their use of resistance strategies. These findings are partially consistent with the current literature. To our knowledge, no study has examined differences in private strategy use for women experiencing CCV versus SCV, but Goodman et al. (2003) and Goodkind et al. (2004) examined private strategy use and proxies of CCV versus SCV and reported inconsistent findings. For example, Goodman et al. (2003) reported that use of all strategy types increased as severity of IPV increased. Goodkind et al. (2004) found the same association when examining overall, private, and public strategies, but resistance strategies were not associated with severity and frequency of physical and psychological abuse, which is consistent with our findings. Thus, our findings highlight the importance of measuring overall, private, and public strategies as well as subcategories to best understand nuances of divorcing mothers’ strategy use.

The Peaks and Declines of Protective Strategy Use After Separation

Our second research question focused on changes in protective strategy use over time. We hypothesized that all divorcing mothers would continue using private strategies following separation but that their use of public strategies would increase after separation, increasing total use of strategies over time. We further posited that mothers who experienced CCV would continue using the most protective strategies, followed by mothers experiencing SCV and no violence. Both hypotheses were partially substantiated, and, similar to our first research question, the findings varied depending on whether the outcome variable was overall strategies, public versus private strategies, or the six subcategories of strategies.

As hypothesized, mothers’ overall protective strategy use peaked after separation due to an increased use of public strategies. Nonetheless, mothers continued using more private versus public strategies across all time periods, which is consistent with previous research. Goodman et al. (2003) and Liang et al. (2005) document that the addition of public strategies does not decrease private strategy use; rather, both types of strategies are used in conjunction. The peak in public strategy use for all mothers in our sample could be explained by the nature of divorce, which may elevate private couple conflict or IPV to a more public platform. For example, mothers may confide in and rely on more formal and informal support to navigate visitation exchanges or arrange childcare to attend court hearings. For mothers who had experienced marital IPV, the decline in public strategy use may be explained by the cyclical nature of IPV, particularly coercive control (Johnson, 2008). Therefore, mothers’ public help-seeking may decrease over time if they perceive the IPV to have stopped when the marital relationship ended.

In contrast to the peak of public strategy use, private strategy use gradually declined over time. Private strategies, such as placating and resisting, are reportedly the least successful in stopping violence or preventing abuse, despite being the most commonly used strategies (Goodkind et al., 2004; Riddell et al., 2009). Thus, private strategy use among abused mothers perhaps decreased over time as they began to view these strategies as less effective in response to their former partners’ behavior. It is also possible that private strategy use decreased between separation and the follow-up interview because mothers had less physical contact with their former partners than they did during marriage and the earlier stages of separation. Research has shown that mothers – especially those who experienced CCV – strategically set rigid boundaries (e.g., supervised visitation exchanges, not responding to contact attempts) with abusive former husbands to minimize intrusion following separation (author citation).

Mothers in the CCV group reported greater use of strategies than the SCV and no violence groups during marriage and at separation, but the differences between the CCV and SCV groups became nonsignificant by the three month follow-up. Mothers in the no violence group still reported using protective strategies at follow-up, although in fewer numbers; this suggests that mothers in the process of divorce are active in seeking help for themselves and their children. Mothers reporting either CCV or SCV continued using more strategies than the no violence group across all time points, which highlights that divorcing mothers who experience IPV are in need of additional supports and resources throughout the divorce process – not just at the time of separation when risk is perceived to be greatest (Campbell, Sharps, & Glass, 2001).

Limitations

Although our findings contribute to the IPV and protective strategies literature, several limitations should be taken into consideration. First, we do not know the specific context in which mothers utilized the various protective strategies. The IPVSI instructions asked mothers to indicate which items they had done in response to their former partner’s behavior. Thus, it is unknown whether the mothers were responding to IPV behaviors or other behaviors such as conflict or alcohol abuse. However, two of the individual strategy items in the IPVSI asked about violence or abuse directly (e.g., “tried not to cry during the violence;” “did whatever he wanted to stop the violence”), which likely prompted responses that were IPV specific. Second, we asked about marital IPV and protective strategies retrospectively – in the last year of marriage and at the time of separation at Time 1, and since the first interview at follow-up – so there is potential for recall bias. However, many studies (e.g., Goodkind et al., 2004) asked participants to recall strategy use over the course of an entire relationship, so our assessment of shorter time periods likely reduced potential recall bias. A related issue is that by design, there were different reporting intervals for the IPVSI at each time period, which may have contributed to variations in numbers of strategies used over time (i.e., a longer reporting period might allow for more strategy use). Future research should consider using a standard reporting period to assess patterns over time. Despite these limitations, our novel application of the IPVSI with a divorcing sample of mothers, measurement of strategies at multiple time points, and comparisons between types of IPV and strategy use address several gaps in the current literature, and, thus, have important research, clinical, and policy implications.

Research Implications

The findings from this study indicate that protective strategy use changes over time as relationships dissolve and that divorcing mothers typically use protective strategies in some form, regardless of experience with IPV. This study also demonstrates that for those who do experience violence, strategy use is affected by violence type. Previous research has documented that women’s public strategy use varies depending on location (e.g., rural, urban; Shannon, Logan, Cole, & Medley, 2006); accessibility, familiarity, and positive (versus negative) experiences with support networks (Shannon et al., 2006); cultural factors (Goodman et al., 2003; Shannon et al., 2006); couple dynamics and severity of IPV (Leone et al., 2007; Riddell et al., 2009; Zanville & Bennett Cattaneo, 2012); and presence or absence of children in the home (Zanville & Bennett Cattaneo, 2012). Our findings contribute to this growing body of literature by documenting that both public and private strategy use varies depending on the mothers’ experiences with IPV and coercive control as well as their marital relationship status. Although, in general, mothers experiencing CCV utilized more protective strategies over time, our findings based on violence type differed depending on whether we analyzed overall strategy use, private versus public strategy use, or strategy use within the six subcategories of the IPVSI. Results point to the importance of examining the complexity in women’s strategy use over time.

Consistent with Liang et al.’s (2005) recommendations, we suggest future research should examine women’s motives for using particular strategies over time, as well as their perception of each strategies’ effectiveness, in order to better understand the process of strategy use for women with different IPV experiences. We also suggest future research simultaneously assess other IPV experiences (e.g., stalking and harassment after separation) that mothers may experience. In addition, future research should compare the strategy use of mothers versus non-mothers to better understand the role of motherhood on type and timing of strategy use.

Clinical and Policy Implications

The findings provide substantial evidence to support the application of Hamby’s (2014) holistic conceptualization of protective strategies and women’s potential risks and losses during the divorce process, which has direct clinical and policy implications. Despite the decrease in overall strategy use at the last time point, mothers who experienced IPV were utilizing more strategies than mothers experiencing no violence. This points to the need for providers and policymakers to recognize that mothers, especially those leaving abusive relationships, will need ongoing support following dissolution. Given the range of protective strategies used by mothers in our study, the findings supportGoodman and Fels Smyth’s (2011) call for a social network-oriented approach to mothers experiencing IPV. We posit that this approach may also be beneficial for divorcing mothers without IPV who sought help from various support networks both during marriage and following separation. In addition, mothers utilized a range of private strategies, some of which (i.e., safety planning) may have been learned from IPV agencies, highlighting the multifaceted nature of protective strategy use. Thus, the findings provide additional credence for the importance of continued funding for advocacy centers such as “Family Justice Centers,” which are multidisciplinary, one-stop-shops for the many services that abused women need while working towards nonviolence or exiting an abusive relationship (Family Justice Center Alliance, 2014). Such organizations are aligned with Hamby’s (2014) definition of protective strategies, risks, and losses, as they attend to a range of needs, which decenters physical violence as the main or only reason women seek help. Given the presence of children, specialized centers and individual practitioners should work with mothers to identify specific resources unique to the experience of co-parenting with abusive former partners.

In the last year of marriage and following separation, all mothers took active steps to protect themselves and their children from their former partners’ behavior. This active engagement with a range of protective strategies points to the importance of increasing research, resources for professionals, and empirically-based support for all mothers experiencing separation and divorce, a time that is often filled with conflict and instability. Future research, practice, and policy should continue differentiating between types of IPV so that divorcing mothers receive individualized resources and support that best meet their needs and potential risks and losses.