Linking Marketing to Nonprofit Performance

Abstract

The critical role of marketing in driving nonprofit performance has been recognized for decades. However, in practice, there has been a disturbingly weak acknowledgment and/or implementation of marketing practices across nonprofits to date. Marketing is often perceived as an avoidable and costly overhead. The issue is complicated by the fact that nonprofit performance is relatively difficult to measure and may often comprise multiple tangible and intangible outcomes with different (linear and nonlinear) functional forms. Furthermore, nonprofit performance outcomes often depend on behavioral and attitudinal changes of the target segment. The authors address these challenges by presenting a methodology to link marketing efforts to nonprofits’ mission-based performance outcome(s). The authors apply their approach with data from a large nonprofit and find empirical support for the notion that marketing can play a pivotal and significant role in improving nonprofits’ mission-based performance outcomes. The findings help present a strong case for nonprofit leaders and policy makers to fund and treat marketing as a critical investment to drive nonprofit entities’ performance.

Keywords

copula method marketing impact mission-based outcome nonprofit performance Poisson regression role of marketing Tobit model

In 1979, Kotler presented a strong case for nonprofits or nonprofit entities¹ (hereinafter referred to as NPs) to introduce marketing in their organizations (Kotler 1979). Kotler’s article was written at a time when most organizations misunderstood or resisted marketing in a nonprofit setting. However, given the obvious benefits of marketing, Kotler was confident that “eventually, out of necessity, marketing ideas will filter into these organizations” (Kotler 1979, p. 44). He concluded with what appeared to be a reasonable vision of the future: “Within another decade, marketing will be a major and accepted function within the nonprofit sector” (p. 44).

More than four decades have passed since Kotler’s provocative article. There are approximately 1.56 million NPs registered with the Internal Revenue Service in the United States today, generating an annual revenue of about $2.54 trillion (McKeever 2018). Yet only a handful of NPs have successfully implemented a comprehensive marketing approach (Akchin 2001; Pope, Isley, and Asamoa-Tutu 2009) or embraced best marketing practices in their organization (Bloom 2009; Newbert 2012). A survey of 232 senior executives of NPs conducted by Accenture (2006) indicated that only 15% of the NPs surveyed were “strongly interested in measuring the real benefit of their development and marketing investments.” Another survey of 1,288 NPs in November 2011 conducted by Nonprofit Marketing Guide indicated that only 24% of the NPs surveyed had a written and approved marketing plan for 2012. A related problem lies in the fact that measuring and assessing the performance of NPs is relatively more complicated than traditional business enterprises (Buchanan 2019). The 2011 State of the Nonprofit Industry Survey of 2,203 individuals from NPs indicated that less than 10% of the NPs surveyed in the United States strongly felt that they had all the impact metrics needed for marketing. These studies collectively show a disturbingly weak acknowledgment, ability, or adoption of marketing as one of the critical drivers of NP performance and a relatively low commitment from top leadership to act otherwise.

In this article, we present an empirical study to link marketing to NP performance outcomes. The underlying goal is to enable NPs and policy makers to quantify the return on marketing dollars. However, our objective is complicated by the fact that NPs (unlike for-profit businesses) may have multiple performance goals ranging from tangible (e.g., revenue from products sold, increase in membership) to intangible (e.g., poverty alleviation, blood donation) outcomes. We overcome this challenge by applying an analytics-based methodology in the context of a major NP. The proposed approach can be generalized by managers and/or policy makers to evaluate the impact of marketing on the performance of virtually any NP. To the best of our knowledge, this is the first major analytics-based study that attempts to empirically link marketing to the mission-based performance outcomes of NPs.

The rest of the article is organized as follows. In the next section, we discuss extant literature on the challenges in measuring NP performance and introduce a conceptual framework for linking marketing to NP performance. We then present the model and data used to quantify the impact of marketing on the outcome-based performance measures. Subsequently, we discuss the findings from the model. We conclude with a detailed discussion that includes comments on this study’s conceptual and practical contributions, substantive economic implications of marketing for NPs, public policy implications, limitations, and future directions.

Literature Review

The application of marketing in NPs has received limited attention in the research literature to date. A review of articles published in the top academic journals of marketing indicates a relatively small number of papers. These studies can be broadly categorized into conceptual and empirical studies. The focus of our research is to conduct empirical analyses to link marketing to the performance of NPs. Extant empirical studies in the context of NPs may be classified on in line with the following three research streams.

The first research stream comprises studies that examine how policies and/or different marketing initiatives of an organization influence consumers’ behavior and/or attitudes (e.g., Bolton, Bhattacharjee, and Reed 2015; Dority, McGarvey, and Kennedy 2010; Moore et al. 2002; Wang, Lewis, and Singh 2016; Webb and Mohr 1998). The second research stream comprises studies that evaluate how marketing efforts can be utilized for fundraising and managing donor behavior and donor relationships (e.g., Anik, Norton, and Ariely 2014; Arnett, German, and Hunt 2003; Hung and Wyer 2009; Kara, Spillan, and DeShields 2004; Khodakarami, Petersen, and Venkatesan 2015; Savary, Goldsmith, and Dhar 2015; Shang, Reed, and Croson 2008; Townsend 2017; Van Diepen, Donkers, and Franses 2009; Zhou, Kim, and Wang 2018). The third research stream comprises studies that have proposed methodological advancements in the context of NPs, with the underlying objective of improving prediction and/or understanding donor behavior (e.g., Aravindakshan, Rubel, and Rutz 2015; Dubé, Luo, and Fang 2017; Gopalakrishnan, Bradlow, and Fader 2017; Schweidel and Knox 2013; Netzer, Lattin, and Srinivasan 2008).

There is a gap in extant research concerning empirical research studies that link marketing to the overall performance of NPs (as defined by the mission of an NP). The underlying reason may be attributed to the fact that measuring NP performance is a complicated task. In the case of for-profit organizations, there are well-established performance metrics such as revenue, profits, market share, and market capitalization that can easily help indicate the overall performance of the firm. In the case of NPs, it is difficult to conceive any standard performance metrics. Nonprofits vary substantially in terms of their set of activities, their organizational mission, and, consequently, their desired goals and end objectives. For example, NPs may be related to human welfare, arts and sciences, education, or environmental and wildlife preservation. The sheer diversity and breadth of activities possible within the realm of NPs is coupled with the fact that the performance reporting requirements are not standardized, stringent, or consistent across different NPs.

A related challenge lies in assigning a monetary value to the performance of NPs. By definition, NPs are “not for profit.” The ultimate objective of an NP is usually to make a “positive difference,” which often involves intangible performance outcomes. In such a scenario, how can NPs effectively link marketing to their performance outcomes?

Measuring Nonprofit Performance

Nonprofits commonly employ performance metrics such as the number of dollars raised during their fundraising events/campaigns, the annual growth of their membership, and the number of people served or accessed through their programs and activities (Sawhill and Williamson 2001). Such metrics are useful indicators of intermediate outcomes from different activities of the organization. However, they typically fail to indicate the extent to which an NP may have succeeded in achieving its end objective(s). For example, an NP learning institution may succeed in increasing enrollment and donations or endowments, but it may fail to improve its ultimate objective of increasing the learning outcomes of the students.

Lately, there has been considerable debate among researchers and NP practitioners alike concerning the choice of appropriate metric(s) for NPs. Much of the discussions have converged toward the idea of moving from an “activity”- or “output”-based metric to an “outcome”-based metric. For example, an NP dedicated to alleviating poverty should not measure its success in terms of the number of people it reached out to or the millions of dollars it raised through its fundraising efforts. Instead, it should evaluate its success on the extent to which it could positively transform poverty-stricken people’s lives. In summary, the performance metric(s) of NPs need to focus on the ultimate desired outcome that stems from its focal mission. Focus on the ultimate outcomes that are more valid indicators of an organization’s impact helps shift the emphasis from activities to results and from how a program operates to what it accomplishes (Bell-Rose 2004; Morino 2011).

Several conceptual frameworks have been proposed in the literature by academics (e.g., Coleman 1987; Gasper 1997; Leeuw and Vaessen 2009; Roche 1999), practitioners, institutions, and organizations (e.g., United Way, Urban Institute) to enable NPs to develop an outcome-based performance measure. We draw on extant literature to show that an NP’s path to performance may be generalized as a sequential chain of five stages (Figure 1). At the beginning of the performance chain are an NP’s resources (labeled “Inputs”) such as people (employees/volunteers), infrastructure, and/or expertise that an NP primarily deploys to generate a set of “Outputs” such as fundraising, products, programs, services, or specific activities. For example, the NP Habitat for Humanity deploys its resources to implement a set of programs and services directed at building homes and communities for people in need.

The “Outputs” of an NP’s efforts may be directly measured by outcomes such as the dollar amount of funds raised, products sold, people reached, and/or members subscribed. These outcomes may be regarded as “Intermediate Outcomes” or the means to enable the NP to achieve its mission-related desired positive change or the “End Outcome(s)” as shown in Figure 1. At the end of the chain of the path to performance lies the “Impact” that an NP may generate by influencing policy changes and regulation and/or generating substantive economic outcomes. The impact of marketing in the context of the performance of NPs has been mainly focused on performance measures related to “Outputs” and “Intermediate Outcomes” of Figure 1. In this study, we focus on the potential impact of marketing on the “End Outcomes” of a real-world NP and discuss how it may be linked to substantive economic outcomes.

Methodology and Data

Context of the Study

Our research takes place in the context of a major NP serving individuals with arthritis (hereinafter referred to as ANP²). Arthritis is a chronic disease of the joints and a leading cause of work disability in the United States (Centers for Disease Control and Prevention 2019). The quality of life of an individual with arthritis is adversely impaired through physical activity limitation. The National Health Interview Survey keeps track of the number of individuals with arthritis-attributable activity limitations (AAAL). Currently, arthritis limits the activities of nearly 23.7 million adults (Barbour et al. 2016). By 2040, this number is expected to increase to 34.6 million (Hootman et al. 2016). Arthritis has a profound economic, personal, and societal impact in the United States. Medical expenditures and earnings losses among U.S. adults with arthritis in 2013 were estimated to be approximately $300 billion or about 1% of the U.S. gross domestic product (Murphy et al. 2018).

The ANP (whose data have been employed in this study) has multiple service points located throughout the country that support patients suffering from more than 100 types of arthritis and related conditions. The organization is committed to raising awareness and reducing the adverse consequences of arthritis, which may prove to be devastating for working adults with severe AAAL. Toward this endeavor, the organization funds research and offers education, products, services, and programs directed at improving the quality of life for individuals living with arthritis. The strategic goal of the ANP is to work toward substantially reducing the AAAL in people with arthritis.

Linking Marketing to Nonprofit End Outcome(s)

We adapt the generalized framework of Figure 1 to quantify the extent to which the marketing function helps drive the ANP’s mission-based performance outcome of “reduction in AAAL” (RAAAL). The ANP employs a portfolio of products, programs, services, and publications to empower individuals with arthritis to take control of their health. Therefore, this portfolio of products/activities will serve as “Output” measures for the ANP. The direct performance of these output measures is assessed in terms of sales revenue, enrollment, subscription, and number of visits and total outreach achieved as a consequence of the ANP’s products, programs, and activities. Collectively, these factors comprise organization’s efforts directed at the desired mission-based performance outcome of RAAAL, as depicted in Figure 2. Furthermore, RAAAL could result in a substantive societal and economic impact by contributing to savings in health care costs and preventing loss of earnings in individuals with arthritis.

Marketing can play an essential role in driving the organization’s performance by generating awareness and communicating the benefits of different programs, resources, and services of the ANP to individuals. Consequently, the impact of marketing may be directly evaluated by assessing the change in RAAAL as a consequence of the ANP’s marketing efforts. In addition, the ANP’s marketing and communication efforts may also contribute to educating and motivating participants with the overarching goal of influencing the desired behavioral change (Rothschild 1999). For example, in the context of our study, the ANP disseminates a lot of relevant information to individuals. This includes articles on its websites and magazines as well as one-on-one interactions of customer service representatives or field officers (located across the country) who educate individuals about different sources and symptoms of arthritis and motivate them to take action. By design, the marketing function is best equipped to disseminate such information to the intended target segment. The consequences can be evaluated in terms of two dimensions of change: attitudinal and behavioral (shown in Figure 2). For the ANP, the attitudinal change manifests in terms of an increase in perceived self-efficacy. The behavioral change manifests in terms of an increase in the number of new actions taken by individuals to combat arthritis. Both changes can contribute to the desired end outcome of RAAAL, as shown in Figure 2.

We model all relationships indicated by the solid black arrows in Figure 2 with relevant statistical models, as described in a subsequent section. We discuss, but do not explicitly model, relationships indicated by dashed arrows. In the next subsection, we describe the data collection procedure undertaken to operationalize Figure 2.

Design of Questionnaire for the Survey

We designed a questionnaire to collect information from each participant associated with the ANP. The key performance outcome metric is RAAAL, measured at the participant level. (For a sample of the survey used to collect the relevant data, see Web Appendix A.) We have modified some of the wording in the original survey to preserve the anonymity of the NP. These include the NP’s name, website URL, and names of specific programs associated with the NP. Next, we provide a brief description of the different types of data collected.

Awareness and engagement

The objective of the first section of the questionnaire was to collect information regarding participant’s awareness and engagement or association with the ANP. The awareness was captured using the question “How did you first hear about the organization?” We measured a participant’s engagement in terms of reading the NP’s magazine (“yes/no”), purchasing products from the organization (“yes/no”), participating in NP events, visiting NP websites/social media, and so on. We measured the length of association with the organization as the number of years an individual was associated with the organization.

Marketing

The marketing reach was measured in terms of frequency of receiving/requesting marketing materials—brochures, newsletters, and e-newsletters from the ANP. These marketing materials focused on information pertaining to different types of arthritis, their treatment, and suggesting lifestyle changes and pain management strategies. Thus, receiving and reading these marketing materials could help participants improve their health conditions.

Use of information sources

An individual associated with the ANP (hereinafter referred to as a participant) has the opportunity to learn about arthritis and how to manage pain and activity limitations attributable to arthritis through content available on different websites managed by the ANP. We measured the frequency of an individual visiting these websites or interacting with the organization through social media by using a scale ranging from “Never” to “At least two times a week.”

Interaction with nonprofit’s employees

Individuals suffering from arthritis could also benefit from interacting with the ANP’s customer service employees. The customer service employees were trained to direct individuals to relevant resources and motivate them to take new actions directed at improving their health conditions related to arthritis. We captured the frequency of interaction with the ANP employees through the mail, email, or in-person/phone using a scale ranging from “Never” to “More than 6 times.”

Participation in programs

The ANP had programs specifically designed to help reduce arthritis pain and activity limitations. The ANP offered four different programs, which we refer to as Program 1, Program 2, Program 3, and Program 4. In these programs, a certified instructor led the sessions designed to reduce pain, increase flexibility, and improve the participant’s overall health. Because these programs could directly affect the outcome metric RAAAL, we measured program participation level on a scale ranging from “Never” to “More than two times.”

Arthritis self-efficacy (ASE)

Self-efficacy is one’s perceived capability to execute a behavior required to produce a situation-specific outcome (Bandura 1986). An individual’s self-efficacy can help predict health-related behavior such as undertaking physical activities regularly, eating healthy food, or likelihood to engage in pain coping strategies (Allegrante and Marks 2003; Brady 2011) and is known to be correlated with changes in pain, function, and depression in people suffering from rheumatoid arthritis and osteoarthritis (Brekke, Hjortdahl, and Kvien 2001). In the context of our study, ASE could help predict differences in individual-level health-related behavioral changes undertaken by participants in response to relevant marketing communications.

The original Arthritis Self Efficacy Scale (ASES) was developed by Lorig et al. (1989) to measure patients’ arthritis-specific self-efficacy, or patients’ beliefs that they could perform specific tasks to cope with the consequences of arthritis. The ASES has 20 items in 3 subscales, with 5 items to measure self-efficacy for managing pain, 9 items to measure self-efficacy for physical function, and 6 items for measuring self-efficacy for controlling other symptoms. A parsimonious eight-item scale has been developed more recently that includes items from three subscales in the original ASES (Lorig 1998). This scale has been used in many arthritis-related studies (e.g., Osborne et al. 2007) and has good psychometric properties. The items in the scale are worded as “How certain are you that you can…,” and the response options range from “very uncertain” (1) to “very certain” (10). We used the eight-item parsimonious scale to measure ASE in this study.

Personality

Personality is defined as a combination of characteristics or qualities that form an individual’s distinctive behavior. In the social psychology literature, the Big Five personality dimension framework has been the most extensively researched and accepted paradigm for classifying the personality types of humans. We measured the Big Five personality dimensions—Extraversion, Agreeableness, Conscientiousness, Emotional Stability, and Open to a New Experience—using a parsimonious Ten-Item Personality Inventory developed by Gosling, Rentfrow, and Swann (2003) to account for individual-level differences of the participants.

Model

Our goal is to empirically model the relationship between marketing and the ANP’s key performance outcome (RAAAL) as discussed previously and illustrated in Figure 2. Consequently, we specify a statistical model to evaluate the extent to which (1) marketing influences the desired behavioral changes in participants and (2) desired behavioral changes and other marketing efforts affect RAAAL. In addition, we specify a measurement model to compute the values of ASE and personality of each participant.

Reduction in AAAL is a consequence of participants making behavioral changes that could help reduce arthritis pain and activity limitations. Some self-directed behavioral changes include taking action directed at lifestyle changes, losing weight, adhering to healthy diet plans, and exercising. Some patients also undergo medical intervention such as surgery, medication, and alternative therapy such as yoga. As discussed previously, marketing can help influence the desired behavioral change(s). For instance, marketing materials such as brochures and newsletters may direct participants’ attention to the organization’s different programs and activities and thus influence the desired behavioral change required to cope with arthritis. In addition, various products and information sources such as magazines, products (e.g., DVDs, books), and websites are all designed to give participants the relevant information to help cope with arthritis pain and activity limitations. Some of these materials contain inspiring testimonials of people who benefitted by making behavioral changes. Such testimonials could motivate others to follow suit. In summary, marketing dollars spent on distributing such products and services can help educate, influence, and/or motivate participants to make positive behavioral changes, which could positively affect RAAAL. Therefore, as a first step, it is essential to understand the extent to which the engagement of participants with the ANP’s marketing materials (such as reading a magazine, buying products and services, or visiting websites) drive participants’ behavioral change.

The number of behavioral changes for each participant i may take values such as 0, 1, 2,…up to a maximum of 8 (i.e., it is a count variable). This variable’s conditional variance is close to its mean, which means that it has a low dispersion parameter. Consequently, a Poisson regression (instead of linear regression) model is suitable for estimating factors related to the number of behavioral changes. Thus, we specify a Poisson regression model with a log link function, as follows:

\begin{array}{l} Log ({Number of behavioral changes}_{i}) = β_{0} + β_{1} \\ \times {Read Magazine}_{i} + β_{2} \times Purchased {products}_{i} \\ + β_{3} \times {Website visits}_{i} + β_{4} \times {Marketing materials}_{i} \\ + …… + β_{n} \times {Personality}_{i} + ∊_{1 i} . \end{array}

In addition to indirectly affecting RAAAL by encouraging behavioral changes, marketing can also increase people’s enrollment in programs offered by the ANP, which are explicitly designed to increase mobility and reduce arthritis-related pain and activity limitations. Along with behavioral changes and participation in programs, personal characteristics such as ASE, personality, and demographics can influence RAAAL. For example, ASE captures a participant’s beliefs that they can take corrective action to cope with the consequences of arthritis. Therefore, one would expect participants with higher ASE to have a higher likelihood of experiencing RAAAL.

It is important to know not only whether a participant experienced any RAAAL but also the extent of RAAAL. The potential savings in medical costs and lost wages are determined by the extent of RAAAL experienced by individuals. We model the extent of RAAAL for each participant i as a function of behavioral changes, participation in programs, ASE, personality, and demographics as follows:

\begin{array}{l} {RAAAL}_{i} = β_{0} + β_{1} \times No {. of behavioral changes}_{i} \\ + β_{2} \times {ASE}_{i} + β_{3} \times Participation {(Programs)}_{i} \\ + β_{4} \times {Length of association}_{i} + …… . + β_{n} \\ \times {Gender}_{i} + ∊_{2 i} . \end{array}

However, a linear regression model is not suitable for estimating this model’s coefficients because of the nature of the dependent variable, RAAAL. There will always be a percentage of individuals suffering from arthritis who may never experience RAAAL. As a result, any interval scale used to capture RAAAL may have a proportion of people reporting “0” or “no RAAAL.” In this particular study, 56% of the participants did not experience RAAAL. In other words, only 44% of the participants who experienced RAAAL will have values higher than zero for the dependent variable. The proposed model should be able to address this data censoring issue. A Type I Tobit model in which a latent variable is observed for values more than zero and censored otherwise would be able to address this issue appropriately. Therefore, we propose the following Tobit model to study the impact of different programs and activities by the ANP on RAAAL:

y_{i} = \{\begin{matrix} y^{*} i f y^{*} > 0 \\ 0 i f y^{*} \leq 0 \end{matrix},

y_{i}^{*} = X_{i} β + ∊_{i},

where $y_{i}^{*}$ represents RAAAL_i as expressed as in Equation 2.

The use of Tobit model gives us the ability to decompose the impact of the variables on (1) the probability of experiencing RAAAL for the segment of the population that has not yet reported RAAAL and (2) the extent of RAAAL experienced by those who report RAAAL. Roncek (1992) illustrates how the marginal effects of independent variables can be obtained based on the derivations provided by Maddala (1983). We use the coefficients obtained from the Poisson and Tobit models to calculate the marginal effects of different variables. Web Appendix B provides the formulas for calculating each variable’s marginal effects on the probability of experiencing RAAAL and the extent of RAAAL.

As the equations show, the number of behavioral changes, which is the dependent variable in Equation 1, is one of the independent variables in Equation 2. As a result, the factors affecting the number of behavioral changes are also indirectly influencing RAAAL. In other words, the error terms of the two models (or the unobserved factors of the two models) are correlated, which increases the need to estimate these equations jointly. When both equations are linear, models can be jointly estimated using seemingly unrelated regression. However, the dependent variables (i.e., number of behavior changes and RAAAL) in Equations 1 and 2 belong to different families of nonlinear distributions, which makes the joint estimation of these models challenging. Such scenarios are common in the context of NPs. That is, performance outcomes of NPs could be multiple and comprise both linear and nonlinear functional forms. For example, an educational institution may want to assess student outcomes based on the extent of improvement in students’ test scores (a continuous measure) and whether the student was able to secure a job within three months of graduation (a binary measure).

A class of multivariate probability models called copula models can be used for joint estimation of such performance outcomes because they allow for joint estimation of any univariate marginal distributions that can come from different distributional families (Danaher and Smith 2011a). Copula models have been used in statistics for several decades. However, it is only recently that it has been widely adopted in several fields, including finance, actuarial sciences, engineering, health economics, and transportation research (e.g., Bhat and Eluru 2009; Cherubini, Luciano, and Vecchiato 2004; Frees and Valdez 1998; Yan 2006; Quinn 2007; Zimmer and Trivedi 2006). These models have been introduced to marketing in recent years (Danaher and Smith 2011a) and are being increasingly used by marketing researchers (e.g., Kumar, Zhang, and Luo 2014; Park and Gupta 2012).

In the context of this study, we use the copula method to capture the dependence of univariate marginal distributions of the number of behavioral changes and RAAAL. Our approach includes (1) fitting a Poisson regression model to identify the factors influencing the number of behavioral changes, which is a count variable; (2) estimating the parameters of a Type 1 Tobit model to identify the factors impacting RAAAL; (3) using the copula method to capture the dependence of univariate marginal distributions of the number of behavioral changes and RAAAL; and (4) calculating the marginal effects to understand the direct and indirect effects of various factors on the probability of experiencing RAAAL and the extent of RAAAL. For a detailed discussion of the copula method, choice of copula models, and the approach used to estimate copula model parameters, see Web Appendix C.

As discussed previously, “ASE” and “personality” are latent constructs measured using an eight-item ASE and the Ten-Item Personality Inventory scale, respectively. We specify a measurement model to test these latent constructs’ validity and reliability using confirmatory factor analysis. We then apply the factor regression coefficients to calculate a value (or score) for the latent constructs so that they can be used in the models specified in Equations 1 and 2.

Data Collection and Results

We first ran a pretest for the questionnaire to check for accuracy, ease of use, and adequacy of questions and determine whether the constructs have acceptable composite reliability. We received a total of 100 responses for the pretest from individuals suffering from arthritis. The pretest revealed that participants did not have difficulty understanding the questions, and the survey questions were adequate to address our main research questions. However, we added one item regarding the length of association with the ANP. This could potentially play a role in participants taking new actions and their subsequent reduction in AAAL. The pretest also showed that the eight-item measure of ASE and ten-item personality scale had Cronbach’s alphas of .93 and .74, respectively, indicating good reliability. Respondents for the finalized survey were identified through the ANP’s database of arthritis patients or participants. A market research firm was employed to send out more than 5,000 surveys. We received completed and acceptable responses from about 1,100 respondents.

Table 1 summarizes the respondent profiles and descriptive statistics of key variables. It shows that 78% of survey respondents are women, which is consistent with the fact that women are more likely to get arthritis, especially certain types of arthritis such as rheumatoid arthritis, osteoarthritis, and fibromyalgia. The survey found that 60% of the respondents are married, 54% have completed at least four years of a college education, and 56% had an annual household income between $36,000 and $100,000. About 44% of survey participants had experienced RAAAL, which is the key outcome metric of this study.

Table 1.

Profile of Survey Respondents and Key Descriptive Statistics.

Characteristics		Descriptive Statistics
Characteristics		Mean/Percentage	SD
Experienced RAAAL		44%
Number of behavioral changes		2.92	1.89
Age (years)		59.6	11.8
Gender	Female	78.3%
Marital status	Married	60.2%
Marital status	Widowed/divorced	24.1%
Race/ethnicity	White	90.5%
	Hispanic	3.6%
	African American	2.7%
Education	At least four years of college	54.2%
Income	<$35,000	21.3%
	$36,000–$100,000	55.6%
	>$100,000	23.1%
Visited at least one website of the ANP		45.6%
Received marketing materials from the ANP		25.4%
Read the ANP’s magazine		26.2%
Purchased the ANP’s products		24.1%
Participated in at least one program offered by the ANP		13.4%

Web Appendix D reports the correlations between the number of behavioral changes, RAAAL, interaction/marketing variables, and key demographic variables. Both the number of behavioral changes and ASE have a significant positive correlation with RAAAL. Most of the participant-initiated interactions with the ANP, such as reading publications, participating in programs, receiving marketing materials, and engaging online (through the ANP’s websites), are significantly correlated with the number of behavioral changes.

Measurement Model

We conducted a confirmatory factor analysis to check for the validity and reliability of the latent construct, ASE, and to get the factor regression coefficients. The confirmatory factor analysis results show a Cronbach’s alpha of .949 for ASE, which is much higher than the minimum level (.7) specified in the literature (Nunnally 1978), suggesting high reliability for the construct. Also, the items loaded highly on the latent construct. Similarly, the personality scale has a Cronbach’s alpha of .75, suggesting good internal consistency.

Checking for Common Method Bias

In the survey, we collect data related to the dependent variable (RAAAL) and other variables that could possibly influence RAAAL from the same individual. Consequently, we need to check for common method bias. One of the commonly used techniques to detect common method bias is the Harman single-factor test, in which all measures are loaded into an exploratory factor analysis. The emergence of a single factor or a general factor accounting for more than 50% of the variance indicates the presence of common method bias (Podsakoff et al. 2003). We conduct a Harman single-factor test by loading all the variables that are suspected to be influenced by a common source such as respondent’s measures of ASE, personality, the frequency of interaction with the ANP, frequency of receiving marketing materials, number of new actions taken, and so on as well as the dependent variable RAAAL to exploratory factor analysis. The total variance explained by a single factor was only 18.7% of the overall variance, which is well below the 50% cutoff. The results imply that the common method is not problematic and does not significantly bias the results.

Number of Behavioral Changes

To understand the influence of marketing materials provided by the ANP on the number of behavioral changes a participant makes, we ran a Poisson regression. Table 2 presents the parameter estimates from Poisson regression and the marginal effects.

Table 2.

Number of Behavioral Changes: Poisson Regression Model.

Variable	Coefficients	Marginal Effect
Intercept	−.395
Read publications	.224***	.621
Purchased products (0/1)	.051***	.141
Engaged online	.110***	.305
Received direct marketing	.111***	.307
Length of association with the ANP	.015***	.042
Female	.108***	.299
Hispanic	–.019*	−.053
Married	.066**	.183
At least four years higher education	−.082**	−.227
Age	.002	.006
Income	−.0002	−.0005
ASE	.003	.008
Personality	.009	.025
Dispersion	.050
−2 log-likelihood	1,457.1

* Significant at 10%.

** Significant at 5%.

*** Significant at 1%.

The results show that the organization’s marketing resources positively influence the participants’ number of behavioral changes. For example, reading publications (.224), purchasing products (.051), and engaging online (.110) significantly and positively affect the number of behavioral changes. The marginal effect for reading publications is .621. This means that the expected number of behavioral changes for an individual who reads the ANP’s publications is .621 more than for someone who has not read the publications. The coefficient for receiving marketing materials is positive (.111) and significant. This shows that requesting or receiving marketing materials such as brochures, newsletters, and e-newsletters helps enhance the number of behavioral changes, and the expected number of behavioral changes for a person receiving such materials is .307 more than for someone not receiving direct marketing. The coefficients and the marginal effects of other variables given in Table 2 may be interpreted similarly. Table 3 shows how the number of behavioral changes, the individual’s participation in different programs offered by the ANP, and other factors affect RAAAL.

Table 3.

RAAAL: Tobit Model Results.

Variable	RAAAL: Tobit	Marginal Effect on RAAAL Level	Marginal Effect on Prob. of Experiencing RAAAL (%)
Intercept	−7.038
Read publications^a		.173	3.88
Purchased products^a		.039	.88
Engaged online^a		.085	1.91
Received direct marketing^a		.086	1.92
Attended the ANP’s programs	1.997***	.661	14.79
Length of association with the ANP	.019**	.006	.14
ASE	.499***	.165	3.69
Number of behavioral changes	.845***	.279	6.26
Female	.476***	.157	3.52
Sigma	5.238
Log-likelihood	−1,871

** Significant at 5%.

*** Significant at 1%.

^a These marginal effects are indirect effects based on the marginal effects of these variables on number of behavioral changes and the marginal effect of number of behavioral changes on RAAAL.

The Tobit model results in Table 3 suggest that participation in different programs (1.997) specifically designed to reduce AAAL has the highest positive (and significant) impact on RAAAL. The results also show that the higher the number of behavioral changes (.845) an individual makes, the higher their RAAAL. As we expected, an individual’s ASE plays a crucial role in reducing AAAL, as indicated by positive (.499) and significant coefficient for ASE. To understand the real impact of these variables on RAAAL for two segments of the population—those who have not experienced RAAAL and those who have—we calculate the marginal effects of the variables using the formulas presented in Web Appendix B. Participating in programs offered by the ANP has the highest marginal effect on the extent of RAAAL (.661) for those who have already experienced RAAAL and on the probability of experiencing RAAAL (14.8%) for those who have not experienced RAAAL. Corresponding marginal effects of ASE are .165 and 3.7%, respectively. Because the number of behavioral changes a person makes is influenced by factors such as reading publications, purchasing products, receiving marketing materials and/or engaging online, we report the indirect marginal effects of these activities on RAAAL. For instance, reading publications has an indirect marginal effect of .173 on the extent of RAAAL and 3.88% on the probability of experiencing RAAAL. The marginal effects of other variables are given in Table 3. These marginal effects reveal the extent to which each activity or program offered by the organization is effective in reducing AAAL. This is a valuable insight for the ANP to enable them to decide (1) what program to offer based on the personal situation of the individual and (2) the amount of resources to be allocated to each program.

We apply the parameter estimates from the Poisson and Tobit models to calculate the marginal distribution functions in the copula likelihood, as explained in the estimation steps for the inference functions of margin method for copula models (see Web Appendix C for details). We fitted five commonly used copula models to select the best model. Table 4 provides the association coefficients for each of these copulas and the Akaike information criteria (AICs) of the models.

Table 4.

Copula Parameter Estimates and Model Fit.

Copula	Correlation Coefficient	Association Coefficient	Log-Likelihood	AIC
T	.2325		188.39	−372.78
Normal	.1811
Clayton		.1396***	31.43	−60.85
Frank		1.3534***	139.29	−276.58
Gumbel		1.1579***	236.32	−470.64

Table 4 shows that the Clayton copula has the best fit, indicated by the lowest AIC of 60.85. It also has a significant association coefficient of .1396, which suggests that the number of behavioral changes and RAAAL are significantly associated with each other. The plot of simulated values of number of behavioral changes (i.e., Poisson_pred) and RAAAL (i.e., Tobit_pred) given in Web Appendix E reveals that the Clayton copula captures the linkage between these variables very well.

Generalization of the Methodology

Evaluating performance outcomes in the context of NPs is relatively more complicated, as there may be a variety of intangible and tangible outcomes and/or outcomes with a widely different functional form (i.e., linear or nonlinear with changes in behavior and/or attitude of the target segment mediating the relationship between marketing and key performance outcomes). Our proposed framework (illustrated in Figure 2) and the methodology implemented in the context of the ANP can be easily generalized in the context of other NPs. The copula-based approach is specifically useful in the context of NPs as it allows performance outcome models of varying relationships (i.e., any combination of nonlinear and linear relationships) to be estimated jointly.

Discussion

Conceptual and Practical Contribution

Nonprofit organizations are often concerned about the financial efficiency of their operations and with monitoring financial performance measures such as the ratio of program expenses to total operating expenses or the share of administrative costs in total operating expenses. Many NPs highlight their fundraising efficiency (i.e., fundraising expenses divided by the total contributions received) in their communication to stakeholders (Epstein and McFarlan 2011). Minimizing overhead costs to increase efficiency and funding for an NP’s core programs and activities makes intuitive sense. However, in practice, such an approach can prove suboptimal, as it assumes that all overhead expenses (e.g., marketing costs) will negatively affect an NP’s performance (Gregory and Howard 2009). Using data from Habitat for Humanity, Coupet and Berrett (2019) find that limiting overhead spending can potentially compromise an NP’s mission and even drive it out of business. Therefore, NPs need to carefully assess the relative impact of overhead on their organization’s performance. A related challenge lies in NP’s ability to assess the factors that affect the organization’s performance. Buchanan (2019, p. 140) finds that “it’s not that nonprofits aren’t committed to assessing and improving performance. It’s that they need much more support and resources to do this work…because business performance assessment for a nonprofit is wholly different and much more complicated undertaking than performance assessment in business.”

In this study, we address this issue in the context of marketing costs. We contribute to the extant literature by making a case for how and why marketing should be regarded as a critical factor for driving the performance of an NP and present a methodology to enable NPs to assess the impact of marketing to their mission-based performance outcome(s). More specifically, we find empirical support for the fact that marketing can play a pivotal and significant role in improving the mission-based performance outcomes of NPs.

Our methodology discussed previously and the framework presented in Figure 2 (in the context of the ANP) can be easily generalized to other NPs. By implementing such a framework, NPs can measure, quantify, and demonstrate the impact of their marketing efforts to their stakeholders. All of this can enable NPs to justify the allocation of marketing resources to drive their performance outcome(s).

Substantive Economic Implications of Marketing for Nonprofits

If marketing positively affects the mission-based performance outcome, it may also affect the substantive economic outcome associated with the NP’s mission (as conceptualized in Figures 1 and 2). For example, in the context of our study, we could empirically establish the impact of marketing on driving the mission-based performance outcome of RAAAL for the ANP. This performance outcome has substantive economic implications. According to the Centers for Disease Control and Prevention, each year, arthritis is estimated to cost $140 billion in medical costs and an additional $164 billion in lost wages (in 2013 dollars). There is an adverse impact on human cost in terms of pain and suffering, indicated by the fact that arthritis is responsible for 750,000 hospitalizations and 36 million outpatient visits each year (Hootman and Helmick 2006). Another study (Murphy et al. 2018) based on the Medical Expenditures Panel Survey finds that the average medical care expenditures attributable to arthritis and other rheumatic conditions were $2,117 in 2013, which is an increase from $1,752 in 2003 (Yelin et al. 2007). Moreover, arthritis and other rheumatic conditions accounted for a loss of earnings of $4,040 per working adult (on average) in 2013. Hootman et al. (2016), drawing on the 2013 National Health Interview Survey data, predict that 78.4 million adults will have doctor-diagnosed arthritis by the year 2040, and 44% of those (34.6 million) will experience AAAL. As the U.S. population ages, arthritis-related costs are likely to soar.

In summary, AAAL costs the nation hundreds of billions of dollars each year. Reduction in AAAL can result in substantive economic impact by reducing costs associated with health care and/or lost wages of individuals suffering from AAAL. One study estimated that implementing arthritis self-help courses among just 10,000 people with arthritis can result in a net savings of more than $2.5 million over four years (Kruger et al. 1998). If marketing significantly affects RAAAL (as shown in this study), it also contributes to the substantive economic outcome associated with RAAAL. Other NPs could employ a similar argument to highlight the critical importance of marketing in their organization.

Public Policy Implications

Policy makers play a vital role in funding and influencing NPs and their operations. One of the most critical policy issues facing NPs is better understanding the different forms of NP financing (Jeavons 2011). More specifically, what are the different areas in which policy makers can fund NPs and help them improve their performance?

With the help of findings from this study, we want to highlight marketing’s potential in driving NP performance and thus present a compelling case for policy makers to fund marketing in NPs and to create more avenues, opportunities, or incentives for NPs to receive marketing resources and/or funds to improve their performance. Our study offers both NPs and policy makers the tools to evaluate NP performance more holistically based on their mission-based performance outcomes (that may include intangible outcomes) rather than just easy-to-measure activity-based outputs or intermediate outcomes. It is also essential for policy makers to fund NPs’ marketing efforts on basis of “operational” rather than “financial” efficiency—to regard marketing as a necessary investment rather than an avoidable cost for driving the mission-based performance outcome(s) of NPs. Policy makers can play a significant role in spearheading this shift and encouraging NPs to invest and fully capitalize on the potential of marketing in their organization.

Limitations and Future Research Directions

The NP sector comprises millions of NPs worldwide that work tirelessly to address some of the biggest challenges faced by our society today. As researchers, we can help contribute by analyzing factors driving NP performance. Kotler (1979) recognized marketing as one of the critical drivers of NP performance decades ago. Yet widespread application of marketing practices in NPs are limited to date. This study proposes a framework, develops a model, and conducts empirical analyses to enable NPs to quantify the impact of marketing on their mission-based performance. Empirical evidence of the role of marketing in driving the mission-based performance of NPs is limited. There is tremendous potential for future research in this space.

Our study relies on survey data, which have inherent limitations. One of the major limitations is that it offers only a snapshot view of the phenomenon being studied. It is difficult to capture any dynamics of the phenomenon with survey data. For example, does the effectiveness of marketing strengthen or weaken over time? Do the desired behavioral changes have a time threshold? Does it make sense for an NP to temporally sequence its marketing efforts? If so, how and to what extent? These questions are difficult to answer with survey data. Future researchers should consider employing longitudinal transaction data from an NP and address these questions and more.

This study is in the context of a specific NP. Future research studies could replicate this study in the context of NPs from other sectors and diverse areas such as climate change, environmental pollution, health care, poverty alleviation, and so on. An interesting extension of this research could entail capturing the attitude of policy makers and leaders of NPs on the role and scope of marketing in driving performance. Recently, the importance of mobile, digital, and social media marketing has been exponentially increasing. Future research studies could also explore the role of digital, social, and mobile in driving NP performance or develop newer, more sophisticated models to link marketing to NP performance. In summary, linking marketing to NP performance is a substantive area of research, and we look to our fellow researchers to contribute further to this research stream.

Supplemental Material

Supplemental Material, sj-pdf-1-ppo-10.1177_0743915620978538 - Linking Marketing to Nonprofit Performance

Supplemental Material, sj-pdf-1-ppo-10.1177_0743915620978538 for Linking Marketing to Nonprofit Performance by Denish Shah and Morris George in Journal of Public Policy & Marketing

Footnotes

Acknowledgments

The authors are grateful to the survey participants and the nonprofit organization for providing data for this study.

Author Contributions

Both authors have contributed equally to this study.

Special Issue Guest Coeditors

Brennan Davis, Dhruv Grewal, and Steve Hamilton

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Online supplement:

Notes

References

Accenture (2006), “Identifying Enablers of Nonprofit High Performance,” (accessed January 1, 2020), http://cdn.volunteermatch.org/www/corporations/resources/docs/Accenture.pdf.

Akchin

Don

(2001), “Nonprofit Marketing: Just How Far Has It Come?” Nonprofit World, 19 (1), 33–35.

Allegrante

John P.

Marks

Ray

(2003), “Self-Efficacy in Management of Osteoarthritis,” Rheumatic Disease Clinics of North America, 29 (4), 747–68.

Anik

Lalin

Norton

Michael I.

Ariely

Dan

(2014), “Contingent Match Incentives Increase Donations,” Journal of Marketing Research, 51 (6), 790–801.

Aravindakshan

Ashwin

Rubel

Olivier

Rutz

Oliver

(2015), “Managing Blood Donations with Marketing,” Marketing Science, 34 (2), 269–80.

Arnett

Dennis B.

German

Steve D.

Hunt

Shelby D.

(2003), “The Identity Salience Model of Relationship Marketing Success: The Case of Nonprofit Marketing,” Journal of Marketing, 67 (2), 89–105.

Bandura

Albert

(1986), Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, NJ: Prentice Hall.

Barbour

Kamil E.

Boring

Michael

Helmick

Charles G.

Murphy

Louise B.

Qin

Jin

(2016), “Prevalence of Severe Joint Pain Among Adults with Doctor-Diagnosed Arthritis — United States, 2002–2014,” MMWR Morbidity and Mortality Weekly Report, 65 (39), 1052–56.

Bell-Rose

Stephanie

(2004), “Using Performance Metrics to Assess Social Impact and Financial Return,” in Generating and Sustaining Nonprofit Earned Income: A Guide to Successful Enterprise Strategies, Chapter 16. San Francisco: Jossey-Bass, 269–80.

10.

Bhat

Chandra R.

Eluru

Naveen

(2009), “A Copula-Based Approach to Accommodate Residential Self-Selection Effects in Travel Behavior Modeling,” Transportation Research Part B: Methodological, 43 (7), 749–65.

11.

Bloom

Paul N.

(2009), “Overcoming Consumption Constraints Through Social Entrepreneurship,” Journal of Public Policy & Marketing, 28 (1), 128–34.

12.

Bolton

Lisa E.

Bhattacharjee

Amit

Reed

Americus

II (2015), “The Perils of Marketing Weight-Management Remedies and the Role of Health Literacy,” Journal of Public Policy & Marketing, 34 (1), 50–62.

13.

Brady

Teresa J.

(2011), “Measures of Self-Efficacy,” Arthritis Care & Research, 63 (11), 473–85.

14.

Brekke

Mette

Hjortdahl

Per

Kvien

T. K.

(2001), “Self-Efficacy and Health Status in Rheumatoid Arthritis: A Two-Year Longitudinal Study,” Rheumatology, 40 (4), 387–92.

15.

Buchanan

Phil

(2019), Giving Done Right: Effective Philanthropy and Making Every Dollar Count. New York: Public Affairs.

16.

Centers for Disease Control and Prevention (2019), “Arthritis: How CDC Improves Quality of Life for People With Arthritis,” (January 30), https://www.cdc.gov/chronicdisease/resources/publications/factsheets/arthritis.htm.

17.

Cherubini

Umberto

Luciano

Elisa

Vecchiato

Walter

(2004), Copula Methods in Finance, Chichester, UK: John Wiley & Sons Ltd.

18.

Coleman

Gilroy

(1987), “Logical Framework Approach to the Monitoring and Evaluation of Agricultural and Rural Development Projects,” Project Appraisal, 2 (4), 251–59.

19.

Coupet

Jason

Berrett

Jessica L.

(2019), “Toward a Valid Approach to Nonprofit Efficiency Measurement,” Nonprofit Management and Leadership, 29 (3), 299–320.

20.

Danaher

Peter J.

Smith

Michael S.

(2011a), “Modeling Multivariate Distributions Using Copulas: Applications in Marketing,” Marketing Science, 30 (1), 4–21.

21.

Dority

Bree L.

McGarvey

Mary G.

Kennedy

Patricia F.

(2010), “Marketing Foods and Beverages in Schools: The Effect of School Food Policy on Students’ Overweight Measures,” Journal of Public Policy & Marketing, 29 (2), 204–18.

22.

Dubé

Jean-Pierre

Luo

Xueming

Fang

Zheng

(2017), “Self-Signaling and Prosocial Behavior: A Cause Marketing Experiment,” Marketing Science, 36 (2), 161–86.

23.

Epstein

Marc J.

McFarlan

F. Warren

(2011), “Measuring the Efficiency and Effectiveness of a Nonprofit’s Performance,” Strategic Finance, 4, 27–34.

24.

Frees

Edward W.

Valdez

Emiliano

(1998), “Understanding Relationships Using Copulas,” North American Actuarial Journal, 2 (1), 1–25.

25.

Gasper

Des

(1997), “Logical Frameworks: A Critical Assessment: Managerial Theory, Pluralistic Practice,” Working Paper Series No. 264. The Hague: Institute of Social Studies.

26.

Gopalakrishnan

Arun

Bradlow

Eric T.

Fader

Peter S.

(2017), “A Cross-Cohort Changepoint Model for Customer-Base Analysis,” Marketing Science, 36 (2), 195–213.

27.

Gosling

Samuel D.

Rentfrow

Peter J.

Swann

William B.

Jr . (2003), “A Very Brief Measure of the Big-Five Personality Domains,” Journal of Research in Personality, 37 (6), 504–28.

28.

Gregory

Ann G.

Howard

Don

(2009), “The Nonprofit Starvation Cycle,” Stanford Social Innovation Review (Fall), https://ssir.org/articles/entry/the_nonprofit_starvation_cycle.

29.

Hootman

Jennifer M

Helmick

Charles G.

Barbour

Kamil E.

Theis

Kristina A.

Boring

Michael A.

(2016), “Updated Projected Prevalence of Self-Reported Doctor-Diagnosed Arthritis and Arthritis-Attributable Activity Limitation Among US Adults, 2015-2040,” Arthritis & Rheumatology, 68 (7), 1582–87.

30.

Hootman

Jennifer M.

Helmick

Charles G.

(2006), “Projections of U.S. Prevalence of Arthritis and Associated Activity Limitations,” Arthritis & Rheumatism, 54 (1), 266–29.

31.

Hung

Iris W.

Wyer

Robert S.

(2009), “Differences in Perspective and the Influence of Charitable Appeals: When Imagining Oneself as the Victim Is Not Beneficial,” Journal of Marketing Research, 46 (3), 421–34.

32.

Jeavons

Thomas

(2011), “ARNOVA’s Symposium on Public Policy for Nonprofits,” Nonprofit Policy Forum, 2 (1).

33.

Kara

Ali

Spillan

John E.

DeShields

Oscar W.

(2004) “An Empirical Investigation of the Link between Market Orientation and Business Performance in Nonprofit Service Providers,” Journal of Marketing Theory and Practice, 12 (2), 59–72.

34.

Khodakarami

Farnoosh

Petersen

J. Andrew

Venkatesan

Rajkumar

(2015) “Developing Donor Relationships: The Role of the Breadth of Giving,” Journal of Marketing, 79 (4), 77–93.

35.

Kotler

Philip

(1979), “Strategies for Introducing Marketing into Nonprofit Organizations,” Journal of Marketing, 43 (1), 37–44.

36.

Kruger

Judy M.S

Helmick

Charles G.

Callahan

Leigh F.

Haddix

Anne C.

, (1998), “Cost-Effectiveness of the Arthritis Self-Help Course,” Archives of Internal Medicine, 158 (11), 1245–49.

37.

Kumar

Zhang

Luo

Anita

(2014), “Modeling Customer Opt-In and Opt-Out in a Permission-Based Marketing Context,” Journal of Marketing Research, 51 (4), 403–19.

38.

Leeuw

Frans

Vaessen

Jos

(2009), Impact Evaluations and Development: NONIE Guidance on Impact Evaluation. Washington, DC: Network of Networks on Impact Evaluation (NONIE).

39.

Lorig

Kate

(1998), “Arthritis Self-Efficacy Scale,” Self-Management Resource Center, https://www.selfmanagementresource.com/docs/pdfs/English_-_self-efficacy_arthritis.pdf.

40.

Lorig

Kate

Chastain

Robert L.

Ung

Elain

Shoor

Stanford

Holman

Halsted R.

(1989), “Development and Evaluation of a Scale to Measure Perceived Self-Efficacy in People with Arthritis,” Arthritis & Rheumatism, 32 (1), 37–44.

41.

Maddala

G.S.

(1983), Limited-Dependent and Qualitative Variables in Economics. New York: Cambridge University Press.

42.

McKeever

Brice S.

(2018), The Nonprofit Sector in Brief 2018: Public Charities, Giving and Volunteering. Urban Institute, National Center for Charitable Statistics (accessed November 28, 2019), https://nccs.urban.org/publication/nonprofit-sector-brief-2018#the-nonprofit-sector-in-brief-2018-public-charites-giving-and-volunteering.

43.

Moore

Jesse N.

Raymond

Mary Anne

Mittelstaedt

John D.

Tanner

John F.

(2002), “Age and Consumer Socialization Agent Influences on Adolescents’ Sexual Knowledge, Attitudes, and Behavior: Implications for Social Marketing Initiatives and Public Policy,” Journal of Public Policy & Marketing, 21 (1), 37–52.

44.

Morino

Mario

(2011), Leap of Reason: Managing to Outcomes In an Era of Scarcity, 1st ed. Washington, DC: Venture Philanthropy Partners.

45.

Murphy

Louise B.

Cisternas

Miriam G.

Pasta

David J.

Helmick

Charles G.

Yelin

Edward H.

(2018), “Medical Expenditures and Earnings Losses Among US Adults with Arthritis in 2013,” Arthritis Care & Research, 70 (6), 869–76.

46.

Netzer

Oded

Lattin

James M.

Srinivasan

(2008), “A Hidden Markov Model of Customer Relationship Dynamics,” Marketing Science, 27 (2), 185–204.

47.

Newbert

Scott L.

(2012), “Marketing amid the Uncertainty of the Social Sector: Do Social Entrepreneurs Follow Best Marketing Practices?” Journal of Public Policy & Marketing, 31 (1), 75–90.

48.

Nunnally

Jum C.

(1978), Psychometric Theory, 2nd ed. New York: McGraw-Hill

49.

Osborne

Richard H.

Wilson

Tanya

Lorig

Kate R.

McCall

Geoffrey J.

(2007), “Does Self-Management Lead to Sustainable Health Benefits in People with Arthritis? A 2-year Transition Study of 452 Australians,” Journal of Rheumatology, 34 (5), 1112–17.

50.

Park

Sungho

Gupta

Sachin

(2012), “Handling Endogenous Regressors by Joint Estimation Using Copulas,” Marketing Science, 31 (4), 567–86.

51.

Podsakoff

Philip M.

MacKenzie

Scott B.

Lee

Jeong-Yeon

Podsakoff

Nathan P.

(2003), “Common Method Biases in Behavioral Research: A Critical Review of the Literature and Recommended Remedies,” Journal of Applied Psychology, 88 (5), 879–903.

52.

Pope

Jennifer A.

Isely

Elaine Sterrett

Asamoa-Tutu

Fidel

(2009), “Developing a Marketing Strategy for Nonprofit Organizations: An Exploratory Study,” Journal of Nonprofit & Public Sector Marketing, 21 (2), 184–201.

53.

Quinn

Casey

(2007), “The Health-Economic Applications of Copulas: Methods in Applied Econometric Research,” HEDG Working Papers 07/22, University of York.

54.

Roche

Chris

(1999), Impact Assessment for Development Agencies: Learning to Value Change. Oxford, UK: Oxfam Publishing.

55.

Roncek

Dennis W.

(1992), “Learning More from Tobit Coefficients: Extending a Comparative Analysis of Political Protest,” American Sociological Review, 57 (4), 503–07.

56.

Rothschild

Michael L.

(1999), “Carrots, Sticks, and Promises: A Conceptual Framework for the Management of Public Health and Social Issue Behaviors,” Journal of Marketing, 63 (4), 24–37.

57.

Savary

Jennifer

Goldsmith

Kelly

Dhar

Ravi

(2015), “Giving Against the Odds: When Tempting Alternatives Increase Willingness to Donate,” Journal of Marketing Research, 52 (1), 27–38.

58.

Sawhill

John C.

Williamson

David

(2001), “Mission Impossible? Measuring Success in Nonprofit Organizations,” Nonprofit Management & Leadership, 11 (3), 371–86.

59.

Schweidel

David A.

Knox

George

(2013), “Incorporating Direct Marketing Activity into Latent Attrition Models,” Marketing Science, 32 (3), 471–87.

60.

Shang

Jen

Reed

Americus

Croson

Rachel

(2008), “Identity Congruency Effects on Donations,” Journal of Marketing Research, 45 (3), 351–61.

61.

Townsend

Claudia

(2017), “The Price of Beauty: Differential Effects of Design Elements With and Without Cost Implications in Nonprofit Donor Solicitations,” Journal of Consumer Research, 44 (4), 794–815.

62.

Van Diepen

Merel

Donkers

Bas

Franses

Philip Hans

(2009), “Dynamic and Competitive Effects of Direct Mailings: A Charitable Giving Application,” Journal of Marketing Research, 46 (1), 120–33.

63.

Wang

Yanwen

Lewis

Michael

Singh

Vishal

(2016), “The Unintended Consequences of Countermarketing Strategies: How Particular Antismoking Measures May Shift Consumers to More Dangerous Cigarettes,” Marketing Science, 35 (1), 55–72.

64.

Webb

Deborah J.

Mohr

Lois A.

(1998), “A Typology of Consumer Responses to Cause-Related Marketing: From Skeptics to Socially Concerned,” Journal of Public Policy & Marketing, 17 (2), 226–38.

65.

Yan

Jun

(2006), “Multivariate Modeling with Copulas and Engineering Applications,” in Handbook of Engineering Statistics, Pham

, ed. London: Springer-Verlag, 973–90.

66.

Yelin

Edward

Murphy

Louise

Cisternas

Miriam G.

Foreman

Aimee J.

Pasta

David J.

Helmick

Charles G.

(2007), “Medical Care Expenditures and Earnings Losses Among Persons with Arthritis and Other Rheumatic Conditions in 2003, and Comparisons with 1997,” Arthritis and Rheumatism, 56 (5), 1397–1407.

67.

Zhou

Xinyue

Kim

Sara

Wang

Lili

(2018), “Money Helps When Money Feels: Money Anthropomorphism Increases Charitable Giving,” Journal of Consumer Research, 45 (5), 953–72.

68.

Zimmer

David M.

Trivedi

Pravin K.

(2006), “Using Trivariate Copulas to Model Sample Selection and Treatment Effects: Application to Family Health Care Demand,” Journal of Business and Economic Statistics, 24 (1), 63–76.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.43 MB