*3.3. Instrumentation*

The research team developed the survey collaboratively with expert consultants from the International Women's House (IWH) in Atlanta, Georgia. IWH has been operating for more than 20 years, providing a safe haven and supportive services for women and children who are victims of family violence, sexual abuse, and human tra fficking and focusing specifically on serving immigrant and refugee women and children with cultural sensitivity. In addition, Calzada and Suarez-Balcazar (2014) model of cultural responsiveness was used to develop assessment questions that queried the executive directors of domestic violence organizations regarding their organization's culturally responsive practices across system levels. Upon Institutional Review Board approval, the survey was administered to executive directors via a Qualtrics online survey. Overall, the survey consisted of 77 questions, some of which were open-ended, and most of which were Likert scale questions related to the frequency with which various culturally responsive policies, practices, and approaches were enacted in the agency. Finally, the survey collected demographic and background information from the executive director and the agency itself.

The majority of survey questions that assessed agency cultural responsiveness fell into the following six domains: (1) agency values and practices (9 items, α = 0.903); (2) agency sta ffing and programmatic practices (9 items, α = 0.899); (3) interagency collaboration and outreach (6 items, α = 0.783); (4) agency systems and processes for assessment (4 items, α = 0.846); (5) program and shelter accommodations and adaptations (5 items, α = 0.688); and (6) agency approaches to language concerns (9 items, α = 0.714). These Likert scale questions were developed using Calzada and Suarez-Balcazar's assessment tool in consultation with IWH expert input. For bivariate analyses, single items in each domain were averaged to reflect a composite score. High levels of reliability indicate the appropriateness of treating these domains as larger constructs for correlation analyses. Specific items can be seen in Table 2 below.




Notes: Respondents rated each item on a scale from 1 = never to 4 = often.

### *3.4. Data Analysis*

Given the exploratory nature of this study, the data analysis presented here is limited to descriptive and bivariate analyses. Frequency distributions and descriptive statistics are reported visually and descriptively below. In addition, correlation analyses that examined the relationship between agency characteristics and the aforementioned six domains of cultural responsiveness are reported.
