In Massachusetts, four dioceses serve an estimated 430,000 Latino Catholics. There are 577 Catholic parishes in Massachusetts, but only 12% have a formal ministry or parish programming for the Spanish-speaking Latino community. To account for both the hierarchical organizational structure in Catholic parish and the uniqueness of parish communities serving Latinos, we implemented a multi-pronged approach for recruitment and survey administration. The Institutional Review Board at the Harvard School of Public Health approved all study procedures.

We were guided by principles of community-based participatory research (Minkler & Wallerstein, 2010) in all study phases. At the start, we formed a Community Advisory Board (CAB) with representatives from faith-based, health, and social service organizations. We also identified contextually appropriate strategies for engaging faith communities by conducting 18 key informant interviews among Latino community leaders and eight focus groups with a total of 67 parishioners (Allen et al., In press). We used this formative research in combination with ongoing input from CAB members to develop and adapt strategies for continuous community and leadership engagement in research activities.

As a means of educating Catholic leaders about the study, we formally launched CRUZA with an informational gathering at the Archdiocese’s Pastoral Center. At this formal event attended by over 30 Catholic leaders from several dioceses in the state, investigators introduced the study, described procedures, elicited opinions, and addressed questions. The Archbishop of Boston, the highest ranking Catholic leader in the State, attended this event and verbalized his support for the CRUZA initiative. The gathering was an extensive event planning effort by our study team made possible by our CAB’s connections. Following the gathering, we presented the study at Catholic regional events and met individually with heads of the four Catholic dioceses in the state, as well as diocesan leaders of Hispanic ministry. Combined, these early efforts established CRUZA’s credibility prior to recruiting individual parishes through the combined effects of formal events, dissemination of written material, support from high-level church leaders, as well as dissemination of information about the study through pastors’ social networks.

Our conceptual framework for the larger CRUZA study guided the selection of constructs to be assessed in organizational surveys. The framework and its application is described in detail elsewhere (Allen et al., Under Review). Briefly, we sought to assess organizational characteristics specified in the Consolidated Framework for Implementation Research (CFIR) (Damschroder et al., 2009). CFIR is a meta-theoretical implementation framework that integrates constructs from relevant theories to understand factors associated with adoption/implementation of innovations or new programs, practices or policies. The framework addresses multiple domains, including internal organizational characteristics and dynamics (“inner organizational setting”), factors external to the organization that may affect willingness to adopt and implement innovations (“outer setting”), characteristics of those charged with adopting and implementing interventions.

Due to the lack of standardized measures relevant for a FBO setting, we adapted items from other studies (Allen et al., Under Review). We conducted cognitive testing among parish leaders (N=5) outside our sample to ensure comprehension and cultural/religious appropriateness of the adapted organizational measures. For each item, parish leaders were asked whether the question was easy or hard to answer, to rephrase the question in their own words, and how sure they were of their answer. Following this process, items were refined and subsequently re-tested in pilot administrations of the full instrument with two additional parish leaders. The revised survey instrument was reviewed by the CAB, our Scientific Advisory Committee, and the Archdiocese of Boston. The instrument was then translated into Spanish by a certified translator employed by the Archdiocese and reviewed by bilingual investigators for accuracy.

The organizational survey instrument consisted of four sections differing by content and intended respondent: A – leadership (Pastor), B – bookkeeping (Business Manager), C – Hispanic ministry (Director of Hispanic Ministry), and D – health/social services (parish nurse or representative from health ministry, if available). We targeted different respondents for each section to maximize response accuracy and reduce respondent burden on any one individual.

Massachusetts parishes eligible for the organizational survey offered at least one mass in Spanish and were not slated for closure or merger during the study period. We compiled initial lists of parishes and pastors by searching print and online archives of the four dioceses, reviewing parish websites, and then making scripted calls to parishes to verify mailing addresses and pastor names. These lists were reviewed by diocesan leaders, who noted additional parish closures and consolidations. Through this process, we identified 70 potentially eligible parishes. We relied on pastors to recommend appropriate personnel to complete each survey section.

Our parish survey recruitment and data collection process is outlined in Figure 1. Between July – August 2012, we mailed recruitment packets in four waves (by region) to pastors of the 70 parishes. Recruitment in waves allowed for more efficient management and promotion efforts. Recruitment packets contained: (1) a letter from the regional bishop encouraging pastors to participate; (2) an excerpt from a blog entry supporting CRUZA written by the Archbishop; (3) a project brochure describing study information tailored with the name of the parish and pastor; and (4) a contact information form and an accompanying postage-paid return envelope on which pastors could identify the appropriate parish representative to complete each survey section. These materials made explicit connections between CRUZA’s goals and Catholic leaders’ commitments to service and holistic care for their parishioners. For example, the bishops’ letters used this common line: “Our call to discipleship includes having concern for the spiritual, physical and mental health of our brother and sisters in the Church and the wider community.” Other materials highlighted: “one of the noblest traditions of the Catholic Church is to care physically and spiritually for its people when they are in need.” Our goal in integrating spiritual themes was to make our messages appropriate for a religious setting.

Approximately two weeks following the mailing of the recruitment packet, interviewers from the Center for Survey Research at UMass Boston and trained bilingual Survey Assistants initiated recruitment calls to pastors. As is standard in survey research, calls were made at different times of day, on different days of the week. When we reached a pastor, we used a standardized protocol and script to: (1) describe the study; (2) inquire if he received the recruitment packet; (3) assess interest in study participation; (4) identify the most appropriate person(s) to complete each survey; and (5) schedule or complete an interview. If the phone was answered but the pastor was unavailable, we asked whether the recruitment packet was received and the best day/time to reach the pastor. If the packet had not been received, we mailed another recruitment packet and called again in 3–7 days.

Our first round of calls were fielded primarily by parish administrative assistants who suggested that pastors were likely to be in the office on the days of regularly scheduled Mass. Beginning in August 2012, Survey Assistants visited pastors. This was done after telephone attempts to contact the pastor were unsuccessful for a one month period. In-person conversations with pastors were scripted and covered the same content as the telephone recruitment protocol. When the pastor encouraged us to contact alternative parish representatives, we used the same telephone scripts and in-person recruitment protocols.

Due to pastors’ responsibilities (e.g., funerals) and the part-time or volunteer status of many parish staff members, potential respondents often did not have established on-site hours. To speak with each potential respondent, we conducted up to five “study staff initiated contacts,” which we defined as any form of direct communication (e.g., face-to-face or phone conversation) but not initiated by the respondent. Responding to an inquiry or confirming an appointment was not categorized as a study staff initiated contact. However, any form of communication was logged as a “contact attempt.”

Bilingual Survey Assistants conducted recruitment calls/visits and interviews between July and December 2012. For each survey section, we aimed to obtain informed consent and complete the interview in less than 20 minutes, or 60 minutes when one respondent elected to complete all survey segments.

Several parish leaders were never reached, despite multiple telephone and in-person contact attempts. A few pastors preferred to complete the survey independently. In these cases, we provided a paper version of the survey. Upon completion of any or all survey segments, a $50 gift card donation was sent to the parish.

Survey Assistants used REDCap (Research Electronic Data Capture) software (Harris et al., 2009) to track call attempts, call dispositions, and in-person contacts. For every contact attempt, we documented date, time of day, contact method (e.g., phone call, visit) and outcome of the contact (e.g., completed survey). For surveys administered in person on site at parishes, they completed field notes documenting overall receptivity to study participation, as well as time required for survey completion. Organizational survey data were entered into DatStat Illume (DatStat, Inc., Seattle, WA), a multi-mode survey platform (DatStat Inc., 2010).

We present descriptive statistics for recruitment and survey administration activities. The final dispositions for each parish were compared by diocese (geographic location). In addition, we examined differences among respondent types (e.g., pastoral versus non-pastoral staff) and the number of contact attempts required to complete a section, as well as the number of sections completed per respondent. Since the number of parishes that completed the survey is small (N=45), we were unable to conduct multivariate analyses to assess specific organizational characteristics associated with survey completion.

After the initial mailing of recruitment materials, we discovered that five parishes in the initial sample were ineligible: three had recently closed and two were designated as chapels, rather than parishes. This reduced our final sample to 65 parishes. We achieved a 75% (49/65) participation rate, with 92% (45/49) of participating parishes completing all four sections of the survey (Table 1). Only three parishes declined to participate (4.6%), citing time and conflicting priorities. While the highest participation rate was among churches in the Archdiocese of Boston (participation rate 78%, n = 29), regional differences were not statistically significant (χ² = 6.42, p = .09).

The number of contact attempts required to complete a single section of the survey ranged from 5 to 34, with an average of 16.6 contact attempts to complete all four survey segments with four different individuals (Figure 1). Pastors were relied on to name appropriate personnel to complete various survey sections, but they were also difficult to reach, requiring an average of 9.3 contact attempts to complete a survey section, compared to 7.8 attempts for other members of the pastoral team, although differences were not statistically significant (p = .22). In-person parish visits by Survey Assistants resulted in relatively high interview rates; nearly a quarter of all interviews were conducted on site, compared with 71% by phone and 5% by mail. Respondents who elected to complete all four components of the survey at one time (n=10) required an average of 64 minutes (SD = 10.1) to complete the process.

Most often, survey respondents were members of the pastoral staff (79%), which included pastors, priests, deacons, lay ecclesial ministers, and other clergy (Table 2). Pastoral staff were also more likely than other types of respondents to complete multiple sections of the survey; they completed an average of 2.0 survey sections, while non-pastoral staff (e.g., administrative assistants, business managers) completed an average of 1.2 survey sections (p < .001). In several instances, members of the pastoral staff completed all four survey sections (32%, n = 14).

Participating parishes were heterogeneous. In particular, parish size ranged from small to very large (60 to 7741 people). The proportion of Latino members ranged from 1% to 100%, with a mean of 46%. The number of weekly masses offered in Spanish also varied from 1 to 11 (M=2.7). Parish size was strongly positively correlated with the number of full-time staff (r = 0.433, p = 0.004) and negatively correlated with the percentage of the congregation that was Latino, although that relationship was not statistically significant (r = −0.05, p = 0.77). Our ability to compare characteristics of participating and non-participating parishes was limited by the lack of comparable, publically available data (see Table 3).

We rigorously identified parishes, made consistent efforts to engage community partners, and utilized multiple recruitment and data collection strategies. This study provides data on the research activities that may be necessary to yield high organizational participation among Latino faith communities. Our detailed description of survey administration procedures may be useful for future research in faith-based settings.