Is Perceived Need for Action Among Women in Violent Relationships Nonlinear and,If So,Why?

Abstract

Despite the prevalence and impact of partner violence, we understand little about women’s action taking except that it seems an unpredictable, nonlinear process. This article determines the degree of nonlinearity in perceived need for help, legal action, or leaving among women in violent relationships. The participants included 143 women who experienced violence in the previous month, enrolled from six primary care clinics. Baseline surveys assessed background characteristics and factors which may affect perceived need for action. Multiple times series assessments of violence and need for action were collected daily for 8 weeks via telephone Interactive Voice Response. Measures of nonlinearity of violence, perceived need for help, legal action, and leaving were computed. Repeated measures ANOVA assessed differences across measures of nonlinearity. To identify factors contributing to nonlinearity, staged multiple regression assessed the relationship between nonlinearity measures and outcomes. Ninety-three women completed sufficient time series for nonlinearity assessment. Measures of nonlinearity were lower for need for legal action compared with needs for help and leaving. Regression analysis suggested that isolation, social networks, and lack of awareness contribute to nonlinearity. Women’s perceived need for legal action and its level of nonlinearity were lowest compared with those of help seeking and leaving. Although its relative linearity suggests that the need for legal action may be the most predictable, its lower mean rating suggests that legal action is a low priority. Although need for help and leaving are of higher priorities, their nonlinearity suggests that intervention will not yield predictable results.

Keywords

battered women domestic violence legal intervention domestic violence domestic violence cultural contexts anything related to domestic violence perceptions of domestic violence

Introduction

Today, 36% of U.S. women experience intimate partner violence (IPV), stalking, or sexual assault at some time in their lives (Black et al., 2011), often with injury requiring medical care (Black et al., 2011; Tjaden & Thoennes, 2000). Compared with nonabused women, victims of violence report poorer physical and mental health overall (Black et al., 2011). The devastating impact of IPV on women and society is well known (Coker, Smith, Bethea, King, & McKeown, 2000; Wisner, Gilmer, Saltzman, & Zink, 1999). However, our ability to intervene in violent relationships is limited by our lack of understanding of women’s action taking to end the violence other than it appears to be an unpredictable, nonlinear process. Using qualitative research methods, others have observed that taking action in IPV appears to be a nonlinear process. This study investigates whether decision making about use of formal services by IPV survivors is indeed quantitatively nonlinear and, if so, identifies the source of that nonlinearity. Nonlinearity may be due to (a) the nonlinearity of partner-perpetrated violence as observed in prior studies (Katerndahl, Burge, Ferrer, Becho, & Wood, 2010, 2014b); (b) the presence of multiple, interdependent predictors; (c) circularly causal predictors, or nonlinearity reflecting (d) a catastrophic phenomenon, in which periods of stable behavior are punctuated by sudden change.

Decision Making in Partner Violence

“Decision making” is the process of making a choice between a number of options and committing to a future course of action. It is a process that may involve a series of steps over time, beginning with a perceived need to make a change. For a decision to be made, more than one option must be available; choice is based upon inputs, values, and constraints. Finally, the decision-making process involves action, even if that action is to delay the change (Higson & Sturgess, 2014).

Help seeking for IPV can include use of both informal and formal services. Because the use of informal social support will be the subject of a future manuscript, this article will focus on decision making concerning use of formal services. Formal actions among women in violent relationships can be grouped into three categories: leaving the household, seeking counseling, or taking legal action. Leaving the relationship is the most studied option. Qualitatively, the process of taking action for these women is described as a nonlinear, fluctuating process (Anderson, 2003; Chang et al., 2006), reflecting a sequential process of “two steps forward and one step back” (Stork, 2008). Leaving an abusive partner may be considered the initial step in recovery (Landenburger, 1998), and the beginning of a difficult and gradual process (Merritt-Gray & Wuest, 1995). However, leaving may not represent the only path to recovery. Bell, Goodman, and Dutton (2007) found that among women seeking general help for IPV, over the course of a 1-year period, 117 of 206 women left the relationship soon after seeking help and never returned, whereas an additional 31 left within 9 months of help seeking. The remaining women either remained in the relationship (n = 31) or spent time in and out of the relationship (n = 27). Cattaneo, Stuewig, Goodman, Kaltman, and Dutton (2007) found that victims of IPV commonly seek both legal and nonlegal help repeatedly over the course of a year. Those who leave soon after seeking help reported the highest quality of life and lowest frequencies of physical or psychological abuse or stalking 1 year later (Bell et al., 2007). The complexity of this decision-making trajectory suggests the presence of a variety of factors in tension.

In general, decision-making processes include social, cognitive, and cultural factors as well as perceptions, interpretation, judgment, motivation, and postaction reflection (Pijanowski, 2009). A woman’s decision to separate from her violent partner or seek help from the health care or legal systems is related to a number of internal and external factors, such as the level of violence and alcohol consumption of both partners in the relationship (Chabot, Tracy, Manning, & Poisson, 2009; McDonough, 2010; Ramisetty-Mikler & Caetano, 2005); the man’s stalking behavior (Logan, Walker, Shannon, & Cole, 2008); pragmatic factors such as financial independence (Kim & Gray, 2008), partner infidelity (Brandt, 2006; Chang et al., 2010), relationship quality, and safety (Ballantine, 2005; Chang et al., 2010; Djikanovic et al., 2012; Duterte et al., 2008; Barrett & St Pierre, 2011); perceived barriers and support (Ballantine, 2005; Burton, 2004); and her prior experience with taking action (Ford, 1983; Koepsell, Kernic, & Holt, 2006). Personal attitudes concerning the violence may also play a role (Gordon, Burton, & Porter, 2004). Perceived or real external family pressures, relationship investment, and alternatives are important (Stork, 2005). The presence of children can be critical to the woman’s decision to take action, creating conflict between concern for the child’s well-being (Ballantine, 2005; Brandt, 2006; Chang et al., 2010; Djikanovic et al., 2012; Petersen, Moracco, Goldstein, & Clark, 2004) and desire to keep the family together (Rhodes, Cerulli, Dichter, Kothari, & Barg, 2010). These stressors and risks may represent important factors in building the perception of need for action, and eventually taking action, contributing to the nonlinearity of decision making. But the complexity of this process differs in minority populations and is particularly important in our study of Hispanic women.

Compared with non-Hispanic Whites, women of color are less likely to leave a violent relationship (Lacey, Saunders, & Lingling, 2011) or seek treatment (Roberts, Gillman, Breslau, Breslau, & Koenen, 2011). Specifically, African American women use mental health services and the justice system less (El-Khoury et al., 2004; Johnson & Zlotnick, 2007; Paranjape, Heron, & Kaslow, 2006), whereas Hispanic women in violent relationships face unique challenges in taking action. Hispanic women are more likely than non-Hispanic White or African American women to use the emergency department (Lipsky & Caetano, 2007), but they are more likely to avoid IPV disclosure to health care professionals (Kelly, 2009). They tend to underutilize formal resources (Bloom et al., 2009) and they are less likely to act once a decision is made (Amanor-Boadu et al., 2012; Flicker et al., 2011; Lipsky, Caetano, Field, & Larkin, 2006). Hispanic women’s action taking may depend upon their need to provide for and protect their children (Kelly, 2009); language barriers; fear and distrust of formal helping resources; cultural barriers (Bloom et al., 2009); lack of knowledge of resource availability, legal rights, and capabilities; social isolation; misinformation; and fear of deportation (Denham et al., 2010). Hence, the temporal variability and multifactorial nature of decision making in IPV may be particularly important for minorities, especially Hispanic women.

Studying Decision Making in Violent Relationships

Understanding the dynamics of such variable, multifactorial processes is challenging. Most studies on action taking among women in violent relationships have used either quantitative methods to document events (i.e., leaving) with cross-sectional correlates of those events or qualitative methods to study factors women consider important in those decisions. Our current understanding of action taking in IPV is limited by our reliance upon linear correlates as measures of stressors, risks, or facilitators. Yet, qualitative research suggests that action taking in IPV is nonlinear (Anderson, 2003; Bell et al., 2007; Chang et al., 2006). Is action taking nonlinear when measured quantitatively? And, if the process is nonlinear, what is the source of that nonlinearity?

The framework most used to study action taking by abused women is the transtheoretical model (Anderson, 2003; Haggerty & Goodman, 2003); the process of moving from no perceived need for action to taking action corresponds to the five stages in the model (Burke, Gielen, McDonnell, O’Campo, & Maman, 2001). The transtheoretical model of change (Prochaska & DiClemente, 1986) states that behavior change moves through five stages: precontemplation, contemplation, preparation, action, and maintenance (Prochaska & DiClemente, 1986). Those in the “precontemplation” stage are characterized by disinterest in change, perceiving that there is no need for change; those in “contemplation” are ambivalent about change, aware of both “pros” and “cons.” In “preparation,” abused women are convinced that change is necessary and are actively preparing to make a change, taking small steps in that direction. In the “action” stage, they make key changes and wrestle to maintain them, dependent upon many forces that push toward and pull away from the change. “Maintenance” differs from “action” in that the change has been maintained for at least 6 months. Those in action and maintenance are at risk for returning to the abusive relationship, especially under stress (Anderson, 2003; Haggerty & Goodman, 2003).

Several contextual factors appear important in fostering the perceived need for action that moves women through these stages, including awareness of the violence and the reality of their situation (Chang et al., 2010; Petersen et al., 2004). Awareness is the primary factor in moving women from precontemplation to contemplation (Burke et al., 2001). Two factors that move abused women further toward action taking are insight into the violence (McDonough, 2010) and a sense of empowerment to act (Ballantine, 2005). Although awareness is important at all stages of the process, self-evaluation is particularly relevant during the contemplation–preparation–action stages (Burke, Denison, Gielen, McDonnell, & O’Campo, 2004). Social support is another factor in action taking. Women with strong support are more likely to seek legal assistance (Wright & Johnson, 2009) and leave the relationship (Koepsell et al., 2006; Wilcox, 2000). Liang, Goodman, Tummala-Narra, and Weintraub (2005) observed that action taking is dependent on problem recognition and support factors. Hence, awareness and appraisal may be important in promoting perceived need for action, whereas coping style and support may be critical in spurring her into action. Taking action to end partner violence depends on multiple, interdependent factors, making the study of dynamical processes of decision making a particular challenge.

Studying Dynamics

Dynamics can be characterized in two ways: (a) dynamical pattern (periodic, chaotic, random) or (b) the degree of nonlinearity (disproportionality between levels of input and output) over time. Periodic dynamics, in which the system cycles its behavior, results when actions and outcomes are tightly coupled, and when current behavior is dependent on previous behavior. Periodic systems have strong attractors (repeating patterns of phenomena) influencing behaviors; they are stable and insensitive to small changes in their state. Periodic systems are predictable and respond predictably to interventions. In chaotic dynamics, the overall pattern of behavior recurs but the specific path is unpredictable. Actions and outcomes are separated in time, and feedback within the system varies in strength and direction. Chaotic systems also have attractors influencing their behavior but they are sensitive to small changes in terms of the specific path they follow. Chaotic systems are unpredictable long term and do not respond predictably to interventions. A type of random dynamics (pink noise or criticality) is common in complex systems. Criticality results from constant stress on a system composed of interdependent components with varying predilections to respond, yielding a random pattern of responses of varying intensity. Systems characterized by criticality have no attractors influencing their behavior, and may or may not be sensitive to changes in the system. Random systems are unpredictable and do not respond predictably to interventions (Morrison, 1991). In summary, dynamical patterns can range from linear and predictable (periodic) to midlevel nonlinear (chaotic) to extremely nonlinear and unpredictable (random).

In addition to dynamical patterns, dynamics can be studied via measures of overall nonlinearity. System dynamics are said to be “nonlinear” if output from such systems is not proportional to input and, hence, unpredictable. Three types of nonlinearity measurements are available (Land & Elias, 2005). One type is “sensitivity to initial conditions” (speed with which two adjacent points diverge over time) such as Lyapunov’s exponent. Another is “algorithmic complexity” (a measure of the amount of information needed to describe the data), assessed by LZ complexity (Ziv & Lempel, 1978). Finally, “irregularity” can be measured by approximate entropy (ApEn; Pincus, 2006). Nonlinearity suggests either chaotic or random dynamics.

The sources of nonlinearity generally derive from the complexity of factors involved in the system. The presence of multiple, interdependent predictors or circular causality (A causes B which, in turn, causes A) can result in nonlinearity. Another source of nonlinearity may be evident in systems characterized by catastrophic phenomena, where relative stability is interlaced with sudden shifts. All these sources create unpredictable dynamics, which cannot be explained by simple cause and effect. Understanding such systems is challenging.

Purpose of Study

The purpose of this study was to determine the degree of nonlinearity in perceived need for help (e.g., counseling or nonshelter programs), need for legal action, or need to leave the relationship among women in violent relationships and, if nonlinear, determine the source of this nonlinearity. Nonlinearity in taking action could be due to nonlinearity of perceived need for action, which may in turn be due to (a) the nonlinearity of partner-perpetrated violence; (b) the presence of multiple, interdependent predictors; or (c) the presence of circularly causal predictors. Or, (d) nonlinearity in taking action could be present if readiness for change is a catastrophic phenomenon. By combining the findings of this study with those of previous analyses of the same data set, we sought to determine which of the four potential explanations of nonlinearity applied to the need for each action.

Method

Sample

Using methods similar to those of a prior study (Burge et al., 2014), women with a recent history of male-to-female partner abuse were recruited from six primary care clinics in San Antonio, Texas. Researchers used two methods for recruitment. They screened women for eligibility in the clinic during a routine primary care office visit, and they left a flyer about the study in the clinic waiting room, instructing interested women to telephone a researcher for more information. Women were considered eligible for the study if they were adult (age = 18-64), nonpregnant, in a married or cohabiting relationship for 5 or more years to ensure a stable relationship, not accompanied by their partners, and reporting verbal or physical abuse in the past 30 days. Researchers screened women for abuse using a six-item brief Conflict Tactic Scale (Straus, Gelles, & Steinmetz, 1980) in the examination room while the women waited to see their physician. Because this study involved longitudinal daily assessment of violent events, it posed considerable risk of discovery by the violent partner and potentially a risk of escalating violence; participant safety was a serious concern. Hence, women who reported male-to-female violence in the past 30 days completed a Danger Assessment Screen to exclude women in severely dangerous relationships. A “yes” to any of the screening questions was reason for exclusion. Researchers referred these women to the Family Justice Center, a city-supported victim assistance program.

Procedure

Baseline assessments

One hundred forty-three participants completed measures at baseline designed to assess background characteristics as well as factors that may affect the decision-making process. Prior work found that religious variables (Katerndahl, Burge, Ferrer, Becho, & Wood, 2015), awareness (Burke et al., 2001), depression and hope, childhood abuse, coping and violence appraisal, and partner’s controlling behaviors contribute to violence dynamics and outcomes, including health care utilization (Katerndahl, Burge, Ferrer, Becho, & Wood, 2014a). In addition to demographics, we assessed religiousness using three scales (seven questions) from the Brief Multidimensional Measure of Religiousness/Spirituality (public religious activities, private religious activities, religious intensity; Fetzer Institute/National Institute on Aging Working Group, 1999). The presence and severity of depression was assessed via the nine-item Patient Health Questionnaire of the Primary Care Evaluation of Mental Disorders (PRIME-MD) (Nease & Malouin, 2003). For violence history, participants reported their history of childhood abuse using the 17-item (Adverse Childhood Experiences Scales; Dube, Williamson, Thompson, Felitti, & Anda, 2004; Felitti et al., 1998) as well as identifying control strategies used by their partner by completing the 17-item Abusive Behavior Inventory (Shepard & Campbell, 1992). Because prior experience with taking action against violence is important, participants were asked whether they have ever left the relationship, sought counseling, or sought legal assistance, and asked to rate that experience from very negative (−3) to very positive (+3).

Finally, to provide contextual information concerning participants’ support systems, we asked participants to describe their social networks (immediate social contacts), using a social network analysis matrix. The effect of social support on women’s actions is complex. Although victims of IPV are more likely to seek legal assistance when they receive adequate social support (Wright & Johnson, 2009), they are more likely to leave the relationship if support is inadequate (Wilcox, 2000), especially if support is sought but not received (Koepsell et al., 2006). Two questions identified members of participants’ social networks: “Who are the people with whom you discuss matters important to you?” and “Who are the people you really enjoy socializing with?” These questions have been shown to reliably identify the significant members of an individual’s social network (Marin & Hampton, 2007). For each member of the network, participants reported their gender, relationship to the person reporting, whether they told the person about the abuse, and reflected on whether the person himself or herself was in an abusive relationship. To construct the social support matrix among members of the network, each participant rated the strength of emotional support given and received by each dyad within the matrix using a 3-point scale (0 = no relationship or support, 1 = weak support, 2 = strong support). Each network was analyzed using UCINET software (Borgatti, Everett, & Freeman, 2002) to describe the network’s exchange of support and structure. The participant’s (“ego”) active personal network (egonet) was described in terms of their centrality, density, reach efficiency (number of members ego can currently reach within two steps divided by the connectedness of the other members), and effective size (number of egonet members minus the average connections among members), and overall egonet efficiency (effective size divided by the number of others in the egonet) and constraints (measure of the extent to which ego is invested in people who are invested in others in the egonet).

Baseline surveys also included assessment of factors affecting the decision-making progress. Appraisal of the violent situation was assessed using the 24-item Appraisal Dimension Scale directed at a primary stressor (violence; Vitaliano, 1985). Hope was measured using the 12-item Herth Hope Index (Herth, 1992). Social support and stress were measured using the 22-item Duke Social Support & Stress Scale (Duke University, 1986). Coping style was assessed with the 53-item, 12-scale COPE Inventory (Carver, Scheier, & Weintraub, 1989). Finally, we used the Acting With Awareness Scale from the 39-item Kentucky Inventory of Mindfulness Skills to assess awareness (Baer, Smith, & Allen, 2004).

Daily surveys

Participants completed a daily assessment using Interactive Voice Response (IVR) via telephone for 8 weeks. Cell phones were provided to women who needed them to participate. About the same time each day, participants telephoned the IVR from a safe location and answered questions concerning the previous day’s experience. In addition to measures of household environment, the daily report included assessment of violence severity by partners as measured by the modified Conflict Tactic Scale and perceived “need for help” (e.g., counseling or nonshelter programs), “need for legal assistance” (e.g., call police, file for divorce, or protective order), and “need to leave” assessed using a 7-point scale.

Weekly reports

To ensure that the women remained safe during the study, researchers asked women to phone in weekly to report ongoing safety and determine whether they had taken any action in the previous week. If participants took action, researcher asked about the type of action, and on what specific day that action was taken.

Analysis

Study completers provided telephone survey responses on an average of 41.81 (74.66%) of days. To compute nonlinearity measures (LZ complexity and ApEn) for this study, complete time series data were needed; Lyapunov exponents were not computed due to their need for longer time series. Most approaches to data imputation assume linear characteristics. To impute missing data in daily violence severity and perceived need for action while retaining nonlinearity, we applied the nstep procedure from the Time Series Analysis (TISEAN) nonlinear package (Heggler, Kantz, & Schreiber, 1999). The nstep approach to imputation has been shown to least distort nonlinear characteristics of time series when compared with traditional methods (Kreindler & Lumsden, 2006). Unlike other approaches to handling missing data, nstep successfully corrected for 25% to 60% missing data if such data were missing at random, 15% to 40% if missing data followed a power distribution in chaotic time series, and 25% to 40% in periodic time series (Kreindler & Lumsden, 2007). When the initial datapoints in the time series were insufficient to apply nstep (generally <4), the mode of the time series was inserted until the time series is long enough to use nstep. To assess the effects of data imputation of nonlinearity assessment, we compared nonlinearity measures using all data sets (n = 93) compared with those using only complete, nonimputed data sets (n = 11) and those with less than 25% of data missing (n = 63). Although measures of both LZ complexity (a measure of algorithmic complexity) and ApEn (a measure of irregularity) increased with increasing use of imputed data, the incremental increases were small for both measures. With 8 weeks of datapoints, we calculated LZ complexity using the Chaos Data Analyzer software, and ApEn using OCTAVE software; higher scores indicated a greater degree of nonlinearity. These calculations were applied to each participant’s time series for partner-perpetrated violence and perceived need for action.

To assess differences between groups in nonlinearity of violence and perceived needs for action, we used repeated measures ANOVA with least significant difference post hoc testing. To determine predictors of nonlinearity of perceived need, we used staged multiple regression to assess the relationship between nonlinearity measures and predictors. First, baseline variables that were significantly (p ≤ .25) correlated with need nonlinearity measures were retained for possible inclusion in regression analysis. Second, the staged regression analysis used stepwise methods to minimize collinearity effects of the predictors on the need nonlinearity measures. In the first stage, we entered demographic variables (age, Hispanic ethnicity, socioeconomic status score; Hollingshead & Redlich, 1958). In the second stage, we entered significant baseline variables. Finally, we entered violence characteristics (frequency, mean episode severity, and the violence nonlinearity measures). A p ≤ .05 was deemed significant with .05 < p ≤ .10 deemed as trending toward significance.

Results

Of the 143 women enrolled, 105 completed the end-of-study interview and 93 provided enough daily reports to be included in nonlinearity analysis. Table 1 shows that participants who completed the study were very similar demographically to those who enrolled. Overall, the sample was predominantly low income and Hispanic. The mean duration of the relationship was 14.8 ± 12.2 (SD) years with the mean duration of abuse being 10.6 ± 11.2 (SD) years.

Table 1.

Demographic Characteristics of Those Who Enrolled in and Those Who Completed the Study.

Descriptor	All (n = 143)Frequencyn (%)	Completers (n = 105)Frequencyn (%)
Language (Spanish)	23 (16)	18 (17)
Race/ethnicity
White	15 (11)^a	10 (10)^a
Hispanic	116 (81)	86 (82)
Marriage
Church wedding	21 (15)^a	17 (16)^a
Civil ceremony	39 (27)	30 (29)
Age (M ± SD years)	44.9 (9.8)	45.7 (9.5)
Children at home	1.55 (1.58)^a	1.56 (1.68)^a
Employment
Part time	23 (16)	16 (15)
Full time	23 (16)	15 (14)
Education^a
< high school graduate	50 (35)^a	36 (34)^a
High school graduate	34 (24)	27 (26)
Some college	56 (39)	41 (39)
Income^a (<US$20,000)	76 (53)^a	58 (55)^a
Socioeconomic class^a,b
I	0 (0)^a	0 (0)^a
II	7 (5)	4 (4)
III	21 (15)	14 (13)
IV	12 (8)	10 (10)
V	95 (66)	73 (70)

Missing data accounts for total ≠ 100%.

Hollingshead and Redlich (1958).

Women provided 4,696 daily reports, which included 1,005 (21%) reports of partner-perpetrated abuse and 622 (13%) reports of wife-perpetrated abuse. Figure 1 presents the histograms for two nonlinearity assessments of perceived need for help, taking legal action, and leaving the relationship, with measures of nonlinearity of violence for comparison. The figures include both LZ complexity and ApEn measures of nonlinearity. Using data sets of known dynamics as benchmarks, most approximate entropies were more nonlinear than known chaotic time series whereas most of the LZ complexities were in the ranges seen in random dynamics. Table 2 compares nonlinearity measures across perceived needs. Both the level of perceived need for legal action and the nonlinearity of need for legal action were less than that for either need for help or leaving.

Figure 1.

Histograms of nonlinearity measures (n = 93).

Table 2.

Comparisons of Nonlinearities of Violence and Need for Action (n = 93).

Nonlinearity	Partner’s Violence	Need for Statistics
Nonlinearity	Partner’s Violence	Help	Legal	Leaving	F (p)
M	—	2.66	2.05	2.75	251.2 (.000)^a
LZ complexity	.861	.853	.693	.885	762.2 (.000)^b
ApEn	.450	.453	.362	.480	583.5 (.000)^b

Note. ApEn = approximate entropy; LZ complexity = Lempel-Ziv complexity.

Legal < coping/leaving.

Legal < others.

Table 3 presents results of multiple regression analyses, predicting measures of nonlinearity for all three perceived needs. General findings show that awareness predicted nonlinearities of need for help and need for leaving, whereas social network variables and partners’ use of isolation predicted nonlinearity measures across all three needs. Specifically, nonlinearity of need for help depended on the number of women in the social network who were victims of abuse as well as the participant’s lack of awareness. Nonlinearity of need for legal action depended on fewer coworkers in her network and lower egonet density. Nonlinearity of need to leave was the only outcome that depended on violence nonlinearity (measured by LZ complexity).

Table 3.

Regression Analyses of Nonlinearity of Needs for Action (n = 93).

Predictors	Help		Legal		Leaving
Predictors	LZ	ApEn	LZ	ApEn	LZ	ApEn
Background
Private religiousness		.33***
Witnessed childhood violence						−.24^†
Social network
Number of in-laws	.30***		.22*
Number of coworkers			−.21*	−.23*
Number probably experiencing abuse			.18^†		.24*
Number possibly experiencing abuse	.24*	.24*
Number whose abuse status uncertain	−.22*
Number you told of abuse		−.24*
Egonet density			−.38****	−.29*
Perceptions and coping
Awareness	−.18^†	−.21*			−.26*	−.26*
Coping
Behavioral disengagement			.22^†
Perceived violence importance
Perceived violence predictability
Partner’s controlling behavior
Isolation	.25*		.33***			.22^†
Total					−.27*
Any prior action taken				.20^†
Partner’s violence
Episode severity				.25*
Nonlinearity (LZ complexity)					.29*
Model
F (p)	5.27****	5.14****	6.10****	4.92****	5.39****	5.76****
Adjusted R²	.209	.166	.317	.229	.215	.229

Note. ApEn = approximate entropy; LZ = Lempel-Ziv complexity.

†

p ≤ .1. *p ≤ .05. **p ≤ .01. ***p ≤ .005. ****p ≤ .001.

Discussion

This study represents the third analysis of the same data set investigating different aspects of nonlinearity of decision making among women in violent relationships. Quantitative assessments of nonlinearity using daily IVR reports of needs for action show wide variation in their nonlinearity measures, but generally indicate that the patterns are nonlinear. The mean level of perceived need for legal action was less than that for either need for help or need for leaving, and its nonlinearity measures were also lower. Yet, our previous analyses (Katerndahl, Burge, Ferrer, Becho, & Wood, 2016) showed that needs for help and leaving exhibited negative feedback from day-to-day, which should minimize nonlinearity (Erdi, 2008). For example, if need to leave increases one day, the tendency is for it to decrease the next day, preventing rapid, progressive increases in perceived need. In contrast, nonlinearity of need for legal action was lower (more predictable) than other needs despite our previous findings of feedforward dynamics (Katerndahl et al., 2016), which should promote greater nonlinearity (Erdi, 2008).

Using regression analysis, the degree of nonlinearity (unpredictability) of need for action generally depended upon three primary baseline factors: social isolation, social networks, and low awareness. Unpredictability was greater when partners attempted to isolate the women, when women had low awareness, and when social networks had more in-laws, fewer coworkers (i.e., family outsiders), more victims of violence, and low density. This combination of influences might be explained by a family culture of conflict and abuse contributing to the unpredictable nature of her need to take action.

Explaining Nonlinearity of Action in IPV

The nonlinearity of taking action may be related to nonlinearity of perceived need for action due to (a) the nonlinearity of the underlying partner-perpetrated violence; (b) the presence of multiple, interdependent predictors; and/or (c) circularly causal predictors; or, it could be due to (a) an underlying cusp catastrophic phenomenon in which the relationship between violence burden and readiness for action is distorted by factors affecting the violence–readiness relationship. By combining this study’s findings with those of prior analyses (Katerndahl et al., 2016; Katerndahl, Burge, Ferrer, Becho, & Wood, 2017), we can determine which of these explanations best explains the nonlinearity of taking action.

Nonlinearity of need for action due to nonlinear nature of underlying violence: If perceived need is linked to his violence, then same-day and prior-day violence should be correlated with perceived need, and the level of nonlinearity (unpredictability) should be similar for both violence and need. In this scenario, perceived need would be nonlinear because his violence is nonlinear. Prior analysis found that, although same-day violence was associated with all three perceived needs, prior-day violence was not (Katerndahl et al., 2016). The levels of nonlinearity of violence and need for help were similar, but regression analysis found that violence nonlinearity was not predictive of nonlinearity of need for help. The level of nonlinearity of need for legal action was significantly lower than that of the violence; that is, need for legal action was more predictable than violence. And, prior-day violence was inversely related to need for legal action (Katerndahl et al., 2016). That is, a day with violence predicted no need for legal action the following day, whereas a day of calm would predict next-day need for legal action. Need to leave was associated with both same-day and prior-day violence (Katerndahl et al., 2016), and the level of nonlinearity was similar to that of violence. Furthermore, the nonlinearity (LZ complexity) of violence predicted the nonlinearity (LZ complexity) of need to leave. Hence, only for need to leave can we make a strong case for its nonlinearity being due to the nonlinearity of the underlying violence.

Nonlinearity of need for action due to dependence upon multiple, interdependent predictors. Based on prior analysis (Katerndahl et al., 2016), all three perceived needs have interdependent prior-day predictors. Of the seven predictors of need for help, 48% of the possible associations among predictors are actually significant predictors. In need for legal action, 38% of the links among its eight predictors are statistically significant. Finally, for the six prior-day predictors of need to leave, 40% of the possible links were significant (Katerndahl et al., 2016). Hence, the case for nonlinearity of need being due to the interdependence of is predictors is strongest for need for counseling, but could be made for all three needs.

Nonlinearity of need for action due to dependence upon circularly causal predictors: Nonlinearity of need could also be a product of circular causality in which predictors of need are themselves predicted by the need. Based on prior analysis (Katerndahl et al., 2016), a strong case for this explanation can be made for need for help in which three predictors (perceived control, his alcohol use, increasing violence) were in circularly causal relationships with the need itself. Needs for legal action and leaving predicted each other. Although need for legal action had two other circularly causal predictors (perceived control, increasing violence), it also had two predictors (stress, his alcohol use) with negative feedback relationships with need for legal action. Such relationships would be expected to dampen nonlinearity, and in fact, need for legal action had the lowest levels of nonlinearity. In the case of need to leave, there was little evidence that circular causality was responsible for its nonlinearity, other than the circularly causal relationship with need for legal action. Four predictors (his alcohol use, desire to move on, concern for effect of violence on children, forgiveness) had a negative feedback relationship with need to leave (Katerndahl et al., 2016). Hence, only for need for help can we make a strong argument that circular causality may explain its nonlinearity.

Nonlinearity of need for action due to catastrophic nature of decision making: Finally, the existence of sudden, reversible changes in readiness for change suggests that action taking may be best modeled as a catastrophic phenomenon using cusp catastrophe modeling. “Catastrophe theory” seeks to explain sudden, large changes in behavior based upon small, continuous changes in one or more control variables; cusp catastrophe theory is one of the simpler catastrophic models of change. Based on prior analysis of readiness for action (Katerndahl et al., 2017), such catastrophic models accounted for more variance in readiness for action than either of the two linear models. For need for help, forgiveness (a key prior-day predictor) was the significant bifurcation variable, distorting the linear relationship between violence and readiness for help seeking. The presence of children at home was an important bifurcation variable, distorting the association between violence and readiness for legal action and leaving. Hope and positive coping were also important as bifurcation variables for readiness to leave (Katerndahl et al., 2017). Hence, the nonlinearity of all three actions could be due to an underlying catastrophic relationship.

Overall, the cause of nonlinearity of help seeking can be strongly made for all explanations except the first one; nonlinearity of help seeking is not associated with nonlinearity of violence. The best explanation for nonlinearity of seeking legal action is that it is a catastrophic phenomenon. Finally, two explanations for the nonlinearity of leaving appear strongly supported: Readiness to leave is a catastrophic phenomenon, and nonlinearity of need to leave is linked to the nonlinearity of the violence.

Implications

Prior studies suggest that women want nonjudgmental, nondirective, individualized intervention from providers (Feder, Hutson, Ramsay, & Taket, 2006). If we want women in violent relationships to make quality decisions, such decision making should involve clarifying values, identifying alternatives, obtaining necessary information, combining them to balance heart-and-head factors to make a sound decision, and then committing to act (Kedin, Shoemaker, & Spetzler, 2009). This framework is compatible with the IPV-specific approaches of Liang et al. (2005) and Chang et al. (2005). However, this study found that need for action among women in violent relationship was nonlinear and unpredictable, making optimal decision making difficult. We should anticipate that their actions will seem sudden and unexpected, and that our interventions will produce unpredictable results. If we choose to intervene in such nonlinear phenomena, then the nature of the intervention should match the potential source(s) of the nonlinearity.

Limitations

This study is subject to several important limitations. First, we excluded women who were at high risk of life-threatening violence. Thus, these findings only apply to women in less violent relationships; this may affect help-seeking results for IPV (Vatnar & Bjorkly, 2013). Table 3 suggests that violence severity increases nonlinearity of need for legal action; previous analysis of this data set also found that violence severity increased subsequent need to leave but decreased actual leaving (Katerndahl et al., 2016). Second, the sample size is small for time series analysis, especially for determining dynamic pattern. However, previous studies suggest that stable measures of ApEn (Yeragani, Pohl, Mallavarapu, & Balon, 2003) and LZ complexity (Zhang et al., 1999) can be obtained with as few as 50 and 30 datapoints, respectively. In addition, the setting may limit the generalizability of the results because women recruited from health care settings may be less isolated than those from a community-based sample, while an urban-based study such as this one may find a greater use of legal services than one conducted in a rural setting (Shannon, Logan, Cole, & Medley, 2006). Finally, the predominance of low-income Hispanics within the sample may limit the generalizability of the findings. Low-income women are less likely to seek help for IPV (Cheng & Lo, 2014). In addition, Hispanic women may avoid IPV disclosure to health care professionals (Kelly, 2009) and lack a willingness to take action (Amanor-Boadu et al., 2012; Flicker et al., 2011; Lipsky et al., 2006). In addition, Hispanic women may depend upon different factors in their decision making (Kelly, 2009). Yet, in one study, minority women report help seeking and legal actions similar to those of Caucasian women (Hyman, Forte, DuMont, Romans, & Cohen, 2009).

Conclusion

Women’s perceived need for legal action and its degree of nonlinearity were lowest compared with needs for help and leaving. Although its relative linearity suggests that the need for legal action may be the most predictable, its lower mean rating suggests that legal action is a low priority. Of the four possible explanations for nonlinearity of need for action, need for help is best explained by its multiple, interdependent predictors and circular causality, whereas need for legal assistance is best understood via cusp catastrophe modeling; nonlinearity of need to leave is best explained by its dependence upon the underlying nonlinear violence itself. Although needs for help and leaving are of higher priorities with women, their nonlinearity suggests that intervention will not yield predictable results.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was funded by a grant from the National Science Foundation (#1260210)

Author Biographies

David A. Katerndahl, MD, is the former director of research in the Department of Family and Community Medicine at the University of Texas Health Science Center at San Antonio (UTHSCSA). He has extensive experience as a principal investigator in conducting research in primary care settings on depression and panic disorder as well as husband-to-wife and child sexual abuse in Hispanic populations. He has undertaken numerous studies using complexity science, leading to presentations at national/international meetings and publications. He is trained in, and has conducted dynamics analyses in, approximate entropy, vector autoregression, and cusp catastrophe modeling.

Sandra K. Burge, PhD, is a social scientist and tenured professor in the Department of Family & Community Medicine at the UTHSCSA. She teaches doctor–patient communication skills to family medicine residents, and teaches research methods to medical students, residents, fellows, and new faculty. For the past 10 years, she has worked with a collaboration of investigators to examine the nonlinear dynamics of intimate partner violence, focusing on women seeking routine health care in primary care clinics. Findings confirm the complexity of violent relationships and suggest new directions for intervention in primary care settings.

Robert L. Ferrer, MD, is a family physician and director of research in the Department of Family and Community Medicine at UTHSCSA. He has also been active in the area of applying complexity science to problems in health and health care, including both theoretical and empirical work. In addition to being a coinvestigator on the violence studies with our study team, he has applied complexity science to health system redesign, unit-level health care quality, and agent-based simulation modeling of primary care and specialist roles in the health system.

Johanna Becho, MA, is a bilingual research specialist in the Department of Family and Community Medicine at UTHSCSA. Since joining the department, she has distinguished herself in the caring and ethical recruitment of subjects and collection of data for several investigations, including studies concerning mental health care in family medicine settings. She is a graduate student in the Family, Couples and Individual Psychotherapy program in the Department of Psychology at Our Lady of the Lake University. She is skilled in the sensitive collection of information from emotionally burdened patients with 10 years of research experience in working with violence victims.

Robert Wood, DrPH, is a biostatistician in the Biometry Unit of the Department of Family and Community Medicine, providing data analysis support in the Department of Family and Community Medicine for research and quality improvement projects as well as collaborating with researchers from other disciplines as well. He provides broad support in data management and statistical methods. He has significant experience with population studies and secondary data, as well as serving as coinvestigator on the team’s prior violence studies.

References

Amanor-Boadu

Messing

J. T.

Stith

S. M.

Anderson

J. R.

O’Sullivan

C. S.

Campbell

J. C.

(2012). Immigrant and nonimmigrant women. Violence Against Women, 18, 611-633.

Anderson

(2003). Evolving out of violence. Research and Theory of Nursing Practice, 17, 225-240.

Baer

R. A.

Smith

G. T.

Allen

K. B.

(2004). Assessment of mindfulness by self-report. Assessment, 11, 191-206.

Ballantine

M. W.

(2005). Decision-making processes of abused women. Dissertation Abstracts International: Section A, Humanities and Social Sciences, 65(11-A), 4346.

Barrett

B. J.

St. Pierre

(2011). Variations in women’s help seeking in response to intimate partner violence. Violence Against Women, 17, 47-70.

Bell

M. E.

Goodman

L. A.

Dutton

M. A.

(2007). Dynamics of staying and leaving. Journal of Family Violence, 22, 413-428.

Black

M. C.

Basile

K. C.

Breiding

M. J.

Smith

S. G.

Walters

M. L.

Merrick

M. T.

. . . Stevens

M. R.

(2011). The National Intimate Partner and Sexual Violence Survey (NISVS): 2010 Summary Report. Atlanta, GA: National Center for Injury Prevention and Control, Centers for Disease Control and Prevention.

Bloom

Wagman

Hernandez

Yragui

Hernandez-Valdovinos

Dahlstrom

Glass

(2009). Partnering with community-based organizations to reduce intimate partner violence. Hispanic Journal of Behavioral Sciences, 31, 244-257.

Borgatti

S. P.

Everett

M. G.

Freeman

L. C.

(2002). UCINET for Windows: Software for social network analysis. Harvard, MA: Analytic Technologies.

10.

Brandt

J. E.

(2006). Why she left. Dissertation Abstracts International: Section B, Sciences and Engineering, 66(8-B), 4473.

11.

Burge

S. K.

Becho

Ferrer

R. L.

Wood

R. C.

Talamantes

Katerndahl

D. A.

(2014). Safely examining complex dynamics of intimate partner violence. Families, Systems & Health, 32, 259-270.

12.

Burke

J. G.

Denison

J. A.

Gielen

A. C.

McDonnell

K. A.

O’Campo

(2004). Ending intimate partner violence. American Journal of Health Behavior, 28, 122-133.

13.

Burke

J. G.

Gielen

A. C.

McDonnell

K. A.

O’Campo

Maman

(2001). Process of ending abuse in intimate relationships. Violence Against Women, 7, 1144-1163.

14.

Burton

S. J.

(2004). To go or not to go. Dissertation Abstracts International: Section B, Sciences and Engineering, 649(9-B), 4605.

15.

Carver

C. S.

Scheier

M. F.

Weintraub

J. K.

(1989). Assessing coping strategies: A theoretically based approach. Journal of Personality and Social Psychology, 56, 267-283.

16.

Cattaneo

L. B.

Stuewig

Goodman

L. A.

Kaltman

Dutton

M. A.

(2007). Longitudinal helpseeking patterns among victims of intimate partner violence. American Journal of Orthopsychiatry, 77, 467-477.

17.

Chabot

H. F.

Tracy

T. L.

Manning

C. A.

Poisson

C. A.

(2009). Sex, attribution, and severity influence intervention decisions of informal helpers in domestic violence. Journal of Interpersonal Violence, 24, 1696-1713.

18.

Chang

J. C.

Cluss

P. A.

Ranieri

Hawker

Buranosky

Dado

. . . Scholle

S. H.

(2005). Health care interventions for intimate partner violence: What women want. Women’s Health Issues, 15, 21-30.

19.

Chang

J. C.

Dado

Ashton

Hawker

Cluss

P. A.

Buranosky

Scholle

S. H.

(2006). Understanding behavior change for women experiencing intimate partner violence: Mapping the ups and downs using the stages of change. Patient and Educational Counseling, 62, 330-339.

20.

Chang

J. C.

Dado

Hawker

Cluss

P. A.

Buranosky

Slagel

. . . Scholle

S. H.

(2010). Understanding turning points in intimate partner violence: Factors and circumstances leading women victims toward change. Journal of Women’s Health, 19, 251-259.

21.

Cheng

T. C.

C. C.

(2014). Domestic violence and treatment seeking. International Journal of Health Services, 44, 735-759.

22.

Coker

A. L.

Smith

P. H.

Bethea

King

M. R.

McKeown

R. E.

(2000). Physical health consequences of physical and psychological intimate partner violence. Archives of Family Medicine, 9, 451-457.

23.

Denham

A. C.

Frasier

P. Y.

Gerken Hooten

Belton

Newton

Gonzalez

. . . Campbell

M. K.

(2010). Intimate partner violence among Latinas in Eastern North Carolina. Violence Against Women, 13, 123-140.

24.

Djikanovic

Wong

Jansen

H. A.

Koso

Simic

Otasevic

Largo-Janssen

(2012). Help-seeking behavior of Serbian women who experienced intimate partner violence. Family Practice, 292, 189-195.

25.

Dube

S. R.

Williamson

D. F.

Thompson

Felitti

V. J.

Anda

R. F.

(2004). Assessing the reliability of retrospective reports of adverse childhood experiences among adult HMO members attending a primary care clinic. Child Abuse & Neglect, 28, 729-737.

26.

Duke University. (1986). Duke Social Support and Stress Scale. Durham, NC: Author.

27.

Duterte

E. E.

Bonomi

A. E.

Kernic

M. A.

Schiff

M. A.

Thompson

R. S.

Rivara

F. P.

(2008). Correlates of medical and legal help seeking among women reporting intimate partner violence. Journal of Women’s Health, 17, 85-95.

28.

El-Khoury

M. Y.

Dutton

M. A.

Goodman

L. A.

Engel

Belmaric

R. I.

Murphy

(2004). Ethnic differences in battered women’s formal help-seeking strategies. Cultural Diversity & Ethnic Minority Psychology, 10, 383-393.

29.

Erdi

(2008). Complexity explained. Berlin, Germany: Springer-Verlag.

30.

Feder

G. S.

Hutson

Ramsay

Taket

A. R.

(2006). Women exposed to intimate partner violence. Archives of Internal Medicine, 166, 22-37.

31.

Felitti

V. J.

Anda

R. F.

Nordenberg

Williamson

D. F.

Spitz

A. M.

Edwards

. . . Marks

J. S.

(1998). Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: The Adverse Childhood Experiences (ACE) Study. American Journal of Preventive Medicine, 14, 245-258.

32.

Fetzer Institute/National Institute on Aging Working Group. (1999). Multidimensional measurement of religiousness/spirituality for use in health research. Kalamazoo, MI: Fetzer Institute.

33.

Flicker

S. M.

Cerulli

Zhao

Tang

Watts

Xia

Talbot

N. L.

(2011). Concommitant forms of abuse and help-seeking behavior among white, African American, and Latina women who experience intimate partner violence. Violence Against Women, 17, 1067-1085.

34.

Ford

D. A.

(1983). Wife battery and criminal justice. National Council of Family Relations, 32, 463-475.

35.

Gordon

K. C.

Burton

Porter

(2004). Predicting the intentions of women in domestic violence shelters to return to partners. Journal of Family Psychology, 18, 331-338.

36.

Haggerty

L. A.

Goodman

L. A.

(2003). Stages of change-based nursing interventions for victims of interpersonal violence. Journal of Obstetrics, Gynecology & Neonatal Nursing, 32, 68-75.

37.

Heggler

Kantz

Schreiber

(1999). Practical implementation of nonlinear time series methods: The TISEAN package. Chaos, 9, 413-435.

38.

Herth

K. A.

(1992). An abbreviated instrument to measure hope. Journal of Advanced Nursing, 17, 1251-1259.

39.

Higson

Sturgess

(2014). Uncommon leadership. New York, NY: Kogan Page.

40.

Hollingshead

Redlich

(1958). Social class and mental illness. New York, NY: John Wiley.

41.

Hyman

Forte

DuMont

Romans

Cohen

M. M.

(2009). Help-seeking behavior for intimate partner violence among racial minority women in Canada. Women’s Health Issues, 19, 101-108.

42.

Johnson

D. M.

Zlotnick

(2007). Utilization of mental health treatment and other services by battered women in shelters. Psychiatric Services, 58, 1595-1597.

43.

Katerndahl

D. A.

Burge

S. K.

Ferrer

R. L.

Becho

Wood

(2010). Complex dynamics in intimate partner violence: A time series study of 16 women. Primary Care Companion of the Journal of Clinical Psychiatry, 12(4), e1-e12.

44.

Katerndahl

D. A.

Burge

S. K.

Ferrer

R. L.

Becho

Wood

(2014a). Do violence dynamics matter? Journal of Evaluation in Clinical Practice, 20, 719-727.

45.

Katerndahl

D. A.

Burge

S. K.

Ferrer

R. L.

Becho

Wood

(2014b). Dynamics of violence. Journal of Evaluation in Clinical Practice, 20, 695-702.

46.

Katerndahl

D. A.

Burge

S. K.

Ferrer

R. L.

Becho

Wood

(2015). Effects of religious and spiritual variables on outcomes in violent relationships. International Journal of Psychiatry in Medicine, 49, 249-263.

47.

Katerndahl

D. A.

Burge

S. K.

Ferrer

R. L.

Becho

Wood

(2016). Predictors of perceived need for and actual action taking among women in violent relationships. Journal of Interpersonal Violence, 31, 1-28. doi:10.1177/0886260516669543

48.

Katerndahl

D. A.

Burge

S. K.

Ferrer

R. L.

Becho

Wood

(2017). Is readiness to take action among women in violent relationships a catastrophic phenomenon? Journal of Interpersonal Violence, 32, 1-25. doi:10.1177/0886260517698280

49.

Kedin

Shoemaker

Spetzler

(2009). Decision quality. Palo Alto, CA: Decision Education Foundation.

50.

Kelly

U. A.

(2009). I’m a mother first. Research in Nursing & Health, 32, 286-297.

51.

Kim

Gray

K. A.

(2008). Leave or stay? Journal of Interpersonal Violence, 23, 1465-1482.

52.

Koepsell

J. K.

Kernic

M. A.

Holt

V. L.

(2006). Factors that influence battered women to leave their abusive relationships. Violence and Victims, 21, 131-147.

53.

Kreindler

D. M.

Lumsden

C. J.

(2006). Effects of the irregular sample and missing data in time series analysis. Nonlinear Dynamics Psychology and Life Sciences, 10, 187-214.

54.

Kreindler

D. M.

Lumsden

C. J.

(2007). Effects of irregular sampling and missing data on largest Lyapunov exponents. Nonlinear Dynamics Psychology and Life Sciences, 11, 401-412.

55.

Lacey

K. K.

Saunders

D. G.

Lingling

(2011). Comparison of women of color and non-Hispanic white women on factors related to leaving a violent relationship. Journal of Interpersonal Violence, 26, 1036-1055.

56.

Land

Elias

(2005). Measuring the “complexity” of a time series. Retrieved from http://people.ece.cornell.edu/land/PROJECTS/Complexity/index.html

57.

Landenburger

K. M.

(1998). Dynamics of leaving and recovering from an abusive relationship. Journal of Obstetric, Gynecologic & Neonatal Nursing, 27, 700-706.

58.

Liang

Goodman

Tummala-Narra

Weintraub

(2005). Theoretical framework for understanding help-seeking processes among survivors of intimate partner violence. American. Journal of Community Psychology, 36, 71-83.

59.

Lipsky

Caetano

(2007). Role of race/ethnicity in the relationship between emergency department use and intimate partner violence. American Journal of Public Health, 97, 2246-2252.

60.

Lipsky

Caetano

Field

C. A.

Larkin

G. L.

(2006). Role of intimate partner violence, race and ethnicity in help-seeking behaviors. Ethnicity & Health, 11, 81-100.

61.

Logan

T. K.

Walker

Shannon

Cole

(2008). Factors associated with separation and ongoing violence among women with civil protective orders. Journal of Family Violence, 23, 377-385.

62.

Marin

Hampton

K. N.

(2007). Simplifying the personal network name generator. Field Methods, 19, 163-193.

63.

McDonough

T. A.

(2010). Policy capturing investigation of battered women’s decisions to stay in violent relationships. Violence and Victims, 25, 165-184.

64.

Merritt-Gray

Wuest

(1995). Counteracting abuse and breaking free. Health Care for Women International, 16, 399-412.

65.

Morrison

(1991). Art of modeling dynamic systems. New York, NY: John Wiley.

66.

Nease

D. E.

Malouin

J. M.

(2003). Depression screening. Journal of Family Practice, 52, 118-126.

67.

Paranjape

Heron

Kaslow

N. J.

(2006). Utilization of services by abused, low-income African-American women. Journal of General Internal Medicine, 21, 189-192.

68.

Petersen

Moracco

K. E.

Goldstein

K. M.

Clark

K. A.

(2004). Moving beyond disclosure. Women and Health, 40, 63-76.

69.

Pijanowski

J. C.

(2009). The role of learning theory in building effective college ethics curricula. Journal of College & Character, 103(3), 1-13.

70.

Pincus

S. M.

(2006). Approximate entropy as a measure of irregularity for psychiatric serial metrics. Bipolar Disorders, 8, 430-440.

71.

Prochaska

J. O.

DiClemente

C. C.

(1986). Toward a comprehensive model of change. In Miller

W. R.

Heather

(Eds.), Treating addictive behaviors (pp. 3-27). New York, NY: Plenum Press.

72.

Ramisetty-Mikler

Caetano

(2005). Alcohol use and intimate partner violence as predictors of separation among U.S. couples. Journal of Studies on Alcohol, 66, 205-212.

73.

Rhodes

K. V.

Cerulli

Dichter

M. E.

Kothari

C. L.

Barg

F. K.

(2010). I didn’t want to put them through that. Journal of Family Violence, 25, 485-493.

74.

Roberts

A. L.

Gillman

S. E.

Breslau

Koenen

K. C.

(2011). Race/ethnic differences in exposure to traumatic events, development of posttraumatic stress disorder, and treatment-seeking for posttraumatic stress disorder in the United States. Psychological Medicine, 41, 71-83.

75.

Shannon

Logan

T. K.

Cole

Medley

(2006). Help-seeking and coping strategies for intimate partner violence in rural and urban women. Violence and Victims, 21, 167-181.

76.

Shepard

M. F.

Campbell

J. A.

(1992). The Abusive Behavior Inventory. Journal of Interpersonal Violence, 7, 291-305.

77.

Stork

E. P.

(2005). Analyzing decision making. Dissertation Abstracts International: Section A, Humanities and Social Sciences, 65(10-A), 3994.

78.

Stork

E. P.

(2008). Understanding high-stakes decision making. Journal of Feminist Family Therapy, 20, 299-327.

79.

Straus

Gelles

Steinmetz

(1980). Behind closed doors. New York, NY: Anchor Books.

80.

Tjaden

Thoennes

(2000). Extent, nature, and consequences of intimate partner violence (Grant# 93-IJ-CX-0012). Washington, DC: National Institute of Justice and Centers for Disease Control.

81.

Vatnar

S. K.

Bjorkly

(2013). Lethal intimate partner violence. Violence and Victims, 28, 772-789.

82.

Vitaliano

P. P.

(1985). Manual for Appraisal Dimension Scale and Revised Ways of Coping Checklist. Seattle: Department of Psychiatry, University of Washington.

83.

Wilcox

(2000). “Me mother’s bank and me nanan’s, you know, support!” Women’s Studies International Forum, 23, 35-47.

84.

Wisner

C. L.

Gilmer

T. P.

Saltzman

L. E.

Zink

T. M.

(1999). Intimate partner violence against women: Do victims cost health plans more? Journal of Family Practice, 486, 439-443.

85.

Wright

C. V.

Johnson

D. M.

(2009). Correlates for legal help-seeking. Violence and Victims, 24, 771-785.

86.

Yeragani

V. K.

Pohl

Mallavarapu

Balon

(2003). Approximate entropy of symptoms of mood. Bipolar Disorders, 5, 279-286.

87.

Zhang

X. S.

Zhu

Y. S.

Thakor

N. V.

Wang

Z. Z.

(1999). Detecting ventricular tachycardia and fibrillation by complexity measure. IEEE Transactions on Biomedical Engineering, 46, 548-555.

88.

Ziv

Lempel

(1978). Compression of individual sequences via variable-rate coding. IEEE Transactions on Information Theory, IT-24, 530-536.