“Catastrophes in the spotlights”

I

“Catastrophes in the spotlights”

A newsworthiness perspective on nearby and faraway

disasters on non-monetary online charity support

Thesis for obtaining the Master of Science in Marketing Management

By:

II

FACULTY OF ECONOMICS AND BUSINESS ADMINISTRATION

CHAIR OF MARKETING

MASTER THESIS

Marketing Management

“Catastrophes in the spotlights”

A newsworthiness perspective on nearby and faraway

disasters on non-monetary online charity support

Author: Michiel de Vries

Student Number: 1817663

Student e-mail: m.a.de.vries.8@student.rug.nl

Course of studies: Marketing Management (M.Sc.)

Date: June 18th ₂₀₁₈

1ST _{Supervisor: Dr. Marijke Leliveld}

III

Abstract

Charities need to make the public aware of their need for resources to help the ones in need. Those resources are essential for causes regarding, for example, medical research, wildlife conservation and disaster relief. This study examines the effects of types of disasters (naturally versus humanly caused) and proximity of the disaster (nearby versus faraway) in a 2x2 factorial design on the willingness to online support charity. The willingness to support is measured in a Facebook context; sign a petition, like the ad, share the ad, and/or comment on the charity’s ad. Past findings often describe effects on monetary charity support. However, this research demonstrates the effects on non-monetary support, which reflects the value of this study. It is expected that the willingness to support charity increases when the disaster is caused by nature instead of men. However, the findings in this research do not show this effect. Also, no expected relation between nearby versus faraway causes on non-monetary charity support was found. The data suggest that nearby disasters do not increase the willingness to non-monetarily support the cause online. In addition, no interaction effects were found. Another contribution to the literature is that the effects of disaster types and its proximity on the willingness to support are explored from a journalistic perspective, in terms of the perceived newsworthiness of the cause. The catastrophes are put in the news spotlights. The results identified that perceived interest and proximity as news indicators may increase the willingness to online support the cause non-monetarily; i.e. sign the petition, like the ad, share the ad and/or comment on the ad. Discussion of the findings, limitations of the study and implications for future research are given.

IV

Preface

This thesis is part of the requirement to graduate for the Master in Science of Marketing

Management at the University of Groningen. The research has been conducted and written in the period from February 2018 to June 2018.

The present study focuses on the effects of disaster types, spatial distance, and its perceived newsworthiness on non-monetary charity support. My motivation for a charitable related topic is because of the fact that I really enjoyed my internship at the NGO ‘Oxfam New Zealand’ in 2014. I found charitable related topics very interesting as I saw that philanthropy actually makes a difference to make the world a better place. Oxfam applied charitable related findings that are documented by researchers in their daily business performance. Therefore, I’d like to contribute to the processes within aid organizations to enhance their performance and so indirect to the beneficiaries that need public support.

I would like to thank my supervisor Dr. Marijke Leliveld for her valuable guidance and constructive feedback during this journey. I also like to thank Dr. Jing Wan in advance for her time and effort reading my thesis as my second supervisor.

Furthermore, I’d like to thank my friends and family who supported me and proofread my work along the way.

I hope you will find pleasure in reading this thesis as I took in writing it.

Michiel de Vries

V

Table of contents

Table of Contents ... V List of abbreviations ... VII List of figures and tables ... VIII

1. Introduction ... 1

2. Literature review ... 3

2.1 Type of disasters: Natural vs. Man-made ...4

2.2 Proximity: Nearby vs. faraway ...5

2.3 Mechanisms that drive donating decisions ...5

2.3.1 Mechanism 1: Awareness of need ...6

2.3.2 Mechanism 2: Solicitation ...7

2.3.3 Mechanism 3: Costs and benefits ...7

2.3.4 Mechanism 4: Altruism ...8

2.3.5 Mechanism 5: Reputation ...8

2.3.6 Mechanism 6: Psychological benefits ...8

2.3.7 Mechanism 7: Values ...9 2.3.8 Mechanism 8: Efficacy ...9 2.3.9 Conclusion ... 10 2.4 Newsworthiness of a story ... 10 2.4.1 Timeliness ... 11 2.4.2 Proximity... 11 2.4.3 Importance or impact ... 12 2.4.4 Human interest ... 12 2.4.5 Conflict or controversy ... 13 2.4.6 Sensationalism... 13 2.4.7 Prominence ... 13 2.4.8 Novelty or unusualness... 14 2.4.9 Conclusion ... 14

2.5 Online non-monetary charity support ... 14

VI

4. Research methodology ... 17

4.1 Experimental design ... 17

4.2 Procedure ... 18

4.3 Sample ... 19

4.4 Data: factors and reliability of measures ... 20

5. Results ... 21

5.1 Descriptive statistics ... 21

5.2 Hypotheses testing ... 22

5.3 Analysing underlying process of newsworthiness ... 24

5.3.1 Factorial ANOVA ... 24

5.3.2 Binary Logistic Regression ... 25

6. Conclusion and discussion ... 26

7. Limitations and future research ... 28

References ... 29

Appendices ... 35

VII

List of abbreviations

CA Cronbach’s alpha

e.g. example given

VIII

List of figures and tables

Figure

Title

Page

Figure 1 Conceptual model of the research 16

Figure 2 Survey conditions US–oil spill and US–flood disaster examples. 18

Table

Title

Page

Table 1 Overview of donation mechanisms and its definitions provided by Bekkers & Wiepking (2010)

6

Table 2 Items in the survey to measure unusualness 19

Table 3 Sample characteristics; school level and employment status (n=401) 20

Table 4 Measures and reliability 21

Table 5 Frequencies and Pearson Chi-Square test of type of disaster and proximity on charity support

22

Table 6 A Binary Logistic Regressions’ output of type of disaster on the charity support options

23

Table 7 A Binary Logistic Regressions’ output of type of proximity on the charity support options

23

Table 8 A Binary Logistic Regressions’ output of type of disaster’s and proximity’s interaction on the charity support options

24

Table 9 Two-way ANOVA’s with interaction for type of disaster and proximity on newsworthiness indicators

25

Table 10 A binary logistic regression for conflict, interest, sensationalism and proximity on the charity support options

1

1. Introduction

Messages from charitable organizations containing a call for help are communicated to the public via different media channels on a daily basis. In a competing media environment, charities such as Unicef, Oxfam, and the International Committee of the Red Cross aim for the audience’s attention. Via internet (e.g. corporate website, online ads, e-mail, and social media), TV (e.g. commercials), direct mail, and print ads in newspapers, they attempt to inform, convince and persuade people to donate (Unicef, 2018; Oxfam; 2018; ICRC, 2018). After all, raising money is the way charities can help society. For example with finding a cure for cancer via medical research, saving animals and places via wildlife conservation, or helping victims of disasters via disaster relief, varying from naturally caused disasters like earthquakes and tsunamis, to man-made disasters, like chemical explosions or war (CAF, 2017; International Disaster Database, 2018; Esghi & Larson, 2008).

When it comes to the type of disasters, there are also some studies suggesting it may influence the willingness to donate (Bennett, & Kottasz, 2000; Oosterhof, Heuvelman, & Peters, 2009; Zagekfa, Noor, Brown, Randsley de Moura, & Hopthrow, 2010) Giving statistics of the USA show an upward trend in the total amount of individual contributions that the American charitable causes gain (Giving USA, 2017). The annual report of philanthropy show an estimated increase of 3.9% to $282 billion in 2016 compared to prior year. Despite the fact that there is an increasing trend in raising money, there are variances in the amount of donations between causes (CAF, 2017).

Regarding those giving differences, the perceived distance to a disaster has been proven to have an effect on donations. Past findings show a positive effect of psychological distance on

monetary charity contributions (Touré-Tillery & Fishbach, 2017; Chen, 2016). Although the chance of one disaster may be greater than the other, disasters may happen all over the world. Every individual can face a disaster, both close and faraway to someone. However, human beings experience a psychological distance of disasters that influence their thoughts and behaviour. This psychological distance is known as the Construal Level Theory (CLT) (Trope, Liberman, & Wakslak, 2007). The authors state that far events (e.g. another country) come with high level of construal and near events (e.g. own country) with low level of construal. Specifically, disasters that happen faraway are

processed by their central, abstract and global features, whereas disasters that happen nearby are processed by their peripheral, concrete and local feature.

2 those monetary donations. For example, people can also give goods, volunteer, or sign an online petition as a way of help (CAF, 2017). As an illustration to non-monetary support, organizations may set up online petitions to generate leads. Those online petitions can be signed and shared via social media, such as Facebook, by individuals who want to show their support for a particular cause. Petitions can easily go viral, even if it is just within a few individual’s network (Corrado, 2017). Note how compared to an actual monetary donation, signing petitions is an easy act of support. However, petition signers may be contacted by the charity organization with a follow-up question for a

monetary donation or another act of help. As little is known about non-monetary charity support, this current study contributes to the literature by examining charitable supporting behaviour via petition signing.

Important to realize is that individual charity contributions are dependent on the

effectiveness of the offline and online marketing messages of the charitable organizations (Wenham, stephens & Hardy, 2003). The composition of charity ad appeals influence donating decision making (e.g. Hibbert, Smith, Davies, & Ireland, 2007; Brunel, & Nelson, 2000). How people process such messages and form an attitude about them is therefore important. It teaches us how marketing messages can be shaped to convince consumers of the value of a product or service. Ad elements determine how the ad is perceived and so the effect of the desired outcome of the message.

Continuing on ad elements, several papers describe the power of a story’s content in terms of news value indicators (Shoemaker, Danielian, & Brendlinger, 1991; Brighton, & Foy, 2007). Specifically, the newsworthiness of a story has an effect on the audience’s attention and interest in the topic (Lee, 2009). Therefore, it can be assumed that the way news value indicators of a charity’s communication are presented, influences the interest and so decision making of potential donors to support charity. Shoemaker et al. (1991) argues that the newsworthiness of a message is determined by eight indicators like proximity, impact, and interest. This research will have a closer look on the effects of the news indicators on charity support. Catastrophes will be put in the spotlights to examine the newsworthiness of nearby and faraway disasters on charitable support.

Goal and outline of the study

3 charity causes will be investigated in a Facebook context; i.e. sign the charity’s petition, like the ad, share the ad or comment on the ad.

This study contributes in two ways to the existing literature. First, not many academic findings investigated donating effects in non-monetary terms compared to monetary donations. Second, the willingness to support is explored from a journalistic perspective via the perceived newsworthiness of the charity’s message, which has not been researched for yet.

The research gives additional insights on the existing findings why people do or do not support certain charity causes. To understand the underlying concept of giving why people do and do not donate, an extensive overview of the drivers to support charity causes by Bekkers and Wiepking (2010) will be provided in the theoretical framework. The outline of this paper is as follows. Chapter 2 presents the past findings of existing literature on the topic, followed by the discussion of the conceptual model in chapter 3. Chapter 4 reports the research methodology. Chapter 5 discusses the empirical results of the research extensively. Chapter 6 reports the discussion and conclusion of the thesis. Lastly, the final chapter in this thesis critically reflects the limitations of this research and suggests future research.

2. Literature review

Many academic research is conducted on the donor decision process. What drives human beings to donate to charity is studied extensively by social-psychologists, economists, and

sociologists (e.g. Andreoni & Petrie, 2004; Vesterlund, 2006; Clark, 2002). Moreover, charitable organizations have looked into new ways to persuade (potential) donators to donate (Hibbert & Horne, 1996; Agg, 2006). This attention to charity’s marketing strategies is very important to maximise the funding and support from the public (Wenham et al., 2003).

This section discusses the literature on the donating decision making process. The thesis builds on these studies to answer the following research question: what are the effects of disaster types and proximity on the willingness to support a charity cause via Facebook? Therefore, the documented findings of several researchers on the main topics will be discussed before testing this study’s hypotheses. First, the present study starts with past findings on the classification of the types of disasters and its documented effects on charitable giving. Second, the theory regarding the effects of perceived spatial distance of disasters will be explained. Third, to understand the underlying theory of charitable giving motivations, an overview of donating drivers will be elaborated. Fourth, as this study explores the effect of the newsworthiness of charity appeals, the concept of

4 petition signing and online lead generation will be given to answer this thesis’ research question.

2.1 Types of disasters: natural vs. man-made

International non-profit organizations cover a broad range of causes that can be classified into two categories (International Disaster Database, 2017). The first type is natural disasters, with main types of hazards such as earthquakes, storms, droughts and epidemics. The second type is technological or man-made disasters, which involves main types as industrial accidents, such as chemical explosions, transport accidents and conflicts. It is not always clear if a disaster falls into only one of the two categories. Storms may seem as a natural caused hazard, but human factors may have contributed. For example, Katrina is seen as evidence that social induced emissions of carbon caused global warming and consequently this large size natural disaster. However, people perceive a disaster caused by either one of the two categories (Smith, 2006).

The degree to which a disaster is perceived as caused by humans or natural factors, has a significant effect on the willingness of people to donate money to charity (Zagefka et al., 2010). An example is the striking amount of donations after the 2004 tsunami that hit Asia and the less popular humanitarian crisis in Darfur. They had a comparable impact, e.g. in lives that were lost and

environmental degradation, but the amount of donations were higher in the natural caused disaster (Steele, 2007; Baker, 2005). However, more factors may have contributed to this donating difference, as for example, the number of Western tourists that were involved made the tsunami very relevant.

Zagefka and her colleagues (2010) found that people are more willing to donate to disasters that are caused by nature, rather than to the ones that are caused by human beings. They measured the mediation effect of perceived victim blame and self-help on donations. Their findings suggested that victims of natural disasters cannot be held responsible for the tragedy that happened, while victims can be blamed more for the disaster if it is perceived as humanly caused. The extent to which they tried to help themselves also explained the result. This was perceived higher for victims of natural disasters compared to the ones that suffer from a man-made disaster.

5

2.2 Proximity: nearby vs. faraway

It is a fact that a disaster may happen in your own neighbourhood, as well as somewhere in another country. According to Trope et al. (2007), spatial distance to events has an influence on mental construal and guide people their expectation, evaluation and behaviour. The four psychological dimensions that the authors categorize are time, space, social and hypothetical distance context. In this research, we treat spatial distance in particular. The psychological distance of close events are construed more concretely and at a lower level, whereas psychologically faraway events are construed more abstractly and at a higher level. In addition, actions of spatial distant events are rather seen as ends than as means to an end (Fujita et al. 2006; Liberman & Trope, 2008).

Spatial distance corresponds with social resemblance. For instance, People from the same neighbourhood have more similarities than people from different countries. Past findings show that the likelihood to help is higher when they perceive the other as similar to oneself (Levine, Cassidy, Brazier & Reicher, 2002; Levine & Thompson, 2004). Also proximity influences emotional response as negative incidents trigger less negative emotional arousal when they are faraway compared to when they are near (Wieser & Pauli, 2008; Williams & Bargh, 2008). The authors found that people were more emotionally triggered (e.g. empathy and concern) by near vs. far events, which resulted in more prosocial actions.

The most important finding regarding proximity is from Touré-Tillery & Fishbach (2017). The authors researched for the spatial distance effect on charity donating. They found that people are more willing to donate to nearby versus far away causes. Another finding of their study is that people expect nearby donations as more impactful. This current research specifically test the effect of proximity of the two disaster types on prosocial behaviour in terms of supporting charity in non-monetary terms instead of non-monetary donations.

In summary, it can be stated that the type of disaster as well as the spatial distance to a disaster effect donating behaviour. Before the hypotheses in this study will be tested, an overview of the drivers that motivate donating behaviour will be discussed in the next section, to understand the underlying concept of charitable behaviour.

2.3 Mechanisms that drive donating decisions

6 to their well-designed review, there are eight mechanisms that contribute to donating (see table 1 for an overview).

Mechanism Definition

Need The actions of individuals or charitable organizations who communicate a need for help to (potential) donors

Solicitation The mere act of (potential) donors that are being solicited

Costs/benefits The perceived costs and benefits for (potential) donors associated with giving

Altruism The pure care of (potential) donors for the charitable organization’s output and consequences of giving to the ones who need it (beneficiaries)

Reputation The social consequences of donations for the (potential) donor Psychological costs

and benefits

The psychological costs and benefits of the donor associated with giving, e.g. increase of a positive mood and/or reduction of the feeling of guilt

Values The influence of personal attitudes and values that (potential) donors endorse on donating

Efficacy The perception of (potential) donors that their donation actually makes a difference

Table 1: Overview of donation mechanisms and its definitions provided by Bekkers & Wiepking (2010).

Below, the mechanisms that drive charity donations will be briefly explained in more detail.

2.3.1 Mechanism 1: Awareness of need

The first determinant of giving that Bekkers and Wiepking (2010) provide is the awareness of need. People can have material needs (e.g. shelter or food), social needs (e.g. family or friends), and psychological needs (e.g. autonomy or relatedness). The public have to become aware of the need for relief funds resources for disaster victims. A study of Lee & Farrel (2003) argued that perceived need is related positively with monetary giving. Actions of charity organizations are required to show the public this beneficiary need.

7

2.3.2 Mechanism 2: Solicitation

The second mechanism that drives philanthropy is solicitation. It is the mere act of being solicited to donate (Bekkers and Wiepking, 2010). Most donations are a result of a response to an individual or charity’s solicitation. Bekkers (2005) and Slaughter, Kang, and Tax (2003) found that solicitations greatly increase the contributions to a cause. The likelihood of giving increases as well when active solicitating is applied, rather than passive solicitating, which implies many vs. only one opportunity to donate (Lindskold, Forte, Haake, & Schmidt, 1977). Moreover to the article of Bekkers and Wiepking, Glazer (1996) argued that when the potential donor has the opportunity to signal their wealth when they respond to a solicitation, the likelihood to make a contribution increases.

2.3.3 Mechanism 3: Costs and benefits

The third philanthropy driver is the cost and benefits that are linked to donating and mostly described in economic journals (Bekkers and Wiepking, 2010). The authors provide an extensive number of papers that studied the effects of the costs or price of giving empirically (e.g. Khanna & Sandler, 2000; Simmons & Emanuele, 2004; Duncan, 2004; Eaton, 2001). Giving monetary resources to charity obviously costs money. When the costs of a donation lowers, the amount of donations increases (Karlan & List, 2007). Despite monetary costs, donators may perceive other obstacles, such as physical discomfort, and timing constraints (Jiobu, & Knowles, 1974; Pharaoh & Tanner, 1997). Moreover, social capital may lower the transaction costs of donating (Bryant, Slaughter, Kang, & Tax, 2003). The authors define social capital as the social network that people participate in, which can be used to gain information about donating opportunities and the access to donating markets. As a result, the associated giving costs will be lowered.

The benefits refer to the incentives that people gain for donating and are characterized as an exchange as the contribution of giving has similarities with buying (Bekkers & Wiepking, 2010). Several economists researched the effects of gifts. These additional benefits of gifts increase donation numbers (Andreonie & Petrie, 2004). For instance, lottery related giving with a chance of winning money (e.g. Landry, Lange, List, Price, & Rupp, 2006; Apinunmahakul & Devlin, 2004) and material gifts that are provided by the aid organization to the donor (e.g. Alpizar, Carlsson, &

Johansson-Stenman, 2007; Chen, Li, & MacKie-Mason, 2006). However, Bekkers and wiepking (2010) concluded that there is no strong foundation that these gifts enhance philanthropy as some

8

2.3.4 Mechanism 4: Altruism

Altruism is categorized as the fourth giving mechanism. It is defined as the care of individuals for the causes of the charitable organizations and/or the consequences of the contributions for beneficiaries (Andreoni, 2006). The positive consequences of altruism can be solely attributed to the beneficiaries and reside outside individuals. Fong (2007) stated that there is unconditional altruism on the one hand, which is defined as the motivation of individuals to help others at a cost to themselves, without regard of the characteristics or intentions of the recipients. On the other hand there is conditional altruism, whereby donating motives (rewards vs. punishments) are based on what people deserve. Many empirical studies describe the “crowding out” effect of which individuals act completely unselfish in regard to the need of others (Kingma, 1989; Brooks, 1999; Diamond, 1999; Marcuello & Sallas, 2001). However, these studies show that people never act entirely unselfish. Therefore, the altruism effect exist, but is never perfect.

2.3.5 Mechanism 5: Reputation

The fifth determinant of donating according to Bekkers and Wiepking (2010) is reputation and is most often investigated in economic and psychological journals. The researchers refer this mechanism to the social consequences of donations for the (potential) donor. In the social environment of charitable contributions, donors will be verbally or non-verbally rewarded or punished for their (non) aid. Those individuals who do give gain status and are socially respected by their peers (Wiepking, 2008). Therefore, people favour that their donations to relief funds can be seen or are known by others (Andreonie & Petrie, 2014). Because of that, face-to-face solicitations are more effective compared to e.g. phone call requests or monetary donations in envelopes (Brockner, Guzzi, Kane, Levine & Shaplen, 1984; Thorton, Kirchner & Jacobs, 1991). In addition, Harbaugh (1998) found that two types of benefits derive from donating, namely ‘intrinsic benefit, that is derived from a person’s own knowledge of what he/she has given, and ‘prestige benefit’, which the donor will get when other people know what he/she has given. The author also states that large donations may create positive reputation as well as newsworthy events.

2.3.6 Mechanism 6: Psychological benefits

Wiepking and Bekkers (2010) argue psychological benefits as the sixth mechanism and is often investigated by (social) psychologists. They authors state that a majority of conducted studies found that giving contribute to one’s self-image in terms of being, for example, an altruistic,

9 positive mood, decreases feelings of guilt, satisfies desires to show gratitude and increases feelings of doing moral good. In literature, you may find these mechanism labelled as “warm glow” or “joy of giving” (e.g. Andreoni, 1989; Batson & Shaw, 1991; Crumpler & Grossman, 2008). For example, Anik, Aknin, Norton, & Dunn (2009) found that helping others indeed lead to a higher level of positive self-image. On the other hand, the authors found in their second study that advertising these

psychological benefits diminishes the intrinsic motivation, which consequently decreases charitable giving. For instance, asking people to donate in order to be happy results in lower donation numbers.

2.3.7 Mechanism 7: Values

Bekkers and Wiepking (2010) labelled the seventh donating driver mechanism values. According to the authors, giving to charity will be more or less attractive based on the personal attitudes and values individuals endorse. People are more likely to donate when they want to make the world a better place. For example, people with altruistic values (Bekkers & Schuyt, 2008), prosocial values (Van Lange, Van Vugt, Bekkers, & Schuyt, 2007), or with moral principles of care (Schervish & Havens, 2002; Wilhelm & Bekkers, 2010) have higher giving intentions. Moreover, Bennett (2003) argued that personal values have a strong influence on the selection of a charity cause to donate to. Furthermore, he found that people donate to organizations that they admire and to the ones that correlate with their own personal values. For instance, people who value animal welfare because they have a deeply beloved pet are more likely to give to animal welfare causes. Another example is when someone has a beloved friend or family with a specific illness like cancer, that person is more likely to donate to those related health charities instead of other aid causes.

2.3.8 Mechanism 8: Efficacy

The last mechanism according to Bekkers and Wiepking (2010) is efficacy. They define this mechanism as the extent to which the donator think that their monetary gift actually makes a difference to the cause they support. Several studies reveal that when people have the idea that their contribution makes no difference, the likelihood that they will donate decreases (e.g. Duncan, 2004; Wiepking, Madden, & McDonald, 2010; Diamond & Kashyap, 1997). It is called

pseudoinefficacy when people are unwilling to help the ones in need when they know that not everybody in that situation can be helped (Västfjäll, Slovic, & Mayorga, 2015). The authors state that donors get demotivated to give and their feeling of “warm glow” to the ones they can help

10

2.3.9 Conclusion

In sum, the above overview provides the eight mechanisms that drive charitable giving according to Bekkers and Wiepking (2010). Those eight determinants bolster the underlying concept of why people do and do not give money to charitable organizations. Therefore, those mechanisms can always be kept in mind why people choose or do not choose to donate. It can be assumed that those mechanisms, just like in monetary donations, also affect charitable giving in non-monetary terms as it is a theoretical framework of intrinsic and extrinsic donor motivations.

Now we know how the types of disasters and proximity affect monetary charitable behaviour and how these underlying donating mechanisms influence this donating decision making. In order to explore if the newsworthiness of a disaster affects non-monetary charitable giving, the concept of newsworthiness will be discussed in the next section. Note that a disaster’s newsworthiness is closely linked to the first mechanism awareness of need. The question is how (potential) donors become aware of the need for help? Therefore, literature on the field of journalism is examined to look for necessities of stories to make charity appeals more interesting to the public.

2.4 Newsworthiness of a story

As mentioned in the introduction, charitable organizations compete for the donator’s help via different communication channels. In their advertising, they tell the story of a cause for which resources are needed. As well as news stories created by journalists, those emergency aid marketing expressions also consist of news factors that influences people’s behaviour. However, the influence of newsworthiness of a cause on supporting that charity has not been academically researched for yet. Therefore, newsworthy indicators from a journalistic perspective will be used in a charity setting to research for its effect on people’s charitable behaviour. This study explores how the types of disasters and its proximity influences the newsworthiness of the cause and consequently how this newsworthiness affects non-monetary charitable support.

11 influenced by eight indicators, namely timeliness, proximity, importance or impact, interest, conflict or controversy, sensationalism, prominence and novelty or unusualness and influence each other which determine a story’s news value (Shoemaker, Danielian, & Brendlinger, 1991). Other studies show similar indicators that determine the newsworthiness of a story (e.g. Brighton, & Foy, 2007; Palmer, 2000). A story must satisfy a high level of at least one of these indicators, but preferably more of those indicators to be selected and attract the interest of the public (Harcup & O’Neill, 2017).

As people have to become conscious of the problem where aid organizations need money for, a high level of newsworthiness may be beneficial to charities in their communication of the need of help to (potential) donors. Therefore, examining the disaster type and its proximity presented in the charity’s advertisement may influence the newsworthiness of the message and henceforth the willingness to support charity. Note that in this study, a cause its proximity will be investigated as an independent variable on non-monetary charity support and that proximity happens to be a news indicator as well. The eight indicators of Shoemaker et al. (1991) will be elaborated in the next sections respectively to give an extensive definition of newsworthiness.

2.4.1 Timeliness

Whether and how a story is covered in the media, may depend on the first characteristic of news, namely timeliness. It refers to whether the event is a recent happening or fresh story

(Shoemaker et al., 1991). Stories of events must be published quickly with respect to the occurrence of an event so it does not lose its news value (Berkowitz, P85. 1997). The author states that the level of attention to a cause may be less, when more significant global events are getting news coverage at the same time.

According to Smith et al. (2012), informing and empowering the public is critical during public health emergencies. Gaining and retaining media attention over time is essential to stop emerging hazards, such as infectious illness. If the level of high newsworthiness has decreased, which for this indicator means that the content is not be seen as new anymore, it is important that charities create story strategies that make the cause a fresh story to continue and retain the audience’s attention.

2.4.2 Proximity

12 eight news indicators. Therefore, the direct effect of proximity will be investigated on the willingness to support charity, as well as the exploration of the types of disasters and their proximity on the ad’s perceived proximity as a news indicator and consequently the effect of this perceived spatial distance on the willingness to support.

2.4.3 Importance or impact

The third news value indicator is importance or impact. It is about the perceived impact of the cause. There are two modes of thought in relief fund communication that determines the perceived impact, namely the affective mode and the calculative mode (Chaiken & Trope, 1999). The affective mode may be more central when the target of thought is personal, specific and vivid

(Epstein, 1994). Small, Loewenstein and Slovic (2007) found that donations to charity increased when identifiable lives are used in the communication, compared to statistical lives or the combination of identifiable with statistical lives. Therefore, the affective mode compared to the calculative mode can be stated as more effective in generating donations. For example, a message telling that the 7-year old Ethiopian girl Rokia is starving and needs help will be more effective than telling that 11 million people in Ethiopia need food assistance. Identifiable lives trigger a more emotional response

compared to statistical ones. This finding shows that the way the impact of a cause is framed affects the amount of raised donations.

2.4.4 (Human) interest

13 sympathy and giving. In a nutshell, a human interest angle can be very useful to attract the

audience’s attention and evokes emotional responses in a charity cause.

2.4.5 Conflict or Controversy

The fifth news value indicator that influences the audience’s attention is conflict or controversy. According to Shoemaker et al. (1991), human beings are interested in stories that consist of conflicts. Stories of conflicts are for example often about religion, politics, war, sports, and/or human rights (Mudd, 2014). As stated by the author, events that are in conflict with people’s ideas engages them emotionally. It is in the nature of human beings to choose a side and let their voice heard for their choice. Conflicts are about the experienced dissatisfaction with the

consequences of the situation shown to the public and as a result triggers people’s opinions.

2.4.6 Sensationalism

The sixth characteristic of news values is sensationalism. It is characterized as sensational, dramatic stories that amuses, arouses and/or entertain (Grabe, Zhou, & Barnett, 2001). Sensational news triggers more sensory and emotional reactions (Graber, 1994). Stories regarding celebrities, crime, sex, accidents as well as disasters are commonly seen as sensational (Davie & Lee, 1995).

Credibility is very important for relief funds organizations to gain the public’s trust (Gibelman & Gelman, 2004). Therefore, charities should be careful in the extent to which they sensationalize the cause. Bias and overhyping the truth of the story may result in less credibility and so support.

2.4.7 Prominence

The seventh indicator of newsworthiness is prominence. This value is about who is telling the story and the authority of the communicator (Schoemaker, Danielian, Brendlinger, 1991). Nowadays, celebrity endorsement is often implemented in firm’s marketing strategies to communicate a firm’s story and generate desirable campaign outcomes (Choi, Lee, & Kim, 2013). Celebrity endorsement is a form of brand promotion where a well-known person is used in a marketing campaign by using his or her fame and place in society (Keller, 2012). Such a marketing strategy increases audience’s attention, ad recall, and positive brand attitudes (Wann-Yih et al., 2012). Also charitable

14

2.4.8 Novelty or unusualness

The last newsworthiness indicator is novelty or unusualness. The characteristic of this indicator is the perception of a story that is different, astonishing or unusual (Shoemaker et al., 1991). Rare and uncommon events are more newsworthy than others (Lundman, 2003). Therefore, unique disasters, like the monstrous and devastating Asian tsunami of 2004, gain more media attention, which consequently influences charitable giving (Brown & Minty, 2008).

2.4.9 Conclusion

To summarize, all the independent variables in this study are discussed and we now know that type of disaster and proximity affects charitable giving. We also know which mechanisms drive donating behaviour. Furthermore, charities need public awareness for their causes so people participate in charitable behaviour. This awareness of need is the first donating driver and can be viewed from a journalistic perspective in terms of newsworthiness of the story. We saw that

newsworthiness consists of eight indicators that determine the attention and interest of the public in a story and so may influence prosocial behaviour. As the documented past findings are mainly about what influences donations in monetary terms, this research focuses on online charity support in non-monetary terms. Therefore, theory regarding online non-non-monetary charity support will be discussed in the next section to be able to answer this study’s research question on how the types of disasters and proximity influences non-monetary online charity support.

2.5 Online non-monetary charity support

Online fundraising is applied by most aid organizations to raise money. Social media is a critical factor in the success of online charity donations. According to the main findings of an analysis of over 40 sources on online giving statistics by the digital marketing agency Nonprofit Source (2018), 21% of donations went through social media in 2016. In addition, the agency argues that 55% of people who engage with charities on social media take some form of action towards the appeal of the charitable organization. This research dives deeper into those other non-monetary forms of individual online actions towards charities as most literature on charitable donations already describe the relation between drivers to support charity in monetary terms.

15 terms on Facebook is via petition signing. It is one of the most popular forms of online activism (Earl, 2006; Earl & Kimport, 2011). Zarett (1999) noted in his research that petitions are a form of a long-lasting social movement tactic. When we talk about online petitions, it can generate leads to

charities’ website or landing page of the cause (Elliot & Earl, 2018). As mentioned in the introduction, signing petitions are an easy act of support and can easily go viral in individuals’ online networks (Corrado, 2017). After signing a petition, the charitable organization may approach the supporter with a request to donate or to support the causes of the organization in the future. Therefore, the potential donor who shows some level of interest in the cause may be turned into an actual donor.

The effects of type of disaster and its proximity in an online charity’s appeal on Facebook on supporting the cause; i.e. sign the charity’s petition, like the ad, share the ad, and/or comment on the ad, will be investigated in this research. Moreover, the effect of perceived newsworthiness of the cause on this form of charity support will be explored.

3. Conceptual model and hypotheses

16

Figure 1: Conceptual model of the research.

A direct effect between ‘type of disaster’ and ‘willingness to online support a charity cause’ is hypothesized. Zagefka et al. (2010) found that human beings are more willing to donate and higher amounts are elicited when disasters occurred in a natural form, versus humanly caused disasters. Therefore, it is hypothesized that people are as well more willing to support charity in non-monetary terms when the disaster is naturally caused. Thus:

H1: A disaster that is caused by nature (vs. man-made), generates a higher willingness to online

support the cause (sign the Facebook petition, like the ad, share the ad and/or comment on the ad).

In addition to the effect of disaster type, the direct effect of ‘proximity’ on ‘willingness to online support a charity cause’ is also hypothesized. Touré-Tillery and Fishbach (2017) stated that people are more willing to take action when donations are needed for causes that happened nearby, versus faraway. In detail, donations increases when perceived distance decreases. In this research, this outcome is also hypothesized for non-monetary charity support. From the second theoretical finding, the following hypothesis derived:

H2: A disaster that is perceived as nearby (vs. faraway), generates a higher willingness to online

17 The interaction effect of the type of disaster and proximity will be investigated as well. Based on the findings of Touré-Tillery and Fishbach (2017), who state that people are more willing to help nearby versus faraway causes, and of Wieser and Pauli (2008), and Williams and Bargh (2008) who found that nearby causes trigger a higher emotional response which results in more prosocial

actions, it is hypothesized that there is no significant difference in the willingness to support between the type of disasters when the disaster is nearby. It is therefore proposed that:

H3: The willingness to online support a natural or man-made disaster (sign the Facebook petition,

like the ad, share the ad and/or comment on the ad) does not differ when the disaster is nearby.

Finally, the underlying process via the newsworthiness dimensions will be explored. An exploratory hypothesis with regard to the newsworthiness of the charity’s ad is drawn to examine the effect of the news indicators on the willingness to online support charity on Facebook. As the documented findings regarding the news indicators of Shoemaker et al. (2010) mention that stories must satisfy a high level of the indicators to attract the public’s interest (e.g. Harcup & O’Neill, 2017; Smith et al., 2012), it is expected that higher perceived levels of the indicators increases the

willingness to online support the charity cause. Hence:

A higher perceived value of the newsworthiness indicators (vs. low), increases the willingness to online support charity (sign the Facebook petition, like the ad, share the ad and/or comment on the ad).

4. Research Methodology

4.1 Experimental design

18

Figure 2: Survey conditions US

–

–

4.2 Procedure

The questionnaire was created with the help of the Qualtrics survey platform. It opened with an introduction to participants (pps) to inform them of the purpose of this study and to assure them that the data will be treated anonymously and confidentially. After, participants had to fill in some questions regarding their attitude, thoughts, and behaviour about this research its topic.

First, the participants signed the informed consent. Second, the pps were asked if they were Facebook active, which means that they should have logged in to their account at least once a month. Third, they were presented with one of the four conditions and asked to imagine themselves with a Facebook message of the International Committee of the Red Cross on their timeline. The ad communicated that the viewer can help victims of a disaster in need by signing the petition. After, the pps were asked if they would support this message by liking the ad, commenting on the ad, share the ad, and/or go the petition (for all in which, 1 = Yes, I would; 2 = No, I would not). Fourth, pps were asked to indicate on a seven point Likert-scale (1 = strongly disagree; 7 = strongly agree) how they perceived the situation in the advertisement. They were confronted with 23 statements that corresponded with the eight news indicators explained by Shoemaker et al. (1991). The items that were used were partly derived from the theory around the news indicators in the study of

19 the complete survey with all the items can be found in appendix 1. Sixth, pps were asked if they found The International Committee of the Red Cross trustworthy, credible, authentic and important on a 7-point Likert Scale (1 = strongly disagree; 7 = strongly agree) (M = 5.4). At last, demographic characteristics were asked regarding their gender (male vs. female), age (1-100), degree of

completed school level, and current employment status. The sample characteristics are provided in paragraph 4.3 of this thesis.

Construct To what extent do you perceive the situation

in the advertisement as:

Indicator unusualness To be rare

To be uncommon To be unusual

Table 2: Items in the survey to measure unusualness.

4.3 Sample

To test the hypotheses, the experiment was conducted among 418 adults from an Mturk consumer panel in the U.S. Therefore, the population under study is all American citizens. Next, they were invited to take part in a 3-min study in exchange for $0.45. Seventeen respondents did not pass the attention check and were eliminated from the analysis. As a result, a total of 401 participants formed the sample in this study on which to run the analysis. The sample consisted of 181 (45.1%) men and 220 (54.9%) women with a mean age of 39 years (M = 39.10, SD = 11.95). An assessment for relations between gender on charity engagement was conducted in form of a Chi-Square test. The result showed no significant gender difference on support as X² (1, N=401) = 2.569, p = .109. A Binary Logistic Regression also showed no effect of age on charity engagement p > .672 with an odds ratio Exp (B) of .996.

The vast majority completed their bachelor' degree (42.4%) and are employed full time (63.8%). Characteristics of the sample concerning their highest degree of school level participants completed and their current employment status can be found in table 3. Analysis of relations between the highest degree of completed school level on charity engagement and current

20

Highest completed degree of school level

Frequency in % Current employment status Frequency in %

Less than a high school diploma

0.5 Employed full time (35 or more hours per week)

63.8 High school degree or

equivalent (e.g. GED)

12.0 Employed part time (up to 34 hours per week)

10.0 Some college, no degree

(e.g. AA, AS)

31.7 Unemployed and currently looking for work

3.2 Bachelor’s degree (e.g. BA,

BS)

42.4 Unemployed and currently not looking for work

0.2 Master’s degree (e.g. MA,

MS, Med)

10.2 Student 1.2

Professional degree (e.g. MD, DDS, DVM)

2.2 Retired 4.2

Doctorate (e.g. PhD, EdD) 1.0 Homemaker 6.7

Self-employed 9.0

Unable to work 1.5

Table 3: Sample characteristics; school level and employment status (n=401).

4.4 Data: factors and reliability of measures

The dataset provided a variety of factors for the purpose of this investigation. First, to check whether the eight indicators of Shoemaker et al. (1991) were well represented within the news survey items, a Factor Analysis (FA) was performed. The FA analysis revealed four factors with eigenvalues greater than 1, a KMO index greater than .5, and communalities greater than .4. They accounted for approximately 60% of the cumulative variance. The scree plot and rotated factor loadings are provided in appendix 2. It can be concluded that the 23 items correspond with four news indicators instead of eight. However, from a theoretical perspective, this research will be based on the eight news indicators of Shoemaker et al. (1991). Therefore, timeliness, proximity, impact, interest, conflict, sensationalism, prominence and unusualness will be individually used to explore the effect of the news indicators on the willingness to online support charity.

Second, to create eight news indicators, the internal consistency of 23 newsworthiness survey items were measured with Cronbach’s Alpha (CA). Table 4 reports the measures and reliability of the factors. All indicators have a CA > 0.7, except for sensationalism and prominence. As

21 the Red Cross) was also measured. The items trustworthiness, credibility, authenticity and

importance were internally consistent and its CA is as well provided in table 4.

Scale Number of items Cronbach’s Alpha Mean SD

Timeliness 2 .784 4.86 1.51 Proximity 3 .911 2.70 1.59 Impact 3 .888 5.21 4.41 Interest 4 .897 4.41 1.53 Conflict 3 .739 4.51 1.48 Sensationalism 3 .672 4.09 1.34 Prominence 2 .239 5.14 1.75 Unusualness 3 .875 3.58 1.53 Reliability of ICRC 4 .969 5.37 1.61

Table 4. Measures and reliability.

Finally, the four options of the willingness to online support the cause were computed into the variable engagement. In detail, when a participant would supported the cause via only one or more of the options, they showed a form of engagement and so support. When pps would not support the cause via any of the options, they did not show a form of engagement (0 = not engaged ; 1 = engaged).

5. Results

5.1 Descriptive Statistics

Before the hypotheses were tested, the percentage of people engaging, signing the petition, liking the ad, sharing the ad, and commenting on the ad for the type of disaster and proximity conditions were analysed. Chi-Square results are also provided to show the relations between disaster type, proximity, and the support options. After, Binary Logistic Regression were performed to test the hypotheses and is provided in paragraph 5.2.

22 only frequencies and significant relations between those variables and its interaction on charity support will be tested for with binary logistic regression in paragraph 5.2.

The table also shows that more people are willing to engage than the one who do not (51.4%). The highest willingness to support is via signing a petition (44.1%), followed by liking the ad (33.9%), sharing the ad (21.9%) and with commenting on the ad as least popular (8.5%).

Type of disaster / proximity

Charity support

Engagement % Petition % Like % Share % Comment %

Yes Chi² Yes Chi² Yes Chi² Yes Chi² Yes Chi²

Oil spill 54.0 X2 = X, p = .294 45.0 X2 = X, p = .729 36.0 X2 = X, p = .379 24.5 X2 = X, p = .218 10.5 X2 = X, p = .147 Flood 48.8 43.3 31.8 19.4 6.5 US 52.2 X2 = X, p = .728 45.8 X2 = X, p = .510 37.3 X2 = X, p = .150 24.9 X2 = X, p = .155 10.0 X2 = X, p = .289 Africa 50.5 42.5 30.5 19.0 7.0 Overall 51.4 44.1 33.9 21.9 8.5

Table 5: Frequencies and Pearson Chi-Square test of type of disaster and proximity on charity support.

In order to analyse whether type of disaster (natural vs. man-made) and proximity (nearby vs. faraway) statistically differ on every support option, repeated cross tables with Chi-Squares tests were performed and also reported in table 5. The Chi-Square tests were not significant, e.g. X² (1, N=401) = 2.101, p > .147 for proximity on commenting on the ad as smallest p-value. The data suggests that the type of disaster (natural vs. man-made) and proximity (near vs. far) do not significantly differ on the five support options. The specific influence of the disaster types and proximity were tested in the next section with Binary Logistic Regressions.

5.2 Hypotheses testing

Binary Logistic Regressions were performed to predict the relationships between the categorial variables type of disaster, proximity, and its interaction on the five dichotomous DV options of the willingness to support charity, i.e. engagement; petition signing, liking the ad, sharing the ad, and commenting on the ad. In order to analyse the main effects and the interaction effect of type of disaster and proximity on every support option, a total of five regressions were performed.

Binary Logistic Regression

23 conclusion, if the type of disaster is natural (flood) vs. man-made (oil spill) does not lead to a higher willingness to online support charity via Facebook. It can be stated that hypothesis 1, ‘a disaster that

is caused by nature (vs. man-made), generates a higher willingness to online support the cause (sign the Facebook petition, like the ad, share the ad and/or comment on the ad’, is not supported.

Type of disaster (IV)–

Charity support (DV)

B Sig. Exp (B) Effect

(non)significant Overall engagement -.102 .717 .903 n.s. Petition -.216 .447 .806 n.s. Like -.090 .757 .914 n.s. Share .174 .596 1.190 n.s. Comment .417 .386 1.517 n.s.

Table 6: A Binary Logistic Regressions’ output of type of disaster on the charity support options.

Second, the output of the regressions in order to investigate the effect of proximity on the five support options are reported in table 7. The results of the regressions do not show a significant effect for all support options (p > 0,051). The odds ratios (Exp B) vary between 0.549 and 0.685. However, when a disaster is near (vs) far, people are 0.549 times less likely to like the Facebook ad is statistically significant when a confidence interval of 90% is applied (α < 0.1). In this study, the conventional standard of 0.05 is kept. Therefore it can be stated that disasters that happen nearby (US) compared to faraway (Africa) do not increase the likelihood to online support charity on

Facebook. In conclusion, hypothesis 2, ‘a disaster that is perceived as nearby (vs. faraway), generates

a higher willingness to online support the cause (sign the Facebook petition, like the ad, share the ad and/or comment on the ad’, is not supported.

Proximity (IV)–

Charity support (DV)

B Sig. Exp (B) Effect

(non)significant Overall engagement -.378 .182 .685 n.s. Petition -.419 .143 .658 n.s. Like -.599 .051 .549 Significant* Share -.487 .178 .615 n.s. Comment -.534 .364 .586 n.s. *Significant at 1% level

Table 7: A Binary Logistic Regressions’ output of type of proximity on the charity support options.

24 Facebook. Therefore, hypothesis 3, ‘the willingness to online support a natural or man-made disaster

(sign the Facebook petition, like the ad, share the ad and/or comment on the ad) does not differ when the disaster is nearby’, is also not supported.

Type of disaster* Proximity (IV)–Charity

support (DV)

B Sig. Exp (B) Effect

(non)significant Overall engagement .627 .119 1.871 n.s. Petition .573 .156 1.773 n.s. Like .575 .177 1.777 n.s. Share .268 .583 1.308 n.s. Comment .257 .732 1.293 n.s.

Table 8: A Binary Logistic Regressions’ output of type of disaster’s and proximity’s interaction on the charity support options.

5.3 Analysing underlying process of newsworthiness

As pointed out in the last paragraph, it can be concluded that the regressions did not show support for the hypotheses in this study. Moreover, to analyse the underlying process of the ad’s newsworthiness on the online charity support options, a total of eight factorial ANOVA’s with interaction of type of disaster and proximity on the newsworthiness indicators were performed. After, Binary Logistic Regressions of the news indicators, that were significantly affected by the type of disaster and proximity, were performed to explore the effects of these news indicators on the willingness to online support the cause on Facebook.

5.3.1 Factorial ANOVA’s: disaster type and proximity on newsworthiness

In order to analyse the influence of type of disaster and proximity on newsworthiness, a 2 (type of disaster: nature vs. man-made) x 2 (proximity: near vs. far) ANOVA’s with interaction on the eight news indicators were performed. The results of the repeated measures ANOVA are provided in table 9.

The main effect of type of disaster on the news indicator conflict proved to be significant. An oil spill leads to higher perceived conflict (M = 4.67) than a flood (M = 4.35), F(1,397)=4.569, p < .033. No other main effects for proximity were significant (p’s > .117).

25 .033, as well as higher sensationalism (M = 4.24 vs. 3.94), F(1,397)= 5.148, p < .024, and higher proximity (M = 3.35 vs. 2.05), F(1,397)= 78.883, p < .001. No other main effects for proximity were significant (p’s > .162). In addition, none of the interaction effects proved to be significant, p’s > .061.

Newsworthiness Indicator (DV)

p-values and means

Type of disaster (IV) M (natural vs. man-made) Proximity (IV) M (near vs. far) Type of disaster * proximity (IV) Timeliness .546 4.82 vs. 4.91 .956 4.86 vs. 4.87 .807 Proximity .458 2.64 vs. 2.77 .000 3.34 vs. 2.01 .135 Impact .117 5.27 vs. 5.16 .885 5.21 vs. 5.23 .117 Interest .129 4.29 vs. 5.51 .033 4.57 vs. 4.24 .462 Conflict .033 4.35 vs. 4.67 .425 4.57 vs. 4.45 .086 Sensationalism .953 4.08 vs. 4.10 .024 4.24 vs. 3.94 .061 Prominence .169 5.02 vs. 5.26 .801 5.12 vs. 5.16 .117 Unusualness .210 3.48 vs. 3.68 .162 3.69 vs. 3.47 .108

Table 9: Two-way ANOVA’s with interaction for type of disaster and proximity on newsworthiness indicators.

From this, one can conclude that a man-made disaster (oil spill) leads to higher perceived conflict and a disaster that happened nearby (US) leads to higher perceived interest, sensationalism and proximity. Because of this reason, only those four significant indicators will be tested for its effect on the willingness to support a charitable organization.

5.3.2 Binary Logistic Regression: newsworthiness on online support options

Although there is no overall effect of type of disaster and proximity on charity support, the effect of those variables on newsworthiness and so on the willingness to online support charity is explored. In order to analyse whether the news indicators conflict, interest, sensationalism and proximity lead to a higher willingness to support charity, Binary Logistic Regressions were performed on the four support options, petition signing, liking, commenting, sharing as well as on engagement. Thus, a total of five regressions were performed. The results of the tests are provided in table 10.

The results of this regression show a significant positive effect of interest on engagement (Exp B = 2.846, p < .001). As well as a significant effect of interest on the other support options; petition signing (Exp B = 3.079, p < .001), liking (Exp B = 2.602, p < .001), sharing (Exp B = 2.406, p < .001), and commenting (Exp B = 1.375, p < .001). This means that people are 1.375 to 3.079 times more likely to show some form of online support when the level of interest increases with one unit.

26 Facebook ad when the perceived level of proximity increases with one unit. No other effects of conflict, sensationalism, and proximity were significant (p > .155).

News indicator

p-values

Engagement Petition Like Share Comment

Conflict .200 .399 .172 .778 .306

Interest .001 .001 .001 .001 .001

Sensationalism .266 .343 .373 .171 .552

Proximity .155 .303 .001 .001 .001

Table 10: A binary logistic regression for conflict, interest, sensationalism and proximity on the charity support options.

The results indicates that a higher perceived value of the news indicator interest increases the likelihood to support charity in all support forms. In addition, a higher perceived proximity increases the likelihood to like, share, and comment. Therefore, the exploratory hypothesis ‘a higher perceived value of the newsworthiness indicators (vs. low), increases the likelihood to support charity’ only counts for the news indicators interest and proximity’.

6. Conclusion and discussion

The present article examined the relationship between the type of disaster (natural vs. man-made) on charity support in non-monetary terms (sign the petition, like the ad, share the ad and/or comment on the ad) via a charity appeal of the International Committee of the Red Cross to sign a petition on Facebook. Engagement was created as the fifth dependent support variable, which means that a participant showed support via either one or more of the support options. The

moderating role of proximity (near vs. far) between the type of disaster and charity support was also examined. In addition, the role of eight news indicators according to Shoemaker et al. (1991) were explored to test for the mediation effect of newsworthiness on charity support.

The prediction was that natural disasters generated a higher willingness to support the charity cause. However, our results were not in line with the hypothesis. The data suggest that it cannot be concluded that a natural disaster ensures a higher willingness to engage; sign the petition, like the ad, share the ad and/or comment on the ad. This result is in contradiction with the finding of Zagefka et al. (2010), who argued that natural disasters vs. man-made disasters increase the

27 not signal much about the prosocial identity of a person, while costly prosocial behaviour (paying money) does. People are more consistent in prosocial behaviour when money is involved, which may explain difference in outcomes when monetary versus non-monetary support is investigated.

Another point to make is that we saw from the finding of Andreonie & Petrie (2014) that people prefer that their prosocial action can be seen by others, which is very much the case via Facebook support as people can show their prosocial behaviour to their peers in their online network. Also, Corrado (2017) stated that supporting causes via Facebook is an easy act of support compared to monetary giving. Because of these facts, the different in outcomes between the past finding of type of disaster on monetary support and the current study of type of disaster on non-monetary support is surprising.

Moreover, in contrast with the study of Touré-Tillery and Fishbach (2017), no evidence was found that disasters that happened nearby versus faraway increased the willingness to engage; sign the petition, like the ad, share the ad and/or comment on the charity’s Facebook ad. The data

suggest that the spatial distance of disasters does not influence the willingness to support a cause. As well as with the difference in finding regarding type of disaster, the difference between monetary and non-monetary support in terms of signalling someone’s prosocial identity may explain the difference.

In addition to the main effects on charity support, the results did not reveal the predicted interaction of type of disaster and proximity on the willingness to engage; sign the petition, like the ad, share the ad and/or comment on the ad. The willingness to support by participants that were faced with either a natural or a man-made disaster ad did not differ between when that disaster was either nearby or faraway. Therefore, it can be suggested that the effect of nearby causes do not drown out the effect of disaster type on the willingness to support.

This study also explored the effect of the type of disaster and proximity on charity support via the perceived newsworthiness of the ads. According to Shoemaker et al. (1991), eight indicators determine the perceived news value of a message, namely timeliness, proximity, impact, interest, conflict, sensationalism, prominence, and unusualness. The catastrophes used in this research were put in the spotlights of those eight determinants that make events newsworthy. The results

supported a positive effect of type of disaster on the news indicator conflict. A disaster that was caused by man (oil spill) lead to higher level of perceived conflict compared to a naturally caused disaster (flood). Proximity showed a positive effect on the news indicators interest, sensationalism, and proximity itself. A disaster that was nearby (US) lead to higher level of perceived interest,

28 proximity showed a positive effect on the willingness to engage; sign the petition, like the ad, share the ad and/or comment on the ad. These findings suggest that when a charity’s story of a cause has a higher perceived level of interest and proximity, people are more willing to online support the cause via Facebook. As interest is mainly about evoking emotional responses (Shoemaker et al. 1991; Verrochi, 2009), these findings can be supported by the finding of Cho and Gower (2006) who argued that an emotional angle in a message influences positive behaviour. Also the finding of Trope et al. (2007) supports the outcome as they stated that perceived spatial distance also influences

behaviour. Note that no main or interaction effects were found between type of disaster and proximity on the willingness to online support. Therefore, the results only provide further exploratory evidence for our hypothesis about the effects of the news indicators.

In conclusion, this thesis contributed to the existing literature in two ways. First, by the finding that the type of disaster and proximity do not influences the willingness to online support charity in non-monetary terms, while past findings show that they do influence the willingness to support charity in monetary terms. Second, by highlighting that the perceived newsworthiness indicators interest and proximity may beneficially influence the willingness to support charity in non-monetary terms.

7. Limitations

and future research

As well as all other studies, this research has got limitations that are worth mentioning. First, the findings on the willingness to support charity was related to a Facebook message context.

29 Second, this study was an experiment in which people knew that they took part in.

Therefore, hypothetical and actual behaviour on charity appeal responses may differ in this study from the real world. People are not always congruent in what they say they will do and what they actually do, known as the theory of planned behaviour (e.g. Azjen, 1991). Future A/B testing in a real setting can test for the actual effects compared to the indicated behaviour by the participants. To illustrate, the four ads that were used in this study could be randomly presented to followers of the International Committee of the Red Cross to test the influence of the type of disaster, proximity, and the news indicators in a real environment.

Third, this research did not test if the natural and man-made caused disasters were actually perceived as natural and humanly caused. This thesis concluded that a flood is natural caused and an oil spill is caused by human beings. However, participants in the research may perceive this

distinction differently and may put other labels on the type of disaster which may have influenced the outcome. Future research should control for this validity aspect.

References

Alpizar, F., Carlsson, F., & Johansson-Stenman, O. (2007). Anonymity, reciprocity and conformity: Evidence from voluntary contributions to a natural park in Costa Rica. Journal of Public

Economics. 92 (5-6): 1047-1060.

Andreoni, J. (1990). Impure altruism and donations to public goods: A theory of warm-glow giving.

Economic Journal. 100: 464-477.

Andreoni, J. (2006). Philanthropy. In L.-A. Gerard-Varet, S.-C. Kolm & J. M. Ythier (Eds.), Handbook of giving, reciprocity and altruism (pp. 1201-1269). North-Holland: Elsevier.

Andreoni, J., & Petrie, R. (2004). Public goods experiments without confidentiality: A glimpse into fund-raising. Journal of Public Economics. 88 (7): 1605-1623.

Anik, L., Aknin, L. B., Norton, M. I., & Dunn, E. W. (2009). Feeling good about giving: the benefits (and costs) of self-interested charitable behaviour. Harvard Business School. 10-12.

Azjen, I. (1991). The theory of planned behavior. Organizational behavior and Human Decision

Processes. 50 (2): 179-211.

Bekkers, R., & Wiepking, P. (2010). A literature review of empirical studies of philantropy: eight mechanisms that drive charitable giving. Nonprofit and Voluntary Sector quarterly. 40 (5): 924-974.

Bennett, R. (2003). Factors underlying the inclination to donate to particular types of charity.

International Journal of Nonprofit and Voluntary Sector Marketing. 8 (1): 12-29.