Abstract
Nonprofits offer services to disadvantaged populations, mobilize collective action, and advocate for civil rights. Conducting this work requires significant resources, raising the question: how do nonprofits succeed in increasing donations and volunteers amid widespread competition for these resources? Much research treats nonprofits as cold, rational entities, focusing on overhead, the “price” of donations, and efficiency in programming. We argue that nonprofits attract donors and volunteers by connecting to their emotions. We use newly available administrative IRS 990 e-filer data to analyze 90,000 nonprofit missions from 2012 to 2016. Computational text analysis measures the positive or negative affect of each nonprofit’s mission statement. We then link the positive and negative sentiment expressed by nonprofits to their donations and volunteers. We differentiate between the institutional fields of nonprofits—for example, arts, education, social welfare—distinguishing nonprofits focused on social bonding from those focused on social problems. We find that expressed positive emotion is often associated with higher donations and volunteers, especially in bonding fields. But for some types of nonprofits, combining positive sentiment with negative sentiment in a mission statement is most effective in producing volunteers. Auxiliary analyses using experimental and longitudinal designs provide converging evidence that emotional language enhances charitable behavior. Understanding the role of emotion can help nonprofit organizations attract and engage volunteers and donors.
Keywords: nonprofits, volunteers, giving, emotion, text analysis
Nonprofits provide vital functions in the United States, including job training, education, childcare, social support, housing, healthcare, and disaster relief (Berrone et al. 2016; Marquis, Davis, and Glynn 2013; Salamon 1987; Weisbrod 1988). Nonprofits also mobilize collective action, advocate for underprivileged populations, and work to protect basic civil rights (Marwell 2004; McCarthy and Castelli 2002). They enrich lives through the arts and encourage citizens to think beyond their own narrow interests (Tocqueville [1835] 1972). Conducting this work requires significant resources, raising the question of how nonprofits can attract the necessary labor and capital to fulfill these third-sector duties.
Since the 1960s, the government has invested in a public-nonprofit partnership model that undergirds social service provision in the United States. However, government agencies at the local, state, and federal level are cutting funding, leaving nonprofits to find alternative sources of revenue (Pettijohn et al. 2013). This shift means nonprofits are increasingly reliant on other forms of support, namely private donations and volunteers, and it raises an important question: How do nonprofits succeed in increasing donations and volunteers amid widespread competition for these resources?
In this article, we draw from growing research in psychology, the sociology of emotions, and organizational theory to better understand how nonprofits attract donors and volunteers by connecting to their emotions. A significant body of theory and research, dating back to Aristotle’s pathos and continuing through affective neuroscience, demonstrates that emotions are foundational to understanding motivation and behavior, generally, and prosocial behaviors (e.g., volunteering and donating), specifically. Yet, despite knowing the importance of the “heart” to donor choice, few studies turn this back to nonprofits themselves to assess these organizations’ deployment of emotion. Instead, research on how nonprofits attract external audiences typically treat them as cold, rational entities, focusing on overhead and administrative costs, the “price” of donations, and how well they efficiently convert external engagement into demonstrable results (Calabrese 2011; Charles 2018; Weisbrod and Dominguez 1986). In focusing on these elements, these studies implicitly assume rational audiences of donors or volunteers seeking to maximize their investment. Although perhaps true for some, this emphasis disregards research demonstrating that emotions are powerful motivators and that organizations carry with them their own cultures and tools to emotionally connect to different audiences.
We investigate use of emotion for nearly 90,000 nonprofits over the period 2012 to 2016. We hypothesize that nonprofits whose mission statements contain more emotional messages will induce emotion in individuals, yielding a higher number of volunteers and more donations, and that this will vary based on the type of emotion (positive/negative) and the institutional field in which a nonprofit is located. Across disciplines, organization scholars emphasize the significance of an organization’s mission as reflective of its culture and for understanding decision-making, behaviors, performance, and the behavior of its clients, employees, and supporters (Kim and Lee 2007; McDonald 2007; Pandey, Kim, and Pandey 2017). Certainly, “the goals or agendas attached to a mission serve to rally, engage, and enroll workers, volunteers, and donors” (Minkoff and Powell 2006:591). Missions play an active role throughout all aspects of firms, and even more so for nonprofit organizations that are inherently mission driven and held accountable by the extent to which they fulfill their stated mission.
Until recently, social scientists lacked the ability to evaluate aspects of nonprofit missions on any large scale. Most of the work done in this area was necessarily limited to small subsets of nonprofits (Berlan 2017; Pandey et al. 2017; Patrick and Caplow 2018). In contrast, to test our hypotheses and provide previously unattainable insights, we utilize the new Internal Revenue Service (IRS) release of more than 2 million nonprofit tax filings. The new IRS administrative data release provides financial information for all e-filing nonprofits plus other information related to governance policies, number of volunteers, and mission. Before this release, research on the nonprofit sector using this major source of information was severely hampered by the fact that the IRS only made non-searchable image files available. It is difficult to overstate the importance of this data release for understanding the U.S. nonprofit sector.
In this article, we use computational text analysis techniques to measure use of emotion by nonprofits in their mission statements using Linguistic Inquiry and Word Count (LIWC) classifications of sentiment (Pennebaker et al. 2015). We examine the relationship between use of emotion and volunteers and donors while accounting for more traditional, financial-based explanations (Calabrese 2011; Charles 2018; Weisbrod and Dominguez 1986). We address potential issues of sampling bias using Frank and colleagues’ (2013) test. We find that positive mission statements are often associated with more volunteers and higher donations, but this pattern varies depending on the institutional field of the nonprofit (e.g., arts, education, healthcare, or housing), with fields stressing community and social bonding being particularly responsive to positive sentiment. Negative sentiment alone is rarely an effective strategy for nonprofits to gain volunteers or donors. For some nonprofit categories, however, the combined use of positive and negative emotions is associated with higher levels of donations and numbers of volunteers.
We supplement our main analyses with experiments and longitudinal models to further establish the relationship between emotions and donations/volunteers. Our experiments examine whether people would be more willing to donate to a charity when a mission-based appeal is couched in emotional language. We focus on two areas of charitable work: holiday gift provision to children and food banks. In each, we pair the experiment with a subset of the IRS administrative data and find converging results. For a set of nonprofits with administrative data over time, we also provide a longitudinal investigation of the influence of sentiment change on donations and volunteers, again producing converging results. The insights gleaned from our main and supplemental analyses are particularly salient to a nonprofit sector trying to attract resources during a time of increased scarcity and competition.
HOW NONPROFIT USE OF EMOTION MOTIVATES VOLUNTEERS AND DONATIONS
To incorporate an emotional dimension is unusual. Prior research on nonprofit donations focuses exclusively on organizations’ financial characteristics (e.g., Calabrese 2011; Weisbrod and Dominguez 1986). This “donor-demand” model stresses a purely economic approach: “Donors give contributions of money in return for an implicitly agreed-upon level of provision and quality of output. We postulate that the market demand function for a particular type of collective-good output depends—as in the case of purely private goods—on price, quality . . .” (Weisbrod and Dominguez 1986:85). Research continues to focus on the use of financial measures of efficiency in predicting contributions to nonprofit organizations.
In contrast, philanthropists, foundations, nonprofit staff, and fundraisers recognize that people also weigh other factors when making the decision to donate to a nonprofit, either financially or through their time—the “heart” can be as important as the “head” (Paxton 2020). Research from psychology and affective neuroscience outlines the importance of emotions to guide decisions, in both conscious and unconscious ways (Damasio 1994; Gray 1990; Zajonc 1980). Affect is the principal mechanism by which preconscious cognition influences behavior. Social psychologists and sociologists of emotion stress how emotion encourages actors to commit to and invest in continued relationships (Lawler and Yoon 1996; Price and Collett 2012; Turner and Stets 2006). Although this literature makes important distinctions between affect, mood, emotions, and sentiment, we follow Gaudine and Thorne (2001:176) in defining emotion as a broad “affective feeling state that may vary in intensity from mild to intense.”
Furthermore, organizational scholars demonstrate that organizations are both rational and affective institutions, using both types of logics to conduct business and make decisions (Bail, Brown, and Mann 2017; Isen and Baron 1991). Organizations have their own unique cultures and affective climates they can use to appeal to and engage audiences (Alvesson 2012; Deshpande and Webster 1989; Schein 2010). By not accounting for the emotional features that operate in tandem with rational aspects, we are left with an incomplete understanding of the features of nonprofits that can be leveraged to maximize engagement from external audiences.
How Emotions Structure Cognition and Behavior
Emotion motivates.
Aristotle lists emotion, or pathos, as one of three fundamental modes of rhetorical persuasion that can induce an audience to make a desired judgement or action. Evidence from multiple literatures continues to underscore the power of emotions in shaping cognition and decision-making and in motivating behavior. Moreover, scholars of organizations and collectives have applied these insights to show how emotions can be effectively deployed to increase performance and cohesion and to gain support from external audiences (Bart 1997; Brown and Yoshioka 2003). We draw on insights from several areas of research to understand why the emotional frames exuded by nonprofits matter for attracting donations and volunteers.
Affective neuroscience and dual process theory emphasize that emotion structures cognition in an automatic, even preconscious, process (Damasio 1994; Gray 1990; Lizardo et al. 2016; Vaisey 2009; Zajonc 1980). The sociology of emotion initially focused on the “nonreflective stream of primary emotive experience” (Hochschild 1979:552). However, dual process models help demonstrate that although emotional responses may have an axiomatic, automatic response to stimuli, emotions also spur deliberate action (Lizardo et al. 2016; Vaisey 2009). That is, affect is involved in both modes of judgment, during preconscious appraisal of stimuli and executive reasoning (Bechara et al. 1997). For example, Wollschleger (2017) finds that religious services that have more emotional energy have greater participation, in part, because churchgoers are motivated to keep coming back to participate in the experience—participants make a deliberate decision motivated by an unconscious euphoria during the program. McDonnell, Bail, and Tavory (2017:6) illustrate how organizations can exploit this emotion and decision-making link:
. . . heightened emotion may make objects or messages resonate, when they might not otherwise, by priming people to find solutions that justify their feelings. In this vein, Bail (2015) shows that anti-Muslim organizations resonated in the aftermath of the September 11 attacks not because their discourses resonated with prevailing views about Muslims or terrorism but because they were charged with palpable fear and anger that focused public attention on their peripheral claims. Audiences who witness such emotions not only focus on those who voice them more ardently than calm or dispassionate messages—but the experience of emotional arousal also impacts the cognitive processes people use to gain further information or the problem-solving processes described previously.
One tool organizations have to create internal coherence, increase identification among employees and supporters, and motivate external audiences is to enhance the emotional frame of their work. Briefly, to frame something is to “select some aspects of perceived reality and make them more salient in the communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation and/or treatment recommendation for the item described” (Entman 1993:55). The frame of a message provides a “schemata of interpretation” for the audience to “organize experiences and guide action” (Snow et al. 1986:464). Certainly, scholars have long recognized that political candidate advertisements routinely appeal to emotion and that emotional frames motivate voters (Brader 2006). Emotions such as anger, when captured and channeled by social movement organizations, can produce remarkable social change (Goodwin, Jasper, and Poletta 2009). Emotional public health campaigns, especially those that induce fear, have been more successful than rational appeals (Witte and Allen 2000). A recent study of advocacy organizations shows that nonprofits get more attention when they use emotion in a rational environment (Bail et al. 2017).
Emotion is relevant to all types of organizations, but it is particularly significant for the nonprofit sector—a sector rife with emotion-laden issues. Individual donors and volunteers are often seeking experiences and organizations that help fulfill their passions (Bronfman and Solomon 2010), even in the face of evidence of ineffectiveness (Berman et al. 2018). Nonprofit leaders face a normative expectation that they display emotion and passion in their line of work (Silard 2018). And employees who are otherwise dissatisfied with working conditions and pay can still be retained when they feel positively about the nonprofit’s mission (Kim and Lee 2007). Individual nonprofits that emotionally connect to these passions, as opposed to focusing solely on highlighting the utility of a donation, may be positioned to become beneficiaries of time and treasure. In short, when nonprofits stress emotion in their mission, promotional materials, or other communicating texts, potential donors and volunteers will experience an unconscious, affective response that spurs a more deliberative decision to engage with the nonprofit.
Hypothesis 1: Use of emotion in describing a nonprofit’s work will be associated with a higher number of volunteers and more donations.
Positive Emotions, Negative Emotions, or Both?
People are unlikely to donate or volunteer unless they feel they can make a difference (Kollock 1998; Olson 1965). Positive emotional framing of mission may help a potential donor or volunteer feel that their investment of time or money can make that difference, signaling an ability on the part of the nonprofit to overcome social dilemmas and attain the charity’s goals. Affective Intelligence Theory, a dual process model, argues that when goals are being met it creates the emotion of enthusiasm and the motivation to act (Marcus, Neuman, and MacKuen 2000). Research suggests that a sense of potential impact increases helping and the anticipated “warm glow” of helping (Cryder, Loewenstein, and Seltman 2013). And in for-profit companies, research shows that employees who can identify their work as advancing the mission of an organization are more likely to report higher job satisfaction and stay with a firm (Brown and Yoshioka 2003).
Negative emotion can also be motivating.
Anxiety plays a pivotal role in triggering when people depart from reliance on automatic responses to enter a cognitive reasoning mode. Positive and negative emotions are signals from information-processing brain systems that encourage reward-seeking (approach) or danger-avoidance (Gray 1990). A negative emotional state triggers behavior as individuals attempt to cope with or resolve the emotion (Lazarus 1991), and it increases solidarity among ingroup members that affirms and enables collective action (Marcus 2013). This implies that nonprofits that stress negative valence could also produce a motivational helping response as audiences act to alleviate negative emotions such as distress or guilt (Bagozzi, Gopinath, and Nyer 1999). Some studies of nonprofit fundraising vignettes show that negative framing can indeed be effective (e.g., Fisher, Vandenbosch, and Antia 2008; but see Das, Kerkhof, and Kuiper 2008). Nonprofits may generally feel a need to emphasize the positive (Silard 2018), but they vary in the extent to which they use emotional language.
For nonprofits, the combination of positive and negative emotions may be particularly powerful in motivating donors and volunteers. A best practice in fundraising appeals is to combine negative and positive (Brooks 2015). Fundraisers argue that potential donors respond best when a charitable nonprofit defines a social problem (negative), demonstrates harm (negative), identifies an opportunity for amelioration (positive), promises to affect change through their programs (positive), and suggests the power of donor participation in this process (positive) (Merchant, Ford, and Sargeant 2010).
Hypothesis 2: The use of both positive and negative emotional language by nonprofits to describe their work will be associated with a higher number of volunteers and more donations.
Variation in Emotions across Institutional Fields
Like other organizations, nonprofits are embedded in fields that carry with them their own norms, constituencies, and institutional logics (Barman 2016, 2017; Powell and DiMaggio 1991). An organizational field is defined as all actors connected to a recognized arena of social life and subject to the same institutional environment (Barman 2017; DiMaggio and Powell 1983). The nonprofit sector can be divided into a set of such fields, each of which has its own normative culture and institutional practices (“rules of the game”) that promote philanthropy (Marquis, Glynn, and Davis 2007). Similar practices and expectations develop because nonprofits in a given institutional field are subject to coercive, normative, and mimetic isomorphic pressures (Barman 2016). And in each field, smaller nonprofits mimic larger nonprofits to obtain solutions to commonly-shared problems (Marquis 2003). All these pressures toward isomorphism mean nonprofits in an established field will come to look alike in terms of their formal structures and policies and the way they interact with external constituencies (Galaskiewicz and Bielefeld 1998; Peyrefitte and David 2006).
Organizational culture is one important characteristic shaped by institutional fields (Martin et al. 1983). Although each nonprofit may have a unique culture, these internal cultures are developed by drawing from symbols, values, and meaning systems already present within the broader environment. The regulation and promotion of emotion is a core aspect of organizational culture (Hochschild 1983; Smollan and Sayers 2009). Both internally and externally, organizations work to create affective climates that mimic their mission or ethos (Smollan and Sayers 2009). Nonprofits within a field may thus cohere around certain emotional themes or practices in their missions.
Nonprofit organizations are typically categorized in accordance with their primary purpose, type, or major function (Fyall, Moore, and Gugerty 2018). Each individual nonprofit may have their own expectations for presentation and behavior, but nonprofits within an area of focus, such as arts, education, or human services, are also likely to face issue-area institutional logics that influence how they construct a mission statement and the level of emotional valence that is deemed appropriate (Marquis 2003; Marquis et al. 2007).
Donors and volunteers may interact with these broader institutional fields prior to any specific nonprofit. Because volunteers and donors are advised to choose their area of passion before selecting a nonprofit to support (Bronfman and Solomon 2010), they come to a nonprofit with a priori expectations of how an organization within their passion area, or field, operates. As Berlan (2017) outlines, the sense-making process individuals go through upon encountering a nonprofit’s mission is conditioned, in part, by these a priori expectations and differing institutional logics. Moreover, different fields may attract different types of donors and volunteers. Civil rights or other social change organizations may recruit and retain activist donors and volunteers who seek challenge and change, whereas other institutional fields may support people who want to serve in more traditional volunteer roles (Eliasoph 2013; Ganz 2009; Han 2014). As such, we argue that the expectations for use of emotion will vary across nonprofit classification—and, in turn, will influence how volunteers and donors respond to these emotions.
The operation of emotions is going to differ along the specifics of each institutional field, but there are broad similarities among fields. In particular, we argue that nonprofits in institutional fields broadly focused on social bonding and interaction (e.g., sports and recreational clubs and arts organizations) will be particularly likely to use and be successful with positive emotions, whereas nonprofits in institutional fields associated with identifiable social problems or challenges (e.g., environmental and poverty organizations) will more often focus on and be successful when highlighting negative emotions.1
Positive Emotions and Social Bonding
We argue that nonprofits in institutional fields broadly focused on social bonding and interaction will be particularly likely to use and be successful with positive emotions. Fredrickson’s (2001) “broaden and build” theory outlines how positive emotions expand people’s sense of self and broaden their repertoires of action so they are more encompassing of others. Positive emotions like pride and love, for example, drive people to want to engage and share with others (Lewis, Haviland, and Barrett 2010), and positive emotions are also tied to creativity, play, and community (Frijda 1986). In performing arts organizations, groups whose missions stress positive dimensions of commonality like friendship, cooperation, and sharing are able to attract more external engagement (Pandey et al. 2017). Going back to Wollschleger’s (2017) example of church services, it was the collective euphoria generated from the social bonding that motivated congregants to return. In short, nonprofits associated with social bonding may be particularly rewarded for invoking positive emotion.
Hypothesis 3: The relationship between positive emotional language and volunteers/donations will be stronger for institutional fields stressing community and social bonding.
Negative Emotions and Social Problems
Nonprofits associated with institutional fields centered around identifiable social challenges may be particularly persuasive with negative emotions. As Kandrack and Lundberg (2014:58) state, “negative mood priming (e.g. sadness) is mediated by an attempt to comfort oneself, to engage in self-therapy.” In other words, when potential donors or volunteers encounter challenges, like feeding the hungry, couched in emotionally negative ways, they should be motivated to alleviate the discomfort this presents by donating or volunteering. The helping response is triggered to alleviate or resolve the negative emotions—anxiety, distress, anger, or guilt (Bagozzi et al. 1999; Lazarus 1991; Marcus 2013).
Experiments suggest that stressing victims, itself a term evoking the negative, yields greater donations with a greater sense of urgency (Small and Loewenstein 2003). How these victims are conceptualized conditions these effects, with a key distinction between identifiable versus statistical victims (Schelling 1968). Identifiable victims are specific individuals, whereas statistical victims are impersonal, like “humanity” or “Americans.” Research suggests identifiable victims produce greater engagement precisely because this language results in greater anxiety, sadness, and depressing feelings (Schelling 1968). Following from this, nonprofits in institutional fields focused on social challenges at the collective (e.g., the environment and civil rights) or individual (e.g., food banks, housing, employment, and crime) level should have greater success with negative emotions, as these issue areas play into external publics’ desire to reduce harm and restore emotional comfort though action.
Hypothesis 4: The relationship between negative emotional language and volunteers/donations will be stronger for institutional fields stressing social challenges.
Differential Influence of Emotion on Donations and Volunteers
Studies demonstrate high correlation between donating and volunteering, suggesting the two phenomena stem from similar forces, such as a propensity for social participation and a pro-social identity (Lee, Piliavin, and Call 1999). Therefore, we might expect a nonprofit’s use of emotion would similarly influence both donations and volunteers. However, there are important distinctions between each act that might make donors and volunteers differentially susceptible to a nonprofit’s emotional content.
A salient distinction for a nonprofit in attracting donations versus volunteers is the level of social engagement and effort needed for an individual to complete either task. Nonprofits have an incentive to lower any and all barriers that can potentially disrupt the process of donating, by, for example, accepting online donations (Waters 2007). Nonprofits do hold donating events such as galas, luncheons, and charity walks, but donating can be completed with minimal effort, in private, anonymously, and according to an individual’s own schedule and timeframe (Lee et al. 1999). Therefore, donations require low levels of social engagement and effort and, when motivated, can be accomplished almost instantaneously. Volunteering, on the other hand, takes more effort, is often more social, and is an embodied experience (Borgonovi 2008; Ringmar and Mast 2018).2 Volunteering tasks often involve directly helping others, or working with others to complete a task. The most typical tasks include preparing food, collecting and delivering clothing and other goods, providing direct care, teaching, and counseling and mentoring, which can take considerable time and mental or physical effort (McKeever 2015). Thus, volunteering is more social, meaning it is more linked to social connections and subject to social expectations, than is donating (Borgonovi 2008).
We noted that positive emotions expand people’s sense of self, create a feeling of community, and encourage engagement with others (Fredrickson 2001; Frijda 1986; Lewis et al. 2010). As volunteering is the more social of the two acts, it should be especially responsive to positive emotion compared to donating. Positive emotion, when used by a nonprofit, links naturally to sociality, cooperation, and sharing, which should attract more engagement by volunteers (Pandey et al. 2017). As volunteering is also the more effortful and embodied of the two acts, this positive emotion may be exceptionally important to stimulate the deliberative decision to engage with the nonprofit. Positive emotion suggests social and embodied rewards in tandem with others that would help align conscious and unconscious dual processes (Pagis 2010; Ringmar 2020; Ringmar and Mast 2018). In short, nonprofits may be particularly rewarded for invoking positive emotion in the case of volunteers. Comparing the effect of emotion used in describing a nonprofit’s work on donations and volunteers will thus provide insights into whether emotion more easily produces low or high effort and social engagement with nonprofits.
MISSION IN ORGANIZATIONS AND NONPROFITS
We consider emotion in nonprofit missions. Whether in the for-profit, nonprofit, or public sectors, missions are an important organizational characteristic that “capture an organization’s unique raison d’etre.” An organizational mission outlines “where a firm is headed; how it plans to get there; what its priorities, values, and beliefs are; and how it is distinctive” (Williams 2008:96). Consequently, statements to convey this mission are now a ubiquitous feature of organizations; they are necessary to convey legitimacy and communicate organizational purpose and identity to internal and external audiences (Pope et al. 2018).
An organization’s mission reflects its culture and the priorities that frame its decisions (Alvesson 2012; Deshpande and Webster 1989; Schein 2010). Missions serve a range of internal purposes. They define the common symbols and meanings of an organization’s activities (Alvesson 2012), motivate employees (Kim and Lee 2007), and direct decision-making and action (Koch, Galaskiewicz, and Pierson 2015). For these reasons, missions are central documents of organizational culture. But missions are also public facing (Brinckerhoff 2009; McDonald 2007; Vandijck, Desmidt, and Buelens 2007; Young 2001). Service and advocacy organizations are aware that they need to signal the right messages in their mission to communicate to external audiences and acquire resources (Campagna and Fernandez 2007; Eng, Liu, and Sekhon 2012). Missions also legitimize an organization among its institutional peers (Koch et al. 2015; Morphew and Hartley 2006). Koch and colleagues (2015) conclude that nonprofit missions are partly cultural, defining a priori the tasks and values of an organization, but also partly adapted to be competitive and close to similar organizations.
Among organizations, mission is especially important to nonprofits. Compared to private entities that maximize profits or governmental organizations that answer to taxpayers, nonprofit organizations are held accountable by the extent to which they fulfill this mission. Missions define the very purpose of a nonprofit (Minkoff and Powell 2006; Pope et al. 2018) and motivate members toward a goal (Bart, Bontis, and Taggar 2001; Frumkin 2002). Stated more succinctly, “nonprofits only exist to pursue the specific public purposes that are expressed in their missions” (Berlan 2017:413). Consequently, the mission of the nonprofit takes center stage when making decisions and communicating goals and organizational identity to external publics. The centrality of mission to nonprofits is evidenced by Pope and colleagues (2018), who find that among the 100 largest corporations and nonprofits, 91 percent of nonprofits displayed mission statements on their homepage and only 58 percent of corporations did.
A mission statement is a relatively short statement reflecting this more generalized, overall mission. Mission statements serve as “legitimizing myths” that hold an organization together but also signal organizational features to external constituencies (Koch et al. 2015; Morphew and Hartley 2006). Potential donors or volunteers do not necessarily read the mission statement, but they ought to encounter the organization’s general mission, which, as both aspirational and practical, should imbue all of a nonprofit’s activities, reveal its values, and reflect and create its culture. A nonprofit that takes a more emotional approach to its goals, work, and clients should have a mission, and mission statement, that reflects this approach.
As both reflective of and generating organizational culture, missions play an active role throughout all aspects of organizations. Missions and mission statements appear and reappear throughout organizational materials such as newsletters, annual reports, brochures, posters, and business cards (Fairhurst, Jordan, and Neuwirth 1997). Reading text is one pathway through which affective responses may be elicited (Gross 2008), and research demonstrates donors are likely to encounter a group’s mission prior to donating (Balsam and Harris 2014) and “mission can attract or repel volunteers” (Nesbit, Christensen, and Brudney 2018:505).
DATA AND METHODS
We evaluate our hypotheses about whether the emotional language expressed by nonprofits is associated with donations and volunteering by examining nonprofit mission statements and other information as recorded in IRS nonprofit administrative reporting forms—the Form 990. The Form 990 is an annual return required by the IRS for most nonprofit organizations and is an immense source of data on nonprofits: their finances, expenditures, governance, mission, compliance with federal requirements, compensation paid to certain persons, and numbers of staff and volunteers.
In 2016, the IRS released 1.3 million Forms 990 through Amazon Web Services for all nonprofits covering the period 2010 to 2015, and it has continually updated the data since.3 The new IRS e-filer data release provides complete 990 financial information for all e-filing nonprofits (about 60 to 65 percent of all 990 and 990-EZ filers). Having such a large number of Forms 990 opens up many new avenues for research. But note that nonprofits with revenues less than $50,000 are not required to submit these forms. And, although churches are tax-exempt entities, they are not required to file Forms 990. This is a notable omission given that churches are an important institution through which people donate and volunteer. The Form 990-EZ does not ask about volunteers, so we limit our analysis to nonprofits that file Forms 990, meaning they tend to have gross receipts above $200,000.
We collected our data from Amazon Web Services in April 2018 and our analyses include tax filings for 501(c)3 organizations between 2012 and 2016. In our main analyses we take the most recent filing for a nonprofit so we do not have multiple returns for the same Employment Identification Number (EIN) across years (88 percent of our observations are 2015 or 2016 filings). Initially, our sample consisted of 130,393 unique 501(c)3 nonprofit 990 filings in 12 NTEE categories between 2012 and 2016. In the online supplement, we discuss the constraints we applied to yield a final set of 89,528 unique 990 filings. In auxiliary longitudinal analyses, we analyzed multiple nonprofit years of data.
Dependent Variables
We measure donations by aggregating four revenue sources from the Form 990: membership dues (Part VIII 1B), contributions from fundraisers (Part VIII 1C), non-cash contributions (Part VIII 1G), and other contributions that exclude government grants, federated campaigns, and revenue from related organizations (Part VIII 1F).4 We drop nonprofits who never receive donations (see the online supplement). Due to a skewed distribution, donations are winsorized at the 99th percentile and then logged.
To measure volunteers, we use the total number of volunteers reported by nonprofits on Part I Line 6 of the Form 990. The IRS provides the following guidance to organizations on how to determine this number: “Make a reasonable estimate of the number of persons that did any type and amount of volunteer work for your organization during the tax year, not including your employees who may have done volunteer work in their spare time.” Organizations have the option, but are not required, to provide further clarity in how they define “volunteer” in Schedule O. We winsorize volunteers at the 99th percentile because investigation of the highest reported numbers of volunteers revealed some irregularities. For example, the American Heart Association appears to count anyone as a volunteer who watched a video about CPR on their website. Once winsorized, volunteers are logged.5
Positive and Negative Emotions through Computerized Text Analysis
To assess the emotional valence used by nonprofits, we use computerized text analysis to determine positive and negative sentiment within each organization’s mission statement. Specifically, we process self-reported mission statements from Part III Line 1 of the Form 990, which instructs nonprofits to “Describe the organization’s mission as articulated in its mission statement or as otherwise adopted by the organization’s governing body, if applicable. If the organization does not have a mission that has been adopted or ratified by its governing body, enter ‘None.’”6
We processed mission statements to remove some punctuation and convert to lowercase, correct misspellings, change any Britishisms in the text to American spelling, and isolate words. As is common in computerized text analysis, we removed stop words, which are frequently used words such as articles, pronouns, and prepositions (e.g., “a,” “is,” or “and”), because they generally do not carry direct relevance to the analysis at hand.
To see how the mission of an organization imbues both the mission statement and outward-facing materials, consider the similarities (italicized) in the following example. The Form 990 reported mission statement for Helping Hand Home for Children states: “To provide a nurturing and therapeutic home for children and to restore each child to a healthy family setting. Children, whose young lives were once filled with fear, pain, and chaos, are learning to trust adults to take care of them and reclaim their childhood.” The mission as it appears on the front page of the nonprofit’s website is a condensed version: “Our mission is to provide a nurturing and therapeutic home for children and to restore each child to a healthy family setting.” An example of a fundraising appeal also uses language pulled from the mission statement: “When you support Helping Hand Home for Children, you are helping severely abused children rebound from trauma and rediscover their childhood. Thanks to community support, Helping Hand Home is able to provide a place to heal for these children. Children whose lives were once filled with fear, pain, and chaos leave here with hope and a chance to have a bright future.”
To determine the validity of nonprofit mission statements as a measure of nonprofit presentation of mission to external constituencies, we compared mission statements as reported on the Form 990 with missions or similar statements on nonprofit webpages for more than 1,000 nonprofits across five fields (reported in the online supplement). We found that most mission statements as reported on the Form 990 are either exactly represented on nonprofit webpages or summarized using very similar language.
Finally, we eliminated mission statements that consisted of fewer than five words. Our pilot study of these mission statements suggested that short mission statements on the Form 990 are not mission statements as typically defined, for example, “charitable organization” or “services for the elderly.” Such short mission statements may not reflect how an organization presents itself on its website as assessed during our pre-study. Therefore, using mission statements as reported on the 990s that are longer than four words (after removing stop words) is an appropriate indicator for nonprofits’ frame of their work to general audiences.
We use Linguistic Inquiry Word Count (LIWC) to measure the positive and negative emotion of mission statements. Pennebaker and colleagues (2015) created the LIWC language dictionaries as a tool to assess mental states and psychological characteristics. They used teams of four to eight human judges to generate lists of words that conceptually match a given dictionary topic, such as “positive emotion,” aided by standard dictionaries, Roget’s Thesaurus, and “hundreds of thousands of text files from multiple studies and sources” (Pennebaker et al. 2015:6). If a majority of judges could not agree on a word’s appropriate fit it was discarded. Once sets of words were generated, they were psychometrically evaluated for internal consistency. If any word was detrimental to the internal consistency for the overall category, two to eight human judges reassessed its fit. LIWC considers both formal language and naturally occurring language (e.g., “b4”) and is a widely used source of information in social and psychological research (Bail et al. 2017; Pennebaker et al. 2015).
To measure the positive and negative emotions within each piece of text, LIWC generates a score that is the percentage of positive or negative emotional words over the total number of words in the statement. Negative emotions include words such as hopeless, poor, and victim. Positive includes comfort, dignity, and well-being. To illustrate this, in Table 1 we highlight two paired missions of organizations that are working on the same issue—(1) support for individuals experiencing neglect and abuse, and (2) free food distribution programs—but that discuss their work in opposing ways. Words that LIWC categorized as positive and negative are bolded.
Table 1.
Examples of Use of Emotional Language in Nonprofit Mission Statements
NTEE Subcategory | Emotion | Original Mission Statement | Stop Words Removed | Score |
---|---|---|---|---|
| ||||
Crime and Legal Related: Protection Against Neglect, Abuse, Exploitation | Negative | the abused women’s fund assists women who are or have been abused; battered; underestimated; unloved; shunned; afraid; and crying out | abused women’s fund assists women abused battered underestimated unloved shunned afraid crying | 38.46 |
Positive | vanessa behan crisis nursery is dedicated to improving the lives of children by providing immediate safety; refuge; and ongoing family support in an environment of unconditional love | vanessa behan crisis nursery dedicated improving lives children providing immediate safety refuge ongoing family support environment unconditional love | 22.22 | |
Agriculture, Food, and Nutrition: Food Service, Free Food Distribution Programs | Negative | to provide food to organizations that feed the poor; distressed and the underprivileged | provide food organizations feed poor distressed underprivileged | 28.57 |
Positive | we provide food at no cost in a caring and respectful manner to those in need in our community while preserving the dignity of those we serve | provide food cost caring respectful manner need community preserving dignity serve | 27.27 |
Note: Words that LIWC categorized as positive and negative are bolded.
We generate two measures of positive and negative emotion for each mission. The first measure is the percent of positive and negative words over total words as initially reported by LIWC. These values are logged. We also create a series of dummy variables based on these scores. Mission statements that have no words coded as either negative or positive are categorized as “no sentiment”; statements with only positive words are categorized as “positive”; statements with only negative words are categorized as “negative”; and statements containing both positive and negative words are coded as “both.” In our analyses, “no sentiment” is the reference category.
Nonprofit Institutional Fields
To account for varying institutional logics, we separately analyze nonprofits across 12 institutional fields that are reflective of either social bonding or social problems. The National Taxonomy of Exempt Entities (NTEE) groups entities into 26 categories by purpose, type, or major function, including arts and culture, education, health, and human services. These NTEE codes are a standard and widely-used classification system within nonprofit research. We chose 12 of the 26 NTEE categories as a set of diverse yet compact, relevant, and interesting institutional fields that reflect a social bonding or social problems tendency based on our conceptualizations. The categories under social bonding are arts, culture, and humanities; education; recreation and sports; youth development; and religion-related. Categories under social problems are environment; healthcare; crime and legal related; employment; agriculture, food, and nutrition; housing and shelter; and civil rights and social action. Table A1 in the online supplement provides information on the subcategories for each of our 12 selected NTEE categories and includes example mission statements to provide context on the type of work these organizations do and variation in mission statement styles.7 The 12 NTEE categories we include cover 61 percent of all nonprofits.
Descriptive statistics and differential use of emotional words (see Table 2) suggest institutional fields should be investigated separately. This strategy is indeed common in the nonprofit literature (e.g., Calabrese 2011). In the analyses, we report results for each of the 12 institutional fields. Within each field we further control for subcategories to recognize that use of emotion, donations, and volunteers may differ across the subcategories of an institutional field. Although we broadly categorize each field under social bonding or social problems, subcategory variation needs to be taken into account. Indeed, some subfields include organizations that are known to have difficulty recruiting volunteers, such as groups that work with people who use drugs or who have chronic illnesses (Hager and Brudney 2011). Such differences must be controlled to better assess the broader institutional field.
Table 2.
Understanding Positive and Negative Emotion by Field
Percent of Mission Statements Expressing a Particular Emotion |
Percent of Emotional Words out of Total (Non-Stop) Words |
Top Five Emotional Words |
|||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Positive |
Negative |
||||||||||
Institutional Field | Just Positive | Just Negative | Both | Mean | Min. | Max. | Mean | Min. | Max | Positive | Negative |
| |||||||||||
Social Bonding | |||||||||||
Arts, Culture, and Humanities | 61.75 | 1.46 | 5.38 | 7.64 | 0 | 60.00 | .42 | 0 | 50.00 | support*, opportun*, inspir*, encourag*, engag* | diss*, low, war, critical, risk |
Education | 63.83 | 2.00 | 8.14 | 8.26 | 0 | 80.00 | .61 | 0 | 50.00 | support*, opportun*. excel, care, excellence | emotional, low, critical, risk, problem |
Recreation and Sports | 59.83 | 1.55 | 5.21 | 8.17 | 0 | 71.43 | .41 | 0 | 42.86 | opportun*. support*, play, encourag* . well | emotional, low, risk, disadvantag*, sever |
Youth Development | 60.68 | 3.69 | 15.50 | 9.52 | 0 | 80.00 | 1.40 | 0 | 60.00 | support*, caring, inspir*. opportun*. care | risk, disadvantag*. emotional, low, advers* |
Religion-Related | 53.43 | 4.17 | 11.70 | 7.85 | 0 | 75.00 | 1.29 | 0 | 42.86 | support*, charit*, faith, care, encourag* | poor, needy, low, emotional, abuse |
Social Problems | |||||||||||
Environment | 59.89 | 2.86 | 7.89 | 7.18 | 0 | 50.00 | .68 | 0 | 60.00 | support*, improve, benefit, trust, energ* | low, problem, critical, threat, diss* |
Healthcare | 67.63 | 2.34 | 13.72 | 10.67 | 0 | 80.00 | 1.05 | 0 | 42.86 | care, support*, improve, well, compassion | low, poor, emotional, sick, critical |
Crime and Legal-Related | 31.96 | 16.54 | 34.50 | 6.98 | 0 | 55.56 | 5.80 | 0 | 57.14 | support*, free, safety, improve, safe | abuse, victim, violence, low, neglect |
Employment | 51.30 | 4.59 | 11.61 | 6.70 | 0 | 60.00 | 1.04 | 0 | 40.00 | opportun*. support*, care, improve, good | disadvantag* . low, risk, sever, critical |
Agriculture, Food, and Nutrition | 46.93 | 11.70 | 14.63 | 6.60 | 0 | 42.86 | 2.40 | 0 | 40.00 | support*, healthy, care, improve, well | needy, low, poor, fight, insecur* |
Housing and Shelter | 31.90 | 26.30 | 24.08 | 6.13 | 0 | 62.50 | 5.20 | 0 | 50.00 | support*, safe, care, profit, opportun*. decent | low, disadvantag*. poor, needy, trauma |
Civil Rights and Social Action | 47.60 | 6.90 | 25.58 | 8.07 | 0 | 50.00 | 3.27 | 0 | 50.00 | support*, interest, free, safe, best | abuse, neglect, defend, low, violence |
Note: Asterisks indicate stemmed words. Unique sentiment words are bolded.
Alternative Explanations
Emotional missions are not the only characteristic of a nonprofit we would expect to influence volunteers or donations. Indeed, previous investigations of nonprofit donations stress a range of financial characteristics, such as administrative expenses, wealth hording, and the strategic deployment of resources as important to donations (e.g., Calabrese 2011; Charles 2018; Weisbrod and Dominguez 1986). This line of research demonstrates that donors are cognizant of and responsive to the financial stewardship of an organization, wanting assurance their donations will be put to good use. The 990 e-filer release is the first time information on volunteers is available across a wide range of nonprofits for investigation; we thus do not have prior research to draw on for alternative, nonprofit-level explanations. For volunteers, Nesbit and colleagues (2018) theorize several organizational characteristics that should influence volunteer involvement, including financial resources, the number of paid staff, government contracts, and reliance on commercial income. Volunteers, and donations, may also be associated with a nonprofit’s level of professionalization. Betzler and Gmür (2016) find that nonprofits that professionalize their fundraising strategies, in part by investing in fundraising capacities and incorporating advice from fundraising experts, are better able to attract donations.
Therefore, in our models, we control for the following: (1) price: total expenses divided by program expenses, logged; (2) fundraising: fundraising expenses of the organization, winsorized and logged; (3) outside funds: “outside” funding from government grants and program service revenues, winsorized and logged; (4) assets: the end-of-year total asset balance, winsorized and logged; (5) employees: the total number of recorded employees, winsorized; and (6) age: the number of years since the organization received tax exempt status. We also control for (7) word count, that is, the number of words in the nonprofit’s mission statement, and, for analyses divided by institutional field, we control for subcategory across each institutional field (see Table A1 in the online supplement).8
Methods: Regression with p-value Adjustment and Robustness to Sample Bias
Linear regressions predict logged donations and logged volunteer counts with our sentiment variables, controls for alternative explanations, and NTEE subcategory codes. Because we used multiple tests, we use a p-value adjustment to control for the Type I error rate. As we outline, to verify the robustness of our results we tested various sample limitations and variable constructions. In all models, arts, culture, and the humanities serve as the reference category. These well-studied organizations are generally emotionally expressive and operate with greater reliance on donations compared to other fields (Kim 2017; Salamon 2011).
Due to the fact that we only have a sample of nonprofits out of the entire population—nonprofits that e-file Forms 990—we test the robustness of our results in relation to potential sampling bias using an approach developed by Frank and colleagues (2013). Based in Rubin’s causal model, this analysis determines how much bias in the design components there must be to invalidate an inference (Frank et al. 2013). Here, our target population—all nonprofits—contains those represented in our sample—e-filers—as well as those not directly represented by our sample. How much of our sample would have to be replaced with other cases, under the limiting condition of no effect in those cases, to invalidate our inference? Put another way, how many nonprofits in our sample would have to be replaced by nonprofits in which there is no association between use of emotion and volunteers and donations to invalidate our inferences? To preview our results: we find that to invalidate our significant coefficients, we would need to replace 70 to 89 percent of nonprofits in our sample, depending on the model, with ones with no association between emotion and volunteers and donations, suggesting the results presented here are robust to the sampling bias.
Auxiliary Analyses: Experiments and Longitudinal Analyses
We supplement the cross-sectional design of our main analyses with experiments and longitudinal models. Doing so allows us to move beyond significant patterns of association and better infer the relationships between emotions and donations/volunteers. First, our experiments examine whether people are more willing to donate to a charity when a mission-based appeal is couched in emotional language. We focus on two areas of charitable work: holiday gift provision to children and food banks. We pair the experiments for each type of nonprofit with a corresponding subset of IRS administrative data to see if results converge. Second, for nonprofits with administrative data over time, we investigate the influence of sentiment change on donations and volunteers longitudinally. These auxiliary analyses bolster the broad findings established in our main analyses.9
RESULTS
Nonprofits use emotion-laden words in their mission statements. Table 2 shows the average percent of positive or negative emotion words in mission statements across 12 nonprofit fields, organized by social bonding or social problems. The table also presents the percentage of nonprofit missions in each category that use positive language, negative language, or both. Differences across nonprofit fields are immediately apparent. On average, all nonprofit fields use more positive emotion words as a percent of their mission statement than negative emotion words. But some categories use negative emotion at much higher rates, for example, crime and legal-related, housing and shelter, and civil rights—all of which fall under our social problems categorization. Some of this is related to the areas in which they work. Crime and legal-related nonprofits are naturally more likely to use victim, abuse, or violence (all negative words in LIWC) than are arts nonprofits. However, these categories also use fewer positive words, suggesting an overall more negative focus.
Social bonding fields typically use more positive emotion words; however, healthcare uses positive emotion words most extensively (67 percent use at least one positive word in their mission statement). The most common positive words used by these nonprofits are care, compassion, and well. A high percentage of arts nonprofits also use positive words in their mission statements.10 The most common words in this category differ, with a focus on encourage, engage, and inspire. The most common positive and negative words are similar within but different across the nonprofit fields, suggesting overarching isomorphism pressures driven by coercive, normative, or mimetic pressures (Barman 2016). Nonprofits in established fields come to look alike in terms of their formal statements of mission and presentation to external constituencies. This descriptive isomorphism supports separating analysis by institutional field and by social bonding versus social problems.
Is use of emotion-laden words by nonprofits associated with higher levels of donations or more volunteers? Table 3 shows the coefficients for regressions predicting logged donations and volunteers with full controls, dummies for all NTEE categories, and dummies for all subcategories within NTEE. For donations, across all institutional fields, using a higher percent of positive and negative emotion words in a mission statement is associated with higher levels of donations. A 10 percent increase in positive emotion words is associated with close to a one-half percent increase in donations ((1.10.041 – 1) × 100). The effect size for negative emotion is slightly larger at .07, such that a 10 percent increase in negative emotion words is associated with about two-thirds of a percent increase in donations. These results are confirmed when use of emotion is considered as a series of dummy variables. Compared to mission statements that do not include any emotion, nonprofits using any positive words are associated with about 13 percent ((exp(.124) – 1) × 100)) higher donations. Donations are associated with an almost 22 percent increase when a nonprofit’s mission statement contains a negative emotion word, compared to nonprofits with missions containing no sentiment. Finally, the combination of negative and positive emotion is particularly positive: donations are expected to be 29 percent higher when nonprofits use both positive and negative emotion words in their mission statement.11
Table 3.
Full Models Predicting Nonprofit Donations and Volunteers
1 |
2 |
3 |
4 |
|
---|---|---|---|---|
Donations |
Volunteers |
|||
Model | Logged Measure | Dummy Indicators | Logged Measure | Dummy Indicators |
| ||||
Positive Emotion | .041*** (.005) | .124*** (.015) | .115*** (.006) | .273*** (.017) |
Negative Emotion | .073*** (.008) | .197*** (.033) | .018 (.01) | .000 (.04) |
Both | .258*** (.023) | .324*** (.027) | ||
Controls | ||||
Price (logged) | −.350*** (.023) | −.350*** (.023) | −.332*** (.027) | −.332*** (.027) |
Fundraising (logged) | .173*** (.001) | .173*** (.001) | .114*** (.001) | .115*** (.001) |
Outside Funds (logged) | −.043*** (.001) | −.043*** (.001) | .035*** (.001) | .035*** (.001) |
Net Assets (logged) | .215*** (.003) | .215*** (.003) | .067*** (.003) | .068*** (.003) |
Employees | .001*** (.000) | .001*** (.000) | .001*** (.000) | .001*** (.000) |
Age | −.003*** (.000) | −.003*** (.000) | .010*** (.000) | .010*** (.000) |
Word Count | .006*** (.000) | .005*** (.000) | .003** (.000) | .001** (.000) |
Institutional fields | ||||
Social Bonding | ||||
Education | −.066*** (.02) | −.068*** (.02) | −.286*** (.024) | −.287*** (.024) |
Recreation and Sports | −.303*** (.025) | −.302*** (.025) | .457*** (.029) | .463*** (.029) |
Youth Development | .062 (.032) | .063 (.032) | .553*** (.038) | .559*** (.038) |
Religion-Related | .638*** (.026) | .638*** (.026) | −.334*** (.031) | −.336*** (.031) |
Social Problems | ||||
Environment | −.077 (.033) | −.078 (.033) | .403*** (.039) | .399*** (.039) |
Healthcare | −.294*** (.023) | −.292*** (.023) | −.081** (.027) | −.071** (.027) |
Crime and Legal-Related | −.048 (.038) | −.041 (.038) | −.192*** (.045) | −.192*** (.045) |
Employment | −.203*** (.044) | −.207*** (.044) | −.706*** (.052) | −.713*** (.052) |
Agriculture, Food, and Nutrition | .603*** (.044) | .602*** (.044) | 1.193*** (.052) | 1.191*** (.052) |
Housing and Shelter | −.307*** (.031) | −.308*** (.031) | .146*** (.036) | .147*** (.036) |
Civil Rights and Social Action | .308*** (.053) | .312*** (.053) | −.097 (.062) | −.100 (.062) |
Constant | 8.699*** (.043) | 8.697*** (.043) | .609*** (.051) | .632*** (.052) |
Observations | 89,529 | 89,529 | 89,529 | 89,529 |
R-Squared | .366 | .366 | .194 | .193 |
Note: These models control for NTEE subcategories. Arts, culture, and humanities serves as the reference category. Standard errors are in parentheses.
p < .017
p < .008
p < .001
p < .00017 (two-tailed tests). (Our p-value adjustment acknowledges multiple tests and controls the Type 1 error rate.)
Table 3 shows different results for volunteers. Model 3 indicates that the effect of positive emotion is higher, and negative emotion has a negligible effect. A 10 percent increase in positive emotion words in a mission statement is associated with over a 1 percent increase in volunteers. The standardized effect size for percent positive emotion (.06) is stronger than the effect of price (–.04) and similar in size to the influence of assets (.08). In contrast, the effect of using a higher percent of negative emotion words in a mission statement is insignificant. A lack of effect of negative emotion on volunteers is echoed in Model 4, where the inclusion of any negative emotion word by a nonprofit is not significantly different from no sentiment. In contrast, use of any positive emotion in a mission is associated with a 31 percent (exp(.271) – 1) × 100) increase in volunteers. Social and effortful volunteering thus appear especially responsive to positive emotion compared to negative emotion. However, use of both positive and negative emotion in a mission statement is slightly higher—producing a 38 percent increase in volunteers; this is a significant increase over positive emotion alone.
Table 3 includes some financial variables common in prior research on donations. Our results generally mirror those reported in prior research. Price, fundraising, outside funds from government grants and program service revenues, and assets all significantly predict donations. The higher the price of a donation, for example, the lower donations observed. In contrast, as we would expect, more fundraising expenses yield more donations and is one of the strongest effects in the model. Effects are similar for these variables in predicting volunteers. Age has a very small negative effect on donations and a very small positive effect on volunteers. Employees are expected to increase the capacity for volunteers, and, although the estimate is positive, it is very small. A 1 percent increase in employees is only expected to increase volunteers by .001 percent. Variation in volunteers appears to be influenced by factors other than employees. The total length of a mission statement, word count, does not independently influence donations or volunteers to any large extent. Also, the differences in R-squares across the models are important to note. Namely, the emotion variables along with the controls account for roughly 36 percent of the variation for the donation models but only 19 percent in the volunteer models. Incorporating the rational dimensions of a nonprofit seems to explain financial investments better than time investments, considering that time investments come with much stronger social expectations (Borgonovi 2008).
Table 3 also shows differences in donations and volunteers across institutional fields, as measured by NTEE codes. Arts, culture, and humanities (a social bonding field) is the omitted category, and we see significant variation in donations and volunteers across fields. Arts organizations tend to receive more donations than many other categories of nonprofits. For example, healthcare, housing and shelter, and recreation and sports all receive roughly 26 percent less in donations compared to arts organizations. In contrast, religion-related and agriculture, food, and nutrition both receive, on average, over 82 to 89 percent more in donations compared to arts organizations, and civil rights receives 36 percent more. Food, agriculture, and nutrition nonprofits have, on average, 229 percent more volunteers than do arts organizations, which makes sense considering the number of food banks that heavily use volunteers within this category. On the opposite end, employment organizations are associated with 51 percent fewer volunteers than the arts. These divergent trends underscore the importance of attending to institutional fields in these analyses.
Accounting for these important differences, Figures 1 and 2 provide the coefficients for the emotion variables across the 12 institutional fields, organized by social bonding and social problems. These come from two regressions (logged emotions or dummy variables) predicting logged donations and volunteers in each of the 12 fields. Each model includes controls for all other variables: price, fundraising, outside funds, employees, NTEE subcategories, word count, age, and assets. Thus, the figures present the results for 48 regressions: 2 dependent variables x 12 nonprofit fields x 2 alternative measures of emotion. Significance thresholds were appropriately modified to account for multiple tests.
Figure 1.
Coefficients from Models Predicting Donations and Volunteers with Logged Emotion
Note: These models control for price, fundraising, outside funds, net assets, employees, and age. Confidence intervals set to 99.2 percent.
Figure 2.
Coefficients from Models Predicting Donations and Volunteers with Dummy-Coded Emotion
Note: These models control for price, fundraising, outside funds, net assets, employees, and age. “No sentiment” serves as the reference category. Confidence intervals set to 99.2 percent.
In particular, as the percent of logged positive words increases in a nonprofit’s mission, arts, culture, and humanities; education; and healthcare nonprofits receive more donations. For volunteers, all categories under social bonding are associated with more volunteers, whereas only healthcare, housing, and employment under social problems are associated with increases. Use of positive emotion by nonprofits is therefore strongly associated with higher numbers of volunteers for social bonding fields. Increases in the logged percent of negative words is significant across a larger number of fields for donations but a smaller range of fields for volunteers. More negative language is associated with higher levels of donations for arts, culture, and humanities; education; and religion-related in the social bonding fields and healthcare, agriculture, and food and nutrition in the social problems fields. Thus, we do not see a premium on donations for using negative emotion by nonprofits in the social problems fields. Among the social bonding fields, negative emotion is related to volunteering only for religion-related nonprofits. But it is associated with more volunteers for healthcare, employment, and civil rights among the social problems fields. Interestingly, although the direction of most coefficients suggests increases in negative language are more likely to be associated with a reduced number of volunteers, especially among social bonding nonprofits, only environment approaches our conservative significance threshold.
Turning to the models that include dummy indicators for the presence of positive sentiment, negative sentiment, or both, similar trends appear: use of positive language or a mixture of positive and negative are generally more likely to be associated with increases in engagement than is just negative language, compared to statements with no sentiment. This is true for four of five social bonding fields: arts and culture, education, recreation and sports, and religion-related. In the social problem fields, employment and agriculture, food, and nutrition see higher donations with the use of only negative language, with employment also associated with larger number of volunteers. Notably, using just negative emotional language drives down volunteers for youth development organizations—the only field with this type of association that reaches significance. In short, we do see evidence that nonprofits are embedded in fields (as are other organizations) that have their own norms, constituencies, and institutional logics. Of particular interest is the differential relationship between the nature of emotion displayed by nonprofits and donations/volunteers across nonprofit fields broadly defined as social bonding oriented versus social problems oriented.
Robustness of Inference to Sample Bias
Because our set of nonprofit organizations does not reflect the entire population, we use Frank and colleagues’ (2013) technique to assess the vulnerability of our results to sampling bias. Specifically, what percent of our sample of nonprofits would have to be replaced with nonprofits in which there is no relationship between use of emotion and donations or volunteers to invalidate our results? To invalidate the significant coefficients in our logged emotion models predicting donations, we would need to replace over 70 percent of the sample with nonprofits that have no association between either positive or negative emotion and donations. For our volunteer model, over 89 percent of our sample would need to be replaced to invalidate the significant association between volunteers and positive use of emotion (we do not perform this analysis on negative sentiment because it is insignificant). In effect, this means sampling bias would have to be so egregious that a significantly high threshold of our sample would have to be replaced with nonprofits having no association to invalidate our inference. Therefore, although our sample of nonprofits only comes from e-filers, the results from these sensitivity analyses suggest our results are more generalizable than not.
AUXILIARY ANALYSES USING EXPERIMENTAL AND LONGITUDINAL DESIGNS
The results reported thus far are novel but observational, limiting inference. Here, we first report on two experiments, paired with relevant subsets of our administrative data, to address this issue. Our experiments examine whether people would be more willing to donate to a fictional charity when a mission-based appeal is couched in emotional language. We focus on two areas of charitable work: holiday gift provision to children and donations to food banks. In each, experiments can be paired with subsets of our administrative data. The children’s gift provision experiment focused on a nonprofit called “Holiday Wishes,” and the food bank experiment discussed the “Centerville Food Bank.” To summarize our results, we see similar findings across experiments and relevant administrative subsets in both cases. We next report results from several longitudinal analyses that assess whether positive change in emotion from 2012 to 2014 is associated with donations or volunteers in 2015 and change in donations and volunteers from 2014 to 2015. These auxiliary analyses, using both experimental and longitudinal designs, improve confidence in the larger administrative analysis.
Experimental Design
We recruited an online sample (N = 584 Holiday Wishes, December 2019; N = 593 Centerville, January 2020) through Amazon Mechanical Turk (MTurk). Data collection was restricted to adults living in the United States.12 The mean age of participants across the two experiments was 39, and 46 percent identified as women. We based our design loosely on experiments 1 and 4 in Sussman, Sharma, and Alter (2015). All participants were told the following: “Imagine that you get a flyer in the mail for [Holiday Wishes, a charity that distributes toys to children during the holidays / Centerville Food Bank, a charity that distributes food to people]. Donating to charity is important to you, and you view this as a worthy cause, so you read the flyer closely. The flyer describes the mission of the charity. The flyer says:”
Then, respondents were randomly assigned into one of three conditions to describe the mission of the charity: positive emotional appeal (bold), negative emotional appeal (italicized), or a neutral control (plain text). The mission for Holiday Wishes was described as follows:
Holiday Wishes believes that at this time of year children [should be cherished with unconditional love / who are poverty-stricken and disheartened are crying out / shall get presents purchased and delivered]. Therefore, the mission of Holiday Wishes is to [bring hope and joy / fight adversity / deliver presents] by providing toys to [support / economically disadvantaged / individual] children during the holidays.
The mission for Centerville was described as,
Centerville Food Bank believes that [helping people enhance their quality of life is possible through the loving provision of food / vulnerable people suffering the pain of hunger is an unfortunate reality / permitting people access to an affordable food supply through a distribution network is efficient]. Therefore, the mission of Centerville Food Bank is to provide food at no cost [in a caring and supportive manner to / to poor, distressed, and needy / to people who are] members of our community.
The nonprofits themselves are fictional, but the language used in the positive and negative emotion conditions was informed by mission statements that appear in the administrative data. The text was presented alongside a corresponding banner ad (available upon request).
On the following page, with the flyer no longer in view, participants indicated whether they would donate to Holiday Wishes/Centerville and how much money they would donate. Participants were asked, “How likely would you be to agree to this donation request today?” and responded on a scale from 1 (very unlikely) to 7 (very likely). They were then asked, “In addition to your regular giving, you typically give $100 to new charities during this time of year. How much would you donate to Holiday Wishes / Centerville Food Bank?” Using a slider, they provided a number between 0 and 100. Participants also responded to demographic questions, openended items to assess understanding, and instructional manipulation checks designed to ensure participants were reading the instructions carefully and improve statistical power (Oppenheimer, Meyvis, and Davidenko 2009).13
Figure 3 presents the results for Holiday Wishes on the left-hand side. For Holiday Wishes, participants reported being significantly more likely to donate in the positive condition (5.1) compared to the neutral condition (4.8; Mann-Whitney p = .067), and average donation amounts were also significantly higher in the positive ($41.70) compared to the neutral condition ($35.19; Mann-Whitney p = .023). These are substantively significant findings. Fundraisers working with postal mailings are delighted to get responses with donations of 2 to 3 percent (Brooks 2015). A manipulation that increased the likelihood of donation by 6 percent, as here, would be celebrated. Similarly, increasing donations received by an average of $6.50, across the hundreds or thousands of donors that such mailers try to reach, would provide a real boost to nonprofit revenues. For Holiday Wishes, the negative emotion condition was not significantly different from the neutral condition in the likelihood of donation or amount of donation.
Figure 3.
Experimental and Administrative Models for Children’s Gift Provision and Food Banks
Does this correspond to similar administrative data? The bottom half of Figure 3 provides this comparison. Considering only the N = 1,879 nonprofits working in the closest NTEE subfield, “human services, children or gift provision,” positive emotion is significantly related to donations and negative emotion is not. In short, in both the Holiday Wishes experiment and Form 990 analyses limited to nonprofits related to children gift provision, positive emotion is successful (compared to no emotion) and negative emotion is not.
Interestingly, the Centerville experiment, presented on the right-hand side of Figure 3, shows a different pattern. For Centerville, participants were no more likely to donate in the positive emotion condition than in the no emotion control condition. Respondents in the negative condition, however, were significantly more likely to donate (5.2) than were respondents in the neutral condition (4.9; Mann-Whitney p = .08). Average donation amounts were not significantly higher in the negative condition compared to the neutral condition, however. Again, the difference in the likelihood of donation is a substantively important result.
It is worth pointing out that the experimental results as presented are the most conservative possible, with only participants who failed instructional manipulation checks excluded. If we also exclude participants who could be classified as nonprofit “skeptics” (e.g., individuals who wrote “This is a company getting donations from people that children will never see” or “It’s a scam”) or those who provided answers with apparently low relation to the topic (e.g., “Jesus Came For You” or “I want some money”) the results are far stronger. For example, for Holiday Wishes, participants who were not skeptics or off-topic were even more likely to donate in the positive condition (5.3) compared to the neutral condition (4.9; Mann-Whitney p = .036) and average donation amounts were also significantly higher in the positive ($42.30) compared to the neutral condition (34.54; Mann-Whitney p = .0084). Coefficients for Centerville were similar, with respondents in the negative condition more likely to donate (5.2) than those in the neutral condition (4.9), but the Mann-Whitney p is reduced to .046.
We see a somewhat similar correspondence with paired administrative data, as shown in the bottom of Figure 3, when looking at Centerville and food banks. Considering only the N = 1,422 nonprofits working in the closest NTEE subfields, “food banks and programs, soup kitchens, etc.”14 positive emotion is significantly negatively related to donations (logged measure: –.19** [.07] / dummy indicator: –.46* [.19]), and negative emotion is positively and significantly related to donations (logged measure: .21** [.07] / dummy indicator: .42 + [.23]).15 The Centerville experiment replicates and extends findings in the Holiday Wishes experiment by showing that negative emotion can also enhance charitable behavior, in this case measured by donation likelihood.
In brief, in the paired Holiday Wishes experiment/administrative data subfield analysis, positive emotion is related to donations in both cases and negative emotion is not. In contrast, in the paired Centerville Food Bank experiment/administrative data subfield analysis, negative emotion is related to donations in both cases but positive emotion is not. The similar findings in both cases improve confidence in the larger administrative analysis. Participants making donation decisions chose to give more often and greater amounts when the charitable appeal was framed using positive emotion in gift giving to children, and they chose to give greater amounts when negative emotions were used for food banks. Our paired lab and administrative studies provide converging evidence that emotional language enhances charitable behavior.
Longitudinal Design
Of the 89,529 nonprofits in our main analyses, 43,610 have multiple observations over time. Of these, approximately 20 percent change their mission. This allows us to ask if positive change in emotion is associated with donations or volunteers. The online supplement presents an alternative longitudinal design using the administrative data and we briefly describe it here. To evaluate change in emotion over time, we measured the percent of positive emotion words in each mission statement in each year and predicted 2015 levels of donations and volunteers with change in positive sentiment from 2012 to 2014. We also predict change in donations and volunteers from 2014 to 2015 with change in mission emotion 2012 to 2014. (See the online supplement for how we assessed whether a change in a mission statement was a substantive change.) Three measures of emotional change include a dummy indicator for whether a nonprofit’s mission statement became at least 10 percent more positive from 2012 to 2014; a categorical variable measuring increased positivity, no change, and decreased positivity (10 percent threshold); and a continuous measure of positive change.
The online supplement presents the full results. Briefly, we find that increasing positive emotion within the mission of an organization between 2012 and 2014 is associated with higher levels of donations and volunteers in 2015. Similar to the primary results presented in Table 3, these effects are more pronounced for volunteering than for donations. And change in positive emotion influences change in volunteers, although significant associations, at our conservative adjusted p-values, no longer hold for donations. Overall, our longitudinal models suggest increasing positive emotion is associated with more donations and more volunteers as well as a positive increase in volunteers. Altogether, these findings help reinforce our primary conclusions that (1) emotion matters and (2) there is a stronger empirical and theoretical link between emotionality and volunteering than with donating.
DISCUSSION AND CONCLUSIONS
In this article, we knit together scholarship from nonprofit studies, organizational and institutional theories, the sociology of emotion and cognition, and affective neuroscience to determine whether use of emotional language yields returns for donations and volunteers across a wide range of nonprofits. We argued that it should, and that there was reason to believe both positive and negative emotions could be beneficial. Using automated text analysis to determine the emotional valence of nonprofit mission statements, we investigate our question by utilizing an untapped data source of newly released IRS Forms 990 for tens of thousands of nonprofits. Results demonstrate that, whether positive or negative, emotions matter; negative emotions alone are generally less effective than positive emotions or the combination of the two; and substantive subfields draw on or benefit from different emotional repertoires. These outcomes hold even with controlling for other organizational characteristics shown to be important in prior research on donations and probing the robustness of these results through additional experimental and longitudinal analyses.
We also found wide variation in the association between emotion and volunteers/donations across institutional fields designated as broadly social bonding oriented or social problems oriented, while accounting for important variation. The nonprofit sector itself is an institutional field, but the size and diversity of this sector has allowed unique niche structures and logics to emerge. Use of positive emotion is associated with higher donations, and especially volunteers, in many social bonding organizations, and a few social problems nonprofits such as healthcare organizations. Negative emotion is useful in promoting donations and volunteers only for a few social problem fields (and one social bonding: religion-related), and this is replicated in our experimental setting. Solely focusing on the negative may be a dangerous strategy, though—coefficient signs for several fields suggest this may be associated with fewer volunteers, with youth development nonprofits reaching a conservative significance threshold. As advocated by fundraisers, the combined use of negative and positive sentiment is particularly positive for arts, education, healthcare, employment, and religious organizations. Interesting differences across categories abound. Knowing these broad trends opens new opportunities for additional research to continue to tease out the specific logics within each field that may drive divergent trends. Indeed, we probed the associations for two specific subfields within our experiments, but these experiments can and should be replicated across all areas to confirm and nuance the administrative differences.
Donating and volunteering both manifest from similar prosocial motivations (Lee et al. 1999), but our results open up interesting insights into distinctions between the acts that render them differentially susceptible to a nonprofit’s emotional content. Generally, emotion, especially positive emotion, had stronger effects on volunteering than on donations. One reason for this may be that volunteering is generally a more social, effortful, and bodily immersive experience than donating, which can occur in the privacy of one’s own home and at a distance. We find that emotion can certainly spur more surface-level engagement via donating, but positive emotions more strongly encourage more intimate and social engagement acts like volunteering. Ringmar (2020:39) argues that “although meaning certainly can be discursively constructed and culturally elaborated on, it is originally an embodied event”—one that should better align conscious and unconscious dual processes during decision-making. One recommendation to boost donations, then, is for nonprofits to design fundraising events to create embodied experiences. Our study identifies one distinction, but future research should consider a wider variety of positive and negative emotion, as well as specific emotions such as pride or shame (Turner and Stets 2006:47), and dive deeper into the underlying cognitive processes.
This work has strong implications for research on nonprofit organizations. Sociologists consider philanthropy to be “fundamentally social in both its determinants and its directions” (Barman 2017:272), and prior work has shown how institutions, networks, and other features of communities interact with nonprofit finances (Galaskiewicz 1985; Galaskiewicz, Bielfield, and Dowell 2006). Theory on individual donations and volunteering makes it clear that both these actions have a strong social basis: individuals’ characteristics, identities, and social networks all influence whether or not they decide to donate or volunteer (Bekkers 2010; Bekkers and Wiepking 2011). However, prior research on nonprofit donations has hewed to the perspective advanced by Weisbrod and Dominguez’s (1986) foundational work that holds utility-maximizing rationality supreme. By drawing on and more carefully engaging with research in psychology, the sociology of emotions, and organizational theory, and the latest research related to decision-making, motivations, and dual process models, we hope this study provides a needed correction to the overemphasis on rational motivation on the part of nonprofits. Instead of ignoring the more social, subjective characteristics of nonprofits, such as the emotional language used within missions, we explicitly incorporate it. By bringing emotion in, we wish to demonstrate that, whether nonprofits recognize it or not, the frames they choose to use in their mission statements and promotional materials, or the visual imagery they use on their websites, trigger emotional pathways and guide donor/volunteer decision-making.
The lack of research on the framing of nonprofit missions is surprising considering the importance of this organizational artifact to this sector. Although some scholars downplay the importance of missions and highlight issues with decoupling, especially in the forprofit sector, (Bromley and Powell 2012), missions are the central focus of nonprofit organizations (Berlan 2017). As such, these texts are an important window into nonprofit organizations’ internal cultures and relationships to external actors. Nonprofit professionals give significant weight and attention to the precise wording and meaning of their mission statements, which “oftentimes take months to produce, involving iterative discussions among executives, board members, and workers, and agonizing debates over the precise language” (Pope et al. 2018:1303). Our insights demonstrate that nonprofits could also make strategic choices, within the parameters of their institutional field, to use emotional language within their mission statements and promotional materials. Future research must more deeply consider the field in which a particular nonprofit is located to understand its operations and their consequences.
A fruitful example of this approach would be to incorporate comparative dynamics. McDonnell and colleagues (2017) propose that the success of a message depends on whether that message resonates. Resonance occurs when a message is not so familiar to audiences that it is rendered banal but is just distinct enough to arouse a new interpretation and extension. Thus, emotional cues might work best when they both propose something that is new and amplify existing emotional currents in a field. On the other hand, Bail and colleagues (2017) find that emotional approaches that enter when cognitive approaches (technical, evaluative, or scientific) have saturated public discussion are most effective, and vice versa. Ideas such as these can be tested in future research, and researchers could use other dictionaries that would measure “cognition” in mission statements.
A key reason why we are able to contribute these new insights to research on nonprofit donations is that researchers have new access to data on nonprofit mission statements. Previously, researchers were limited in the types of questions they could ask about nonprofits, in part due to insufficient data. Prior to the IRS 990 release, key sources of data for nonprofit research were either purposeful samples, contained limited financial information on nonprofits, which necessarily restricted analyses, or were from the late 1990s. The few studies that did consider nonprofit missions were restricted to a generally small set of nonprofits due to an arduous process of having to collect each statement by hand. In short, until now, scholars have never had the ability to analyze nonprofit missions in any widespread way. Furthermore, our understanding of volunteering has heretofore been limited by a focus on individual volunteers through self-reports in surveys. Due to a lack of data, research never considered volunteers from the nonprofit side: how nonprofits attract and retain volunteers. The 990s represent a fundamentally different way to approach volunteer rates in the United States, which allows for new insights centered around how nonprofits influence volunteering, rather than the socioeconomic characteristics of the individual survey-takers.16 For example, although we do not discuss it in detail, the volunteer findings are novel even for the controls, as this is the first time any nonprofit characteristic was used to predict volunteers; this underscores the significance of the 990 data. Past the 2016 release of data from 2010 to 2016, the IRS is continuously releasing new tax forms, providing a vast new trove of rich, longitudinal data.
Finally, the availability of nonprofit mission statements, received as text data, opens the door for text analysis and other similar computational methods to enter research on organizations such as nonprofits. Certainly across the social sciences, the explosion of text-based data is allowing new and pressing questions to be addressed in novel ways. In our case, we used a straightforward measure with a dictionary-based approach. Text analysis remains a field in progress, however, and a range of alternatives to dictionary-based approaches exist. Alternative methods using semi-supervised machine learning (Davidov, Tsur, and Rappoport 2010) or that account for higher-order sentiment are becoming more advanced (Hirschberg and Manning 2015) and may better understand irony, sarcasm, and humor (Bharti et al. 2016). These new approaches create new opportunities to explore sentiment, emotion, and meaning beyond positive/negative, which can and should be explored.
This article provides new insights for both social scientists and practitioners, but it is not without limitations. The most significant limitation of this study, and the larger 990 data release, is that only entities that e-file (about 60 to 65 percent of all 990-PC and 990-EZ filers) are included in the sample. Furthermore, a set of nonprofits are not required to file Forms 990 of any kind: any tax-exempt organization can choose to file a full Form 990, but tax-exempt organizations are only obligated to file a full Form 990 if they have more than $200,000 in gross receipts, a Form 990-EZ if they have gross receipts less than $200,000, or a Form 990N if they have $50,000 in gross receipts. Although our Frank and colleagues’ sensitivity analysis suggests the extent to which sampling bias may influence our results is low, it is still important to note that insights may not be generalizable to the smallest nonprofits. One part of this limitation, however, will soon be overcome: in 2019, new legislation required all nonprofits to e-file, with minor exceptions. Certainly, the cross-sectional nature of our main design limits causal interpretation. However, our auxiliary experimental and longitudinal designs provide additional evidence to support our primary findings and theoretical causal pathways.
In measuring donations, we do not know the number of donations—only the overall amount donated. Consequently, we do not know whether a nonprofit relies on a high number of small, grassroots donors or fewer, but more sizable, contributions. It is possible that the emotional valence of missions would differently affect donations from these two sources, which likely use differing donation strategies. Similarly, nonprofits do not report how volunteer experiences are structured, the tasks volunteers do, or the average length of service. In the results by institutional field, we generally see volunteers in social bonding categories are more responsive to emotion than are those in social problems organizations. One plausible explanation is that social bonding organizations have more social, embodied, and immersive experiences, akin to the “cultivation model” of volunteering (Ganz 2009; Han 2014). An extension of this study, therefore, would investigate how qualitative volunteer experience modifies the relationship between emotions and engagement.
In the United States, the nonprofit sector plays a vital role in carrying out social services, administering government programs, and providing the social cohesion necessary to sustain a modern democracy (Salamon 1987; Tocqueville [1835] 1972; Weisbrod 1988). Consequently, researchers are continuously trying to evaluate the effectiveness of this sector, finding that nonprofits are associated with decreasing crime rates, the mitigation of neighborhood poverty, and the promotion of subjective well-being (Ressler et al. 2016; Sharkey, Torrats-Espinosa, and Takyar 2017; Small, Jacobs, and Massengill 2008; Velasco et al. 2019). To support the work of nonprofits, in 2017, Americans donated $410 billion to charitable organizations, of which 70 percent came from individual contributions (Giving USA 2018). Moreover, roughly a quarter of people in the United States formally volunteer their time to a nonprofit organization monthly (Cnaan and Handy 2005).
As the U.S. government increasingly moves away from the traditional public-nonprofit partnership model and as new social challenges emerge, the nonprofit sector is in a precarious situation—a situation that challenges the sector’s ability to provide social goods for society. Therefore, information that can help this sector attract external engagement and resources is of utmost importance. Through this study, we demonstrate that by attending to the emotional language within their mission statements, nonprofits have a new tool they can utilize to help them navigate these challenging and deep social currents.
Supplementary Material
Acknowledgments
We are deeply indebted to Nicholas Reith for early research assistance. We are grateful to Marta Ascherio, Ken Hou-Lin, Jennifer Glanville, Wesley Longhofer, David Pedulla, Chantal Hailey, and Elizabeth Vandewater for useful feedback.
funding
The authors gratefully acknowledge support from the Corporation for National and Community Service (201502185 PI: Paxton) and the National Institute of Child Health and Human Development (R24 HD42849, PI: Mark Hayward; T32 HD007081-35, PI: R. Kelly Raley) to the Population Research Center at the University of Texas at Austin.
Biographies
Pamela Paxton is the Linda K. George and John Wilson Professor of Sociology at The University of Texas at Austin. She has research interests in nonprofits, politics, gender, and methodology. She is the author of articles on social capital, women in politics, and quantitative methodology. Her research appears in a variety of journals, including the American Sociological Review, American Journal of Sociology, and Social Forces. With Melanie Hughes and Tiffany Barnes, she is co-author of the 2020 book, Women, Politics, and Power: A Global Perspective. She is also an author of Nonrecursive Models: Endogeneity, Reciprocal Relationships, and Feedback Loops (2011).
Kristopher Velasco is a PhD candidate in sociology at the University of Texas at Austin. Lying at the intersection of political sociology, organizations, culture, and global and transnational sociology, his research investigates the cultural and social dimensions and consequences of organizations. For more information see krisvelasco.com.
Robert W. Ressler, named an emerging scholar in nonprofit research by ARNOVA, received their PhD in sociology from the University of Texas at Austin in 2019. They currently work as Sr. Research Associate for the Institute of Child, Youth and Family Policy at Brandeis University focusing on community organizations and childhood opportunity. Their work can be found in the Journal of Marriage and Family, Social Science Research, and Children and Youth Services Review among others. Robert’s newer work investigates how community investments create “symbolic realities” that challenge oppressive social systems, theorizing an achievable path toward greater social change.
Footnotes
Dividing organizations along broad lines is not uncommon. For example, other research divides voluntary associations between expressive and instrumental organizations (e.g., Gordon and Babchuk 1959) or distinguishes elite and welfare nonprofits (e.g., Marquis et al. 2013). Of course, variation exists. For example, different organizations in the same field may exhibit social bonding or social problems regardless of the overall classification of the field. One environmental nonprofit, for example, might focus on building a community of bird watchers, whereas another frames their work as tackling climate change. And some organizations may mix both orientations, confronting a social problem through focusing on social bonding. We view a “social bonding” or “social problem” orientation as a general but useful tendency linking institutional fields that likely influences how external audiences respond to emotional language. We acknowledge variation within category in our models.
Typically, nonprofits take two different approaches to arranging volunteer programs: a “plug-in” model (Eliasoph 2013) or a cultivation model (Ganz 2009; Han 2014). The plug-in model mimics donating: nonprofits attempt to make participation as simple as possible with low-stakes opportunities to volunteer that produce emotionally satisfying experiences with minimal effort (Eliasoph 2013). A cultivation model, in contrast, takes more investment on the part of the nonprofit to develop volunteers who become central to its very operation. Cultivated volunteers have deep emotional ties to the nonprofit and meaningful leadership responsibilities (Han 2014). Regardless whether a volunteer is plugged in or cultivated, the act of volunteering is more social than donating (Borgonovi 2008).
Many thanks to the Aspen Institute’s Nonprofit Data Project, along with partners GuideStar, the Urban Institute, The Foundation Center, the Lilly Family School of Philanthropy, and the Johns Hopkins Center for Civil Society Studies and funders the Charles Stewart Mott Foundation and the Bill and Melinda Gates Foundation, who advocated (and ultimately sued) for the release of the data in machine-readable form.
Although we recognize the importance of these sources of revenues, we do not include contributions obtained through federated campaigns (Part VIII Line 1a), as these are received indirectly from federated funders such as the United Way. We cannot distinguish the donor-directed portion of these contributions from other factors. Nor do we include contributions from related organizations (Part VIII Line 1d), as related organizations include a diverse group of supporting, supported, and employee organizations that do not reflect individual donations as we try to measure.
Results do not substantially differ if donations or volunteers are logged but not winsorized.
Organizations can also list their mission statement in Part I line 1. In certain cases, the mission statement as reported in Part III was cut off, possibly due to an e-filing error. If the statement in Part III line I was shorter than the statement in Part I, we replaced it with the statement from Part I.
The NTEE categories we do not include are animal-related; mental health and crisis intervention; diseases, disorders, and medical disciplines; medical research; public safety, disaster preparedness, and relief; human services; international, foreign affairs, and national security; community improvement and capacity building; philanthropy, voluntarism, and grantmaking foundations; science and technology; social science; public and societal benefit; mutual and membership benefit; and unknown.
Auxiliary analyses include additional measures of a nonprofit’s organizational, managerial, and professional capacities. We include the percent of expenses dedicated to administration as a measure of organizational capacity; number of board members as managerial capacity; and a professionalization index, per Harris, Petrovits, and Yetman (2015), as the sum of four possible good governance policies: whistle blower, conflict of interest, document retention and destruction, and meeting minutes (see also Hwang and Powell 2009). Results from auxiliary analyses are similar to the main findings (results available from authors upon request).
Another important issue is the possibility of reciprocal effects. The theoretical argument strongly supports our causal pathway (i.e., organizational culture influences emotionality of mission statements which then influences external engagement), but reciprocal effects are nevertheless possible. Nonprofits that are successful at attracting donations and volunteers may have more positive cultures and subsequently use more positive emotion. Moreover, due to common expectations that members of a nonprofit’s board of directors donate to the organization themselves or fundraise, the very board members who write the mission statement may also be the ones gathering donations. Therefore, future research is needed to fully adjudicate the possibility of these reciprocal effects.
The high use of “war” in arts, culture, and humanities organizations is due to the number of museums and monuments dedicated to particular wars and local war reenactment groups.
Table A2 in the online supplement, which provides results that include nonprofits that receive zero donations, shows negative coefficients for logged negative emotion and for the negative emotion dummy. That is, when nonprofits that receive zero donations are included in the model, negative emotion in mission statements reduces donations and volunteers. Our research questions pertain to nonprofits that attempt to acquire donations. Nonprofits that do not receive any donations at all rely on fees for service or on government contracts that have their own funding strategies. For these nonprofits, language that stresses need (negative) may be a more successful strategy. Future research needs to tease out how different presentation strategies may be differentially successful across nonprofits with varied funding streams.
MTurk workers are not a random or representative sample, but they are more representative than a typical college sample. See Abascal (2020:308) for a recent review of studies of Amazon Mechanical Turk.
Prior to analysis, 69 Holiday Wishes and 34 Centerville participants were excluded for failing the instructional manipulation checks. Respondents failing the checks for whom we have demographic information are a couple years younger, and a little more diverse, than those who passed the checks. Among those included, respondents were similar across demographic groups in the conditions for both experiments. Only in Centerville, the proportion of self-identified men (.41) was significantly lower in the neutral condition than in the other conditions (.51, .53).
NTEE codes used for Holiday Wishes are P30, P31, P32, P58. Codes for Food Banks are K30, K31, K34, K35, K36. All models include the same controls as the primary analyses.
The negative relationship between positive emotion and donations in Food Banks runs counter to the finding for the food and agriculture category as a whole (see Figure 2) and is nonintuitive. This unique field may be worthy of further study.
Nonprofits are given much leeway in defining and reporting volunteers. Because of the newness of this data source, investigations into potential sources of bias are not as extensive as known bias from surveys of volunteering (Abraham, Helms, and Presser 2009). Nonprofits may have an incentive to over-estimate volunteer numbers to demonstrate public support to government or foundation funders. Nevertheless, 990-reported volunteers is the most extensive source of data on nonprofit volunteers to date.
References
- Abascal Maria C. 2020. “Contraction as a Response to Group Threat: Demographic Decline and Whites’ Classification of People Who Are Ambiguously White.” American Sociological Review 85(2): 298–322. [Google Scholar]
- Abraham Katharine G., Helms Sara, and Presser Stanley. 2009. “How Social Processes Distort Measurement: The Impact of Survey Nonresponse on Estimates of Volunteer Work in the United States.” American Journal of Sociology 114(4):1129–65. [DOI] [PubMed] [Google Scholar]
- Alvesson Mats. 2012. Understanding Organizational Culture. Thousand Oaks, CA: Sage. [Google Scholar]
- Bagozzi Richard P., Gopinath Mahesh, and Nyer Prashanth U.. 1999. “The Role of Emotions In Marketing.” Journal of the Academy of Marketing Science 27(2):184–206. [Google Scholar]
- Bail Christopher A., Brown Taylor W., and Mann Marcus. 2017. “Channeling Hearts and Minds: Advocacy Organizations, Cognitive-Emotional Currents, and Public Conversation.” American Sociological Review 82(6):1188–213. [Google Scholar]
- Balsam Steven, and Harris Erica E.. 2014. “The Impact of CEO Compensation on Nonprofit Donations.” The Accounting Review 89(2):425–50. [Google Scholar]
- Barman Emily. 2016. Caring Capitalism: The Meaning and Measure of Social Value. Cambridge, UK: Cambridge University Press. [Google Scholar]
- Barman Emily. 2017. “The Social Bases of Philanthropy.” Annual Review of Sociology 43:271–90. [Google Scholar]
- Bart Christopher K. 1997. “Industrial Firms and the Power of Mission.” Industrial Marketing Management 26(4):371–83. [Google Scholar]
- Bart Christopher K., Bontis Nick, and Taggar Simon. 2001. “A Model of the Impact of Mission Statements on Firm Performance.” Management Decision 39(1):19–35. [Google Scholar]
- Bechara Antoine, Damasio Hanna, Tranel Daniel, and Damasio Antonio R.. 1997. “Deciding Advantageously before Knowing the Advantageous Strategy.” Science 275(5304):1293–95. [DOI] [PubMed] [Google Scholar]
- Bekkers René. 2010. “Who Gives What and When? A Scenario Study of Intentions to Give Time and Money.” Social Science Research 39(3):369–81. [Google Scholar]
- Bekkers René, and Wiepking Pamala. 2011. “A Literature Review of Empirical Studies of Philanthropy: Eight Mechanisms That Drive Charitable Giving.” Nonprofit and Voluntary Sector Quarterly 40(5): 924–73. [Google Scholar]
- Berlan David. 2017. “Understanding Nonprofit Missions as Dynamic and Interpretative Conceptions.” Nonprofit Management & Leadership 28(3):413–22. [Google Scholar]
- Berman Jonathan Z., Barasch Alixandra, Levine Emma E., and Small Deborah A.. 2018. “Impediments to Effective Altruism: The Role of Subjective Preferences in Charitable Giving.” Psychological Science 29(5):834–44. [DOI] [PubMed] [Google Scholar]
- Berrone Pascual, Gelabert Liliana, Massa-Saluzzo Federica, and Rousseau Horacio E.. 2016. “Understanding Community Dynamics in the Study of Grand Challenges: How Nonprofits, Institutional Actors, and the Community Fabric Interact to Influence Income Inequality.” Academy of Management Journal 59(6):1940–64. [Google Scholar]
- Betzler Diana, and Gmür Markus. 2016. “Does Fundraising Professionalization Pay? The Impact of Organizational Fundraising Capability on a Charity’s Net Revenue from Private Donations.” Nonprofit Management and Leadership 27(1):27–42. [Google Scholar]
- Bharti Santosh Kumar, Vachha Bakhtyar, Pradhan RK, Babu Korra Sathya, and Jena Sanjay Kumar. 2016. “Sarcastic Sentiment Detection in Tweets Streamed in Real Time: A Big Data Approach.” Digital Communications and Networks 2(3):108–21. [Google Scholar]
- Borgonovi Francesca. 2008. “Divided We Stand, United We Fall: Religious Pluralism, Giving, and Volunteering.” American Sociological Review 73(1):105–28. [Google Scholar]
- Brader Ted. 2006. Campaigning for Hearts and Minds: How Emotional Appeals in Political Ads Work. Chicago: University of Chicago Press. [Google Scholar]
- Brinckerhoff Peter C. 2009. Mission-Based Management: Leading Your Not-For-Profit in the 21st Century. New York: John Wiley & Sons. [Google Scholar]
- Bromley Patricia, and Powell Walter W.. 2012. “From Smoke and Mirrors to Walking the Talk: Decoupling in the Contemporary World.” Academy of Management Annals 6(1):483–530. [Google Scholar]
- Bronfman Charles, and Solomon Jeffrey. 2010. The Art of Giving: Where the Soul Meets a Business Plan. San Francisco, CA: Jossey-Bass. [Google Scholar]
- Brooks Jeff. 2015. The Fundraiser’s Guide to Irresistible Communications. Medfield, MA: Emerson & Church Publishers. [Google Scholar]
- Brown William A., and Yoshioka Carlton F.. 2003. “Mission Attachment and Satisfaction as Factors in Employee Retention.” Nonprofit Management and Leadership 14(1):5–18. [Google Scholar]
- Calabrese Thad D. 2011. “Do Donors Penalize Nonprofit Organizations with Accumulated Wealth?” Public Administration Review 71(6):859–69. [Google Scholar]
- Campagna Claudio, and Fernandez Teresita. 2007. “A Comparative Analysis of the Vision and Mission Statements of International Environmental Organisations.” Environmental Values 16(3):369–98. [Google Scholar]
- Charles Cleopatra. 2018. “Nonprofit Arts Organizations: Debt Ratio Does Not Influence Donations—Interest Expense Ratio Does.” The American Review of Public Administration 48(7):659–67. [Google Scholar]
- Cnaan RA, and Handy F. 2005. “Towards Understanding Episodic Volunteering.” Vrijwillige Inzet Onderzocht 2(1):29–35. [Google Scholar]
- Cryder Cynthia E., Loewenstein George, and Seltman Howard. 2013. “Goal Gradient in Helping Behavior.” Journal of Experimental Social Psychology 49(6):1078–83. [Google Scholar]
- Damasio Antonio R. 1994. Descartes’ Error: Emotion, Rationality and the Human Brain. New York: HarperCollins. [Google Scholar]
- Das Enny, Kerkhof Peter, and Kuiper Joyce. 2008. “Improving the Effectiveness of Fundraising Messages: The Impact of Charity Goal Attainment, Message Framing, and Evidence on Persuasion.” Journal of Applied Communication Research 36(2):161–75. [Google Scholar]
- Davidov Dmitry, Tsur Oren, and Rappoport Ari. 2010. “Enhanced Sentiment Learning Using Twitter Hashtags and Smileys.” Proceedings of the 23rd International Conference on Computational Linguistics: Posters. Association for Computational Linguistics. [Google Scholar]
- Deshpande Rohit, and Webster Frederick E. Jr. 1989. “Organizational Culture and Marketing: Defining the Research Agenda.” Journal of Marketing 53(1):3–15. [Google Scholar]
- DiMaggio Paul J., and Powell Walter W.. 1983. “The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields.” American Sociological Review 48(2):147–60. [Google Scholar]
- Eliasoph Nina. 2013. Making Volunteers: Civic Life after Welfare’s End. Princeton, NJ: Princeton University Press. [Google Scholar]
- Eng Teck-Yong, Liu Chih-Yao Gordon, and Sekhon Yasmin Kaur. 2012. “The Role of Relationally Embedded Network Ties in Resource Acquisition of British Nonprofit Organizations.” Nonprofit and Voluntary Sector Quarterly 41(6):1092–115. [Google Scholar]
- Entman Robert M. 1993. “Framing: Toward Clarification of a Fractured Paradigm.” Journal of Communication 43(4):51–58. [Google Scholar]
- Fairhurst Gail T., Jordan Jerry Monroe, and Neuwirth Kurt. 1997. “Why Are We Here? Managing the Meaning of an Organizational Mission Statement.” Journal of Applied Communication Research 25(4):243–63. [Google Scholar]
- Fisher Robert J., Vandenbosch Mark, and Antia Kersi D.. 2008. “An Empathy-Helping Perspective on Consumers’ Responses to Fund-Raising Appeals.” Journal of Consumer Research 35(3):519–31. [Google Scholar]
- Frank Kenneth A., Maroulis Spiro J., Duong Minh Q., and Kelcey Benjamin M.. 2013. “What Would It Take To Change an Inference? Using Rubin’s Causal Model to Interpret the Robustness of Causal Inferences.” Educational Evaluation and Policy Analysis 35(4):437–60. [Google Scholar]
- Fredrickson Barbara L. 2001. “The Role of Positive Emotions in Positive Psychology: The Broaden-and-Build Theory of Positive Emotions.” American Psychologist 56(3):218–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Frijda Nico H. 1986. The Emotions. Cambridge, UK: Cambridge University Press. [Google Scholar]
- Frumkin Peter. 2002. On Being Nonprofit: A Conceptual and Policy Primer. Cambridge, MA: Harvard University Press. [Google Scholar]
- Fyall Rachel, Moore M. Kathleen, and Gugerty Mary Kay. 2018. “Beyond NTEE Codes: Opportunities to Understand Nonprofit Activity Through Mission Statement Content Coding.” Nonprofit and Voluntary Sector Quarterly 47(4):677–701. [Google Scholar]
- Galaskiewicz Joseph. 1985. Social Organization of an Urban Grants Economy: A Study of Business Philanthropy and Nonprofit Organizations. Cambridge, MA: Academic Press. [Google Scholar]
- Galaskiewicz Joseph, and Bielefeld Wolfgang. 1998. Nonprofit Organizations in an Age of Uncertainty: A Study of Organizational Change. Piscataway, NJ: Transaction Publishers. [Google Scholar]
- Galaskiewicz Joseph, Bielefeld Wolfgang, and Dowell Myron. 2006. “Networks and Organizational Growth: A Study of Community Based Nonprofits.” Administrative Science Quarterly 51(3):337–80. [Google Scholar]
- Ganz Marshall. 2009. Why David Sometimes Wins: Leadership, Organization, and Strategy in the California Farm Worker Movement. Oxford, UK: Oxford University Press. [Google Scholar]
- Gaudine Alice, and Thorne Linda. 2001. “Emotion and Ethical Decision-Making in Organizations.” Journal of Business Ethics 31(2):175–87. [Google Scholar]
- Giving USA. 2018. Giving USA: The Annual Report on Philanthropy. Indianapolis, IN: Giving USA. [Google Scholar]
- Goodwin Jeff, Jasper James M., and Polletta Francesca, eds. 2009. Passionate Politics: Emotions and Social Movements. Chicago: University of Chicago Press. [Google Scholar]
- Gordon C. Wayne, and Babchuk Nicholas. 1959. “A Typology of Voluntary Associations.” American Sociological Review 24(1):22–29. [Google Scholar]
- Gray Jeffrey A. 1990. “Brain Systems That Mediate both Emotion and Cognition.” Cognition & Emotion 4(3):269–88. [Google Scholar]
- Gross Kimberly. 2008. “Framing Persuasive Appeals: Episodic and Thematic Framing, Emotional Response, and Policy Opinion.” Political Psychology 29(2):169–92. [Google Scholar]
- Hager Mark A., and Brudney Jeffrey L.. 2011. “Problems Recruiting Volunteers: Nature versus Nurture.” Nonprofit Management and Leadership 22(2): 137–57. [Google Scholar]
- Han Hahrie. 2014. How Organizations Develop Activists: Civic Associations and Leadership in the 21st Century. Oxford, UK: Oxford University Press. [Google Scholar]
- Harris Erica, Petrovits Christine M., and Yetman Michelle H.. 2015. “The Effect of Nonprofit Governance on Donations: Evidence from the Revised Form 990.” The Accounting Review 90(2):579–610. [Google Scholar]
- Hirschberg Julia, and Manning Christopher D.. 2015. “Advances in Natural Language Processing.” Science 349(6245):261–66. [DOI] [PubMed] [Google Scholar]
- Hochschild Arlie Russell. 1979. “Emotion Work, Feeling Rules, and Social Structure.” American Journal of Sociology 85(3):551–75. [Google Scholar]
- Hochschild Arlie Russell. 1983. The Managed Heart: Commercialization of Human Feeling. Berkeley: University of California Press. [Google Scholar]
- Hwang Hokyu, and Powell Walter W.. 2009. “The Rationalization of Charity: The Influences of Professionalism in the Nonprofit Sector.” Administrative Science Quarterly 54(2):268–98. [Google Scholar]
- Isen Alice M., and Baron Robert A.. 1991. “Positive Affect as a Factor in Organizational Behavior.” Research in Organizational Behavior 13:1–53. [Google Scholar]
- Kandrack Ryan, and Lundberg Gustav. 2014. “On the Influence of Emotion on Decision Making: The Case of Charitable Giving.” Pp. 57–73 in Human-centric Decision-Making Models for Social Sciences, edited by Guo P and Pedrycz W. Berlin: Springer. [Google Scholar]
- Kim Mirae. 2017. “Characteristics of Civically Engaged Nonprofit Arts Organizations: The Results of a National Survey.” Nonprofit and Voluntary Sector Quarterly 46(1):175–98. [Google Scholar]
- Kim Seok Eun, and Lee Jung Wook. 2007. “Is Mission Attachment an Effective Management Tool for Employee Retention? An Empirical Analysis of a Nonprofit Human Services Agency.” Review of Public Personnel Administration 27(3):227–48. [Google Scholar]
- Koch Bradley J., Galaskiewicz Joseph, and Pierson Alisha. 2015. “The Effect of Networks on Organizational Missions.” Nonprofit and Voluntary Sector Quarterly 44(3):510–38. [Google Scholar]
- Kollock Peter. 1998. “Social Dilemmas: The Anatomy of Cooperation.” Annual Review of Sociology 24(1):183–214. [Google Scholar]
- Lawler Edward J., and Yoon Jeongkoo. 1996. “Commitment in Exchange Relations: Test of a Theory of Relational Cohesion.” American Sociological Review 61(1):89–108. [Google Scholar]
- Lazarus Richard S. 1991. Emotion and Adaptation. Oxford, UK: Oxford University Press. [Google Scholar]
- Lee Lichang, Piliavin Jane Allyn, and Call Vaughn R. A.. 1999. “Giving Time, Money, and Blood: Similarities and Differences.” Social Psychology Quarterly 62(3):276–90. [Google Scholar]
- Lewis Michael, Haviland Jeanette M., and Barrett Lisa Feldman. 2010. Handbook of Emotions. New York: Guilford Press. [Google Scholar]
- Lizardo Omar, Mowry Robert, Sepulvado Brandon, Stoltz Dustin S., Taylor Marshall A., Van Ness Justin, and Wood Michael. 2016. “What Are Dual Process Models? Implications for Cultural Analysis in Sociology.” Sociological Theory 34(4):287–310. [Google Scholar]
- Marcus George E. 2013. Political Psychology: Neuroscience, Genetics, and Politics. New York: Oxford University Press. [Google Scholar]
- Marcus George E., Neuman W. Russell, and MacKuen Michael. 2000. Affective Intelligence and Political Judgment. Chicago: University of Chicago Press. [Google Scholar]
- Marquis Christopher. 2003. “The Pressure of the Past: Network Imprinting in Intercorporate Communities.” Administrative Science Quarterly 48(4):655–89. [Google Scholar]
- Marquis Christopher, Davis Gerald F., and Glynn Mary Ann. 2013. “Golfing Alone? Corporations, Elites, and Nonprofit Growth in 100 American Communities.” Organization Science 24(1):39–57. [Google Scholar]
- Marquis Christopher, Glynn Mary Ann, and Davis Gerald F.. 2007. “Community Isomorphism and Corporate Social Action.” Academy of Management Review 32(3):925–45. [Google Scholar]
- Martin Joanne, Feldman Martha S., Hatch Mary Jo, and Sitkin Sim B.. 1983. “The Uniqueness Paradox in Organizational Stories.” Administrative Science Quarterly 28(3):438–53. [Google Scholar]
- Marwell Nicole P. 2004. “Privatizing the Welfare State: Nonprofit Community-Based Organizations as Political Actors.” American Sociological Review 69(2):265–91. [Google Scholar]
- McCarthy John D., and Castelli Jim. 2002. “The Necessity for Studying Organizational Advocacy Comparatively.” Pp. 103–21 in Measuring the Impact of the Nonprofit Sector, edited by Flynn P and Hodgkinson VA. New York: Springer. [Google Scholar]
- McDonald Robert E. 2007. “An Investigation of Innovation in Nonprofit Organizations: The Role of Organizational Mission.” Nonprofit and Voluntary Sector Quarterly 36(2):256–81. [Google Scholar]
- McDonnell Terence E., Bail Christopher A., and Tavory Iddo. 2017. “A Theory of Resonance.” Sociological Theory 35(1):1–14. [Google Scholar]
- McKeever Brice S. 2015. The Nonprofit Sector in Brief 2015: Public Charities, Giving, and Volunteering. Washington, DC: Urban Institute. [Google Scholar]
- Merchant Altaf, Ford John B., and Sargeant Adrian. 2010. “Charitable Organizations’ Storytelling Influence on Donors’ Emotions and Intentions.” Journal of Business Research 63(7):754–62. [Google Scholar]
- Minkoff Debra C., and Powell Walter W.. 2006. “Nonprofit Mission: Constancy, Responsiveness, or Deflection.” In The Nonprofit Sector: A Research Handbook, edited by Powell WW and Bromley P. Stanford, CA: Stanford University Press. [Google Scholar]
- Morphew Christopher C., and Hartley Matthew. 2006. “Mission Statements: A Thematic Analysis of Rhetoric across Institutional Type.” The Journal of Higher Education 77(3):456–71. [Google Scholar]
- Nesbit Rebecca, Christensen Robert K., and Brudney Jeffrey L.. 2018. “The Limits and Possibilities of Volunteering: A Framework for Explaining the Scope of Volunteer Involvement in Public and Nonprofit Organizations.” Public Administration Review 78(4): 502–13. [Google Scholar]
- Olson Mancur. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge, MA: Harvard University Press. [Google Scholar]
- Oppenheimer Daniel M., Meyvis Tom, and Davidenko Nicolas. 2009. “Instructional Manipulation Checks: Detecting Satisficing to Increase Statistical Power.” Journal of Experimental Social Psychology 45(4):867–72. [Google Scholar]
- Pagis Michal. 2010. “From Abstract Concepts to Experiential Knowledge: Embodying Enlightenment in a Meditation Center.” Qualitative Sociology 33(4):469–89. [Google Scholar]
- Pandey Sheela, Kim Mirae, and Pandey Sanjay K.. 2017. “Do Mission Statements Matter for Nonprofit Performance? Insights from a Study of US Performing Arts Organizations.” Nonprofit Management & Leadership 27(3):389–410. [Google Scholar]
- Patrick Patricia G., and Caplow Susan. 2018. “Identifying the Foci of Mission Statements of the Zoo and Aquarium Community.” Museum Management and Curatorship 33(2):120–35. [Google Scholar]
- Paxton Pamela. 2020. “What Influences Charitable Giving?” In The Nonprofit Sector: A Research Handbook, 3rd ed., edited by Powell WW and Bromley P. Stanford, CA: Stanford University Press. [Google Scholar]
- Pennebaker James W., Boyd Ryan L., Jordan Kayla, and Blackburn Kate. 2015. The Development and Psychometric Properties of LIWC 2015. Austin, TX: University of Texas at Austin. [Google Scholar]
- Pettijohn Sarah L., Boris Elizabeth T., De Vita Carol J., and Fyffe Saunji D.. 2013. “Nonprofit-Government Contracts and Grants: Findings from the 2013 National Survey.” Washington, DC: Urban Institute. [Google Scholar]
- Peyrefitte Joseph, and David Forest R.. 2006. “A Content Analysis of the Mission Statements of United States Firms in Four Industries.” International Journal of Management 23(2):296–301. [Google Scholar]
- Pope Shawn, Bromley Patricia, Lim Alwyn, and Meyer John W.. 2018. “The Pyramid of Nonprofit Responsibility: The Institutionalization of Organizational Responsibility across Sectors.” VOLUNTAS: International Journal of Voluntary and Nonprofit Organizations 29(6):1300–14. [Google Scholar]
- Powell Walter W., and Paul J DiMaggio, eds. 1991. The New Institutionalism in Organizational Analysis. Chicago: University of Chicago Press. [Google Scholar]
- Price Heather E., and Collett Jessica L.. 2012. “The Role of Exchange and Emotion on Commitment: A Study of Teachers.” Social Science Research 41(6): 1469–79. [DOI] [PubMed] [Google Scholar]
- Ressler Robert, Paxton Pamela, Pivnick Lilla, Velasco Kristopher, and Weiss Inbar. 2016. “Nonprofits, Tweets, and Subjective Wellbeing.” American Sociological Association Annual Meeting, Seattle, WA. [Google Scholar]
- Ringmar Erik. 2020. “How Do Performances Fuse Societies?” American Journal of Cultural Sociology 8(1):29–44. [Google Scholar]
- Ringmar Erik, and Mast Jason. 2018. “Introduction: Cultural Sociology and Cognitive Neuroscience” (https://portal.research.lu.se/ws/files/41349466/Introduction_Cultural_Sociology_and_Cognitive_Neuroscience_final_and_complete_version.pdf).
- Salamon Lester M. 1987. “Of Market Failure, Voluntary Failure, and Third-Party Government: Toward a Theory of Government-Nonprofit Relations in the Modern Welfare State.” Journal of Voluntary Action Research 16(1–2):29–49. [Google Scholar]
- Salamon Lester M. 2011. America’s Nonprofit Sector: A Primer, 3rd ed. New York: Foundation Center. [Google Scholar]
- Schein Edgar H. 2010. Organizational Culture and Leadership, Vol. 2. New York: John Wiley & Sons. [Google Scholar]
- Schelling Thomas C. 1968. “The Life You Save May Be Your Own.” In Problems in Public Expenditure Analysis, edited by Chase SB. Washington, DC: The Brookings Institute. [Google Scholar]
- Sharkey Patrick, Torrats-Espinosa Gerard, and Takyar Delaram. 2017. “Community and the Crime Decline: The Causal Effect of Local Nonprofits on Violent Crime.” American Sociological Review 82(6): 1214–40. [Google Scholar]
- Silard Anthony. 2018. “Emotions for a Cause: How the Emotion Expression of Nonprofit Leaders Produces Follower Engagement and Loyalty.” Nonprofit and Voluntary Sector Quarterly 47(2):304–24. [Google Scholar]
- Small Deborah A., and Loewenstein George. 2003. “Helping a Victim or Helping the Victim: Altruism and Identifiability.” Journal of Risk and Uncertainty 26(1):5–16. [Google Scholar]
- Small Mario Luis, Jacobs Erin M., and Massengill Rebekah Peeples. 2008. “Why Organizational Ties Matter for Neighborhood Effects: Resource Access through Childcare Centers.” Social Forces 87(1):387–414. [Google Scholar]
- Smollan Roy K., and Sayers Janet G.. 2009. “Organizational Culture, Change and Emotions: A Qualitative Study.” Journal of Change Management 9(4):435–57. [Google Scholar]
- Snow David A., Rochford E. Burke Jr., Worden Steven K., and Benford Robert D.. 1986. “Frame Alignment Processes, Micromobilization, and Movement Participation.” American Sociological Review 51(4):464–81. [Google Scholar]
- Sussman Abigail B., Sharma Eesha, and Alter Adam L.. 2015. “Framing Charitable Donations as Exceptional Expenses Increases Giving.” Journal of Experimental Psychology: Applied 21(2):130–39. [DOI] [PubMed] [Google Scholar]
- Tocqueville Alexis de. [1835; 1840] 1972. Democracy in America. New York: Vintage Books. [Google Scholar]
- Turner Jonathan H., and Stets Jan E.. 2006. “Sociological Theories of Human Emotions.” Annual Review of Sociology 32:25–52. [Google Scholar]
- Vaisey Stephen. 2009. “Motivation and Justification: A Dual-Process Model of Culture in Action.” American Journal of Sociology 114(6):1675–715. [DOI] [PubMed] [Google Scholar]
- Vandijck Dominique, Desmidt Sebastian, and Buelens Marc. 2007. “Relevance of Mission Statements in Flemish Not-for-Profit Healthcare Organizations.” Journal of Nursing Management 15(2):131–41. [DOI] [PubMed] [Google Scholar]
- Velasco Kristopher, Paxton Pamela, Ressler Robert W., Weiss Inbar, and Pivnick Lilla. 2019. “Do National Service Programs Improve Subjective Well-Being in Communities?” The American Review of Public Administration 49(3):275–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Waters Richard D. 2007. “Nonprofit Organizations’ Use of the Internet: A Content Analysis of Communication Trends on the Internet Sites of the Philanthropy 400.” Nonprofit Management and Leadership 18(1):59–76. [Google Scholar]
- Weisbrod Burton Allen. 1988. The Nonprofit Economy. Cambridge, MA: Harvard University Press. [Google Scholar]
- Weisbrod Burton A., and Dominguez Nestor D.. 1986. “Demand for Collective Goods in Private Nonprofit Markets: Can Fundraising Expenditures Help Overcome Free-Rider Behavior?” Journal of Public Economics 30(1):83–96. [Google Scholar]
- Williams Linda Stallworth. 2008. “The Mission Statement: A Corporate Reporting Tool with a Past, Present, and Future.” The Journal of Business Communication 45(2):94–119. [Google Scholar]
- Witte Kim, and Allen Mike. 2000. “A Meta-analysis of Fear Appeals: Implications for Effective Public Health Campaigns.” Health Education & Behavior 27(5):591–615. [DOI] [PubMed] [Google Scholar]
- Wollschleger Jason. 2017. “The Rite Way: Integrating Emotion and Rationality in Religious Participation.” Rationality and Society 29(2):179–202. [Google Scholar]
- Young Robert B. 2001. “Colleges on the Cross Roads: A Study of the Mission Statements of Catholic Colleges and Universities.” Current Issues in Catholic Higher Education 21(2):65–81. [Google Scholar]
- Zajonc Robert B. 1980. “Feeling and Thinking: Preferences Need No Inferences.” American Psychologist 35(2):151–75. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.