The moderating role of religiosity on nonprofit advertising

Journal of Business Research 67 (2014) 23–31

Contents lists available at ScienceDirect

Journal of Business Research

The moderating role of religiosity on nonprofit advertising☆ Christopher D. Hopkins a,⁎, Kevin J. Shanahan b, Mary Anne Raymond a a b

Clemson University, Department of Marketing, USA Department of Marketing, Quantitative Analysis and Business Law, Mississippi State University, USA

a r t i c l e

i n f o

Article history: Received 1 August 2011 Received in revised form 1 December 2011 Accepted 1 April 2012 Available online 26 April 2013 Keywords: Nonprofit Religion Social responsibility Advertising

a b s t r a c t Nonprofits need to compete for donations now more than ever. Maintaining a socially responsible reputation and developing ads that are emotional and likable are tactics used to encourage donations. This study explores the role of religiosity as a moderator of antecedents to intent to donate to the nonprofit sponsor of pro-social ads and finds that religiosity acts as a moderator in the relationship between liking of the ad, perceived corporate social responsibility of the nonprofit, and intent to donate to the nonprofit. Managerial implications suggest that nonprofits employing negative emotional appeals in advertising should (1) feature real victims (not actors) to generate empathy and (2) target more religious individuals through demographic and psychographic segmentations. © 2013 Elsevier Inc. All rights reserved.

1. Introduction Nonprofit organizations have grown from 12,000 in 1940 (Boris, 1999) to over 1.5 million by 2010 (Urban Institute, 2010) and must compete against other nonprofits for donations. Message tactics chosen by these organizations influence the way the public responds to nonprofit messages (Keller & Lehmann, 2008) designed to increase donation intent. Nonprofit marketing therefore resembles marketing for profits in the need to understand both antecedents to the proclivity to donate to a particular nonprofit and the moderating variables that may strengthen the relationships between these antecedents and donation intent. The marketing literature primarily treats religiosity as either a categorical demographic variable or on rare occasion as a direct effect on a dependent variable. Yet, in the non-marketing literature both mediating (Momtaz, Hamid, Ibrahim, Yahaya, & Chai, 2011) and moderating (Momtaz, Ibrahim, Hamid, & Yahaya, 2010) effects of religiosity on other variables have been tested and shown to exist. For instance, religiosity mediates the psychological well-being of widowed elderly people (Momtaz et al., 2011), so a higher degree of religiosity lessens the negative effects of widowhood, and religiosity moderates the relationship between social isolation and psychological well-being (Momtaz et al., 2010). Were religiosity found to play a similar moderating role on

☆ The authors are grateful for comments and insights by Arch Woodside, Boston College, and Brian Engelland, The Catholic University of America. ⁎ Corresponding author at: Department of Marketing, College of Business and Behavioral Science, 257 Sirrine Hall, Box 341325, Clemson, SC 29634-1325, USA. E-mail addresses: [email protected] (C.D. Hopkins), [email protected] (K.J. Shanahan), [email protected] (M.A. Raymond). 0148-2963/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jbusres.2013.03.008

nonprofit advertising, this would be an important finding for both academic research and practitioners. 2. The consequences of not knowing a religion's role on consumers' reactions to advertising Because of a strong relationship between religious persons and a greater concern for maintaining high moral standards (Wiebe & Fleck, 1980), marketers must understand the role religion plays in consumers' reactions to advertising. With advertising often pushing the moral envelope to jockey for the public's share of mind, marketers run the risk of offending stakeholders, particularly those with high moral standards. Further, the public may expect non-profit firms to hold to a higher moral standard in advertising than they expect profit-oriented firms. Hirschman (1983) finds that religious affiliation (e.g., Catholic, Protestant, and Jewish) shapes attitudes toward products and services. However, marketing messages impact more than just Christians and Jews. Muslims also respond strongly to advertising based on their religious tenets. For example, studies by Luqmani, Yavas, and Quraeshi (1987) and Michell and Al-Mossawi (1999) find that a failure to understand Islam in relation to advertising strategies results in an alienation of a wide segment of the conservative Saudi public. Differences in how devout versus lenient Muslims respond to perceived controversial elements in advertisements suggests a need to match creative execution and message content to society's socio-cultural environment (Peebles & Ryans, 1984), in particular where religion plays an important role for target segments. Failure to do so alienates the public and generates a negative attitude toward the advertisement and brand recall (Gardner, 1985; Zinkhan & Martin, 1982).

24

C.D. Hopkins et al. / Journal of Business Research 67 (2014) 23–31

The more religious a person, the more likely that person will be offended by the advertising of controversial products (Fam, Waller, & Erdogan, 2004). Ardent religious people tend to be more conservative, have a greater concern for moral standards and possess more traditional attitudes (Barton & Vaughan, 1976; Wiebe & Fleck, 1980; Wilkes, Burnett, & Howell, 1986). Products such as cigarettes, alcohol, contraceptives and underwear (Barnes & Dotson, 1990; Rehman & Brooks, 1987; Shao, 1993) represent just some of the controversial products that can offend very religious people. Wilson and West (1981) describe categories of controversial products and services including products (birth control, drugs for assisted suicide), services (abortion, artificial insemination), and concepts (political ideas and palliative care). The list provided by Wilson and West (1981) demonstrates a potential problem for nonprofits—in fact any entity targeting the general population some of their products/service/concepts will offend certain segments of the population based upon religious tenets. Therefore, marketers need to develop an understanding of consumers' potential reaction to advertising in order to mitigate the potentially harmful consequences of offending the public. 3. Framework and variables in the study Prior research identifies a number of variables that act as direct or indirect antecedents to increase intent to donate to a nonprofit sponsor of an ad intended to affect pro-social change. These variables include negative emotional response to the ad, social responsibility of the sponsor, and attitude toward the ad (Shanahan & Hopkins, 2007). This study operationalizes affective response to the nonprofit ad as negative emotional response, an approach that is both intuitive and supported by the literature. A perusal of much of the advertising used by nonprofits to promote pro-social change indicates that stimulating an emotional response from the public is a consistent priority. From the haunting music of Sarah McClachlin coupled with tortured animals, to images of starving children needing just pennies a day to survive, nonprofits employ images to generate sadness, compassion and empathy; all identified as negative emotional appeals. Shaver, Schwartz, Kirson, and O'Connor (1987) suggest that people exposed to emotion laden advertising organize their emotions into two categories, both positive and negative, where negative emotions include anger, sadness, and fear (Bagozzi & Moore, 1994). The negative state relief model (Baumann, Cialdini, & Kenrick, 1981) indicates that negative emotions drive individuals to reduce their negative feeling by supporting the advertised cause (i.e., donations). A strong link between attitude toward the ad and nonprofit donations (Brown & Stayman, 1992; MacKenzie, Lutz, & Belch, 1987) suggests but does not test a direct relationship between overall attitude toward the ad and perceived social responsibility of the sponsor. However, overall attitude toward the ad appears to directly influence increased intention to donate to the nonprofit and indirectly influence through the perceived social responsibility of the sponsor. Many nonprofits promote pro-social activities to affect positive societal change for the betterment of the public. Why? Because, acting in a socially responsible manner appears to have a positive effect on nonprofit donations (Brown & Stayman, 1992; MacKenzie et al., 1987) supporting a solid link between social responsibility and intent to donate. 4. Religiosity This study focuses on one aspect of religious beliefs – theism or the belief in an omnipotent God – a facet of the Religious Practices and Attitudes Inventory (RPAI) developed by D'Onofrio, Eaves, Murrelle, Maes, and Spilka (1999), D'Onofrio, Murelle, et al. (1999). The RPAI views religiosity on a continuum from atheist/agnostic to devotional and is strongly tied to another well studied continuum—Hofstede's (1988) individualism/collectivism scale. Hofstede (1988) defines

individualism as a focus on rights over duties, a concern for oneself, personal autonomy, and self-fulfillment. Conversely, deeply religious people focus on community, a concern for others beyond themselves, and duty to God; often rejecting self-gratification in favor of self denial and abstinence. Oyserman, Coon, and Kemmelmeier (2002) suggest that individualism values personal goals and personal control with social goals playing a lesser role (Triandis, 1995). A century ago Durkheim posited that religion and society support one another. Religion offers society a public normative set of acceptable behaviors (Durkheim, Cosman, & Cladis, 2001). Objects of worship exist for the mutual consumption by members of a collective or group. Religion offers the collective a system of ideas and symbols around which and through which members of a collective group interact. Thus, most religion is by definition collectivist (with the exception of some Eastern religions such as Buddhism). The core element of collectivism binds and mutually obligates individuals as groups (Oyserman et al., 2002). In-groups, not individuals, are the basic unit of measure and include religious groups (Hui, 1988). Triandis (1995) further adds that collectivism may refer to a broad range of values and behaviors, well beyond that of the individual. Oyserman (1993) offers that the goals of collectivism include sacrifice for the common good and that life satisfaction derives from carrying out social roles and obligations. These descriptions suggest that more religious individuals tend to perceive that they have a duty to others, are willing to sacrifice to help the greater good (tithing, donating to charity), and will look favorably on the good works of others to mitigate what can be seen as sinful behaviors (smoking, drinking, drug use etc.). Consequently, upon exposure to nonprofit pro-social change-based ads (those depicting the results of drunk driving or smoking), high religiosity groups will exhibit a stronger emotional reaction, maintain a stronger view of social responsibility in the sponsor who is trying to nudge society in a manner consistent with their beliefs, and an increased proclivity to donate to those nonprofits. H1. The high religiosity group has a stronger negative emotional response to nonprofit pro-social change ads than does the low religiosity group. H2. The high religiosity group has a more positive attitude toward the ad than does the low religiosity group. H3. The high religiosity group perceives greater social responsibility of the sponsor than does the low religiosity group. H4. The high religiosity group has greater intent to donate than the low religiosity group. In addition to a direct effect of religiosity on the variables in the study, religiosity will moderate the relationships between the variables. For high religious groups, messages promoting smoking cessation or anti-drunk driving messages resonate strongly based on group beliefs. For low religious groups that tend to be more individualistic, such messages may be seen as an intrusion of their rights: they can take care of themselves without others' interference. Because the basic traits of the high religious group are about social obligation, and for some evangelical message delivery, the nonprofit message seems familiar and welcome. As such, the link between an ad that is liked, a sponsor that is seen as doing their mission, and donations to that nonprofit will be stronger than for low religious groups. H5. The high religiosity group has a stronger relationship between negative emotional response and perceived social responsibility of the sponsor than the low religiosity group. H6. The high religiosity group has a stronger relationship between negative emotional response and overall attitude toward the ad than does the low religiosity group.

C.D. Hopkins et al. / Journal of Business Research 67 (2014) 23–31

H7. The high religiosity group has a stronger relationship between negative emotional response and increased intent to donate than does the low religiosity group. H8. The high religiosity group has a stronger relationship between overall attitude toward the ad and perceived social responsibility of the sponsor than does the low religiosity group. H9. The high religiosity group has a stronger relationship between overall attitude toward the ad and increased intent to donate than does the low religiosity group. H10. The high religiosity group has a stronger relationship between perceived social responsibility of the sponsor and increased intent to donate than does the low religiosity group.

5. Research methodology Governments in many developed countries increasingly support efforts to reduce smoking and the negative effects of alcohol consumption (c.f. Piacentini & Banister, 2009; Shiu, Hassan, & Walsh, 2009). Thus, ad material dealing with the effects of these activities appears relevant and consistent with the purposes of this study. A review of the extant literature indicates print media successfully delivering anti-smoking and anti-drunk driving messages for nonprofits (Shanahan & Hopkins, 2007; Shanahan, Hopkins, & Carlson, 2008; Shanahan, Hopkins, & Carlson, 2010). Therefore, the use of anti-smoking/drunk driving print advertising seems both meaningful and appropriate. This study utilized two separate ads to eliminate potential extraneous variation. Two nonprofit organizations serve as sponsors, the American Cancer Society (ACS) and Mothers Against Drunk Driving (MADD), based on the assumption that respondents would possess some degree of knowledge and awareness about each organization. Shanahan, Hopkins, & Carlson (2008) employ an anti-smoking ad featuring Barb Tarbox; a real-life anti-smoking crusader. As an actual, highly successful and out of market ad, the Tarbox ad presents itself as a perfect candidate for this study. The Tarbox ad generates a strong

25

emotional response that resonates across age groups. Because the Tarbox ad is Canadian and only appeared in the province of Alberta in local magazines and posters, unless respondents traveled to Alberta and specifically read one of those magazines, the ad would not likely be familiar to respondents. The second ad features a young female, killed by a drunk driver. The ad draws upon prototypical MADD type anti-drunk driving messages. Fig. 2 presents a copy of each.

6. Sample and instrument As a first step in attempting to understand one's level of conviction to a religion, this study adopts the definition forwarded by Dube and Wingfield (2008, p. 503) who define religiosity as “a person's degree of adherence to the beliefs, doctrines, and practices of a particular religion”. Consistent with this definition, this study taps into each respondent's personal or latent beliefs and commitment relative to their own level of spirituality, regardless of the particular religion, faith or doctrine. With this working definition of religiosity in mind, surveys and data were collected via a stratified probability sample drawn from an online consumer panel. Utilization of the consumer panel affords the opportunity to apply constraints for the purpose of tapping a sample that consists of both religious and non-religious individuals. In this case, the first question on the survey serves as a filter which asks individuals “Regardless of your chosen faith, do you consider yourself a religious person?” The answer to this question (yes/no), places individuals within each treatment group. Two separate survey instruments (each identical with the exception of the advertisement), are created and distributed randomly to panel members via e-mail solicitation. With a total of 306 usable responses, 161 high in religiosity (replying “yes” to the filter question), and 145 low in religiosity (replying “no” to the filter question), the overall sample consists of 50.3% female, 78.8% Caucasian, 8.8% African American, 5.6% Hispanic, 5.6% Asian and the remaining 1.6% of other ethnicities. Of the total respondents, 73% attended college, 34.3% reports an income greater than $50,000, and 48% are older than 47 years of age with an average age of 45. Overall,

MODEL OF MODERATING EFFECTS

Moderating Effect High Religiosity vs. Low Religiosity

Overall Attitude Toward the Ad

Perceived Social Responsibility of the Sponsor

Negative Emotional Response to the Ad

Intent to Donate

Fig. 1. Model of moderating effects.

26

C.D. Hopkins et al. / Journal of Business Research 67 (2014) 23–31

ADVERTISING TREATMENTS

Fig. 2. Advertising treatments.

the sample represents 40 different states with the top five represented being California (12.78%), New York (6.39%), Ohio (5.11%), Texas (4.47%) and Pennsylvania (4.15%). Relative to religiosity, 85% of the respondents are Christian, 4.5% Jewish, 4.1% Hindu, 2.9% Muslim, 1.3% Buddhists and the remaining 2.2% of other faiths. Of the Christians, 24% are Baptists and 23% are Catholic/Roman Catholic; other denominations include Methodist, Presbyterian, Lutheran, Church of Christ, and Pentecostal. The measures incorporated into the survey instrument assess predispositions to respond in a consistently favorable/unfavorable manner toward an object and employ scales with endpoints such as

good/bad, like/dislike as specified by Jarvis, Mackenzie, and Podsokoff (2003), p. 200. The questionnaire (following the filter question) consists of four, multi-item measures for overall attitude toward the ad (Kamp & MacInnis, 1995); negative emotional response to the ad (Moore & Harris, 1986); perceived social responsibility of the sponsor (Dean, 1999); and willingness to donate to the sponsor (Dodds, Monroe, & Grewal, 1991). See Table 1 for a list of items. These scales consist of multi-item measures with established reliability and validity. The survey contains additional items to assess manipulation and confound influences. Additional scales assess the demographics of sample members.

Table 1 Confirmatory model results. Construct/item

Comp. st. estimatea

t-Value

Negative emotional response to the ad, (α = .91; variance extracted = .82) Worried Frightened Alarmed Angry

.83 .96 .97 .51

– 26.34 26.73 10.80

Overall attitude toward the ad (α = .95; variance extracted = .93) Disliked it very much/liked it very much Not at all positive/very positive Did not like it at all/liked it very much

.96 .85 .97

– 28.53 48.40

Perceived social responsibility of the sponsor (α = .95; variance extracted = .88) (The American Cancer Society or MADD) cares about bettering society (The American Cancer Society or MADD) is socially responsible (The American Cancer Society or MADD) plays a necessary role in society (The American Cancer Society or MADD) contributes to society (The American Cancer Society or MADD) will keep its promises and commitments.

.87 .87 .86 .91 .89

– 23.89 23.46 26.25 25.23

Willingness to donate (α = .94; variance extracted = .92) Not at all likely/very likely Much less probable/much more probable Likely to be greatly reduced/likely to be greatly increased

.89 .96 .90

– 31.73 27.18

a b c

χ2 = 158.58, df = 84, p ≤ .000. GFI = .95, AGFI = .93, RMSEA = .047, St. RMR = .038, IFI = .98, CFI = .98, RFI = .97. All p ≤ .01 for individual items; the first item for each measure was set to a value of 1.00. Significant parameter t-values in bold.

b, c

C.D. Hopkins et al. / Journal of Business Research 67 (2014) 23–31

7. Results 7.1. Measure assessment A structural model assessment of the proposed confirmatory model with LISREL 8.51 demonstrates adequate fit with a χ 2 = 158.58, df = 84, p ≤ .000. GFI = .95, AGFI = .93, RMSEA = .047, St. RMR = .038, IFI = .98, CFI = .98 and RFI = .97. Employing Harman's single-factor test (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003), the χ 2 difference test with one degree of freedom results in Δχ2 ≥ 3.84; implying that common method variance is not a factor. Each scale proves both reliable and valid. In Table 1, the coefficient alpha values range from .95 to .91 while the variance extracted scores range from .93 to .82. This study assesses convergent validity via the t-value associated with the ratio of factor loadings to standard error for each item. Generally, a t-value of greater than 2.00 (Segars, 1996), lends support for convergent validity (see Table 1). Further, variance extracted scores greater than .50 support a case for convergent validity (Fornell & Larker, 1991). As noted, all measures meet these requirements. Tests for discriminant validity are consistent with guidelines forwarded by Bagozzi (1980), Gerbing and Anderson (1988) and Fornell and Larker (1991).

27

assess attitude toward the sponsor (α = .82) and the victim (α = .88) depicted within each ad. Both of these scales employ 7-point items including bad/good, unfavorable/favorable and negative/positive. Two t-tests incorporating the sponsor's ads as the independent and the scale score for each of the attitude measures as dependents provide evidence that there are no confounds. The test of the sponsor confound yields x = 16.62 for the MADD ad treatment and x = 15.92 for the ACS ad treatment with p = .20, t = 1.26. Testing for the victim confound results in x = 12.46 for the MADD ad treatment and x = 13.51 for the ACS ad treatment with p = .077, t = 1.64. Results of these tests suggest that that no degradation to internal validity results from preexisting perceptions held about the sponsors or victims. 8. Hypothesis testing—H1–H4 Hypotheses 1–4 are tested via multivariate analysis of variance (MANOVA) tests, with religiosity serving as the single, categorical independent variable and the scale score for each construct presented in Fig. 1 serving as the dependent variables. Each of the hypotheses posits that a significant difference exists in each criterion variable relative to religiosity, with the mean score being greater for the high religious group. Noting the results in Table 2, the overall model is significant with all of the four posited relationships receiving support.

7.2. Manipulation and confound checks 8.1. Hypothesis testing—H5–H10 The findings support the primary manipulation (high versus low religiosity). To measure religiosity, this study utilizes a 4-item, 7-point Likert (strongly disagree/strongly agree) measure from D'Onofrio et al. (1999); deemed appropriate for studies assessing the influence of religiosity on adolescent substance abuse (D'Onofrio, Eaves, et al., 1999; D'Onofrio, Murelle, et al., 1999; Eaves, Hatemi, Prom-Womley, & Murrelle, 2008), religious affiliation (D'Onofrio, Eaves, et al., 1999) and religiosity's influence on heroin addiction (Isralowitz, 2002). The four items include “I feel I can always count on God,” “I feel that without God, there would be no purpose in my life,” “My faith in God helps me through hard times,” and “I try to live how God wants me to live.” The scale score for this measure (α = .98) serves as the dependent variable and high/low religiosity serves as the independent variable in an independent sample t-test. The results yield x = 29.50 for the high group and x = 14.23 for the low group with p ≤ .000, t = 18.126. Additional tests, run separately for each ad treatment (MADD/ACS), find that for the MADD group, the test with religiosity groups (high/low) as the independent and religiosity scale score as the dependent variable is significant (p ≤ .000, t = 13.168), with x = 30.15 for the highly religious group and x = 13.65 for the low religious group. These same patterns exist for the ACS group with p ≤ .000, t = 12.533 for the overall model and with x = 28.89 and 14.75 for the high/low religious groups, respectively. The religiosity manipulation is supported for the entire sample as well as for each ad treatment taken separately. Confound checks between the ACS and MADD advertisement groups ensure that no threats to internal validity are present based on perceptions of the sponsors. Two 3-item semantic differential measures

To test H5–H10, it is necessary to perform multi-sample structural analysis, which affords the ability to assess the significance of possible differences in causal influence across differing moderating conditions (Bollen, 1986; Joreskog & Sorbom, 1996). In this case, the high/low religious groups, as defined during the data collection phase and supported via the manipulation checks, serve as the dichotomous moderator. This method of moderation assessment is deemed appropriate given that the testing of moderation effects via nonmetric measurement is regarded as a useful approach that simplifies interpretation (Hair et al., 2010; pg. 756). The method requires the testing of each possible pair of models (constrained and unconstrained) and then assessing differences in structural parameters via χ2 difference tests with one degree of freedom. This procedure requires first running a model with the paths under consideration being freely estimated and then running a model with the hypothesized paths constrained as equal. A significant difference between the two models (Δχ2 ≥ 3.84) indicates that a differential influence exists, attributable to the moderation effect. Thus, comparison of parameter coefficients delineates under which level of the moderator the causal effect is stronger. Table 3 indicates that four of the six hypotheses receive support. The influence of negative emotional response on perceived social responsibility is significantly different (Δχ2 ≥ .3.84; p-value ≤ .05) and greater for the high religiosity group, thus supporting H5. This same pattern exists for overall ad attitude/perceived social responsibility, overall ad attitude/intent to donate and perceived social responsibility/ intent to donate parameter estimates, thus lending support to H8, H9

Table 2 Univariate and multivariate results. Main effects

Treatments High religiosity Low religiosity

Dependent variables

a

Negative emotional response to the ad

Overall attitude toward the ad

Perceived social responsibility of the sponsor

Willingness to donate

x

x

x

x

F-ratio

p-Value

6.650

.010

21.43 18.76

Significant parameter p-values in bold. a Wilks' Λ = .889 (f(4) = 8.434; p = .000) power = .999.

15.73 13.80

F-ratio

p-Value

11.29

.001 30.14 27.15

F-ratio

p-Value

17.43

.000 14.50 11.46

F-ratio

p-Value

32.43

.000

28

C.D. Hopkins et al. / Journal of Business Research 67 (2014) 23–31

Table 3 Moderated relationships. Model/ structural pathb

χ2

Negative emotional response → perceived social responsibility Negative emotional response → overall ad attitude Negative emotional response → intent to donate Overall ad attitude → perceived social responsibility Overall attitude toward the ad → intent to donate Perceived social responsibility → intent to donate

559.17

−5.22

.14

564.18

−0.21

563.44

a b

Δχ2a

High religiosity Coefficient

Low religiosity tValue

Hi

Coefficient

t-Value

2.23

.08

1.66

H5

.24

4.53

.27

3.86

H6

−0.95

.28

5.46

.23

5.07

H7

544.85

−19.54

.49

11.42

.28

8.70

H8

560.15

−4.24

.35

7.19

.25

4.59

H9

560.67

−4.12

.57

7.49

.47

5.10

H10

Significant Δχ2 and p-values in bold. Δχ2 = 564.39df = 204 (χ2 value for constrained model—χ2 value for unconstrained model); Δχ2 = ±3.84df = 1 is critical value for α = .05.

and H10 respectively. The Δχ2values relating to H6 and H7 are not significant. In summary, of the ten posited hypotheses, eight are supported. These findings imply that the level of religiosity held by respondents does impact perceptions and behaviors. This impact further appears to moderate the influence that given phenomena have on held intentions. 9. Discussion This study investigates the moderating effect of religiosity on antecedents influencing intention to donate to a nonprofit organization and finds strong supporting evidence that religiosity does moderate the relationships. Of the six moderating effects in the study, five are significant. Religiosity moderates the link between a negative emotional response and intent. Intuitively those scoring higher on the religiosity scale become involved in pro-social activities to affect positive societal change through their churches and other faith based activities more often than those do scoring lower on the religiosity scale. The ad sponsors (ACS and MADD) advocate pro-social change leading to support for them by the high religiosity group. The same type of moderating effect of religiosity appears on overall attitude toward the ad and perceived social responsibility of the sponsor. The moderating effect of religiosity on perceived social responsibility and intent to donate has two facets. First, a congruency occurs in the belief that pro-social change is good and the role of both the nonprofit and high religiosity group. Second, the higher religiosity group already has a history of giving back to society in the form of money (tithing) and time (i.e., mission work, faith based community work, etc.). Combined, it is not surprising to see a religiosity moderating effect on the social responsibility and intent to donate relationship. However, religiosity does not moderate all relationships. Religiosity does not moderate the links between negative emotional response and perceived social responsibility of the sponsor. Both ads have a strong emotional appeal rather than a more typical financial penalty (i.e., click it or ticket for seatbelts; drive drunk go to jail). In the anti-smoking ad Barb Tarbox cannot quit smoking and is going to die. She soon will leave her young daughter and husband behind; celebrating their last Christmas together. In such a case, a human response to the suffering of others leads to a sense of sadness and sympathy. Similarly, in the anti-drunk driving ad, a drunk driver kills a young woman; a young woman with her whole life and future ahead of her. Sadness and sympathy for this girl likely occur regardless of the religious nature of the respondent. The ad moves both groups and has a strong liking of the ad and increased proclivity to donate. The question then is why there exists a moderating effect for social responsibility?

The advertising likely captures the pro-social nature of the high religiosity group. Thus, while both high and low religiosity groups react strongly to the ads, the attempt to change behavior, being congruent with the high religiosity group's belief system, moderates social responsibility but not overall attitude and intent to donate. In addition to the moderating effect of religiosity, all four main effects are significant. That is, for all four variables (negative emotional response to the ad, overall attitude toward the ad, perceived social responsibility of the sponsor and intent to donate) a significant difference exists between the high and low religiosity groups. Clearly, an important factor in nonprofit advertising includes an understanding of the level of religiosity of a nonprofit's public support base, particularly when the sponsor also seeks donations. 10. Religiosity and corporate social responsibility In a study of 152 self-described Christian companies, Ibrahim, McDougall, and Greene (1991) find employee-centered values and behaviors including a golden rule whereby the firms embrace integrity and kindness. This parallels Brammer, Williams, and Zinkin (2007) who cite religious individuals as holding broader conceptions of the social responsibility of businesses than non-religious types. The more deeply held religious inclinations are, the greater an orientation to corporate social responsibility. Sheffield (2006) offers that socially responsible companies are more responsive to community projects, to solving social problems and to applying very high ethical standards in business activities such as advertising. In other words religiosity, whether at the corporate or individual level, is an important moderator variable. CSR advertising is cause-related marketing (Bronn & Vrioni, 2001). In fact, Angelidis and Ibrahim (1993) describe CSR as social actions whose purpose is to satisfy social needs. Angelidis and Ibrahim (2004) find a significant relationship between degree of religiousness and attitude toward economic and ethical components of CSR. When marketing socially responsible products (e.g., green energy or biodegradable products) the religiosity construct seemingly is important. Given that highly religious people tend to be collectivist and want to help others, it is important to demonstrate how using socially responsible products is the socially responsible thing to do. The first step is to identify segments of the population that would score high on the religiosity scale. One approach is to find activities, interests and exhibited behaviors of individuals that have been shown to correlate with religious beliefs and tendencies. For example, looking at values and lifestyles (i.e., VALS) categories, several indicate the importance of worship, religion and responsibility as key indicators of those that fall within these types. So in addition to demographic segmentation based on religion whereby individuals can be identified by church membership

C.D. Hopkins et al. / Journal of Business Research 67 (2014) 23–31

or affiliation, psychographic segmentation methods may be employed as well to find the very publics nonprofits seek to identify. 11. A Roadmap for Nonprofits Employing Negative Emotional Appeals Ranganathan and Henley (2008) find that attitude toward helping others (altruism) by itself does not lead to donation behavior. They suggest altruistic people need to be targeted with a call to action via advertising and they report that religiosity highly correlates with donation behavior. Several studies (Shanahan & Hopkins, 2007; Shanahan, Hopkins, & Carlson, 2008) indicate negative emotional appeals effectively delivers nonprofits' messages to their publics. A roadmap to success would provide a useful tool for nonprofits seeking pro-social change whether it includes appeals targeting antismoking campaigns, AIDS awareness or even save the whales. Based upon our the literature and this study's results, such a roadmap would contain several key elements of success. See Table 4. First, nonprofits benefit by using real victims (Shanahan & Hopkins, 2007) since viewers hold actor portrayals in the same regard as deceptive ads (Shanahan & Hopkins, 2007; Shanahan, Hopkins, & Carlson, 2008). Not only does the viewing public regard the use of undisclosed actor portrayals as deceptive, but the Federal Trade Commission defines such ads as deceptive and a violation of FTC regulations (Shanahan, Hopkins, & Carlson, 2010). Protection Motivation Theory suggests a real victim generates a higher level of empathy from the viewer (Shanahan, Hopkins, & Carlson, 2008) and empathy for the victim is a key antecedent for donations to charities and nonprofits (Hung & Wyer, 2009). Empathy for the victim functions as a strong predictor of donations to the nonprofit (Verhaert & Van den Poel, 2011). A real victim generates empathy, through protection motivation, and leads to a stronger negative emotional response (i.e., sadness, concern), more positive perception of social responsibility of the nonprofit and increased donations. This is true for the general population, which includes both high and low religiosity individuals. However, this study demonstrates that approaching high religiosity individuals is the more effective target market for nonprofits. This study demonstrates that religiosity plays a moderating role in these relationships, increasing further the strength of links between ad message and emotional response, perceived social responsibility and donation proclivity. An anti-drunk driving ad featuring a young girl killed by a drunk driver generates the desired response (Shanahan, Hopkins, & Carlson, 2010). However, this study shows that when targeting the message to people high in religiosity, there is a magnified effect. Further, the relationships between the antecedent variables and

29

donation intent are stronger for those individuals categorized as highly religious. Recall that the religiosity items asked about being able to count on God, without God life would be have no purpose, God helps coping in hard times and religious people try to live how God wants them to live. To score high on the religiosity scale, one does not need to attend church. Thus, in addition to reaching out to church membership, additional methods to identify high religiosity individuals would be helpful to nonprofits to reach other non-church going religious types. One potential approach might employ psychographic segmentation based on lifestyle and values. Such segmentation methods aid in media placement of advertising. VALS (values and lifestyles) divides consumers into eight types. Of these eight types of consumers, two are conservative, traditional with established codes of conduct based on religion, community and social organizations, who organize their social lives around their families and place of worship. These two groups alone represent 30% of the US population over 18 years of age or roughly 71 million individuals. Sargeant and Harscook (2007) find that shared beliefs between the nonprofit mission and donors leads to repeat donations. That is personal beliefs and values which match the nonprofit leads to increased financial and volunteer support. Achievers and believers, the two VALS groups identified as conservative and focused on family and place of worship differ in resources available to assist nonprofits in pro-social change. The first group, achievers, is resources rich and leisure activities and satisfaction often centers on their homes, families and place of worship. Achievers are societally conscious intellectuals and professionals. They are politically conservative leaders in activist publics and see the need for social change (Dozier & Repper, 1992). This combination of wanting to help society improve and having the resources to contribute makes this VALS type ideal as a target market for nonprofits. Politically conservative with strong ties to their place of worship, achievers offer nonprofits a large segment – approximately 20% of the US population – which likely scores high on the religiosity scale. The second group, believers, has far less financial resources than does the achievers type, however believers represents the best VALS type for word of mouth advertising. Advertising can be tailored to reach this group by first identifying opinion leaders and then employing word of mouth to spread the nonprofit's message. Their belief in ideals makes them excellent candidates to donate time to causes; particularly ones that promote pro-social change and with which believers are already familiar. Familiarity is a hallmark of believers who favor established brands. One would expect this tendency toward established brands may manifest itself in the selection of nonprofits to support.

Table 4 Nonprofit donation studies related to intent to donate to sponsor. Authors

Findings

Shanahan and Hopkins (2007)

Use of real victims—most effective with highest proclivity to donate. Use of deceptive advertising—not significantly different from using actor portrayal. Repeat charity giving driven by service quality, shared beliefs, risk, personal link to charity and trust. Truth in advertising fails when using an actor portrayal for anti-smoking ads, particularly when the perception is that real victims could be used. Ironically, many qualitative responses to a “real” victim used in out of market ads were that she did not look like a real cancer victim. Recommendations support Shanahan & Hopkins (2007) in that real victims provide the strongest relationship between antecedents to donate and intent to donate to ACS. When viewers took the victim's perspective this had a positive effect on the amount of money they donated. However, when an a priori disposition to take the perspective of a potential donor at the time they read the appeal, these same characteristics decreased the appeal's effectiveness. The conflict in the image as a victim and potential donor, simultaneously, can decrease the ad's effectiveness. Viewers experience a stronger emotional connection with real victims, along with greater perceptions of sponsor's social responsibility leading to greater intent to donate. Undisclosed actor portrayals—a technical violation of FTC regulations. Implications are that self-regulation may help mitigate negative reactions to PSA messages. Find that empathy, in the form of concern and personal distress, provides value in predicting charitable donations. Empathic concern positively affects the donation decision. However, empathy negatively affects donation to a single charity as highly empathic individuals tend to donate to multiple charities. Religiosity has a strong main and moderating effect on negative emotional appeal, perceived social responsibility of the nonprofit and donation proclivity.

Sargeant and Harscook (2007) Shanahan, Hopkins and Carlson (2008)

Hung and Wyer (2009)

Shanahan, Hopkins and Carlson (2010)

Verhaert and Van den Poel (2011)

Current manuscript

30

C.D. Hopkins et al. / Journal of Business Research 67 (2014) 23–31

For both VALS types, the key is identifying media selection for these groups as past research has indicated these two types have predictable patterns of behavior and consumption—ostensibly in consumption of nonprofit support as well. Correlation between the goals of nonprofits and the media consumption of target markets can be established and then media placement based on readership traits is made simpler. For example, to reach a more family oriented and conservative target, which would include both achievers and believers, publications such as Southern Living or television programming featuring Paul Dean offer an excellent option. An understanding of the media consumption patterns of VALS types aids in ad placement that reaches those high on religiosity. In offering guidance and recommendations for a desired course of action to non-profit organizations, the highly religious segment of the population appears to be a quite viable group to target. An analysis of more than 350,000 interviews conducted by Gallup in 2008 finds Mississippi, Alabama, South Carolina, Tennessee, Louisiana, and Arkansas to be the most religious states in the nation. Vermont, New Hampshire, Maine, and Massachusetts are the least religious states. For this ranking, Gallup uses the responses to a straightforward question that asks: “Is religion an important part of your daily life?” The percentage of each state's adult (18 and older) population that answers in the affirmative provides the basis for rankings, and a starting point for the process of segmenting the market within the United States. Non-profits can reach the religious segment by focusing efforts within the traditional “Bible-belt” region of the United States. Although states such as Alabama, Arkansas, and Tennessee have relatively lower average disposable income levels, residents there contribute substantially more to charity than their wealthier state counterparts (Sauter, Stockdale, & Allen, 2011).

12. Limitations and future research Because nonprofits need to compete with other nonprofits for donations while maintaining a socially responsible reputation, developing ads that are emotional and likable proves to be a useful tactic that encourages donations. If a variable moderates the relationship between the desired outcome and antecedents to that outcome, this should be an important variable to include in any research study or assessment on that outcome. This study provides evidence that religiosity plays a significant moderating role in how consumers react to nonprofit advertising. Religiosity is an excellent construct to use as a segmentation variable for identifying markets for socially responsible products. Further there is a main effect of religiosity for all variables in this study. As such, religiosity is a key variable in nonprofit advertising, particularly for the assessment of social responsibility of the sponsor, intent to donate and negative emotional response to the ad. This study was solely about nonprofit advertising. It would be insightful to see if these patterns replicate using for profit ads and/or sponsors. Perhaps the reaction would be even stronger if a for-profit sponsored anti-drunk driving such as the “know your limit” ads run by the liquor industry a decade ago. The moderating effect may disappear altogether were both groups to feel a for-profit ad was self-serving or mandated by legal rather than for altruistic reasons. Certainly, this limitation provides numerous research opportunities for future studies. Additionally, this study focuses on religiosity based on the theism subscale, which measures the strength of relationship of the individual with God. Other subscales measure sense of community, place in God's plans etc. Each of these may also play a moderating role and help to further segment the population to assist nonprofits to deliver their messages in a more efficient manner. Future studies should include the testing of religiosity in a religion specific (i.e., Catholic versus Baptist) context so as to bolster understanding of this moderating effect particularly if there are strictures against drinking in one denomination and inclusion of alcohol in ceremonies for other denominations.

References Angelidis, J., & Ibrahim, N. (1993). Social demand and corporate strategy: A corporate social responsibility model. Review of Business, 15(1), 7–10. Angelidis, J., & Ibrahim, N. (2004). An exploratory study of the impact of degree of religiousness upon an individual's corporate social responsiveness orientation. Journal of Business Ethics, 51, 119–128. Bagozzi, R. P. (1980). Causal models in marketing. New York: Wiley. Bagozzi, R. P., & Moore, J. D. (1994). Public service advertisements: Emotions and empathy guide prosocial behavior. Journal of Marketing, 58, 56–70. Barnes, J. H., Jr., & Dotson, M. J. (1990). An exploratory investigation into the nature of offensive television advertising. Journal of Advertising, 19(3), 61–69. Barton, K., & Vaughan, G. (1976). Church membership and personality. Social Behavior & Personality, 4, 11–16. Baumann, D. J., Cialdini, R. B., & Kenrick, D. T. (1981). Altruism as hedonism: Helping and self-gratification as equivalent responses. Journal of Personality and Social Psychology, 40, 1039–1046. Bollen, K. A. (1986). Structural equations with latent variables. New York: John Wiley & Sons. Boris, E. T. (1999). Nonprofit organizations in a democracy: Varied roles and responsibilities. In Elizabeth T. Boris, & Eugene Steuerle (Eds.), Nonprofit and Government, 1-12, Washington, DC: Urban Institute Press. Brammer, S., Williams, G., & Zinkin, J. (2007). Religion and attitudes to corporate social responsibility in a large cross country sample. Journal of Business Ethics, 71, 229–243. Bronn, P. S., & Vrioni, A. B. (2001). Corporate social responsibility and cause related marketing: An overview international. Journal of Advertising, 20, 207–222. Brown, S. P., & Stayman, D. M. (1992). Antecedents and consequences of attitude towards the ad: A meta-analysis. Journal of Consumer Research, 19, 34–51. D'Onofrio, B. M., Eaves, L. J., Murrelle, L., Maes, H. H., & Spilka, B. (1999). Understanding biological and social influences on religious affiliation, attitudes and behaviors: A behavior-genetic perspective. Journal of Personality, 67, 953–984. D'Onofrio, B., Murrelle, L., Eaves, L. J., McCullough, M. E., Landis, J. L., & Maes, H. (1999). Adolescent religiousness and its influence on substance use: Preliminary findings from the mid-atlantic school age twin study. Twin Research, 2, 156–168. Dean, D. H. (1999). Brand endorsement, popularity and event sponsorship as advertising cues affecting consumer pre-purchase attitudes. Journal of Advertising, 28(3), 1–12. Dodds, W. B., Monroe, K. B., & Grewal, D. (1991). The effects of price brand and store information on buyers' product evaluations. Journal of Marketing Research, 28, 307–319. Dozier, D., & Repper, F. (1992). In James E. Grunig (Ed.), Research firms and public relations practices, in excellence in public relations and communication management (pp. 185–219). Dube, L. F., & Wingfield, S. S. (2008). Economics, sociology, politics, and religion: Success of marketing students. Atlantic Economic Journal, 38, 503–504. Durkheim, E., Cosman, C., & Cladis, M. S. (2001). The elementary forms of religious life. Oxford University Press. Eaves, L. J., Hatemi, P. K., Prom-Womley, E. C., & Murrelle, L. (2008). Social and genetic influences on adolescent religious attitudes and practices. Social Forces, 86, 1621–1646. Fam, K. S., Waller, D. S., & Erdogan, B. Z. (2004). The influence of religion on attitudes towards the advertising of controversial products. EJM, 38(5/6), 537–555. Fornell, C., & Larker, D. F. (1991). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39–50. Gardner, M. (1985). Mood states and consumer behavior: A critical review. Journal of Consumer Research, 12, 251–267. Gerbing, D. W., & Anderson, J. C. (1988). An updated paradigm for scale development incorporating unidimensionality and its assessment. Journal of Business Research, 25, 186–192. Hirschman, E. (1983). Religious affiliation and consumption processes. Res in Mark. (pp. 131–170): JAI Press, 131–170. Hofstede, G. (1988). Culture's consequence. Beverly Hills, CA: Sage. Hui, C. H. (1988). Measurement of individualism–collectivism. Journal of Research in Personality, 22, 17–36. Hung, I., & Wyer, R. (2009). Differences in perspective and the influence of charitable appeals: When imagining oneself as the victim is not beneficial. Journal of Marine Research, 46(3), 421–434. Ibrahim, N. L., McDougall, R. P., & Greene, G. (1991). Characteristics of Christian based companies. Journal of Business Ethics, 10, 123–132. Isralowitz, R. E. (2002). Religious affiliation of Russian-speaking heroin addicts in Israel. Journal of Social Psychology, 142(6), 791–793. Jarvis, C., Mackenzie, S., & Podsokoff, P. (2003). A critical review of construct indicators and measurement model misspecification in marketing and consumer research. Journal of Consumer Research, 30, 199–218. Joreskog, K. G., & Sorbom, D. (1996). LISREL 8: User's reference guide. Chicago: Scientific Software International. Kamp, Ed, & MacInnis, D. J. (1995). Characteristics of portrayed emotions in commercials: When does what is shown in ads affect viewers? Journal of Advanced Research, 35, 19–28. Keller, P., & Lehmann, D. R. (2008). Designing effective health communications: A meta-analysis. Journal of Public Policy & Marketing, 27(2), 117–130. Luqmani, M., Yavas, U., & Quraeshi, Z. (1987). Advertising in Saudi Arabia: Content and regulation. International Journal of Advertising, 6(1), 58–71. MacKenzie, S. B., Lutz, R. J., & Belch, G. E. (1987). The role of attitude toward the ad as a mediator of advertising effectiveness: A test of competing explanations. Journal of Marketing Research, 2(3), 130–143. Michell, P., & Al-Mossawi, M. (1999). Religious commitment related to message contentiousness. International Journal of Advertising, 18, 427–443. Momtaz, Y., Hamid, T. A., Ibrahim, R., Yahaya, N., & Chai, S. T. (2011). Moderating effect of religiosity on the relationship between social isolation and psychological well-being. Mental Health, Religion and Culture, 14, 141–156.

C.D. Hopkins et al. / Journal of Business Research 67 (2014) 23–31 Momtaz, Y., Ibrahim, R., Hamid, T. A., & Yahaya, N. (2010). Mediating effects of social and personal religiosity on the psychological well-being of widowed elderly people. Journal of Death and Dying, 61, 145–162. Moore, D. J., & Harris, W. D. (1986). Affect intensity and the consumer's attitude toward high impact emotional advertising appeals. Journal of Advertising, 25, 37–50. Hair, J. H., Black, W. C., Babin, B. J., & Andersen, R. E. (2010). Multivariate Data Analysis. Upper Saddle River, NJ: Prentice Hall. Oyserman, D. (1993). The lens of personhood: Viewing the self, others, and conflict in a multicultural society. Journal of Personality and Social Psychology, 65, 993–1009. Oyserman, D., Coon, H. M., & Kemmelmeier, M. (2002). Rethinking individualism and collectivism: Evaluation of theoretical assumptions and meta-analyses. Psychological Bulletin, 128, 3–72. Peebles, D., & Ryans, J. (1984). Management of international advertising. Boston, MA: Allyn and Bacon. Piacentini, M. G., & Banister, E. M. (2009). Managing anti-consumption in an excessive drinking culture. Journal of Business Research, 62, 279–289. Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879–903. Ranganathan, S., & Henley, W. (2008). Determinants of charitable donation intentions: A structural equation model. International Journal of Nonprofit and Voluntary Sector Marketing, 13(1), 1–11. Rehman, S. N., & Brooks, J. R. (1987). Attitudes toward television advertisements for controversial products. Journal of Healthcare Marketing, 7(13), 78–83. Sargeant, A., & Harscook, R. (2007). Building donor loyalty: The antecedents and role of commitment in the context of charity giving. Journal of Nonprofit & Public Sector Marketing, 18(2), 47–68. Sauter, M. B., Stockdale, C. B., & Allen, A. C. (2011, Dec. 15). America's most (and least) charitable states. Wall Street Journal accessed at http://247wallst.com/2011/12/15/ americas-most-and-least-charitable-states/ Segars, A. (1996). Assessing the unidimensionality of measurement: A paradigm and illustration within the context of information systems research. Omega, 25, 107–121.

31

Shanahan, K. J., & Hopkins, C. D. (2007). Truths, half-truths, and deception: Perceived social responsibility and intent to donate for a non-profit using implicature, truth and duplicity in print advertising. Journal of Advertising, 36, 33–48. Shanahan, K. J., Hopkins, C. D., & Carlson, L. (2008). The efficacy of the use of implicature and actor portrayal labels by non-profits in anti-smoking print advertisements. Journal of Current Issues and Research in Advertising, 30(2), 65–78. Shanahan, K. J., Hopkins, C. D., & Carlson, L. (2010). Improving the effectiveness of nonprofit public service announcements through new regulatory requirements. Journal of Public Policy and Marketing, 29(2), 219–231. Shao, A. T. (1993). Restrictions on advertising items that may not be considered decent: A European viewpoint. Journal of Euromarketing, 2(3), 23–43. Shaver, P., Schwartz, J., Kirson, D., & O'Connor, C. (1987). Emotion knowledge: Further exploration of a prototype approach. Journal of Personality and Social Psychology, 52, 1061–1086. Sheffield, T. (2006). The religious dimensions of advertising. Macmillan Publishing. Shiu, E., Hassan, L. M., & Walsh, G. (2009). Demarketing tobacco through governmental policies—The 4 Ps revisited. Journal of Business Research, 2, 269–279. Triandis, H. C. (1995). Individualism and collectivism. Boulder, CO: Westview Press. Urban Institute (2010). Number of nonprofit organizations by state. National Center for Charitable Statistics (http://nccsdataweb.urban.org/PubApps/profileDrillDown. php?rpt=US-STATE, accessed February 7, 2011.) Verhaert, G., & Van den Poel, D. (2011). Empathy as added value in predicting donation behavior. Journal of Business Research, 64(12), 1288–1295. Wiebe, K., & Fleck, J. (1980). Personality correlates of intrinsic and extrinsic and non-religious orientations. Journal of Psychology, 105, 181–187. Wilkes, R., Burnett, J., & Howell, R. (1986). On the meaning and measurement of religiosity in consumer research. JAMS, 14(1), 47–56. Wilson, A., & West, C. (1981, Jan./Feb). The marketing of unmentionals. Harvard Business Review, 51, 91–102. Zinkhan, G., & Martin, C. (1982). Message characteristics and audience characteristics. Advances in Consumer Research, 10, 27–31.