key: cord-0031958-06cu8h6f
authors: Gutowski, Ellen R.; Goodman, Lisa A.
title: Coercive Control in the Courtroom: the Legal Abuse Scale (LAS)
date: 2022-05-19
journal: J Fam Violence
DOI: 10.1007/s10896-022-00408-3
sha: cc663fdc64d8b1dbc207cee1e9dfa1a1393e49b7
doc_id: 31958
cord_uid: 06cu8h6f

Intimate partner violence (IPV) survivors seeking safety and justice for themselves and their children through family court and other legal systems may instead encounter their partners’ misuse of court processes to further enact coercive control. To illuminate this harmful process, this study sought to create a measure of legal abuse. We developed a list of 27 potential items on the basis of consultation with 23 experts, qualitative interviews, and existing literature. After piloting these items, we administered them to a sample of 222 survivor-mothers who had been involved in family law proceedings. We then used both exploratory factor analysis (EFA) and Rasch analysis (RA) to create a final measure. Analyses yielded the 14-item Legal Abuse Scale (LAS). Factor analysis supported two subscales: Harm to Self/Motherhood (i.e., using the court to harm the survivor as a person and a mother) and Harm to Finances (i.e., using the court to harm the survivor financially). The LAS is a tool that will enable systematic assessment of legal abuse in family court and other legal proceedings, an expansion of research on this form of coercive control, and further development of policy and practice that recognizes and responds to it.

Intimate partner violence (IPV) -that is, physical, psychological, sexual (Center for Disease Control and Prevention, 2020), and/or economic (Adams et al., 2008 ) abuse against a current or former partner or spouse -is a pervasive social problem. Over one in four women 1 in the United States experience IPV during their lifetime (Center for Disease Control and Prevention, 2020) and an estimated 38.6% of women homicide victims are murdered by an intimate partner (Stöckl et al., 2013) .

Many, though by no means all, IPV survivors ultimately leave abusive relationships. Yet, doing so opens the door to a host of new challenges: Upon separation, survivors face a heightened risk of violence and diminished social and economic resources (DeKeseredy et al., 2017; Hardesty, 2002; Thomas et al., 2015) . For survivors who have children, separation necessitates addressing child custody, which can be terrifying as it often entails facing the person who abused them in court (Hardesty & Ganong, 2006; Walker et al., 2004) . Further, because the same person who abuses one's partner in a household may also abuse their children (Bancroft et al., 2012) , separation may involve an increased risk of harm to the children (Hayes, 2017) .

Faced with these myriad risks and uncertainties, many survivors enter into the family or civil court systems with hopes of accessing safety and justice for themselves and their children (Hardesty, 2002) . Instead, however, they may face a new form of abuse -here called "legal abuse" -characterized by the abusive parent's use of court processes to further coerce and control them (Coy et al., 2015; Douglas, 2018a; Elizabeth, 2017; Miller & Smolter, 2011; Watson & Ancis, 2013) . A growing collection of qualitative studies and anecdotal accounts suggest that legal abuse may cause severe, adverse consequences for survivors, such as psychological problems and economic hardship (Douglas, 2020;  Increasingly recognized at the state-level in the American criminal-legal system, coercive control is central to the experience of IPV (Dutton & Goodman, 2005; Stark, 2007; Stark & Hester, 2019) . Coercive control is a pattern of behaviors that involves identifying victims, infiltrating their lives, isolating them to gain control, and dominating them --through literal confinement (e.g., forcing one's partner to stay in confined spaces), ongoing surveillance (e.g., frequent phone calls or monitoring whereabouts), or threats (e.g., communicating harmful intentions violently or through a subtle signal, such as a look or a single word) (Duron et al., 2021; Dutton & Goodman, 2005) . As Evan Stark, a foundational thinker in this area, has written, coercive control has a "strategic intelligence" such that apparently discrete acts of physical violence may anchor a broader pattern of domination (p. 198, Stark 2007) . Further, given that coercive control involves one party using their social power to dominate another (Stark, 2007) , the social locations of partners in a relationship are central to making sense of such attempts at control and the degree to which they are effective (Dutton & Goodman, 2005; Stark, 2007) . Each partner's intersectional identities, performance of roles associated with these identities, and social positions (i.e., positions within social structures that uphold systems of power and oppression) are integral to how abuse manifests, how it is experienced, and how it may be recognized or discounted by those outside of the relationship (Anderson, 2009; Crenshaw, 1989) . For example, a Honduran woman who is an English language learner who is partnered with a White, male, US citizen may find that her partner increasingly threatens her with violence after she obtains a green card and employment. Her partner may experience her increase in power as a threat to his identities (and power associated with those identities) (Anderson, 2009) . His violent threats, in turn, may hold greater power to control her behavior given his greater social power within structures of dominance such as white supremacy, xenophobia, and patriarchy. Further, she may hesitate to seek help, fearing that her positionalities as a Latina woman and an English language learner will reduce her credibility. Efforts to maintain control may intensify when a survivor attempts to leave -the ultimate display of autonomy (Stark, 2007) . Family court proceedings can then become a new context for heightened control tactics.

In one of the first papers on the subject of legal abuse, Miller and Smolter (2011) proposed the existence of "paper abuse," defined as the use of "frivolous" legal motions that are an "extension of traditional [intimate partner abuse] tactics" (p. 647). Subsequent scholarship has extended this definition and demonstrated how legal proceedings can serve as platforms for post-separation coercive control as they enable abusive partners to have direct contact with their victims, often over an extended period of time (Elizabeth, 2017) . Indeed, court proceedings may be the only way for abusive partners to legally maintain contact with their former partners (Przekop, 2011) . Face to face confrontation between the parties is typically required when litigating family law matters, such as post-separation parenting, even when the survivor has a no-contact restraining order in place (Hardesty, 2002) .

Multiple qualitative studies have described the range of strategies that abusive partners use to extend coercive control into the legal context. One set of tactics serves to prolong legal cases (Douglas, 2018a; Miller & Smolter, 2011; Przekop, 2011; Watson & Ancis, 2013) -sometimes for multiple years -through actions such as charges of contempt and unfounded accusations, which can lead to extensive investigations (Watson & Ancis, 2013) . Another mechanism of legal abuse involves seeking child custody or parental visitation for the purpose of exerting power and control over survivors during separation (Elizabeth, 2015; Goldstein, 2010; Toews & Bermea, 2017; Ward, 2016; Watson & Ancis, 2013) . For example, abused mothers have reported that their ex-partners sought full custody or increased visitation time out of a desire to cause them distress and not out of a wish to be involved in their children's lives (Watson & Ancis, 2013) . Parents may also use "custody blackmail" (p. 608, Hardesty, 2002) in order to get victims to comply with their demands. The threat of losing child custody and visitation to an abusive partner can be highly distressing for survivor-mothers, not only because of the possibility that they will lose cherished care time with their children, but also because many abusive partners are also abusive parents, and survivormothers often fear for the safety of their children when their children are in the care of an abusive parent (Bancroft et al., 2012; Elizabeth, 2019) . Further, survivors may be required to have on-going communication with an abusive parent to negotiate parenting arrangements or visits (Dekeseredy et al., 2017; Hardesty, 2002; Hardesty & Ganong, 2006; Zeoli et al., 2013) , providing opportunities for continued threats and abuse (Goldstein, 2010; Hardesty, 2002; Toews & Bermea, 2017; Varcoe & Irwin, 2004; Zeoli et al., 2013) .

Ex-partners may also distort information or lie in court as a tactic of control (Miller & Manzer, 2018; Toews & Bermea, 2017; Watson & Ancis, 2013) . In Elizabeth's (2017) qualitative research with mothers involved in custody litigation, participants described their former partners' efforts at character defamation to make them appear unfit to parent. Consistent with past qualitative findings, one survey of 109 attorneys who primarily represent survivors indicated that abusive partners frequently sought to portray the survivor as psychologically unstable (White-Domain & Phillips, 2016) . One can imagine the practically and psychologically damaging nature of this process, given that custody decisions are looming (Meier, 2009) .

Finally, ex-partners can use financial abuse as a tool of control in court (Elizabeth, 2017) . They can seek to modify child support, seek full custody so that they will not be required to pay anything, make excessive court dates to negotiate it, or simply refuse to pay it at all (Przekop, 2011; Toews & Bermea, 2017; Watson & Ancis, 2013) . Parents who use this form of abuse may lie about finances, hide assets, or change jobs in order to avoid sharing resources (Goldstein, 2010; Przekop, 2011; Watson & Ancis, 2013) . Prolonging the legal case can also financially drain survivors, as extending legal negotiations increases the cost of representation (Toews & Bermea, 2017; Watson & Ancis, 2013) .

Thus, qualitative research has documented the mechanisms through which partners use legal abuse including by forcing survivors into distressing face-to-face contact through court proceedings (Elizabeth, 2017; Miller & Smolter, 2011; Przekop, 2011) , attacking the survivor's parental rights or visitation time (Toews & Bermea, 2017; Watson & Ancis, 2013; Ward, 2016) , threatening the children's safety (Toews & Bermea, 2017; Varcoe & Irwin, 2004; Zeoli et al., 2013) , publicly denigrating the survivor's capabilities as a parent (Elizabeth, 2017; Toews & Bermea, 2017; Watson & Ancis, 2013; White-Domain & Phillips, 2016) , and exerting financial abuse against the survivor through the process (Elizabeth, 2017; Przekop, 2011; Toews & Bermea, 2017; Watson & Ancis 2013) . Further, harrowing anecdotal accounts indicate that at least for some, legal abuse may produce grave consequences for survivors' mental health (Douglas, 2018b; Gutowski & Goodman, 2020; Toews & Bermea, 2017; Ward, 2016) . A psychometrically valid measure of legal abuse that includes each of these elements would be a step towards enabling the recognition of coercive control in post-separation legal cases and enhancing current knowledge of its features and costs. Yet, efforts to measure the phenomenon of legal abuse quantitatively are still in nascent stages.

Three measures exist for concepts related to legal abuse (Clemente et al., 2019; Hines et al., 2015; Rivera et al., 2018) . One is a self-report scale of legal and administrative aggression, developed to capture men's experiences of their partners' manipulation of legal and other administrative systems (Hines et al., 2015) . This scale does not conceptualize legal abuse within the context of coercive control. Further, because it was validated with a sample of men, the authors note that this phenomenon likely manifests differently for women (Hines et al., 2015) . A second existing measure is the legal harassment scale, a self-report measure validated with a combined sample of 209 fathers and mothers undergoing family law litigation in Spain (Clemente et al., 2019) . Although this measure offers a novel contribution to the literature, it uses non-specific language (e.g., "I get verbally insulted" without clarification of who -ex-partners, judges, or lawyers -is doing the insulting) and was devised to attend to this issue in a different population and cultural context from the current study. The third relevant measure assesses procedural abuse, defined as "the use of legal means, systems, or procedures to perpetrate abuse" (Rivera et al., 2018 (Rivera et al., , p. 2786 . This measure has shown good internal consistency with a sample of IPV survivor-mothers who had separated or were planning to separate from an abusive partner (Rivera et al., 2018) . However, the sample of 40 survivor-mothers was too small to establish the measure's psychometric validity, and the authors note the need for more research that focuses on legal abuse among this population.

While these three existing measures have advanced the literature, there remains a need for a psychometrically valid measure of legal abuse that is relevant to women who have experienced IPV. Building on existing research and existing measures, the present study developed what we believe is the first psychometrically valid measure to assess legal abuse against IPV survivor-mothers involved in post-separation legal proceedings.

The sample included in the present study was comprised of 222 mothers who met the following eligibility criteria: Parent of a child under age 18 who spoke English or Spanish and was involved in a family law case currently or within the past one to two years with another parent who was abusive (i.e., participants endorsed having been afraid of this other parent). Participants came from 36 different US states located in the Northeastern (n = 55; 25%), Midwestern (n = 49; 22%), Southern (n = 58; 26%), and Western (n = 60; 27%) regions of the country. The sample was 77.5% White, 90.1% heterosexual, 3.6% bisexual, 1.4% gay or lesbian, 1.4% queer, 1.4% "other" sexual identity, and 1% pansexual. All participants endorsed experiencing one or more acts of psychological (97.7%), physical (81.5%), and/or sexual (63.5%) abuse from the other parent. Participants' legal cases lasted from less than six months to over five years with over half (51.8%) lasting two years or longer. Most cases (78.4%) were on-going. For more information on participant demographics, see Table 1 .

After we received approval from Boston College's institutional review board (IRB), we sent email announcements to 908 legal agencies (e.g., legal aid, professional associations, and non-profit organizations), and 478 domestic violence agencies (e.g., direct service agencies and consortiums) throughout the United States. The study announcement stated the study's purpose was "to understand the experiences of mothers in family court." Attorneys and service providers were asked to refer family court-involved mothers of at least one child aged under 18 years. Referring providers were instructed to tell potential participants that they might be eligible for the study and in the announcement, participants were invited to complete screening questions by phone or online to determine eligibility.

Eligible participants were invited to take the full survey in Spanish or English. All potential participants, including those who did not qualify, were provided with information about domestic violence services. Eligible respondents were offered a $20 gift card as a "Thank you" for their participation.

Four hundred and fifty individuals responded to the survey of which 140 either did not qualify (n = 117) or did not complete the screening (n = 23). Among the remaining 310 responses, 75 were determined to be invalid according to a series of validity checks (i.e., examining participants' ip addresses, referral sources, and consistency in responses) and were removed. Of the 235 respondents who remained, 13 had significant attrition and did not complete any of the items on the LAS so were removed casewise prior to conducting analyses. This resulted in a final sample of 222 participants.

Participants responded to general demographic questions and questions about their legal cases as described above under Participants.

Consistent with recommendations for psychological measure development (Clark & Watson, 1995) , we used three sources of information to generate and refine a face valid set of items for the LAS: Interviews conducted in the context of a previous qualitative study involving 19 survivors who had sought child custody (including 13 White, four Black, and two Latina participants aged 34-67 years; see Gutowski & Goodman, 2020), a review of existing research, and a review of prior measures. Once we had developed a draft of the measure, we refined it over the course of the first author's ongoing consultation with 23 experts including eight family law attorneys, three domestic violence advocates, two mental health practitioners, three survivormothers who have been through family court, and seven researchers from social science and legal disciplines (DeVellis, 2017) . This process resulted in 50 items. To reduce the number of items further, the first author sought a second round of consultation with 10 experts who were asked to rate each item as "Essential," "Important" or "Not Important" following Ford-Gilboe and colleagues (2016). We retained only those items that were rated as "Essential" and/or "Important" by the majority of experts. This process resulted in a final list of 27 items that were administered to participants (see Appendix Table 5 ). Participants indicated whether the other parent in their legal case engaged in each behavior on the list by selecting "yes" or "no".

To assess the construct validity of the LAS, we chose the Use of Children Scale (Beeble et al., 2007) , which measures abusive tactics that harass, intimidate, or frighten survivors by using children. Given that this measure captures mechanisms that abusive partners engage to target and control mothers and children, we anticipated that this scale would be positively and significantly correlated with the LAS. This scale is a sevenitem measure using a five-point rating scale of 0 "none" to 4 "very much." A sample item includes, "Used children to keep track of you." This measure showed good internal consistency both in a previous study of 156 women survivors (α = 0.88; Beeble et al., 2007) and in the current study (α = 0.88).

Given the link between legal abuse and IPV, we also assessed the LAS' construct validity using the Composite Abuse Scale -Short Form (CAS-SF) (Ford-Gilboe et al., 2016), a 15-item self-report measure of psychological, physical, and sexual intimate partner abuse. We modified the scale for the present study to be applicable to separated survivors, by asking that each participant respond with a "yes" or "no" to a series of questions about whether their ex-partner had committed a set of specific acts of abuse. A sample item is: "The other parent forced or tried to force me to have sex." The total scale score ranged from a possible 0 to 15, with higher scores indicating more pervasive experiences with prior abuse. This measure, validated in a sample of 6,278 Canadian women, showed strong internal consistency (α = 0.94) with this original sample (Ford-Gilboe et al., 2016) and adequate internal consistency in the current study (α = 0.78).

After a draft of the survey was completed, the items were piloted to identify and eliminate potential ambiguous language (Heppner et al., 2015) . Three research assistants went through the survey in depth to check for clarity of items and to provide an estimate for the amount of time it would take to complete. After another round of editing, three survivors who are also mothers who had sought child custody in family court reviewed an English version of the survey. Based on their feedback on clarity of item wording, the survey was edited a final time before being distributed to participants.

Two Spanish-speaking IPV researchers with translation expertise translated and back-translated (Heppner et al., 2015) the survey from English to Spanish with attention to consistency of meaning in the two languages and across dialects of Spanish.

Missing data were minimal, accounting for 0.65% of overall values (0% on the LAS to 2.3% on the Composite Abuse Scale). Little's MCAR test was nonsignificant, X 2 (248) = 235.48, p = .706, suggesting that data were missing completely at random (Little, 1988) .

Exploratory factor analysis (EFA) using a tetrachoric correlation matrix ascertained latent constructs underlying the legal abuse items and determined the appropriate number of items on the measure. Because extant tested measures for legal abuse are limited, EFA is an appropriate analytic approach for the present study as it determines construct validity during initial scale development (Worthington & Whittaker, 2006) . Quantitative analyses were conducted primarily in STATA 16.0 (StataCorp, 2019) and jMetrik (Meyer, 2014) .

Prior to conducting a factor analysis, we ran correlations in order to determine whether the items were sufficiently correlated to justify using factor analysis (Watkins, 2018) . Item total correlations were significant and ranged from 0.28 to 0.68. Based on preliminary correlational analyses, we did not remove any items prior to conducting an initial EFA, as all correlated highly enough with the total scale to demonstrate the potential for factorability.

Bartlett's test of sphericity was significant (p < .001), suggesting that the correlation matrix was adequate for conducting an EFA (Watkins, 2018) . However, as Bartlett's test of sphericity is sensitive to sample size and likely to be significant in larger samples (Worthington & Whittaker, 2006) , we also examined the Kaiser-Meyer-Olkin (KMO) statistic. KMO (0.79 on a possible range of 0-1) indicated good sampling adequacy (Watkins, 2018) .

We used multiple methods and reasoned reflection to determine the appropriate number of factors to extract (Henson & Roberts, 2006) . Examining a scree plot (Cattell, 1966) , eigenvalues, parallel analysis, as well as theoretical interpretability, we considered various one, two, and three-factor solutions. Ultimately, a two-factor solution was supported. We used principal-axis factoring (PAF), as recommended for EFA and a Promax rotation, given our assumption that factors were correlated with one another (Watkins, 2018; Worthington & Whittaker, 2006) .

We conducted several EFAs with different items included. We examined both factor pattern coefficients (i.e., factor loadings) and internal consistency statistics while also considering theoretical interpretability. We prioritized the retention of items that experts rated as "Essential" during the item development phase, that existing literature highlighted as core features of legal abuse, that exemplified coercive control, and that evidenced factor pattern coefficients of 0.4 or higher (Worthington & Whittaker, 2006) . This process resulted in the elimination of 13 items.

The first factor, Harm to Self/Motherhood (i.e., using the court to harm the survivor as a person and a mother) consisted of 10 items. The second factor, Harm to Finances (i.e., using the court to harm the survivor financially) consisted of four items. The final two-factor solution accounted for 66.49% of the shared variance in the 14 items. Communalities were in the desirable range (0.54-0.80) (Worthington & Whittaker, 2006) . The absolute values of the two factor loadings across each item differed by at least 0.31. The 10 items for Harm to Self/Motherhood loaded most strongly onto the first factor, with factor pattern coefficients that ranged from 0.63 to 0.91. Was dishonest about your character or mental health to professionals on your case was the most strongly loading item. The four items for Harm to Finances loaded most strongly onto the second factor with factor pattern coefficients that ranged from 0.72 to 0.93. The other parent threatened to take control of all assets was the most strongly loading item. (For factor pattern coefficients for the final 14-item measure, see Table 2 ).

Subscale scores and a total scale score were computed by averaging item scores. Harm to Self/Motherhood and Harm to Finances correlated positively and significantly with one another (r = .36, p < .001), as expected. Further, the overall LAS showed good internal consistency (α = 0.84). Harm to Self/Motherhood also demonstrated good internal consistency (α = 0.85). Harm to Finances showed adequate internal consistency (α = 0.71).

Further supporting the measure's construct validity, its total scale score as well as each subscale correlated with other related concepts in the expected directions. Specifically, the Use of Children scale correlated positively and significantly with the total LAS (r = .42, p < .001), Harm to Self/Motherhood subscale (r = .42, p < .001), and Harm to Finances subscale (r = .23, p < .001). Further, prior psychological, physical, and sexual IPV from the other parent correlated positively and significantly with the total LAS (r = .38, p < .001), Harm to Self/Motherhood subscale (0.30, p < .001), and Harm to Finances subscale (r = .37, p < .001).

Rasch analysis (RA) is often used as a complement to classical test theory (CTT) in measure development and measure validation studies to further refine instruments (Heppner et al., 2015) and/or to provide exploratory information about item and measure performance (Linacre, 1994) . RA is advantageous for the present study as it is easily employed with binary data and provides specific information on item-level characteristics (DeVellis, 2017). As such, we analyzed the items to ascertain each item's location on the latent continuum that represents the construct of legal abuse (de Ayala, 2009) . Given that unidimensionality is an important assumption of RA and a multi-scale measure does not necessarily indicate that a measure is multidimensional , we first examined a principal components analysis (PCA) of the residuals with all 14 items in a single model. While multidimensionality always exists to varying degrees, we sought to determine whether the multidimensionality in the data are large enough to justify dividing items into separate tests (Linacre, 2012) . Although there are no set norms for determining unidimensionality (Linacre, 2012) , because the eigenvalue for the first contrast in the residuals was < 3 (Arcuri et al., 2015) and there were no misfitting items (Bond & Fox, 2015) , we present findings from an analysis with all items included in a single model. Infit and outfit statistics fell within the range (0.5-1.5) considered productive for measurement (Linacre, 2002) , indicating that the variance in the data were consistent with the Rasch model's prediction . Item z-scores, representing item endorsabilty (i.e., difficulty) indicated a balance of "easy" and "hard" items (i.e., 2.14 z-scores falling both above and below 0) (Mallinckrodt & Tekie, 2016) . See Table 2 for item z-scores as well as infit and outfit statistics. The item separation index (4.92) and reliability (0.96) indicated that the item difficulties can be rank ordered on the latent trait (Meyer, 2014) . The person separation index (1.51) and reliability (0.69) supported that the items could adequately discriminate between those with high and low levels of legal abuse (Fisher, 1992) . A Mantel-Haenszel differential item functioning (DIF) analysis tested whether participants responded to items differently based on whether their ex-partners identified as men or women. As no items reached significance, DIF was not supported.

Despite mounting concerns about the profound negative consequences of legal abuse for survivors (Douglas, 2020; Gutowski & Goodman, 2020; Rivera et al., 2018; Ward, 2016) , efforts to establish a valid measure that captures survivor-mothers' experiences with this form of abuse have been limited. To fill this critical gap, this study created the Legal Abuse Scale (LAS), a 14-item psychometrically sound measure of legal abuse designed to assess the perceptions of survivor-mothers involved in family law processes (see Appendix Tables 3 and 4 for English and Spanish versions of the LAS). The LAS is a novel contribution to the literature in that it builds on previous measures (e.g., Clemente et al., 2019; Hines et al., 2015; Rivera et al., 2018) , relies on extensive interviewing, piloting, and consultation with interdisciplinary experts, including survivor-mothers, and was evaluated in the context of a survey of survivor-mothers involved in family court. The 14-item LAS conceptualizes legal abuse as a form of coercive control and reflects the specific mechanisms through which ex-partners use post-separation legal processes to exert control over survivor-mothers. Factor analysis supported two subscales. The first, Harm to Self/ Motherhood, includes using in-person proceedings to cause distress, attacking custody and care time, threatening child(ren)'s safety, and denigrating the survivor's ability as a parent. The second, Harm to Finances, includes threats and actual attempts at taking control of assets and withholding finances (e.g., child support). Supporting the measure's validity, legal abuse and its subscales were significantly associated with IPV and use of children to perpetrate abuse.

While this study provides an original contribution to the literature in this area, it has several limitations that demand consideration, specifically regarding the sample. First, 77.5% of participants identified as White. The lack of ethno-racial diversity in this sample parallels an unfortunate trend in emerging research on this topic, as existing studies on survivor-mothers' family court experiences have relied on largely White samples. Many parents opt not to use the courts to settle custody and visitation and it is possible that many people of color favor making informal child custody and visitation arrangements given the history of racial oppression enacted by legal systems in the US. Second, given that only 20 participants in the sample had been partnered with women, it is possible that this measure does not capture the full range of experiences for sexual minority survivors, particularly as DIF is more likely to be detected with larger samples .

Additionally, the sample was a convenience sample and participants were referred by attorneys and practitioners who responded to postings on professional listservs. We attempted to mitigate the resulting potential bias by recruiting from multiple types of agencies and listservs (e.g., we were able to recruit survivors without legal representation by seeking referrals from domestic violence agencies). However, it is possible that our sample systematically leaves out a subset of family courtinvolved survivor-mothers who have not sought services from organizations or attorneys. Moreover, data collection occurred during the COVID-19 pandemic and it is possible that the onset of the pandemic influenced data collection. For example, participants who were overwhelmed with managing pandemic-related stress may have been less likely to complete the survey. Further, some domestic violence or legal services agencies, perhaps those serving more at-risk survivors, may have been less likely to refer survivors as they may have been adapting to the pandemic.

Another limitation is that like other studies of legal abuse, this one relies on self-report data. Self-report methods possess well-known disadvantages including that the data they generate are non-objective and have resulting validity concerns (Barker et al., 2016) . However, this kind of subjectivity is no more present in this study than in any other research that uses self-report assessments to identify abuse in a relationship.

Finally, as we surveyed survivors, many of whom were in the midst of stressful, ongoing family law cases, we took care to ensure that our data collection methods were trauma-informed. For this reason, we sought to limit the number of questions we included on the survey about traumatic experiences, such as past abuse. We were therefore unable to test the convergent and discriminant validity of the LAS more thoroughly.

The LAS makes possible a number of potential research directions. First, to refine the LAS using classical test theory (CTT), a confirmatory factor analysis (CFA) should be conducted to confirm the factor structure of this newly developed measure (Mallinkrodft et al., 2016) . Given that exploratory factor analysis (EFA) supported a two-factor solution and interpretations of unidimensionality indicators vary, future efforts to refine the LAS using item response theory (IRT) might consider models that do not assume unidimensionality (Harvey, 2016) even though proponents of Rasch analysis (RA) assert that it is robust against this assumption (Miles et al., 2016) . Second, it is critical to understand how survivors' intersectional identities shape the extent to which they are vulnerable to legal abuse. The LAS should be validated in the context of a range of survivor communities, including various subgroups of survivors of color and survivors who are trans and gender nonbinary. Third, the LAS' efficacy at detecting the presence of legal abuse and consequently aiding in family court cases should be rigorously evaluated. Important areas of exploration include the extent to which victims and perpetrators of legal abuse identify and report such experiences as well as which additional methods (e.g., interviews; third party accounts; other measures for abuse) may be employed in combination with the LAS to provide the most comprehensive assessment of its presence. Fourth, we need to know more about how the LAS operates outside of the specific family court setting. In some states, post-separation parenting determinations may be made in civil courts and it is important to determine whether legal abuse works similarly or differently in that context. Fifth, we need to know more about the prevalence, correlates, and outcomes of legal abuse. To what extent does it pervade the court system? Do prevalence rates, features, and tactics for managing legal abuse differ across regions? Is legal abuse related to ongoing severity and risk of physical violence towards the survivor or the children? And what is its relationship to specific harms to survivors and their children over time? Addressing such questions would be critical for violence prevention efforts.

Interdisciplinary practitioners such as lawyers, judges, and court evaluators have historically faced challenges identifying and addressing legal abuse (Laing & Heward-Belle, 2020) . By operationalizing the specific behaviors that constitute legal abuse, the LAS offers a tool that legal professionals can use to identify this form of coercive control, a first step toward putting an end to it. As others have recommended (Scott & Crooks, 2006) , supportive advocacy on behalf of survivor-mothers may be crucial to ensuring that they are not coerced into accepting terms that they feel are unsafe or tasked with holding abusive partners accountable. Advocates may use the LAS to detect whether legal abuse is a concern and modify their approach to supporting and advocating for survivors accordingly. Such modifications might include validating survivors' experiences, sharing information with them about legal abuse, and assisting them in taking appropriate action (e.g., finding peer support and mental health resources to manage the court process or assisting survivors in seeking resources to build their defense in states where abusive litigation is against the law). Findings from the RA suggest that using all items when administering the scale is advantageous in order to ensure adequate validity.

Although the establishment of the LAS represents a critical step forward, we want to caution against its use as a stand-alone tool: There is danger that the LAS itself could be used as a tool of legal abuse if, for example, the abusive partner endorses items so as to suggest that they, rather than the survivor, are the true victim of coercive control in the courtroom. Such cross-allegations can leave survivors extraordinarily disempowered, particularly women of color who have reported being criminalized in the context of helpseeking (Durfee & Goodmark, 2021) .

To guard against misuse, we suggest two strategies: First, the LAS should be used in combination with other evaluation methods (e.g., interviews, self-report measures and third-party accounts) to determine whether legal abuse is present. This kind of broad evaluation paves the way for the second strategy -to place LAS results in the context of a clear picture of the dynamics of social power of those involved. As noted earlier, coercive and controlling tactics "work" to establish domination only to the extent that they reflect a power imbalance in the relationship supported by existing social structures (Anderson, 2009; Ashcroft, 2000; Dutton & Goodman, 2005; Stark, 2007) . For example, in heterosexual relationships, patriarchal social structures, gendered role expectations, and stereotypes about women make it easier for men to establish a regime of control than the other way around. In such a relationship, if both partners were to endorse the item "was dishonest about your character or mental health" on the LAS, one would need to view this report in light of a longstanding history of devaluing femininity on the basis of its association with emotionality and irrationality in one's broader social context (Anderson, 2009) . It would also be critical to understand how those associations may be internalized in the parties' identities, performed through their interactions, and interpreted in their environments (including in court) (Anderson, 2009 ). At the same time, it would also be important to hold the awareness that neither abuse perpetration nor victimization are restricted to specific gender identities. Indeed, no research to date has been conducted that establishes the LAS' ability to determine which partner is the primary aggressor.

Ultimately, such an analysis would require careful consideration of multiple sources of information.

Social power plays a critical role in family court: Epstein and Goodman (2019) describe the ways women, and especially women of color, may be subject to dangerous stereotypes such as the hysterical White woman, the angry Black woman, the gold digger, the manipulator in a way that damages their credibility and makes them further vulnerable to a partner's control. Indeed, support for women's vulnerability to mischaracterization in court comes from one study suggesting that mothers, more than fathers, are at risk of losing custody when fathers claim parental alienation (i.e. that the mother is attempting to hurt his relationship with the children) (Meier, 2020) . Given the implication of this finding, -that fathers may be able to effectively leverage their social power to maintain control -it is critical for practitioners to accurately identify legal abuse to avoid potentially devastating outcomes. Thus, it is crucial that professionals evaluate LAS results in the context of the intersecting identities and corresponding social power of both parties involved.

Protective mothers who have survived IPV have been disclosing experiences of legal abuse and its devastating consequences for themselves and their children to journalists and qualitative researchers for decades (Coy et al., 2015; Douglas, 2018a; Elizabeth, 2017; Miller & Smolter, 2011; Rivera et al., 2018; Watson & Ancis, 2013) . However, until now, existing attempts to define and measure legal abuse have been scarce and limited. This study offers a psychometrically valid measure of legal abuse that conceptualizes this construct as a form of coercive control and reflects the specific mechanisms through which the legal process becomes another realm in which partners coerce and control survivor-mothers who have exited the relationship. By illuminating the often unacknowledged abuse dynamics taking place in family courts across the country, this new measure expands opportunities to further uncover the features and costs of legal abuse and address the urgent need to make these dynamics explicit in our courtrooms. Table 3 Legal Abuse Scale (LAS)-English Table 4 Legal Abuse Scale (LAS) -Spanish

Actually went to court to take out a restraining order (i.e., order of protection) against you without cause

Tried to prolong the legal process. (For example, switched lawyers, was late with paperwork)

Tried to scare you in the court house. (For example, by using words, looks, or by getting in your space)

Was dishonest about your character or mental health to professionals on your case (e.g., judges, mediators, evaluators etc

Was dishonest about you using alcohol or drugs to professionals on your case (e.g., judges, mediators, evaluators etc

Was dishonest about your ability as a mother to professionals on your case (e.g., judges, mediators, evaluators etc

Used a social identity of yours (e.g., race, ethnicity, religion, social class, sexual orientation) or your immigration status against you in your case

that women should not work and/or should be home with their kids) against you in your case

Development of the scale of economic abuse

Gendering coercive control

Exploring the measurement properties of the Montreal Cognitive Assessment in a population of people with cancer

Naming knowledge: A language for reconstructing domestic violence and systemic gender inequity. Women and Language

The batterer as parent: Addressing the impact of domestic violence on family dynamics

The scree test for the number of factors. Multivariate Behavioral Research

Intimate partner violence

Constructing validity: Basic issues in objective scale development

Institutional violence against users of the family law courts and the legal harassment scale

It's like going through the abuse again': domestic violence and women and children's (un) safety in private law contact proceedings

Demarginalizing the intersection of race and sex: A black feminist critique of antidiscrimination doctrine, feminist theory and antiracist politics

Abusive endings: Separation and divorce violence against women

Domestic and family violence, mental health and well-being, and legal engagement

Family violence, lawyers and debt

Is there a protection order to prison pipeline? Gendered dimensions of cross-petitions

Observing coercive control beyond intimate partner violence: Examining the perceptions of professionals about common tactics used in victimization

Coercion in intimate partner violence: Toward a new conceptualization

From domestic violence to coercive control: Towards the recognition of oppressive intimacy in the Family Court

Custody stalking: A mechanism of coercively controlling mothers following separation

It's an invisible wound': the disenfranchised grief of post-separation mothers who lose care time

Discounting women: doubting domestic violence survivors' credibility and dismissing their experiences

Reliability statistics

Development of a brief measure of intimate partner violence experiences: The Composite Abuse Scale (Revised)-Short Form

Like I'm invisible": The experiences of mothers who survived intimate partner violence seeking child custody through the family court system

Separation assault in the context of post-divorce parenting: An integrative review of the literature

How women make custody decisions and manage co-parenting with abusive former husbands

Improving measurement via item response theory: Great idea, but hold the rasch

Custody evaluators' beliefs about domestic violence allegations during divorce: Feminist and family violence perspectives

Civil court responses to intimate partner violence and child custody: The role of custody evaluators

Indirect abuse involving children during the separation process

Use of exploratory factor analysis in published research: Common errors and some comment on improved practice

A self-report measure of legal and administrative aggression within intimate relationships

Sample size and item calibration stability

What do infit and outfit, mean-square and standardized mean

Winsteps® Rasch measurement computer program user's guide. Winsteps.com Little, R

Using focus groups and Rasch item response theory to improve instrument development

Item response theory analysis of Working Alliance Inventory, revised response format, and new Brief Alliance Inventory

A historical perspective on parental alienation syndrome and parental alienation

Mapping gender: Shedding empirical light on family courts' treatment of cases involving abuse and alienation

US child custody outcomes in cases involving parental alienation and abuse allegations: What do the data show

Denial of family violence in court: an empirical analysis and path forward for family law

Reconsidering focus groups and Rasch model item response theory in instrument development

Safeguarding children's wellbeing: Voices from abused mothers navigating their relationships and the civil courts

Paper abuse: When all else fails, batterers use procedural stalking

One more battleground: Domestic violence, child custody, and the batterers' relentless pursuit of their victims through the courts

A longitudinal examination of mothers' depression and PTSD symptoms as impacted by partner-abusive men's harm to their children

Factors associated with child custody evaluators' recommendations in cases of intimate partner violence

Intervention for abusive fathers: Promising practices in court and community responses

The global prevalence of intimate partner homicide: a systematic review

I have lost everything": Trade-offs of seeking safety from intimate partner violence

I was naive in thinking, 'I divorced this man, he is out of my life'": A qualitative exploration of post-separation power and control tactics experienced by women

If I killed you, I'd get the kids": Women's survival and protection work with child custody and access in the context of woman abuse

An integrative review of separation in the context of victimization: Consequences and implications for women

In her words: Recognizing and preventing abusive litigation against domestic violence survivors

Exploratory factor analysis: A guide to best practice

Power and control in the legal system: From marriage/relationship to divorce and custody

The mental health factor in domestic violence custody cases: Results from a brief survey of lawyers who represent DV survivors

Scale development research: A content analysis and recommendations for best practices

Post-separation abuse of women and their children: Boundarysetting and family court utilization among victimized mothers

The authors would like to thank Erin C. Miller, David L. Blustein and Jessica L. Shaw for their support and guidance; the many experts and community partners who offered expertise; Meg Riley for her research assistance; Newton Wellesley Hospital, the Larry J. Seigel Fellowship for Victimology Studies, the Center for Human Rights and International Justice, and Boston College for their financial support; the reviewers for their helpful feedback; and the participants who shared their experiences with the hope of helping someone else.

Conflict of interest None of the authors have any conflicts of interest to report.

Please mark whether the other parent did the following actions:1. Threatened to use the court to take custody of your children away from you.□Yes, the other parent did this □No, the other parent did not do this 2. Actually took you to court to take custody of your children away from you.□ Yes, the other parent did this □No, the other parent did not do this 3. Threatened to use the court to get unsafe access to your children.□ Yes, the other parent did this □ No, the other parent did not do this 4. Actually took you to court to get unsafe access to your children.□ Yes, the other parent did this □ No, the other parent did not do this 5. Threatened to use the court to punish you.□Yes, the other parent did this □ No, the other parent did not do this 6. Took you to court repeatedly.□ Yes, the other parent did this □ No, the other parent did not do this 7. Took you to court when the only clear reason was to cause you distress.□ Yes, the other parent did this □ No, the other parent did not do this 8. Was dishonest about your character or mental health to professionals on your case (e.g., judges, mediators, evaluators, mental health professionals etc.).□ Yes, the other parent did this □No, the other parent did not do this 9. Was dishonest about your ability as a mother to professionals on your case (e.g., judges, mediators, evaluators, mental health professionals etc.).□ Yes, the other parent did this □ No, the other parent did not do this 10. Told professionals on your case (e.g., judges, mediators, evaluators, mental health professionals etc.) that you are trying to harm their relationship with the children.□ Yes, the other parent did this □ No, the other parent did not do this 11. Threatened to withhold financial support.□ Yes, the other parent did this □ No, the other parent did not do this 12. Actually withheld financial support.□Yes, the other parent did this □ No, the other parent did not do this 13. Threatened to take control of all assets.□ Yes, the other parent did this □ No, the other parent did not do this 14. Actually took you to court to take control of all assets.□ Yes, the other parent did this □ No, the other parent did not do this

Por favor seleccione si el otro padre/madre hizo las siguientes acciones:1. La amenazó con ir a la corte de familia para quitarle la custodia de su(s) hijo(s). □ Sí, lo hizo □ No, nunca lo hizo 2. La demandó con el fin de quitarle la custodia de su(s) hijo(s).□ Sí, lo hizo □ No, nunca lo hizo 3. La amenazó con usar la corte para contactar y visitar a su(s) hijo(s) de formas no muy seguras.□ Sí, lo hizo □ No, nunca lo hizo 4. Efectivamete la llevó a usted ante la corte para poder contactar y visitar a su(s) hijo(s) sin autorización, poniéndola a usted y a su(s) hijo(s) en riesgo.□ Sí, lo hizo □ No, nunca lo hizo 5. La amenazó con usar la corte para castigarla.□ Sí, lo hizo □ No, nunca lo hizo 6. La llevó ante la corte en varias ocasiones.□ Sí, lo hizo □ No, nunca lo hizo 7. La llevó ante la corte únicamente para causarle angustia.□ Sí, lo hizo □ No, nunca lo hizo 8. Le mintió a profesionales que trabajaban en su caso sobre su carácter o su salud mental (por ejemplo, a jueces, mediadores, evaluadores, etc.).□ Sí, lo hizo □ No, nunca lo hizo 9. Le mintió a los profesionales que trabajaban en su caso sobre su habilidad como madre (por ejemplo, a jueces, mediadores, evaluadores, etc.).□ Sí, lo hizo □ No, nunca lo hizo 10. Les dijo a profesionales que trabajaban en su caso (por ejemplo, a jueces, mediadores, evaluadores, etc.). que usted estaba tratando de dañar la relación de él/ella con los niños. Please mark whether the other parent did the following actions: