Business-nonprofit partnerships for the good : examining the role of commitment in different communication appeals for social alliances to achieve favourable consumer attitudes

(1)

Business-nonprofit partnerships for the good:

Examining the role of commitment in different communication appeals for social

alliances to achieve favourable consumer attitudes

Name: Maxim Bos

Student number: 11152702

Supervisor: Bob Rietveld

Date: 23/06/2017

(2)

Statement of Originality

This document is written by Maxim Bos, who declares to take full

responsibility for the contents of this document.

I declare that the text and the work presented in this document is original

and that no sources other than those mentioned in the text and its

references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the

supervision of completion of the work, not for the contents.

(3)

Table of Figures

TABLE 1: TYPES OF MARKETING INITIATIVES WITH A SOCIAL DIMENSION...9

TABLE 2: TYPES OF RELATED ALLIANCES...11

TABLE 3: EXPERIMENTAL (2X2) FACTORIAL DESIGN...21

TABLE 4: RELIABILITY SCALE FOR ATTITUDE TOWARDS COMPANY MEASURE...25

TABLE 5: DEMOGRAPHICS RESPONDENTS...25

TABLE 6: DESCRIPTIVE STATISTICS DEPENDENT VARIABLE...26

TABLE 7: DESCRIPTIVES DISTRIBUTION OF ATTITUDE IN HIGH COMMITMENT CONDITIONS...27

TABLE 8: TEST OF NORMALITY FOR HIGH COMMITMENT CONDITIONS...28

TABLE 9: DESCRIPTIVES DISTRIBUTION OF ATTITUDE IN LOW COMMITMENT CONDITIONS...29

TABLE 10: TEST OF NORMALITY FOR LOW COMMITMENT CONDITIONS...29

TABLE 11: HOMOGENEITY OF VARIANCES - EEVENE’S TEST...30

TABLE 12: BETWEEN-SUBJECTS FACTOR SAMPLE SIZES: ATTITUDE...31

FIGURE 1: ONE-WAY ANOVA ATTITUDE X (HIGH/LOW) COMMITMENT...32

TABLE 13: TESTS OF BETWEEN-SUBJECTS EFFECTS...33

TABLE 14: PROCESS, MODEL 1, Y=ATTITUDE, X=COMMIT, M=APPEAL...33

FIGURE 2: TWO-WAY ANOVA ATTITUDES X COMMITMENT + APPEAL...33

(5)

Abstract

Social Alliances concern an advanced state of collaboration between companies and social enterprises or nonprofit organisations. Besides the primary objective of improving social welfare, the company engaging in a social alliance can benefit by fostering positive brand attitudes and herewith build its brand equity. Effective communication is necessary to get recognition for one’s social efforts. However this is a tricky domain, as consumers seem to be extra judgemental and suspicious about corporate communication with a social

dimension. This study examines the role of company commitment to a social alliance in particular, when influenced by communication appeal used in a press article. It was found that consumers have more favourable attitudes towards the company when it is 1) highly committed to the social alliance and 2) when a rational communication appeal is used. However there was no interaction effect found; neither communication appeal influenced the strength of the relationship between commitment and attitude. These insights can alleviate some of the doubts that managers may have about committing to a social alliance and how to get the appropriate recognition for it.

(6)

1. Introduction

1.1 Background

The societal aspect of marketing has only gained serious interest in the last decade (Andreasen, 2012), but is considered increasingly important due to its potential of influencing attitudes and changing consumer behaviour (Vock et al., 2013). Traditionally, social

marketing applied marketing knowledge and techniques to enhance social objectives such as driving responsibly (Kotler and Zaltman, 1971), however recent studies suggest the bridging to commercial marketing as corporations are actually well positioned to help tackle some of society’s most pressing issues (Kotler and Lee, 2004; Hastings and Saren, 2003; Dauvergne and Lister, 2013). As such, social alliances have emerged, which describe a long-term partnership between a company and a social enterprise/non-profit (and the public sector), to enhance a societal objective while reaping organizational benefits in the process (Berger et al., 2006; Vock et al., 2013). This paper aims to better understand the indirect value that

social alliances can bring to the company in terms of consumer attitudes, through external

communication.

Even though 80% of global consumers agree that business must play a role in addressing societal issues (Edelman, 2017), communicating socially responsible behaviour is considered a slippery slope, because consumers are extra critical to such messages (Ellen et al., 2006; Forehand and Grier, 2003). The challenge is to appear genuinely concerned with the focal issue, while avoiding consumer suspicion of ulterior, self-serving motives (Du et al., 2010). The Reputation Institute (2016) revealed that 20% of consumers is ‘not sure’ of their perception of some CSR dimensions – signalling an awareness gap. Since generating positive returns from one’s social initiatives obviously requires stakeholders’ awareness, this issue deserves more attention. “Effective communication is required to capitalize on good

initiatives and overcome poor perceptions” (CSR RepTrak® 100, 2016).

Marketing communications specialists use varying strategies to accomplish their objectives (Andreasen and Kotler, 2012, p.269-277) that typically involve persuasion, or the

(7)

influencing of someone’s beliefs, attitudes, decisions or actions (Friestad and Wright, 1994). In the advertising literature, the distinction is often made between rational (or informational) and emotional appeals (Akpinar and Berger, 2017; Yoo and MacInnis, 2005) – targeting cognition and affect (or feelings) respectively. In view of real communication attempts by companies engaged in social alliances it seems that there is no consensus in this regard. This deserves consideration, since both cognition and affect play a considerable role in the processing of, and hence acceptance of, corporate communications with a social dimension (Du et al., 2010; Mohr, Webb and Harris, 2001).

The level of company commitment is a key element that differentiates social alliances from other CSR-related initiatives. Previous studies have suggested that longer commitment to a cause signals sincere concern (Webb and Mohr, 1998) that should garner more

favourable attitudes, however this hasn’t been tested on all dimensions of commitment nor in the context of social alliances.

As social alliances are a relatively new phenomenon and require significant corporate investment: research that considers external communication is vital for building awareness, generating positive attitudes and as such contributing to society in a sustainable way. Several researchers have called for more research on social alliances, as it is considered a promising form of cross-sector collaboration that will grow in importance in the near future (Austin, 2000; Berger et al., 2004; Vock et al., 2013). Literature on social alliances is scarce, as it hinges between social marketing, strategic alliances and corporate citizenship. Few have considered it from a consumer perspective (Vock et al., 2013), even though its importance is acknowledged (Berger et al., 2006). This study also responds to calls for the exploration of potential factors that might influence the (in)effectiveness of CSR

communication (Du et al., 2010), considering the role of commitment and communication appeal in particular.

This empirical study employs a 2 (emotional vs. rational appeal) x 2 (low vs. high commitment) design using real press articles manipulated to fit the study as external communication stimuli.

(8)

1.2 Research Questions

1. To what extent does the expressed company commitment to a social alliance impact consumer attitudes toward the company?

2. What role does the communication appeal (emotional vs. rational) play in influencing those perceptions?

1.3 Outline of Thesis

This chapter provided an introduction to the field that social alliances have emerged from, drawing attention to the importance of conducting research in this area. The research

questions that seem like a worthy next step to explore were presented. The following chapter will contain a review of the literature that will serve as the needed background to this study. This discussion will build up into the conceptual framework and the proposed hypotheses. Then, the methodology and experimental design utilized for this study will be covered, which will be used to test the hypotheses. The final section will consist of the data analysis results, a discussion of what these results imply and finally the managerial and theoretical

(9)

2. Literature Review

2.1.1 Social Alliances

Social Alliances would be categorized as a form of corporate citizenship in the CSR context; therefore some insights will be drawn from the respective literature. A survey of more than 300 industry experts revealed that CSR initiatives create most value by enhancing the company’s reputation (Chernev and Blair, 2015), while corporate citizenship particularly contributes to reputation in terms of trust and attitude building (Fombrun, Ponzi and

Newbury, 2015). This study focuses on the latter – attitudes – from a consumer perspective. Social Alliances stem from the field of social marketing, which was introduced by Kotler and Zaltman (1971) as a means to apply commercial marketing techniques to drive behavioural change for a better society, hinging on the idea that what’s good for society is good for business in the long-run. The industry hasn’t adequately responded to this call (Andreasen, 2012) and it’s only gained real traction in recent years: by 2008, the American Marketing Association decided to update its definition of marketing to include societal impacts (Ama.org, 2017). As with all marketing, the focal emphasis of social marketing is on value creation (Dibb and Carrigan, 2013). The main distinction is that for social marketing the focus is on society and individuals’ long-term well being, while for commercial marketing the focus has been on consumers’ short-term benefits, at all costs. It’s argued that with the increased interest in CSR, growing consumer social consciousness and a changing consumption mindset; commercial and social marketing are coming closer together (Dibb and Carrigan, 2013). Businesses are increasingly involved in social activities and/or adding societal benefits to its offerings, while social marketers focus more on personal benefits to drive positive behavioural change (Andreasen and Kotler, 2008, p.10+p.91). Resulting from these trends, different types of corporate initiatives with a social dimension have emerged in fast pace in the last decades. Table 1 presents the main differences between these

(10)

Table 1: Types of Marketing Initiatives with a Social dimension

Activity type Description Illustrative

studies Empirical findings Cause-Related Marketing (CRM) Transactional-based donation model, directly related to the sales of a product or service

(Varadarajan and Menon, 1988; Nan and Heo, 2007)

CRM messages elicits more favourable consumer attitudes than no CRM offer, high-fit between company and cause is more effective than low-fit.

Cause Promotion/ Sponsorship

Supporting social causes through promotional efforts or sponsorships – usually involving a donation in return for brand associative material

(Becker-Olsen and Simmons, 2002; Menon and Kahn, 2003)

High-fit sponsorship results in more favourable attitudes toward

sponsorship and sponsor. Higher congruence further enhanced CSR ratings when primed to focus attention on sponsor brand.

Corporate (Strategic) Philanthropy

Corporate (strategic) giving, through donations and/or other corporate resources – giving without getting back

(Porter and Kramer, 2006; Kotler and Andreasen, 2008)

Supporting issues in competitive context, aligned with strategy, creates more shared value and can result in win-win. It positively

enhances reputation and can change the image of an organization.

Employee volunteerism

Encouraging employees to volunteer in support of social causes

(Peloza, Hudson and Hassay, 2009)

Strategic fit and focus on core competencies of the firm increases effectiveness.

Corporate Social Marketing

Supporting social behaviour change campaigns, using a company’s unique assets and capabilities

(Kotler and Lee, 2004; Kotler and Andreasen, 2008)

The social good does not come at the expense of the company: “what’s

good for them is good for us”.

Offering a compelling value proposition that makes the desired behaviour easy and rewarding is key.

(11)

All of the initiatives in Table 1 are considered Corporate Societal Marketing Initiatives, or “those marketing activities that have at least one non-economic objective related to social

welfare through the use of the resources of the company and/or one of its partners” (Hoeffler

and Keller, 2002). Most Social Alliances however, can be considered a (collaborative) form of

Corporate Social Marketing (CSM), where the focus is on behavioural change, which benefits

the individual, society and herewith the corporation.

Kotler and Lee (2004) argue that for this very reason “corporate social marketing

outshines the rest” in terms of marketing benefits. Since the change in behaviour is designed

to benefit individuals (e.g. in health, or living standard), it’s likely to create strong and lasting positive associations with the company that spurred this change (Kotler and Lee, 2004). In addition, behavioural or attitudinal change might be observed for consumers that are not directly targeted, but support the company for its engagement in the social initiative; hereby indirectly contributing to the company and society (Vock et al., 2013; Kotler and Lee, 2004). CSM efforts should always aim for an overlap of interests between what society needs and what a company’s goals and objectives are, while avoiding an appearance of having hidden agendas (Kotler and Lee, 2004). This corresponds with related studies that stress the

importance of fit or congruence between a company and its partnering cause (Nan and Heo, 2007; Becker-Olsen and Simmons, 2002; Menon and Kahn, 2003). Furthermore, to reap the full marketing potential, the initiative should target behaviours that directly relate to the company’s product or service (Kotler and Lee, 2004).

For example, Vodafone’s Farmers Club provides a solution to a societal issue through behaviour change that’s directly related to their services. They found that 50% of people living in severe poverty in Turkey are smallholder farmers. By offering those farmers an affordable mobile price plan with a services package that gives them access to relevant and real-time agricultural information (provided by Tabit – social enterprise), SMS alerts as well as a digital marketplace tailored to farmers’ individual needs – they were able to reach an underserved segment while working towards resolving a societal issue (Vodafone Connected Farming, 2015).

(12)

In sum, social alliances stem from the field of social marketing and can be considered an improved form of corporate social marketing, where the collaboration provides for more relevant expertise at the table and for better opportunities to drive change and hereby make more impact. Next, it’s important to point out the differences and similarities between other types of related alliances, to reveal the true identity of social alliances.

Table 2: Types of related Alliances

The main points of difference between the types of alliances are the involvement of a social enterprise as opposed to a familiar charitable cause, the potential extra governmental layer and the objective of directly improving social welfare. Social Alliances seem to represent a melding of strategic and cause-brand alliances, which encompass a combination of different corporate social initiatives (Berger et al., 2004; Porter and Kramer, 2006).

Type of Alliance Cause-Brand Alliance Strategic Alliance Social Alliance

Characteristics Short-term partnership

between company and non-profit (usually well-known charitable cause)

Long-term partnership between two companies

Long-term partnership between company, social enterprise/non-profit (and government)

Objectives Association of a company

with a non-profit: promoting the company while raising money/ awareness for the cause

Strategically utilizing complementary resources to achieve more economic benefits to both partners

Strategically tackling a societal issue using complementary resources, while realizing benefits to partners involved

Example Pampers and Unicef: to

donate 1 dollar per sold pack of Pampers

Starbucks and United Airlines: to offer Starbucks coffee on flights

Vodafone and Tabit: to offer smallholder farmers the opportunity to improve its livelihood through service offer

(13)

Most of the literature explored the philanthropic or transactional stage of

company/non-profit collaborations – representing simpler partnerships such as short-term giving and cause-related marketing programs. Social Alliances are considered the most advanced state of collaboration – the integrative stage – “when partners’ missions, people

and activities begin to experience more collective action and organizational integration”

(Berger et al., 2004). Industry experts define it as: “a partnership between a company and a

nonprofit that has moved beyond cause-related marketing and philanthropy to encompass a close, mutually beneficial, long-term partnership that is designed to accomplish strategic goals for both partners” (Berger et al., 2006). A definition from the management literature

differentiating between the different stages sounds: “cross-sector projects formed explicitly to

address social issues and causes that actively engage the partners on an ongoing basis. Projects may be “transactional” – short-term, constrained, and largely self-interest oriented – or “integrative” and “developmental” – longer-term, open-ended, and largely common-interest oriented” (Selsky and Parker, 2005). Existing studies on alliances in the integrative or

developmental stage are usually still conceptual in nature or based on case studies and elite

interviews (Álvarez-González et al., 2017; Berger et al., 2006).

Strategic objectives from the company’s perspective are not perfectly defined and may differ per alliance, but in general can be considered widely-ranging and may include: increasing public awareness, gaining market share and enhancing reputational or social capital through improving brand and trust perceptions (Berger et al., 2006; Selsky and Parker, 2005). As they are long-term and significant investments, they are often related to growing or transforming the business in a way, while pursuing a social-welfare objective. In all cases however, they intend to be strongly associated with how organizations are

perceived both internally (e.g. employees – identity) and externally (consumers – reputation). The majority of extant literature focused on the internal dimension, contributing to initial management complexities. These studies concern factors explaining how to choose partners (Liu et al., 2016; Vock et al., 2013; Berger et al., 2006) and how to effectively manage this relationship with partners (Lavie et al., 2012; Knox and Gruar, 2007).

(14)

Congruence, or fit on different levels between partners continues to be crucially important for social alliances to flourish internally (Berger et al., 2004; Berger et al., 2006). Mission-, resource- and evaluation-fit are deemed most important; which means sharing a similar vision or mission, having a shared perception and measure of success, and the

interdependency of resources (Berger et al., 2004). Successfulness further depends on effectively managing relational mechanisms – stressing the importance of mutual trust and relational commitment (Liu et al., 2016; Lavie et al., 2012). Once a successful partnership is reached, the benefits are significant and can have a transformative effect on an organization (Berger et al., 2006). Social alliances further have shown to resonate with many employees, resulting in improved morale and feeling of being more closely aligned with the organization and its values – resulting in more overall internal enthusiasm for the company.

Considering the external dimension, successful social alliances can potentially: enhance a company’s reputation, improve loyalty and impact purchase behaviour (Berger et al., 2006; Kotler and Lee, 2004). The authors argue that most benefits depend on consumer company identification, or “the degree of overlap in a consumer’s self-concept and his or her perception of a corporation” (Berger et al., 2006; Sen and Bhattacharya, 2001). Vock et al. (2013) similarly suggest that the successfulness of social alliances in terms of consumer perceptions is dependent on both personality and the level of fit between partners – individuals’ social value orientation helped explain this. They named it a “double-edged sword”: where some consumer segments (pro-socials) identify and appraise social alliances, others (pro-selfs) disidentify and can even punish the company for it (Vock et al., 2013).

Social Alliances have mainly been investigated from an internal perspective. It’s argued that once management and relational practices are better understood internally, more attention will be given to the external dimension. As one company manager of a successful social alliance already pointed out: “our biggest challenge right now is to get recognition for

our efforts” (Berger et al., 2004). Considering the limited attention that has been devoted to

this in the literature and given its potential to impact brand perceptions and generate favourable attitudes, this focus has been chosen for this study.

(15)

2.1.3 Attitude – towards company

Many researchers in the field have used the perceptual construct of brand attitude, as it is considered a main predictor of consumer behaviour, developing brand preference and one of the main independent constructs that builds brand equity (Faircloth, Capella & Alford, 2001; Fishbein and Ajzen, 1975). Brand equity is derived from consumer brand knowledge, which consists of two main constructs: brand awareness and brand image. (Keller, 1993) A brand’s image consists of the total brand associations, as well as its strength, uniqueness and favourability. In other words: ‘anything linked in memory to a brand’ (Aaker, 1991). Brand attitude is considered a type of brand association; therefore it has an indirect effect on brand equity through brand image.

Brand attitude is defined as a consumers’ overall evaluation of a brand (Faircloth, Capella & Alford, 2001). The consumer forms an image of a brand by all forms of brand interactions and then judges whether the image is beneficial and personally relevant to them, which results in an overall brand attitude (Wu and Wang, 2011). A positive brand attitude can then increase the likelihood of that consumer considering the brand, talking about the brand and eventually choosing the brand (Wu and Wang, 2011; Fazio, 2007). An overall attitude is considered the sum of beliefs about a given behaviour weighted by evaluations of those beliefs (Fishbein and Ajzen, 1975).

Brand attitudes can also be related to beliefs about non-product attributes and symbolic benefits, given that they serve as a value-expressive function. “The success of a

marketing program is reflected in the creation of favourable brand associations” (Keller,

1993) – that is, consumers believe the brand has (product & non-product) attributes and benefits that satisfies their needs and wants such that a positive overall brand attitude is formed (Keller, 1993). Brown and Dacin (1997) were one of the first to empirically validate that certain CSR associations can positively influence company evaluations, measured in terms of overall attitudes towards the company. This measure has been adopted for this study, because the evaluation of a company based on given information reflects the new association that would transfer to more general brand attitudes in the real world. This

(16)

prediction is aligned with previous research in advertising (McKenzie Lutz and Belch, 1986), sponsorship (Becker-Olsen and Simmons, 2002) and brand extensions (Aaker and Keller, 1990). For example, attitude towards an advertisement is considered a causal mediating variable in the process through which advertising influences brand attitudes (McKenzie-Lutz and Belch, 1986). In other words, the evaluation of a certain message/behaviour is typically transferred to the evaluation of that brand.

This effect is expected to be similar in the context of social alliances. Therefore a measure of consumers’ overall attitude towards the fictitious company engaging in the social alliance is considered a reasonable predictor of brand attitude formation.

2.3.1 Commitment

Commitment is one of the key elements that differentiate social alliances from its preceding cause-related marketing programs, which should reduce the potential burden of consumer perceptions of “exploiting the cause” (Ellen et al., 2006). Commitment is defined as “an

implicit or explicit pledge of relational continuity between exchange partners” and consists of

three aspects: amount of input, durability and consistency of support (Dwyer et al., 1987). Social Alliances score rather high on all dimensions: there’s significant input in terms of corporate resources and human capital, durability usually surpasses years and support shouldn’t weaken over time for a social alliance to be successful. As Berger et al. (2006) mentioned: “all partners have to be fully committed to it, or it is likely to fail”. This presents an interesting avenue for research: what is the role of this increased commitment on all three dimensions, in consumers’ attitudes towards the company engaging in a social alliance?

Drumwright (1996) was one of the first to study advertising with a social dimension and found that commitment plays an important role in internal (i.e. employees) evaluations of such campaigns. Those lasting 6 months or less “invariably spelled doom” (Drumwright, 1996), while campaigns of multiple years were often perceived as more successful. Webb and Mohr (1998) further revealed that for cause-related marketing programs the durability of support was used as a cue for consumers’ attributions of a firm’s motives – longer-term

(17)

commitments were more likely to be seen as genuine, while short-term commitments were perceived as more exploitative. Ellen et al. (2006) more recently tested this again and found that short-term commitment led to higher stakeholder-driven attributions of company motives, which is reactive in nature and therefore perceived negatively by consumers.

Bearing in mind that in these previous studies the commitment variable was only measured in terms of durability and the treatments were rather extreme: 1 month vs. 7 years, this issue needs a fresh perspective that’s applied in a more realistic context for social

alliances. Does putting in more corporate resources and adding years of consistent support to a social alliance affect consumers’ attitudes in any way, or is it not considered more important? It is proposed that expressing a high commitment on all three dimensions will positively impact consumer attitudes towards the company. Low commitment will more likely be regarded as reactive, resulting from stakeholder pressure, or exploiting an issue for PR or image enhancement (Ellen et al., 2006; Du et al., 2010). Also, as the newsworthiness of social alliances diminishes over time (Berger et al., 2006), while corporate input stays consistent and considerable: high commitment should signal genuine concern for the issue, which should result in more favourable attitudes.

H1: A positive relationship exists between commitment and attitude towards company; where high commitment leads to higher attitude towards company.

2.3.2 Communication Appeal – Emotional vs. Rational

Drawing from the advertising literature, communication appeal is a primary strategic consideration (Andreu et al., 2015) and is typically distinguished as either emotional or

rational (or informational) Stafford, 2005; Akpinar & Berger, 2017). In nonprofit marketing

communication strategies similar message structures were identified (rational, emotional and moral) and the authors suggest choosing one to encode a specific desired message

(18)

are used to communicate about the same social alliances (e.g. Appendix A + B) through press articles, signalling that there’s no agreement on which one is more effective in getting the desired message across. To stay within the scope of this study, only appeal is taken into consideration for communication strategies that aim at influencing target audiences’

attitudes, perceptions and beliefs.

Emotional appeals are designed to appeal to the receiver’s emotions, or to stimulate affective responses, such as feelings and moods (Yoo and MacInnis, 2005). Studies have shown that emotional elements enhance brand preference (Zajonc, 1980), whereas attitudes are not exclusively formed on beliefs about the product/brand attributes, but also on attitude towards the ad (MacKenzie and Lutz, 1989). Friestand and Wright (1994) stressed that emotional appeals were developed to avoid consumers’ growing sceptical resistance to rational arguments, arguing that for this reason emotional appeals are generally more effective. This supports the more recent “storytelling” approach that’s often applied in branded video: where a brand communicates its values through a story, turning the brand into the hero (Lundqvist et al., 2012). This is considered more convincing and memorable than presenting information and facts. However, such emotional appeals are only effective when they’re considered authentic and particularly not deceiving (Lundqvist et al., 2012). In support of this argument, Cotte et al. (2005) revealed that if consumers infer too much manipulative intent in an emotional (guilt) appeal, they are unlikely to feel guilty but instead elicit negative feelings that cause negative attitudes towards the ad and the sponsor. It appears that the effectiveness of emotional appeals hinges on the primary feelings that it evokes. Positive feelings may directly form positive brand attitudes through the affect transfer mechanism (Janiszweski, 1988), but can also create cognitive flexibility, which may influence meaningfulness and beliefs about the message and hence influence attitudes (Yoo and MacInnis, 2005).

Rational appeals on the other hand, are designed to appeal to cognition by presenting objective information and hereby attempting to change consumers’ beliefs and judgments (Zajonc, 1980; Yoo and MacInnis, 2005). The primary response of the viewer should thus

(19)

concern rational evaluative thoughts, usually regarding the message credibility and

meaningfulness (Yoo and MacInnis, 2005). This judgment in turn affects consumers’ beliefs and feelings toward the ad, which are predictors of brand attitude formation (Yoo and

MacInnis, 2005). For example: if an ad is perceived as credible and relevant, consumers are less likely to feel sceptical or show resistance, which should positively influence their attitude toward the ad.

It seems that both cognition and affect play a role in the brand attitude formation process for both communication appeals, such that feelings evoked in emotional appeals enhance evaluations of the ad’s credibility and evaluative thoughts in rational appeals influence feelings to some extent, which impact attitudes toward the ad (Yoo and MacInnis, 2005). Even though the primary responses targeted (cognition/affect) are crucially important, success of both appeals is likely contingent on providing both favourable feelings and

evaluations of the ad as credible and meaningful (Yoo and MacInnis, 2005). In addition to that, the relative importance of both affect and cognition is dependent on context, which is defined by e.g. category, competitive environment and target audience (Vakratsas and Ambler, 1999).

It is proposed that because the context of both communication appeals is a

corporations’ involvement in a societal issue: this should directly trigger cognitive responses, regardless of appeal, as consumers are extra critical to corporate CSR communication (Du et al., 2010). In addition to that, consumers generally automatically infer motives as to why a company is engaged in a specific CSR-related behaviour (Rifon et al., 2004; Ellen et al., 2006; Becker-Olsen et al., 2006) and especially when it involves a serious societal issue. Is the corporation genuinely concerned, or simply exploiting the cause? It is proposed that consumers are more likely to highly elaborate on the message, as they are more inclined towards effortful cognitive activities (Becker-Olsen et al., 2006; Petty and Cacioppo, 1986). This means that the relative importance of the strength of the argument would increase, which is better suited to the rational appeal. Rational appeals should furthermore be judged as more credible and transparent (Akpinar and Berger, 2017), due to its factual and direct

(20)

approach. It is therefore proposed that consumers are more willing to accept rational appeals and will evaluate them more favourable, relative to emotional appeals. As the level of

commitment is considered an argument for companies to act genuinely and the argument strength is increased by high elaboration; a moderation effect is expected here. It is thus argued that when the rational appeal is used, the relative importance of commitment increases and therefore has a stronger effect on attitude towards company.

For the emotional appeal, consumers are expected to be more aware of the

manipulative attempt and therefore perceive the message relatively more negatively. Also, the articulation of commitment is more likely to be perceived as another way to convince them, in the emotional appeal, which could seem like “bragging” and backfire (Sen and Bhattacharya, 2009). In the current form it is expected that the emotional appeal won’t elicit the right emotions strongly enough and won’t be perceived as very credible, therefore garnering less favourable attitudes. If on the other hand, a video was used in which the farmer tells his own story; it would likely elicit stronger emotions, increase message credibility and hereby improving the chances of garnering favourable attitudes. It’s thus proposed that, in the domain of external textual communications of social alliances; using a rational appeal would be a better strategy.

H2a: A positive relationship exists between rational communication appeal and attitude towards company.

H2b: The relationship between commitment and attitude towards company is positively moderated by the rational appeal; the effect of commitment on attitude towards company is stronger in the rational appeal.

(21)

2.5 Conceptual Framework

H2b H2a H1

This model displays the design of this study as constructed by the proposed hypotheses. The lower arrow depicts the first main effect (H1); the second arrow pointing to Attitude towards Company (H2a) relates to the second proposed main effect; the final arrow pointing down concerns the proposed moderation effect (H2b).

3. Methodology

An experimental survey approach was chosen to assess the conceptual model created and to test the proposed hypotheses. A 2 (high vs. low commitment) x 2 (emotional vs. rational appeal) factorial design was used to test the role of commitment, moderated by

communication appeal on the dependent variable: attitude towards the company engaged in a social alliance.

3.1 Research Method

This study has a deductive and exploratory approach, with a chosen quantitative research method. This is exemplified by the use of structured questions with predetermined answers, reaching a high amount of participants to quantify responses to a specific research problem

Communication appeal

Rational

Emotional

Commitment

Attitude towards Company

High

(22)

(Saunders et al., 2012). The nature of the proposed research questions suggests combining an experiment with a survey; different experimental conditions were deemed necessary to manipulate the independent variables and the survey element is needed for distribution and collecting consumer responses.

3.2 Research Design

Hypotheses were tested using a 2 (high vs. low commitment) x 2 (emotional vs. rational appeal) factorial design. Responses towards experimental stimuli were collected through a short online survey distributed through Qualtrics. Participants were randomly assigned to 1 of 4 experiment conditions. Two versions of real press releases were used and manipulated to fit this study, in each manipulating the commitment variable.

Table 3: Experimental (2x2) factorial design

Low Commitment

Rational Appeal High Commitment

Low Commitment Emotional Appeal

High Commitment

3.3 Stimuli Development

Twenty-five social alliances have been analyzed online, looking for differences in stakeholder communication. The first thing that was noticed is that the amount of internal reporting is not aligned with the external output: companies make beautifully explicit and transparent reports internally, but are somewhat afraid to show it externally. This corresponds with the notion that CSR efforts shouldn’t seem like “bragging” (Sen and Bhattacharya, 2009) and that PR may be a better option to facilitate the positive impact of CSR rather than using company’s own advertising (Chernev and Blair, 2015). Some companies use blog posts, videos or infographics to communicate about its social alliance, however the main way of reaching consumers seems to be through press releases: local, national and international newspapers

(23)

of varying sizes report on it (The Guardian, 2012; BizTech Africa, 2015; Zee News, 2015). This form of communication is therefore used and manipulated to fit this experimental study – differentiating between emotional and rational press articles.

This study aims to control for company- and stakeholder-specific variables that may cause confounding effects. Firstly, by using a fictitious company that respondents have no prior knowledge or perception of, operating in a single industry with a moderate reputation (Reputation Institute, 2016). Companies with good reputations usually find the positive effects of their CSR communications amplified, whereas opposite effects occur for companies with poor reputations (Brown and Dacin, 1997; Yoon et al., 2006). The same holds for industries such as tobacco or oil, for which stakeholders’ reactions will be more suspicious and sceptic (Du et al., 2010).

Next, a single stakeholder type (consumer) is examined and a company type that participants are similarly involved with (telecom services). Since communicating about CSR to shareholders or investors requires a considerably different approach, this research will be tailored to the “general public” – specifically consumers whom don’t proactively seek CSR information (Dawkins, 2005), but become aware of it through independent channels such as editorial coverage in newspapers, TV or online channels. (Du et al., 2010)

Furthermore, confounding effects for involvement with the cause is minimized by choosing a societal issue that is likely not highly personally relevant to any of the participants (‘extreme poverty’). However, as stable personality differences still may influence results, social goodwill – adopted from Chernev and Blair (2015) – is included as a pre-measure.

Finally, the fit construct, or congruence theory, has widely been used to explain effects in consumer reactions to CSR initiatives with a social dimension. However, as Berger et al. (2006) argued that partnering organizations are likely match in terms of core values, company culture and they should complement each other in terms of what they bring to the table; otherwise they won’t commit to such a long-term partnership. Austin and Seitanidi (2012) similarly point out that organizations should carefully look for early indications of

(24)

partnership compatibility, as “the correct partnership is everything” (Weiser et al., 2006, p.6). Therefore the fit construct is taken for granted and controlled for by using a single, highly fitting (complementing) social alliance.

3.4 Pre-test – manipulation checks

A pre-test was conducted to check whether the experimental stimuli were manipulated sufficiently for usage. A scale was adopted from Yoo and MacInnis (2005) that asked participants’ perception of the communication appeals as different: one being emotional and the other rational. 16 students were randomly selected for this in a university canteen and were asked to rate on a seven-point scale, the extent to which they viewed the message as emotional (‘this message appeals to my emotion’, ‘this message creates a mood’), rational (‘this message provides a lot of information’). Results of a one-way ANOVA showed

significant mean differences between both communication appeals in terms of emotionality and rationality scores (p<.01); the pre-test was successful. For the commitment variable, a scale was adopted from Ellen et al. (2006) that asked participants’ perception of company commitment on two 7-point semantic differential items (‘Alphone is highly committed to the social alliance’). The same 16 students were exposed to either one (high vs. low) of the treatments and two significantly different levels of commitment were attained, p<.05. Those exposed to the low commitment treatment rated commitment significantly lower (M=4.08) than those who received the high-commitment treatment (M=5.43).

3.5 Subjects, procedures and measures

An online experimental survey was created in Qualtrics and distributed by means of

convenience sampling – an email list – to a population of (ex) colleagues, relatives, friends, fellow students and other contacts. This sampling method was chosen as it gives most freedom to the researcher and is easiest in terms of finding sufficient respondents under time and resource constraints (Saunders et al., 2012). A total of 105 participants participated in the experiment, all of Dutch nationality. 56% are male, 31% is in between 26-34 years old

(25)

and 21% in between 18-25. 37% finished a higher level of applied education (HBO), 23% finished a bachelor’s degree (WO) and 27% a master’s degree. In total 6 participants had missing values, of which 3 had to be discarded, leaving n=102.

Participants were first of all given a short introduction to the kind of research being done and to Social Alliances in general. Since most people are most likely not familiar with the term, this was deemed necessary background information. Then, participants were asked to fill out some general information (gender, age, highest level of education, social goodwill), which could be used as control variables. After, participants were randomly assigned to one of the four experimental conditions and were exposed to the stimulus material (e.g. High commitment, rational press article). Finally, participants were asked to give their perceptions of the fictitious company engaged in the social alliance, being the dependent variable.

The social goodwill pre-measure was taken from Chernev and Blair (2015) and was phrased as the following statement: “I think it’s important for companies to give back to

society”. A 7-point semantic differential scale was then used to assess this single-item

constructs in terms of how much participants agree with it.

The dependent variable measure ‘attitude towards the company’ was adopted from Vock et al. (2013) and was phrased as “My perception of Company Alpha is….” to solely find what we’re interested in: consumers perception of the company’s engagement in the social alliance. The measure reflects a general attitude measure that many researchers in the field of Marketing have used and quantify as three constructs: ‘favourable vs. unfavourable’, ‘good vs. bad’ and ‘positive vs. negative’. A 7-point semantic differential scale was used and the sum of all items divided by 3 formed the ‘attitude towards company’ measure.

4. Results

(26)

This chapter presents the obtained results, which were analyzed in SPSS, version 24. First of all, a reliability test was performed for the continuous variables as constructed by values of Cronbach Alpha. Secondly, the sample demographics, relevant descriptive statistics and frequency checks are presented. Given that the dependent variable is continuous, the two independent variables are categorical and there’s independence of observations: a two-way ANOVA test is considered appropriate for testing the hypotheses. First the assumptions for two-way ANOVA are accounted for: detecting possible outliers, conducting normal

distribution and homogeneity of variances analyses. Once these assumptions are met, the hypotheses are tested through two-way ANOVA tests.

4.1 Reliability

Reliability determines the degree to which multiple scale items measure a single construct. Both independent variables are pre-determined and categorical; so don’t need to be tested on reliability. The dependent variable items for attitude (n=3, α=0.934) indicate high reliability (>0.7) and thus all items have good internal consistency that allowed for the merging into the single variable construct: attitude towards company. The corrected item-total correlations indicate that all items have a good correlation with the total score of the scale (>0.3); none of the items were to be deleted from the scale.

Table 4: Reliability scale for Attitude towards company measure

Cronbach’s Alpha N of items

,934 3

4.2 Descriptive Statistics

(27)

Frequency Percent Valid percent Cumulative percent Gender Male Female Missing Total 56 43 6 105 53,3% 41% 5,7% 100% 56,6% 43,4% 100% 56,6% 100% Age 18-25 26-34 35-42 43-51 51-58 Missing Total 22 33 17 15 13 5 105 21,0% 31,4% 16,2% 14,3% 12,2% 4,8% 100% 22% 33% 17% 15% 13% 100% 22% 55% 72% 87% 100% Education Secondary School MBO HBO Bachelor’s WO Master’s WO Doctorate PhD Missing Total 2 5 39 24 28 1 6 105 1,9% 4,8% 37,1% 22,9% 26,7% 1% 5,7% 100% 2% 5,1% 39,4% 24,2% 28,3% 1% 100% 2% 7,1% 46,5% 70,7% 99,0% 100%

Table 6: Descriptive statistics dependent variable

(28)

Table 6 shows the descriptive statistics of the overall data for Attitude towards the company. Scores range from 1,67 to 7 with a mean of 4,66 and a standard deviation of 1,149. The negative Skewness (-.233) and Kurtosis (-.446) mean a moderate negatively skewed distribution of data and indicate a longer left tail, a flatter shape and little scores in the tails.

4.3 Normality of Distribution

Normality tests were conducted and related to skewness and kurtosis to assess the

distribution of the dependent variable for each of the two levels of the independent variables. As the sample n>50 the statistics for the Kolmogorov-Smirnov test should be most accurate, however both measures will be taken into account for higher reliability.

Table 7: Descriptives distribution of Attitude in High Commitment conditions

High Commitment Statistic Std. Error

Attitude

RATIONAL

Mean

95% confidence interval for mean 5% Trimmed mean Median Variance Std. Deviation Minimum Maximum Range Interquartile range Skewness Kurtosis Lower bound Upper bound 5,464 5,164 5,764 5,462 5,667 ,482 ,694 4 7 3 1 -,014 ,129 ,145 ,481 ,935

Statistic Statistic Statistic Statistic Statistic Statistic Std. Error Statistic Std. Error

Attitude Valid N

102

(29)

EMOTIONAL

Mean

95% confidence interval for mean 5% Trimmed mean Median Variance Std. Deviation Minimum Maximum Range Interquartile range Skewness Kurtosis Lower bound Upper bound 4,507 4,130 4,884 4,463 4,667 ,834 ,913 3 7 4 1 ,551 1,307 ,145 ,464 ,902

Table 8: Test of Normality for High Commitment conditions

ª. Lilliefors Significance Correction

The rational appeal in the high commitment condition presents data that’s slightly negatively skewed (-0,14) with a positive kurtosis (1,29), indicating a relatively symmetrical distribution with many scores in the tails and a pointy shape. For the emotional appeal: the data is moderate positively skewed (0,551) with a positive kurtosis (1,31), indicating a pointy and longer tail to the right with more scores in the tails. The dependent variable seems to be normally distributed for both the emotional and the rational appeal in the high commitment treatment, as the assumptions of Kolmogorov-Smirnov and Shapiro-Wilk are not violated: all significant values for high commitment are higher than the p value of 0.05. One outlier (29) was observed in the emotional condition. Given that the mean (3,682) and the 5% trimmed mean (3,611) don’t differ much, it’s suggested that the extreme scores have no strong influence on the mean. Therefore the outlier was not removed for further analysis.

Kolmogorov-Smirnovª Shapiro-Wilk

Appeal _Statistic _df _Sig. _Statistic _df _Sig.

Attitude Rational Emotional ,139 ,150 23 25 ,200 ,148 ,970 ,942 23 25 ,685 ,165

(30)

Table 9: Descriptives Distribution of Attitude in Low Commitment conditions

Low Commitment Statistic Std. Error

Mean

95% confidence interval for mean 5% Trimmed mean Median Variance Std. Deviation Minimum Maximum Range Interquartile range Skewness Kurtosis Lower bound Upper bound 4,875 4,5368 5,213 4,898 5,667 ,880 ,938 3,33 6 2,67 2 -,080 -1,351 ,145 ,414 ,809 Attitude RATIONAL EMOTIONAL Mean 95% confidence interval for mean 5% Trimmed mean Median Variance Std. Deviation Minimum Maximum Range Interquartile range Skewness Kurtosis Lower bound Upper bound 3,682 3,088 4,276 3,611 3,333 1,793 1,339 1,67 6 5,33 1,67 ,933 ,527 ,145 ,491 ,953

Table 10: Test of Normality for Low Commitment conditions

ª. Lilliefors Significance Correction

Kolmogorov-Smirnovª Shapiro-Wilk

Appeal _Statistic _df _Sig. _Statistic _df _Sig.

Attitude Rational Emotional ,197 ,148 32 22 ,003 ,200 ,883 ,927 32 22 ,002 ,107

(31)

The data for the rational appeal in the low commitment condition is very slightly negatively skewed (-0,08) with a negative kurtosis (-1,351), which means that it’s almost a completely symmetrical distribution that is shaped flatter than usual. The emotional appeal on the other hand, is positively skewed (0,933) with a positive kurtosis (0,527), indicating a pointy shape with many scores in the tails. The dependent variable in the emotional appeal in the low commitment condition seems to be normally distributed, as the values for

Kolmogorov-Smirnov and Shapiro-Wilk are again not significant. For the rational appeal the assumption of normality is not met, as both the Kolmogorov-Smirnov (p<.05) and Shapiro-Wilk (p<.05) values are significant. By analysing the plotted graph for this specific treatment, a reasonably straight line was found (see Appendix D). This means that it’s still acceptable as normally distributed data along a reasonably straight line. One outlier was found in the emotional condition. However, the closeness between the mean (3.682) and the 5% trimmed mean (3.611) indicates that the extreme score(s) don’t affect the mean a lot, so the outlier doesn’t have to be removed from the dataset.

Table 11: Homogeneity of variances - Levene’s test Dependent: Attitude

F df1 df2 Sig.

2,644 3 98 ,053

Levene’s test revealed that the error variance of the dependent variable is equal across all groups. The assumption for homogeneity of variances is supported, because p>0,05. Finally, Table 12 shows that the between-subjects factor sample group sizes are close to being equal and N>25, which successfully meets ANOVAs last assumption and is considered acceptable for the recognition of patterns (Saunders et al., 2012).

All assumptions for two-way ANOVA were accounted for. There were no extreme violations of the assumptions; therefore the data is appropriate for use in further analyses.

(32)

Table 12: Between-subjects factor sample sizes: Attitude

Appeal Commitment Mean Std. Deviation N

Rational High Low Total 5,464 4,875 5,121 ,694 ,938 ,887 25 27 52 Emotional High Low Total 4,507 3,681 4,121 ,913 1,340 1,195 25 25 50 Total High Low Total 4,965 4,389 4,660 ,941 1,255 1,150 48 54 102

Lastly, a frequencies analysis revealed that only 3 of the total n=102 somewhat disagreed with the social goodwill statement “I think it’s important for companies to give back to

society.” 11 neither agreed nor disagreed and the other 88 either somewhat agreed, agreed

or strongly agreed. Not enough variance was observed to create low vs. high social goodwill measures, as more than 86% cares, at least to a certain extent, about companies’

involvement with society. Therefore, unfortunately the pre-measure of social goodwill was discarded for further analysis.

4.4 Hypotheses Testing

Hypothesis 1 stated that subjects would respond more favourable in terms of attitude towards the company under high commitment conditions, relative to low commitment conditions. A two-way ANOVA was performed to assess the effect of commitment on

attitude, regardless of communication appeal. A significant difference was found for attitudes under high commitment (M=4,97, SD=0.94) and low commitment (M=4,39, SD=1,26),

(33)

of moderate size, of commitment F(1, 98) =12,79, p<.05, η² = .115) on attitude. Therefore H1 is supported: communicating high commitment to the social alliance leads to more favourable attitudes. Figure 1 graphically displays this main effect in a one-way ANOVA, taking into account commitment only at both levels of communication appeal. After controlling for gender, age and education: no significant effects were found in any of the between-subjects tests. Meaning that neither of the variables had confounding effects on the dependent variable Attitude towards company in any of the conditional groups. Thus the following interpretation of results may be considered accurate representations of the proposed measured effects.

Figure 1: one-way ANOVA attitude x (high/low) commitment

Hypothesis 2a stated that the rational appeal would garner more positive attitudes relative to the emotional appeal, regardless of commitment. A two-way ANOVA test between-subjects effects revealed that this main effect is present too. A significant difference in means was observed between attitude towards the company after being exposed to emotional appeals (M=4,12, SD=1,19) and rational appeals (M=5,12, SD=0.89), regardless of the level of commitment. This means that there was a significant main effect, of high size, of

communication appeal F(1, 98) = 29.57, p<.05, η² = .232) on attitude, which is in support of H2a. Table 13 presents the between-subjects effects test and reveals the presence of both

(34)

main effects in terms of significant values of Appeal and Commitment. Figure 2 illustrates these observations in a graph.

Table 13: Tests of Between-Subjects Effects Dependent Variable: Attitude

Type III Sum

of Squares df Mean Square F Sig. Partial Eta Squared (η²) Source Corrected Model Intercept Appeal Commitment Appeal * Commit. Error Total Corrected Total 37,976a 2142,921 28,865 12,475 ,348 95,575 2348,667 133,551 3 1 1 1 1 98 102 101 12,659 2142,921 28,865 12,475 ,348 ,975 12,980 9197,298 29,598 12,792 ,357 ,000 ,000 ,000 ,001 ,552 ,284 ,957 ,232 ,115 ,004

a. R squared = ,284 (Adjusted R Squared = ,262) b. Computed using alpha = ,05

(35)

Hypothesis 2b proposed that the rational appeal would moderate the interaction effect between commitment and attitude, such that a change in commitment would cause a stronger effect on attitude towards the company in the rational appeal.

As Table 13 and Figure 2 already give away and was revealed by the two-way ANOVA test results: there was a non-significant interaction effect between commitment and appeal on attitude F(1, 98) = 0.36, p = .55). This indicates that there is no significant

difference in strength of effect of commitment on Attitude towards the company for any of the two communication appeals. PROCESS by Hayes (2012) further confirmed, when

bootstrapping 5000 times at confidence intervals level of 95%, that there was no interaction (p = .57) between appeal and commitment. Given that p>.05 at the interaction, no further analysis of the output was necessary. See the following Table 14. Therefore H2b is not supported.

Table 14: PROCESS, Model 1, Y=Attitude, X=Commit, M=Appeal

Given that there is no significant interaction effect; the two observed main effects could be safely interpreted. The significant main effect for communication appeal is of high effect size (η² = .232), meaning that 23.2% of the variance on Attitude towards the company can be explained by communication appeal. The significant main effect for commitment is of

moderate effect size (η² = .115) as 11.5% of the variance on Attitude towards the company is explained by level of commitment.

5. Discussion

The findings of this study contribute to the literature on social marketing, social alliances and CSR communication in a number of ways.

Model constant Appeal Commit Int_1 coeff 6,7736 -,7210 -,3527 -,2361 se ,8299 ,5836 ,5669 ,4129 t 8,1620 -1,2356 -,6221 -,5718 p ,0000 ,2196 ,5353 ,5688 LLCI 5,1267 -1,8791 -1,4777 -1,0554 ULCI 8,4204 ,4370 ,7723 ,5833

(36)

First of all, this study supports the notion that social alliances don’t simply create direct value to the target population (i.e. smallholder farmers) and society, but can also create indirect value to external populations (i.e. consumers) that get acquainted with such initiatives and hereby form positive attitudes (Kotler and Lee, 2004) and build brand equity.

Next, the role of commitment has shown to be significantly important in the context of social alliances. This contributes to the knowledge on social alliances and social marketing in particular, as those alliances that aim for behavioural change should always have long-term, high commitment objectives. This study suggests that an increased commitment would directly improve consumer attitudes towards the company engaged in the alliance. Previous studies (Berger et al., 2004; Berger et al., 2006; Liu et al., 2016; Lavie et al., 2012) posited that commitment is a major factor determining the internal success of social alliances; this paper suggests that it is of similar importance in terms of external perceptions.

This is useful knowledge for managers doubting whether or not to commit to a long-term, potentially open-ending partnership with a social enterprise. It reassures, to a certain extent, that consumer will react more positively to higher, longer-term and consistent commitment to a social alliance. This could imply that a higher commitment inhibits consumer scepticism to a certain extent, which is one of the main pitfalls of external CSR communication (Du et al., 2010). Previous studies have suggested that this effect occurs because of an increased perception of genuineness (Webb and Mohr, 1998) or consumers attribute more strategic and values-driven motives (Ellen et al., 2006) to it. A major

shortcoming of this study is that it fails to explain why this effect occurs. Future research could explore this mediation effect. Also it could be interesting to explore at what point commitment would reach a saturation level, such that an increase wouldn’t affect outcomes significantly anymore.

This study also contributes to the CSR communication literature (in the context of social alliances). The relatively large variance in attitudes among different treatments signals the importance of carefully designing and tailoring external communication appeals to the right channel and audience. Rational appeals were perceived significantly better than

(37)

emotional appeals in a press article, however would the same effect occur in other domain(s) such as a video on YouTube or a blog post? This could be interesting to explore in future research. The relatively positive reactions to the rational appeal (M=5,12) further indicate that companies should communicate about its citizenship – engagement in social alliances – through external sources and especially in an objective and transparent way, which aligns with findings that PR is the better option (Chernev and Blair, 2015) as long as it doesn’t seem like bragging (Sen and Bhattacharyan, 2004). It could be argued that the emotional appeal lacks authenticity (Lundqvist et al., 2012) due to its message design – textual, without visuals – and due to the fact that it’s not taken from a ‘real’ newspaper with certain credibility (Yoo and MacInnis, 2005). Future research could resolve this issue by using real press articles from reputable news sources constituting real companies with moderate reputations.

Even though both high commitment and rational appeals lead to significantly higher attitudes towards the company, no interaction effect took place. This means that both variables were each considered important, however one variable did not affect the other. This disconfirms the proposition that the increased relative argument strength (Petty and Cacioppo, 1986) would affect the evaluation of commitment particularly more in the rational appeal, relative to the emotional appeal. It seems more that the emotional appeal lacks authenticity and credibility in the domain of a textual press article, which is enhanced by the feeling of deception (Akpinar and Berger, 2017) that arises from the evident attempt of targeting emotions. This seems a more plausible explanation as to why the emotional appeal is perceived significantly less positive than the rational appeal, because no moderation effect is observed.

To sum up; this study serves well as a first step towards explaining how external social alliance communications (in form of a press article) can be perceived more positively by consumers, in terms of attitude towards the company. The findings are aligned with and add to the research on the role of commitment in particular. Also the proposed higher effectiveness for the rational appeal in the context of a press article about social alliances is

(38)

supported, however it remains unclear why these effects are taking place. Due to the scope of this study and the limited available literature on social alliances, this presents its main shortcoming: even though some valuable insights are provided, it falls short on explaining why. Given that the interaction effect was not significant; different reasoning is needed in future research.

6. Conclusions

This study presents an initial effort at explaining how social alliances could be communicated externally through press releases to achieve favourable consumer attitudes. This study builds on what we know of social alliances until now and attempts to contribute to that knowledge by exploring communication appeals in press releases and consumers’

perceptions of it, as a way for companies to build brand equity. An exploratory, experimental survey approach was chosen to find out what the effect of communication appeal and commitment is on consumers’ perception of companies engaging in social alliances. This study emphasized internal validity as a way to minimize confounding effects and purely get the insights that was looked for. The findings are to a large extent aligned with related research and proposed theories, however it fails to explain why some of the effects are taking place, which paves the way for future research.

Potential mediating variables could be explored, such as consumer attributions of motives or feelings/emotions vs. meaningfulness/credibility judgments. Does company commitment influence consumer attributions of motives for the alliance, which in turn influences attitude formation? Or what feelings or credibility judgments arise from different communication appeals that potentially transfer to attitudes?

Furthermore, besides looking at the role of commitment in external communications, other important message variables could be considered too: societal impact, motives or fit (Du et al., 2010). Also, this study took into account press releases containing text only, but

(39)

what would results be different when studying other forms of communications and other channels? It could be argued that for example the emotional appeal is better suited to video and pictorial forms of communication, as it’s authenticity and credibility would be higher and therefore have a higher chance of eliciting the right emotions (Lundqvist et al., 2012).

6.1 Limitations and Future Research

Due to the scope and nature of this empirical study, there are some limitations to the generalizability and external validity of the findings.

First, only a single industry and company type was considered as a way to minimize confounding effects, however this strongly limits the generalizability of the findings.

Considering different types of companies in different industries would be necessary in future research to resolve this issue. Also, this study chose to use a fictitious company so that existing reputations wouldn’t affect the results and evaluations would only be based on commitment and appeal. Future research could look at the effects of companies engaging in social alliances with good vs. bad reputations, as a way to resolve this external validity issue.

In addition to that, this study chose not to include visuals in either communication appeal to minimize potential confounding effects, however this may have influenced the effectiveness of the emotional appeal. In a real-world context companies that choose for an emotional appeal are likely to include visuals as a way to enhance the proposed effect.

Furthermore, this study mainly collected respondents with high social goodwill, which could have enhanced the results. A broader and more differentiated population might better reflect an actual population of different consumers with different worldviews. As Vock et al. (2013) revealed that proself respond quite differently to social alliances than prosocials and CSR interpretations in general are often dependent to some extent on personality:

considering such differences may be useful for future research. In addition to measuring attitudinal constructs, future research should also consider behavioural measures such as purchase or switching intentions.

Business-nonprofit partnerships for the good : examining the role of commitment in different communication appeals for social alliances to achieve favourable consumer attitudes