Diversity and Inclusion Practices in Nonprofit Associations: A Resource- Dependent and Institutional Analysis

Since the early 1990s, diversity and inclusion (D&I) efforts have received increasing attention among management scholars. Although the benefits and challenges of implementing D&I practices are now well established, few studies have explored the extent to which nonprofit associations, in particular, engage in D&I efforts. As such, we have no knowledge of the role that associations play in assisting with the diffusion of these practices throughout their respective professional fields or trades. Therefore, using a national survey of over 150 executives of nonprofit associations, this study explores the institutional and resource-based challenges associations face when seeking to implement D&I practices, both within their organizations and throughout their professional fields and trades. While the findings from this study suggest that nonprofit associations only engage in D&I practices to a modest degree, there is also evidence of institutional entrepreneurship. Implications of these findings for research on D&I practices in nonprofit associations as well as for association practitioners seeking to improve their D&I programming are provided.

23 associations, in particular, develop and support D&I programs and activities, both within their organizations and within their broader professional fields and trades.
Nonprofit associations are member-based organizations that help to educate, train, and credential individuals and organizations in a particular industry, trade, or profession . These organizations are also repositories of best practices since they often provide resources and technical support as well as advance policy, technology, and operational practices (Balla, 2001;Gazley, Tschirhart, & Hager, 2014;Greenwood, Suddaby, & Hinings, 2002;Hager, 2014).
Despite being recognized as important information brokers and leaders in their professions, little is known about D&I efforts in nonprofit associations and even less is known about their D&I influence in their respective fields and industries. Although there has been recent research on various aspects of nonprofit associations (e.g., see Gazley et al., 2014;Hager, 2014;Nesbit & Gazley, 2012), to date there has been no research assessing how these organizations have an impact on D&I practices throughout their workforce.
Most extant literature on diversity in nonprofits, writ large, has focused primarily on board member characteristics and the relationship between diversity and board outcomes (e.g., see Bernstein & Bilimoria, 2013;Brown, 2015;Brown, 2002;Buse et al., 2016;Gazley et al., 2010;Gündemir et al., 2017). This focus has emerged, in part, because nonprofit organizations have often struggled with closing their "diversity gap" (Medina, 2017) where they, themselves, are primarily homogenous while serving diverse populations.
Using data from a nationwide survey of association executives, the present study is the first of its kind to measure D&I practices within nonprofit associations specifically; more broadly, this study is one of the few to measure D&I within nonprofit organizations in general. The study builds on earlier research that has explored D&I practices in human resources management (Konrad, Yang, & Maurer, 2016;Yang & Konrad, 2011). This earlier work defined various D&I practices and described institutional and resource-based constraints and opportunities that influence the adoption of these practices.
Findings from the present study suggest that while nonprofit associations tend to engage in D&I practices at modest levels, the role of leadership is crucial to D&I adoption. Leadership, defined as being when an "individual influences a group of individuals to achieve a common goal" (Northouse, 2018, p. 5), is a well-known factor that has been shown to either facilitate or prevent organizational change (Kotter, 2012).
The findings from this study also indicate that nonprofit associations are able to overcome both resource-based and institutional barriers in their efforts to implement D&I practices. This finding provides evidence of institutional entrepreneurship by showing that some individuals seek opportunities for change and leverage organizational resources in order to make this change occur (DiMaggio, 1988).

Literature Review
While the terms diversity and inclusion may seem self-explanatory, these terms often have contested meanings among scholars and practitioners (Weisinger et al., 2016). For the purposes of this study, diversity is defined broadly as "any significant difference that distinguishes one individual from another" (Kreitz, 2008, p. 102).
Although research has traditionally focused on diversity in terms of race and/or gender, diversity can be conceptualized differently depending upon context. For example, there may be male-or female-dominated professions that are ethnically diverse but lack gender diversity. Inclusion, on the other hand, is defined as "the degree to which an employee perceives that she is an esteemed member of the work group through experiencing treatment that satisfies her needs for belongingness and uniqueness" (Shore et al., 2011(Shore et al., , p. 1265. To further provide clarity into the differences between diversity and inclusion, Roberson (2006, p. 228) outlined practices that differentiate the concepts. Diversity practices, she suggested, are those that affect the "demographic composition of groups or organizations." Inclusion, she argued is "employee involvement and the integration of diversity into organizational systems and processes."

The Role of Nonprofit Associations in the Diffusion of D&I Practices
As organizations that connect individuals, groups, and other organizations across the workforce, associations are a unique subsector of the nonprofit sector. Indeed, nonprofit associations are member-based organizations that are created to "educate, train, and credential people" (Nesbit & Gazley, 2012, p. 560) in diverse fields such as acting or performing, engineering, education, nursing, and publishing (to name just a few). Some nonprofit associations are composed of other organizations (e.g., "institutional members," such as businesses or hospitals), others are composed of primarily individual members (e.g., teachers), and some are composed of both.
Studies have shown that nonprofit associations can act as "mediating structures" (Berger & Neuhaus, 1996;Couto, 1999) or facilitators of change, both among their membership as well as in the broader professional community they represent. Greenwood, Suddaby, and Hinings (2002) found that professional associations were critical to "theorizing" about change by developing and advancing innovations and assisting in the diffusion of those innovations across industries. Nonprofit associations have also been shown to support innovations in human resources practices (Farndale & Brewster, 2005), environmental sustainability (Dickson & Arcodia, 2010), economic development (Kshetri & Dholakia, 2009), public policy (Balla, 2001), and the use of technology (Newell, Robertson, & Swan, 1998;Swan & Newell, 1995). Nerland and Karseth (2015) argued that the knowledge work of associations is more valuable now than ever because of the world's globalized workforce. Perhaps not surprisingly, then, there has been extensive scholarship focusing on ways that nonprofit organizations, including associations, can partner with the for-profit sector to support outcomes that increase overall social value (e.g., see Austin, 2000;Rivera-Santos, Rufín, & Wassmer, 2017;Selsky & Parker, 2005;Villani, Greco, & Phillips, 2017).

Institutional and Resource Based Approaches to D&I
While the business case to support D&I within organizations has generally been wellestablished, organizations often struggle with effective policies and practices that support more diverse and inclusive workplaces. Kossek and Pichler (2007) suggested that successful diversity management programs should have a broad mandate to create a sense of justice within an organization, reduce discrimination, and increase financial competitiveness.
In an attempt to provide greater guidance for organizations as they incorporate D&I practices, Yang and Konrad (2011) identified and defined the types of practices that are likely to support D&I. These practices are internal hiring, training and mentoring policies (particularly for individuals from under-represented groups), and offering benefits to a wide range of employees (such as policies that allow for working remotely). The authors also identified institutional and resource-dependent factors that support or constrain D&I efforts.
In general, institutions can be defined as having "regulative, normative, and cognitive structures and activities that provide stability and meaning for social behavior" (Scott, 1995, p. 33). These structures and activities can be laws, regulations, customs, culture, ethics, social, and/or professional norms that organizations must respond to in order to maintain a competitive advantage (DiMaggio & Powell, 1983;Miles, 2012). Adhering to or adopting new structures supports an organization's legitimacy, allowing it to gain support and acceptance from a wider range of stakeholders (Kostova, Roth, & Dacin, 2008;Yang & Konrad, 2011).
Institutional change, however, is not without challenges. For one, organizations tend to resist change (Oliver, 1991). Indeed, out-of-date rules, practices, norms, and regulations often limit an organization's interest or ability to move forward with new ideas and practices. Eisenhardt (2000) argued that people and organizations are often fundamentally ambivalent about what progress means, which can stifle innovation. Additionally, Freisel and Kwon (2017) have discussed the impact that leadership resistance has on change. If leadership fails to support or champion change then it is unlikely that change will succeed (Kotter, 2012).
In one of the only studies of D&I practices within nonprofit associations, Solebello, Tschirhart, and Leiter (2016) suggested that using internal structures and institutions may help the implementation of these practices, particularly in the face of opposition. It is likely, then, that existing institutions may either facilitate or constrain an organization's behavior. Based on this literature, the following hypothesis is offered: Hypothesis 1: Nonprofit associations have more (fewer) D&I practices in the presence of institutional support (barriers).
Implementing D&I practices within organizations can undoubtedly be difficult. However, implementing these practices across organizations, industries, or even professional fields can prove even more challenging. Although there has been some research on the transfer of human resources practices within multinational organizations (often drawing from Oliver's [1991] theory of organizational response to pressure applied by external stakeholders), Ferner, Almond, and Colling (2005, p. 308) argued that D&I practices exist in a "contested institutional terrain," which can make it difficult to successfully support the implementation across subsidiaries.
While much of the D&I literature centers on the outcome of an industry's response to external pressures, little is known about how perceived institutional constraints can affect nonprofit associations in supporting the adoption of D&I practices throughout their professions or trades. It stands to reason, though, that if a firm rejects changes at its headquarters then those changes are unlikely to be adopted by its subsidiaries. Therefore, the second hypothesis offered is: Hypothesis 2: Nonprofit associations are more (less) likely to support D&I practices in their field when there are institutional supports (barriers).
An alternative viewpoint to institutions as constraining behavior is that in certain circumstances individuals are those who strive to influence and change institutions. This viewpoint is generally conceptualized as institutional entrepreneurship, where "new institutions arise when organized actors with sufficient resources see in them an opportunity to realize interests that they value highly" (DiMaggio, 1988, p. 657). Thus, individuals and organizations can influence the environment in which they exist (Battilana, Leca, & Boxenbaum, 2009).
Institutional entrepreneurship is likely crucial to the implementation of change within organizations and professional fields. Individuals proposing changes, however, are not immune from the resistance of others (Eisenhardt, 2000;Seo & Creed, 2002;Solebello et al., 26 2016). Indeed, some D&I efforts can make people feel uncomfortable; and, those experiencing discomfort are likely to push back against these efforts. As such, two additional hypotheses are: Hypothesis3: Institutional entrepreneurs in nonprofit associations are more likely to support D&I practices within their organizations despite institutional barriers.
Hypothesis4: Institutional entrepreneurs in nonprofit associations are more likely to support D&I practices within their field despite institutional barriers.
Resources often help to support nonprofit legitimacy (Barney & Clark, 2007;Bies, 2010;Guo & Acar, 2005;Yang & Konrad, 2011). Indeed, nonprofit organizations can leverage their (often limited) resources in support of priorities and opportunities the organization may have available. As such, the degree to which a nonprofit implements D&I practices may be indicative of its overall capability and capacity (Solebello et al., 2016;Yang & Konrad, 2011). While much of the literature focuses on financial resources, how nonprofit managers perceive and prioritize (or resist and prevent) D&I efforts can guide how their (often limited) financial resources are spent. Ultimately, this can help to further their D&I goals (Yang & Konrad, 2011).
Nonprofit associations, specifically, are highly dependent on their members for financial resources. These members generally pay an annual fee in order to obtain benefits from the association. Although in some fields pursuing D&I practices might be encouraged by members and stakeholders (such as women in STEM fields), in other fields D&I efforts may risk the support of their membership. Solebello et al. (2016), thus, highlighted a "paradox" in conducting research on D&I in associations: that is, spending money (i.e., resources) on D&I efforts tends to limit an association's ability to reach broader, more general audiences.
These types of tensions are likely to cause executives of nonprofit associations to scale back D&I efforts so as not to alienate existing members. Especially if members join an association based on a set of pre-existing standards and values, shifting those standards and values may cause members to feel as if the organization no longer represents them (Knoke, 1986(Knoke, , 1990Solebello et al., 2016). The threat of shrinking membership, then, could jeopardize the survival of nonprofit associations, which could discourage association executives from engaging in unsupported activities. As such, the final three hypotheses are: Hypothesis 5: Nonprofit associations with more (fewer) financial resources will prioritize (de-prioritize) D&I work.
Hypothesis6: Nonprofit associations with greater (lesser) support from their stakeholders will be more (less) likely to engage in D&I practices within their organization.
Hypothesis7: Nonprofit associations with greater (lesser) support from their stakeholders will be more (less) likely to engage in D&I practices in their trades or professional fields.

Methods and Data
A national survey of executives of nonprofit associations was conducted from January to March 2018. The population for the survey was all registered 501(c)(6) membership associations and 501(c)(3) professional and trade associations as determined by the National Taxonomy for Exempt Entity (NTEE) subsection code "3." To identify the population of associations, the August 2016 Business Master Files (BMF) of IRS-registered associations was obtained from the National Center for Charitable Statistics (NCCS). Of those in the BMFs, 36,873 organizations were classified as 501(c)(6) associations. Another 12,518 organizations were classified as 501(c)(3) professional and trade associations. In total, this resulted in a population of 49,391 nonprofit associations.
A random sample of the total population was drawn (n=2,099), of which approximately 25% were 501(c)(3) associations. Two organizations were removed from the sample because they had a non-U.S. mailing address. A sample analysis was completed to ensure that those included in the final sample were not systematically different from those not included. None of the observed variables were statistically significant indicators of selection into the sample.
A survey instrument was developed using extant literature on D&I practices in organizations. Focus groups with association executives were also organized in order to identify key themes and practices. The survey was then refined through piloting and cognitive interviews with a small group of association executives who reviewed questions systematically for clarity and shared meaning (Sudman, Bradburn, & Schwartz, 2010). The final survey instrument was administered online using Qualtrics and consisted of 55 questions that asked association executives about their perceptions of D&I practices within their ("organization") as well as their perceptions of D&I practices within their professional field or trade ("the field"). There were also questions that provided descriptive information about the organization and demographic information of the respondent (e.g., role in the organization, gender, level of education, and race/ethnicity).
Adapting Dillman's (2008) and Millar and Dillman's (2011) surveying technique, all identified executive directors or board chairs were mailed a letter that described the purpose of the study. The letter also invited each individual to participate and offered them a $5 Amazon gift card for completing the survey. This initial letter was followed by a postcard approximately one month later. Attempts were also made to contact individuals by e-mail and phone.
Slightly over 300 association executives were not able to be reached either by phone, e-mail or postal mail, resulting in a final sample of 1,712. Of those who were contacted, 278 individuals responded to the survey. This represents a response rate of 16.24%. Those without completed surveys were removed from the analysis, for a total of 152 completed and useable responses. While the response rate is lower than desired, and attrition was high, the nature of the survey (i.e., questions about D&I practices, which can make some people uncomfortable) likely attributed to the low response rate (Dillman et al., 2009;Fink, 2002). Smith (1997) also found that response rates can suffer when surveying at the organizational level, where issues of authority to speak on behalf of an organization, a respondent's knowledge of the content area, or simply a lack of time can weaken response rates.

Dependent Variables
There are six dependent variables in this study. These variables were created using 32 questions on D&I practices developed by Konrad, Yang and Maurer (2016). Specifically, respondents were asked to first state "whether or not these practices are used in your organization (among board, staff, and volunteer leaders)." Respondents were then asked if their association "has worked to implement the practice in your field (members or potential members, employers, other organizations), either through requested guidance, providing information or resources, or through collaboration." • Dependent care resources and referral service? 10.14% 20.59% X(bar)=mean value. Notes: Percentage of organizations engaging in items included for each dependent variable, for both organization and field. Cronbach's alpha (α) and mean factor scores are included for each measure. Question text for survey questions 14-19: "The first option is whether or not these practices are used in your organization (among board, staff, and volunteer leaders). The second is if your organization has worked to implement the practice in your field (members or potential members, employers, other organizations), either through requested guidance, providing information or resources, or through collaboration." The six D&I practices included as dependent variables are: 1) Linking Diversity to Strategy (composed of eight items); 2) Recruiting a Diverse Workforce (composed of five items); 3) Selecting a Diverse Workforce (composed of four items); 4) Training and Developing a Diverse Staff (composed of four items); 5) Monitoring the Effectiveness of Staffing for Diversity (composed of four items); and 6) Providing Work-Life Flexibility (composed of seven items). The first three practices focus on ensuring that there are programs and policies in place to support D&I in the workforce. All questions had response options of either "No" (coded as "1"), "Unsure" (coded as "2"), or "Yes" (coded as "3").
A confirmatory factor analysis was conducted to ensure construct validity. Factor scores were generated for each category. Item descriptions, mean factor scores, and Cronbach alpha scores are presented in Table 1. Although factor scores above 0.80 are generally considered strong, factor scores of 0.70 and higher are considered adequate for social science research (Nunnally & Bernstein, 2017).

Independent Variables
The independent variables were drawn from a list of 18 perceived obstacles and challenges in conducting D&I work. This list was generated during the survey design process in consultation with several association executives. Respondents checked off items they recognized as "obstacles or challenges to implementing new or ongoing diversity and inclusion programming or policies" in, both, their organization and their field.
These obstacles and challenges were then grouped into either institutional (11 items) or resourced-based categories (five items). Positive responses were coded as "Yes" ("1"). All other responses were coded as "0" since a non-answer can reasonably be assumed to be not "yes." In addition, several control variables were used in the analysis. These controls were obtained from each organization's Form 990 (i.e., their annual nonprofit financial filing form). Controls included their annual revenue (logged to account for outliers), whether or not the organization was a 990N e-postcard filer (indicating gross revenues under $50,000 annually) and organizational age (also logged to account for outliers). A variable was also included to assess respondents' perception of their association's diversity and the diversity within their field. These responses were measured on a Likert scale ranging from one to seven, with 1="Not at All" and 7="To a Great Extent."

Descriptive Statistics and Sample Analysis
Of the organizations that were able to be matched to their Form 990 (n=138), 31% are 501(c)(3) associations and 69% are 501(c)(6) associations. The average organizational age is 30.35 years (with a standard deviation (SD) of 21.31) and ranges from two to 91. Average organizational income is $744,391 (SD $3,800,930) with a median income of $56,899. It should be noted that the average income value is clearly influenced by some high-income organizations, as most organizations in the sample are small.
Fifty-two percent of the organizations are professional associations, 11% are trade associations, 10% are chambers of commerce, and 9% are combined trade and professional associations. Fifty-six percent of the organizations represent individual members only, while 14% represent institutional members (i.e., other organizations), and nearly 23% represent both institutional and individual members.
A review of the individual characteristics of survey respondents shows that 49% are male, while 47% are female. One individual indicated that they are gender-nonconforming. Three percent of respondents chose not to respond. Approximately 83% of respondents self-identify as white. Just under 6% self-identify as Black or African-American, approximately 3% selfidentify as Asian, approximately 1% self-identify as Native American or Alaskan Native, and approximately 4% of respondents preferred not to respond. Ethnically, 2% of respondents also indicate that they are Hispanic or Latino. These characteristics are consistent with a recent national survey of association members, which found respondents to be predominately white (75%), with only 3% identifying as Black, and 3% identifying as Hispanic. This national survey also showed that 43% association members are female (Nesbit & Gazley, 2012).
Fifty-two percent of respondents in the present study are executive directors or CEOs of their association. The next largest group of respondents indicated that they are "other directors." Only one respondent indicated they are a "D&I manager or director." Findings from a sample analysis suggest that executives from associations that have higher incomes were more likely to respond to the survey(p<0.05), indicating that there are possibly more professionalized organizations (based on revenues) in this study.
Respondents indicated that they believed their organizations are doing slightly better than the average (4.50) when it comes to diversity (mean=4.26). Approximately 49% of respondents indicated that they engage in D&I practices in both their organization and their field. Table 2 provides a breakdown of responses to the question asking respondents what they thought diversity meant for their organization and professional field or trade. Although respondents could choose more than one of 12 responses, most identified race and/or gender as the "single most important dimension.
The frequency of specific D&I activities ranges from a low of approximately 5% (when it comes to "tracking the diversity of employees receiving promotions") to just over 60% (when it comes to "Supporting employee participation in professional associations targeting diverse groups, such as women's professional associations"). The latter item was the only item practiced by more than half of responding organizations in both their association and their field. The mean factor score for all measures is relatively low, with only three over the midpoint at a score of 2, indicating a relatively low level of engagement in D&I practices in that category. Figure 1 presents a graphical representation of the mean factor scores for each measure by both the association and the field.

Figure 1. Distribution of D&I Factor Scores
Notes: Dashed line=Within organization; Solid line=Within field. Kernel density with higher scores (xaxis) indicates higher factor scores as values increase, or more activity within each category.

Analysis of Institutional and Resource Based Obstacles
Multivariate analysis using ordinary least-squares (OLS) regression was used to estimate the relationship between institutional and resource-based barriers and associational practices at both the organization and field levels. This method is useful when dependent variables are continuous, not categorical, and the independent variables are homoscedastic and exogenous (Jarque & Bera, 1980). A total of 12 models were estimated: six pertaining to D&I practices within associations and an additional six pertaining to D&I practices within respective fields. All models included the same control variables. Findings for these estimates are displayed in Figures 2 through 5.
For the first item (i.e., Is the organization diverse? and Is the field diverse?), a positive and significant relationship indicates that the more diverse an association is (as perceived and rated by the respondent) the more likely that association is to engage in specific D&I practices. For the resource-based and institutional constraints, negative and significant values indicate that those obstacles prevent participation in D&I practices. Positive and significant values, however, indicate that the association engages in those activities despite facing resource-based and institutional constraints.

Internal Association Practices ("In the Organization")
As shown in Figure 2, the self-reported measure of association diversity is positively related to an association's efforts to link diversity to strategy. That is, the more diverse an association is, the more likely it is to engage in this practice.
When considering Hypotheses 1 and 5, they are partially supported by the data. Specifically, there is evidence that leadership has an impact on the adoption of D&I practices-even when facing institutional and resource-based constraints. Among institutional constraints, there is clear evidence that leaders either promote or stifle D&I practices. Indeed, the obstacle labeled "staff leader opposition" is significant and negative, particularly when it pertains to recruiting and selecting a diverse workforce, monitoring effectiveness, and work-life flexibility.
Figure 2 also shows that the board being uncomfortable is also significant and negative in linking diversity to strategy and in training a diverse workforce. This finding may lend support to previous research showing that managers are more likely to hire individuals who are culturally similar to themselves (Rivera, 2012). Since the respondents in this study are 85% white, without giving deliberate attention to supporting diversity in hiring, training, and mentoring, associations may very well struggle in this area. In other words, the lack of support among leadership may prevent associations from engaging in D&I practices. This is consistent with literature in the field of change management, which has argued that organizational and institutional change must be driven by a champion and that a champion should be a leader within the organization (if not the director or CEO) (Kotter, 2012).
Hypothesis 4 on institutional entrepreneurship was also partially supported. Only volunteer opposition was positively related to three categories of activities within associations, i.e., recruiting, selecting, and training a diverse workforce. Indeed, associations seem to be engaging in these activities despite opposition from these key organizational supports. With the constraining role of leadership, however, it may not be surprising that there is only limited support for institutional entrepreneurship.
It should be noted that choosing "not applicable/doesn't apply to us" was significant and negative in all but one of the six categories. This suggests that executives of associations who believe that engaging in D&I practices is not relevant to their work do not engage in most practices. Resource-based Hypotheses 5 and 6 receive limited support. For Hypothesis 5, the relationship between the organization's revenues and its participation in D&I practices is unclear. Revenues are significant and positive for some activities, but interestingly, it is also significant and positive (when it comes to training a diverse staff and monitoring effectiveness) for small organizations (i.e., for organizations that file a Form 990N e-postcard, indicating that they have gross revenues of under $50,000 annually). Taken together, this provides evidence of a U-shaped relationship. Thus, larger associations may have the resources and capacity necessary to act as information brokers and leaders in their respective fields, but smaller organizations may also see this work as part of their primary mission. Other resourcebased variables were not significant, providing no evidence for support of Hypothesis 6.

External Association Practices ("In the Field")
When considering the role of association activity in promoting D&I efforts in their respective fields, there were different patterns. For one, resource-based variables outside of revenues neither constrain nor facilitate associations engaging in specific activities. However, the Ushaped difference between larger and smaller groups is also evident here. Additionally, association age is positively related to engaging in D&I practices. Indeed, older, assumedly more established, associations take a more aggressive role in D&I efforts than younger associations.
Hypotheses 3 and 5 are partially supported. That is, there is evidence suggesting the existence of institutional entrepreneurship. When facing institutional constraints, there is limited influence of leadership opposition in an association's behavior. In fact, the findings suggest that associations overcome institutional barriers to engage in D&I practices in their fields. For example, several practices are carried out despite resistant leadership and member or volunteer opposition. This could mean that association staff believe that they have an imperative to act on D&I. Perhaps some are compelled to engage in these practices in order to educate their members, volunteers, and stakeholders.
It is important to note that staff being uncomfortable while working on D&I issues is positive and significant in all but one category. This likely means that association staff often operate outside of their comfort zones. Finally, despite a heightened political climate around issues of D&I, the findings show that associations continue to engage in many practices, including linking diversity to strategy and training a diverse workforce. While some association employees may find institutional constraints difficult to overcome, these findings highlight the importance of institutional entrepreneurship, where individuals take it upon themselves to support changes in institutional norms and behaviors.

Discussion and Conclusion
This study explored the relationship between engaging in D&I practices and institutional and resource-based obstacles. The findings of the study were mixed. However, the role of leadership in associations, the findings suggest, is crucial to providing a barrier to, or supporting, D&I practices. Importantly, the impact of leadership was found to be a constraint on D&I practices but only when attempting to promote these practices inside of the association. As associations looked to influence their overall field, the impact of leadership was lessened. Thus, various factors affect the promotion of D&I within associations and their respective fields differently.
While many associations struggle to include D&I practices in their policies and procedures, this study provides evidence of institutional entrepreneurship of associations within their fields. Indeed, within their fields, associations are more likely to engage in some D&I practices despite facing (what is perceived as) opposition from their members, leaders, and volunteers.

39
This indicates a degree of tenacity among association executives to do what they believe is right for their field, regardless of what some stakeholders may feel. Association staff may feel that it is necessary to promote these practices in their field specifically because of this opposition. That is, they may feel obligated to educate and inform those in their field about the value of D&I in their industry.
Associations, however, should bear in mind that staff may be uncomfortable doing this work, and association managers may benefit from providing additional staff training or professional development to employees, board members, and key volunteers regarding issues of D&I. This word of caution is important given the finding that more D&I practices occurred at the industry level than at the organizational level. Although this could be due to lack of internal capacity, implementing D&I practices can undoubtedly be challenging. Future research can help to clarify these explanations.
The findings regarding resources were also mixed. Indeed, the findings indicated that small associations (with one or no staff) seem to be engaging in similar patterns of D&I work as larger associations. It should be noted, however, that this study did not actually measure the scale of these efforts. Thus, future research can help to provide insights into this relationship.
Although the first of its kind, this study is not without limitations. For one, the response rate is lower than desired. Selection bias is always a consideration in studies with a low response rate. Considering that those who responded were voluntarily willing to take a survey on D&I practices in their associations, those opting not to participate may be facing D&I challenges at either the organizational or industry level. Thus, it is plausible that the constraints outlined in this study may be more pronounced in associations that opted not to participate.
In addition, this study did not measure the existence of respondent bias. If individuals inside organizations resist policies and processes that create diverse or inclusive environments, this could reasonably be expected to have an impact on their responses. Future research that includes measures of bias would add additional insights to the impact of individual preferences on organizational D&I outcomes. Moreover, qualitative research and case studies may add greater depth and understanding about organizational culture and processes, allowing the opportunity to evaluate how organizations value D&I for themselves and their stakeholders. Despite these limitations, this study provides evidence of the types of barriers associations face when attempting to engage in D&I practices.

Implications
Findings from this study have implications for both research and practice in associations as well as in the nonprofit sector more broadly. With ever-growing attention to D&I practices in organizations of all types, acknowledgment of the importance of leadership support in D&I activities and programs is certainly timely. While the findings from this study are consistent with most literature on change management in organizations and institutional entrepreneurship, overcoming deeply felt discomfort or opposition to these issues may affect decisions to assign often scarce resources to the work of becoming more diverse and inclusive (Solebello et al., 2016). Yet, individuals committed to change can engage in these practices and help change the norms of their associations and their fields. Nonprofit leaders, however, should take stock of their own commitment to D&I as well as the commitment of their organization's before implementing programs and practices relating to D&I.