Initial Literature Screening, Assessment, & Organization

Each of the 106 articles returned from the database searches was assessed according to three successive levels of inclusion criteria. The first and most basic inclusion criterion is that the articles must present U.S.-based empirical research. Second, the content of each article should be directly applicable to intimate partner violence interventions (e.g., intervention program development, implementation, and evaluation research; applied community-based research methods; etc.). To assess the second criterion, I coded each article using keywords from the authors’ description of substantive research topics covered. This information was obtained by first examining each article’s title and, if available, abstract or keywords list; and then reviewing the stated purposes, research questions, or hypotheses provided in the main body of each article. The latter step was done to ensure that the coding of each article’s focal research topics was both accurate and comprehensive, as this coding was subsequently used as a “filter” for excluding articles not relevant to IPV interventions. Specifically, I examined the list of codes generated from this process (see codes listed under the “Research Topic” category in Table 4) and selected the following codes that fit within the scope of this review’s previously-described goals and research questions:

• Macro Contexts of IPV
• Coordinated Community Response to IPV
• Community Capacity
• Key Stakeholders’ Perspectives
• Measures
• System Response
• IPV Intervention Evaluation
• IPV Intervention Efficacy
• IPV Intervention Proposal
• IPV Intervention Development
• Public/Program Policy
• Research/Evaluation Methods
• IPV Intervention Description
• Approach Evaluation
• Practitioners’ Perspectives
• IPV Perpetrator Interventions
• IPV Victimization Intervention
• Prevention
• Intervention
• Bystander Intervention

Finally, a particularly crucial inclusion criteria for literature obtained from the fourth through sixth database searches (see Table 1) is that the presented research must include sexual minority women as a focal population. For example, among the two empirical research articles returned from the fifth and sixth literature searches, which targeted SMW-inclusive IPV intervention and prevention research (see Table 1), one examined male and female college students’ perceptions of the prevalence of IPV perpetration among their peers (Witte, Mulla, & Weaver, 2015), and the other detailed an item-response theory approach to measuring teens’ attitudes about dating violence among heterosexual adolescents (Edelen, McCaffrey, Marshall, & Jaycox, 2009). While both of these studies somewhat reflect research aligned with community psychological values and methods, their substantive foci failed to meet this review’s inclusion criteria for articles returned from the three database searches specific to sexual minority women.

Any articles not meeting the above-described three levels of inclusion criteria were excluded from remainder of analyses conducted for this review. This initial assessment and filtering analytic phase yielded a final set of 29 empirical studies included in the below-described analyses and later-provided synthesis and methodological critique. Among these studies, 22 are exclusively focused on or applicable only to heterosexual, and typically male-identified, populations; whereas 7 studies are specifically inclusive of female-identified lesbian, gay, bisexual, transgender, and/or queer groups and individuals (i.e., sexual minority women [“SMW”]).

Synthesis & Methodological Evaluation

The second phase of deductive analysis served to examine the similarities, differences, and anomalies among the set of included articles according to the above-described two categories determined in the previous data reduction and organization process. Following a qualitative comparative analytic (QCA) approach (Leech & Onwuegbuzie, 2007; Onwuegbuzie & Weinbaum, 2017; Onwuegbuzie, Dickinson, Leech, & Zoran, 2009), I first reviewed each article in its entirety using a literature description and data extraction form (available here) that I created for this project. I developed this form with specific guidance from the systematic literature review and data extraction standards, protocols, templates, and guidelines provided by The PRIMSA Group (Moher, Liberati, Tetzlaff, Altman, & The PRISMA Group, 2009) and the Cochrane Collaboration’s resources library and handbook for systematic reviews and meta-analyses (Higgins & Green, 2011). The literature description form consists of sections for recording each reviewed study’s substantive research topic(s) (determined by the stated purpose(s), research question(s), and/or hypotheses, as provided in each article), target population(s)), target population(s), sampling frame setting(s), sampling frame(s), sampling methods, research design (e.g., experimental, cross-sectional, descriptive, etc.), overarching methodology (i.e., qualitative, quantitative, or mixed-methods), specific data collection methods (e.g., measures, interview questions, archival data sources, etc.), data analytic approaches and procedures, and key findings. The final piece of information included in the form is the ecological levels of analysis involved in a study’s design, methods, and analysis. The ecological levels of analysis section was intentionally placed at the end of the description form so that the author could determine which levels of analysis were involved in each study based on the characteristics of the full array of information and study characteristics recorded in the prior sections of the form. That is, the ecological levels of analysis involved in each of the reviewed studies were inferred based on the other data points recorded in the literature description form for each reviewed study, all of which were directly extracted from the information provided in each article. The summative data across the reviewed studies were then compiled and restructured to form a methodologically-focused codebook, presented in Table 4, consisting of (1) information categories corresponding to the research design and methods-related sections covered in the literature description form, and (2) the codes generated in each information category from the discrete characteristics recorded in the description form across the reviewed studies.

Methodological Review & Evaluation

Research conducted within the subset of community psychology focused around intimate partner violence was initially evaluated according to the level of inclusion and exclusion of the historically marginalized population of particular interest for the purposes of this review: sexual minority women (SMW). The implementation of action-oriented community-psychology methodologies and analytic approaches was then reviewed within each of these categories (i.e., inclusion or exclusion of SMW) in terms of (1) the appropriateness of the methods to the research question(s), (2) how the methods facilitated the inclusion or exclusion of sexual minority women, and (3) whether and how (where applicable) exclusion of sexual minority women is justified.

Defining and Evaluating Methodological Rigor

The notion of “scientific rigor” is particularly important in evaluating a set of applied social science research studies employing varying methodologies. This is because “rigor” is not necessarily clearly defined across methodologies, and some research philosophies do not consider certain methodologies as capable of achieving scientific rigor at all. For the purposes of this review, methodological rigor is broadly defined as the extent to which a given study employs research methods that best address the overarching research question(s) and/or hypothesis/hypotheses, and the quality of the implementation of those methods. This definition of methodological rigor therefore does not prioritize a particular method nor category of methods (e.g., quantitative over qualitative, mixed-methods over single-methods, etc.), but rather prioritizes the choice and implementation of methods employed to address a given study’s purposes. In addition, this definition of methodological rigor allows for more specific and in-depth consideration of the limitations presented, and not presented, in each study’s report.

Rigor in action-oriented community-based research methods was evaluated according to the choice, description, justification, appropriateness, and execution of each reviewed study’s design (i.e., experimental or cross-sectional) and overarching methodology (i.e., quantitative, qualitative, or mixed-methods). Each of these aspects of the reviewed literature was identified in terms of each evaluated study’s purpose, hypotheses, and/or research questions description, sampling frame definition, sampling and data collection methods, analytic approach(es), and description of findings and limitations. Because the underlying goals of this systematic review are motivated by the continued relative absence of sexual minority women in the larger body of IPV-related empirical research literature, especially in terms of IPV perpetrator interventions research, particular attention was given to the sampling frame definitions and sampling methods employed among the empirical studies reviewed here. That is, the evaluated research included in the present review was specifically evaluated in terms of how the methods facilitate the inclusion or exclusion of specific populations, particularly sexual minority women, and whether and how the exclusion of specific populations is justified in each empirical study’s report.

In addition, to assess transparency and reproducibility of the research methods and findings, the overall presentation, dissemination mechanism(s), application, and accessibility of the research was noted where available or applicable. These latter assessments of research transparency and reproducibility also incorporated considerations regarding the incorporation of key and/or distal stakeholders’ input. The accessibility of the research to primary and distal stakeholders would be reflected according to whether and how stakeholders are provided information about and access to reports of a given project’s progress and findings. The role of stakeholder input in the evaluated community-psychology literature was noted in terms of the extent to which efforts made to ensure that all available stakeholders’ and informants’ voices are considered throughout the research process, and that certain voices are not unjustifiably privileged over others. These considerations include evaluation of whether feedback is explicitly solicited, or at least accepted when offered, from key stakeholders and informants to the research, and whether such feedback is genuinely considered by the primary researchers of a given investigation.

This document was created using R-v3.3.3 (R Core Team, 2016), and the following R-packages:

base-v3.3. (R Core Team, 2016), bibtex-v0.4. (Francois, 2014), cairoDevice-v2.24. (Lawrence, 2015), car-v2.1. (Fox & Weisberg, 2011), DBI-v0.7. (Databases, 2014), dplyr-v0.5. (Wickham & Francois, 2015), extrafont-v0.17. (Chang, 2014), ggplot2-v2.2. (Wickham, 2009), gWidgets-v0.0. (Verzani, 2014), gWidgetsRGtk2-v0.0. (Lawrence & Verzani, 2014), igraph-v1.0. (Csardi & Nepusz, 2006), knitr-v1.16. (Xie, 2015), pander-v0.6. (Daroczi & Tsegelskyi, 2015), plyr-v1.8. (Wickham, 2011), RGtk2-v2.20. (Lawrence & Temple Lang, 2010), rmarkdown-v1.5. (J. Allaire et al., 2016), RQDA-v0.2. (Huang, 2014), RSQLite-v1.0. (Wickham, James, & Falcon, 2014), scales-v0.4. (Wickham, 2016a), tidyr-v0.6. (Wickham, 2016b), ggthemes-v3.4. (Arnold, 2016), kableExtra-v0.2. (Zhu, 2016), tufte-v0.2. (Xie & Allaire, 2016), devtools-v1.13. (Wickham & Chang, 2016), sysfonts-v0.5. (Qiu & others, 2015), showtext-v0.4. (Qiu, 2015), Riley-v0.1. (Smith-Hunter, 2017), and stats-v3.3. (Team & worldwide, 2017)