WHETHER YOU “LIKE” IT OR NOT: THE EFFECT OF
“LIKING” A CHARITY ON PEOPLES’ WILLINGNESS TO
DONATE
by RIK HUTTINGAUniversity of Groningen Faculty of Economics and Business
MSc Marketing 17 January 2014 Spaanse Aakstraat 85 9741 CV Groningen (06) 21175011 h.j.huttinga@student.rug.nl Student number: 1891081 Supervisor: Dr. M.C. Leliveld
2 Preface
Hardly an hour after the introduction session I explained the topic of Cause Related Marketing to my girlfriend Bernice. She came up with some examples of Cause Related Marketing on Social Media, and from that moment on, the idea to assess the possible substitution effect of liking was born. I was very enthusiastic from the start till the end, however I also experienced troubles during the last half year. Due to a torn ligament and the subsequent surgery, both my mobility and concentration were reduced a lot, causing me to miss a number of feedback sessions with Ms. Leliveld and causing me to slightly fall behind. Fortunately, I could have feedback sessions with Ms. Leliveld over the phone, which helped me a lot with writing my thesis and returning back to schedule. Therefore, my first word of thanks goes to Marijke Leliveld for her support and flexibility during the last few months. Her feedback, remarks and explanations were very constructive and helped me a lot.
A second word of thanks goes to all the respondents who took time and effort to fill in my questionnaire. I could not test my hypotheses and realize this Master Thesis without the help of these 114 respondents.
3 Table of contents
ABSTRACT ... 4
1 INTRODUCTION ... 5
2 THEORETICAL FRAMEWORK ... 9
2.1 Moral Licensing or Moral Self Image ... 9
2.2 Liking on a Social Media platform ... 10
2.3 Willingness to Donate ... 11
2.4 Description of relations and hypotheses ... 12
3 RESEARCH DESIGN ... 15
3.1 Design and Participants ... 15
4 Abstract
5 Introduction
Recently, together with the World Food Program, the Unilever Foundation launched a campaign on Facebook. People were asked to like a certain page, and per like, one meal would be donated to the World Food Program. In this way, the like seems a very nice way to support a good cause. However, might it be possible that it prevents you from donating real money to charity? That is, in small characters at the bottom of the page was written: There is a maximum of 1.000.000 meals, corresponding with a donation of €200,000 of Unilever to the World Food Program””. In other words, a like would only support the charity by 20 cents, an amount which is probably much lower than when you would make a direct donation to the charity. In this thesis, I will focus on the influence of a like on a social media platform on peoples’ willingness to donate in general.
Nowadays, the behavior of people on social networks or the internet in general has become very important. Over the last decade, the media environment has changed a lot with the introduction of the internet. Due to the digitalization, online marketing has become a great medium to reach a large amount of people or potential customers. According to Stephen and Galak (2012), three types of media can be distinguished in online marketing: owned media (media channels that the company owns), paid media (advertising) and earned media. Earned media refers to media activity that a company does not directly generate, but which is generated by other entities like consumers (Stephen & Galak, 2012). Earned media is not directly controlled by the company itself, but is important to consider because it can have impact on consumer actions and the sales of a company (Trusov, Bucklin & Pauwels, 2009; Stephen & Galak, 2012).
6 known as customer enragement (Leeflang et al., 2013). On the other hand also positive word-of-mouth could be achieved.
Word-of-Mouth and social media
Word-of-mouth (WOM), or even better electronic word-of-mouth (eWOM) can be both an opportunity and a challenge for companies at the same time. One big advantage of eWOM over other types of WOM is the “constant opportunity to interact with others, regardless of time or place” (Gruen, Osmonbekov, & Czaplewski, 2006, p. 452). In the past, in order to engage in WOM people needed to meet or to call each other. However, nowadays the opportunity to interact and to exchange know-how is assumed to be always available, due to the internet (Gruen et al., 2006). Furthermore, according to Trusov, Bucklin and Pauwels (2009), the WOM of Social Media could have greater impact on consumer action than for example paid media and traditional earned media (reviews in newspapers). Receiving information from a social source may be more influential in influencing consumer opinions and consumer behavior in comparison to messages pushed by marketers (Fitzsimons & Lehmann, 2004). Customer-to-customer interactions can serve as an information source that affects the value of the firm’s offering and future purchase intentions of the customer regarding specific firms (Gruen et al., 2006). Messages sent by other consumers can be seen as more reliable. As a consequence, WOM on Social Media could have a strong impact on new customer acquisition (Trusov et al., 2009).
Another advantage of Social Media is that besides hearing from others about a brand, people can just see other people supporting a brand because these others follow the company on Twitter, or like the company on Facebook, it makes the supporters of the brand transparent to prospective customers (Naylor, Lamberton & West, 2012). Berger and Heath (2007) argue that when people see similar others supporting a brand it will lead to greater affinity with that particular brand. Because of the fact that people voluntarily adopt the brand on Social Media (Naylor et al., 2012), people perceive the consumer-generated message as more credible and show less reactance in comparison with marketer-driven recommendations (Fitzsimons & Lehmann, 2004). Social Media makes the supporters of the brand transparent to prospective customers (Naylor, Lamberton & West, 2012). According to Stephen and Galak (2012), the small impact of each time someone shares or likes a brand can accumulate to have a substantial impact on the sales of a company.
7 brand due to a negative experience or a damaged self-image (Argo, White & Dahl, 2006; Alexandrov, Lilly & Babakus, 2013). The negative messages therefore negatively influence the purchase intentions of other consumers and their perceived value of the firms (Gruen et al., 2006). The challenge for the companies is to encourage the sharing of positive messages while they discourage the sharing of negative messages concerning their products or brands.
In sum, the advantages of WOM encourages firms to actively engage in social media marketing by trying to influence the preferences and purchase decisions through social networks (Kumar, Bhaskaran, Mirchandani & Shah, 2013), by encouraging their fans to engage in positive WOM (Alexandrov, Lilly & Babakus, 2013). Kumar et al., (2013) state that companies are focusing an increasing part of their marketing budgets on engaging customers through social networks like Facebook and Twitter. At the same time, negative WOM needs to be discouraged to prevent from the potential negative outcomes of Social Media (Alexandrov, Lilly & Babakus, 2013).
Social Media and good causes
According to Hung, LI and Tse (2011), the credibility of information from social sources can make social media more appropriate for “green advertising” and other social campaigns. The focus in current research is on the activity of charities, especially Dutch charities, in social media marketing. Most research regarding the use of Social Media focuses on profit making organizations instead of the non-profit organizations, fostering the relevance of current research.
8 Table 1: the likes on Facebook and followers on Twitter for the top 25 Dutch charities ranked on yield (source: RTL Nieuws, 2013), number of likes and followers reported on 18th September 2013.
Charity Yield in € Likes on Facebook Followers on Twitter
1. KWF Kankerbestrijding €137.200.000 26.664 9.217
2. Unicef NL €53.445.000 12.526 15.714
3. Artsen Zonder Grenzen €43.733.000 27.583 4.382
4. Wereld Natuur Fonds €42.295.000 90.188 20.798
5. Hartstichting €39.100.000 10.961 5.808 6. Rode Kruis €37.284.000 21.262 8.365 7. Natuurmonumenten €31.788.000 39.080 14.294 8. Plan Nederland €31.653.000 19.837 5.994 9. Oxfam/Novib €31.158.000 26.526 8.245 10. Cordaid €30.400.000 6.608 5.086
11. Leger des Heils €26.100.000 6.467 2.636
12. Kerk in actie €25.771.000 819 2.091 13. Dierenbescherming €21.876.000 17.335 10.180 14. De Zonnebloem €20.299.000 6.240 1.609 15. Amnesty International €19.640.550 33.864 10.744 16. Greenpeace €18.796.013 23.111 20.277 17. Woord en Daad €18.354.352 905 985 18. Liliane Fonds €18.130.047 3.820 1.276 19. SOS Kinderdorpen €17.876.739 6.945 2.767 20. Nierstichting €16.852.000 9.502 1.503 21. Reumafonds €13.285.000 7.423 1.235 22. Cliniclowns €13.265.931 16.472 3.878 23. Longfonds €12.942.402 2.509 2.432 24. Compassion NL €12.886.549 8.748 3.668
25. War Child Holland €12.367.366 10.321 11.883
As table 1 shows, lots of people support the social network pages of charities. This popularity of the social media pages of charities seems to be a good thing according to the research on Social Media marketing (Trusov et al., 2009; Steven & Galak, 2012). Liking and sharing on social networks could for example have positive effects on acquisition of customers and sales (Gruen et al., 2006). An important issue for charities is to discover how they should be active on social media to encourage people to donate. Reed, Aquino and Levy (2007) argue that exploring the conditions under which people would donate to charity is becoming very important for the charities to satisfy their increased need for support. So, in a way, social media could have benefits for regular firms but could also have benefits for charities. People like and share the social media pages of charities, and therefore possibly affecting the purchase or donate intention of other consumers.
9 due to people perceiving “liking” as a substitute for a monetary donation. This reasoning is related to the research on moral licensing (Khan & Dhar, 2006), which in essence implies the following. When people behave altruistic, they could give themselves a license to behave egoistic as result of the prior good deed. Whether liking can be seen as a good deed and therefore licenses the liker to behave bad afterwards is investigated in current research.
Research have shown that the likes on Facebook could affect other consumers. According to Naylor et al. (2012) a visual numerical representation of the likes could already be a mini-connection with consumers that can yield positive effects on how the consumers evaluate their products or brands, and even can affect their purchase intention. So according to Naylor et al. (2012) it is important for the companies to show their fan base on Facebook. On the other hand, the public visibility of the like could possibly also have negative effects. According to White and Peloza (2009), when in public people are more prone to do something to benefit others, like doing a good deed. In contrary, when concerning a more private situation, people are more eager to behave more egoistically and choose in favor of themselves (White & Peloza, 2009). Especially the possible substitution effect of liking as some kind of donation to a charity could have big effects for the cash flows of charities using social networks.
In sum, the research question in current research is: “can “liking” the social network page of a charity foundation lead to negative effects through moral licensing and decreased donations, and is this relation moderated by the public visibility of the “like”.
Moral Licensing or Moral Self-Image
10 behavior in a prior task is likely to activate and establish altruistic credentials and therefore fosters people to act more self-indulgent in a latter task” (Khan & Dhar, 2006, p. 260). In the research of Khan and Dhar (2006), the respondents who helped a foreign student subsequently felt licensed to spend less money to charity and keep more money for themselves. An explanation for this behavior is that an altruistic act would boost the self-concept of the respondent and thereby reducing negative self-attributions, fostering self-indulgent choices (Khan & Dhar, 2006).
Related to this, Monin and Jordan (2009) describe the Moral Self-Image. They argue that the individuals’ morality is dynamic and malleable and can be influenced by social and situational factors. The moral self-image is extensively addressed by Jordan, Gino, Tenbrunsel and Leliveld (2013), with an important conclusion: “not only the moral self is malleable but it also presents a way to gauge this malleability” (Jordan et al., 2013, p. 28). People wish to have a positive moral (Dunning, 2007), and they take action in order to rebuild this sense of a positive moral when they behaved immorally (Mazar, Amir & Ariely, 2008). After individuals act immorally, they look for opportunities to rebuild their moral self-concept by, for example, engaging in moral actions (Jordan, Mullen & Murnighan, 2011). In the same way, Sachdeva, Iliev and Medin (2009) focus on moral compensation, showing that priming with positive or negative traits can affect the moral behavior by influencing the moral self-concept. An increase in the moral self-concept reduced pro-social behavior (Sachdeva et al., 2009).
The visibility of the behavior might possibly moderate this effect as visibility can possibly give an extra boost to your moral self-image. According to Jordan et al. (2011) the moral self consists of a personal experience. This personal experience compared with the ideal moral self-image can change the moral self-image of the person or how Jordan et al. (2013, p. 4) mention: “This apparent gap between one’s unethical actions and one’s desire to maintain a positive moral self-image leads to a dynamic and malleable perception of one’s morality”. According to Khan and Dhar (2006), participants are not aware of the connection between the prior and latter behavior, but there definitely is a change in their self-concept, which resembles the MSI. According to Aquino and Reed (2002), it is likely that there would be some variation on how people see their moral selves. This variation in how people see themselves is supported by the finding of Monin and Jordan (2009) that the individual’s morality is dynamic and malleable. However, despite these individual differences, there is a large consensus that all individuals have the universal desire to be moral (e.g. Dunning, 2007; Reed, Aquino and Levy, 2007).
Liking on a Social Media Platform
11 exposure of the “Like” are only passive, there is no active persuasion involved (Naylor et al., 2012). Friends on Facebook could see that they are “liking” something, but for strangers your like is ambiguous, it does not give any information regarding preferences or demographics, it is just a number (Naylor et al., 2012). A short example explains previous part. When Jack Jackson likes the Facebook page of Unicef NL, the people in his Facebook network (called “friends”) can get a message indicating “Jack Jackson liked Unicef NL”. While people outside the Facebook network of Jack, only see a number: “12.526 people like Unicef NL”, they do not exactly know who liked Unicef NL.
According to White and Peloza (2009), donating could be tangible (e.g. money) or intangible (e.g. support or promotion), using this definition “liking” could be seen as an intangible non-monetary donation, because the presence of the like could lead to higher brand evaluations when evaluated by other people (Naylor et al., 2012). The like could be seen as a non-monetary donation of exposure, your like causes others to be exposed to the brand or company you just liked. A like could be more explaining when the like transfers in a picture of the one supporting the company, where others can identify the fan (Naylor et al., 2012). This picture can show some demographic characteristics of the fan (Naylor et al., 2012), and thereby making the “like” more visible to others since they can see who “liked” the page and recognize characteristics of the particular person in the picture. Therefore, they can feel connected to the person who liked the page, when he or she matches the group they want to be in.
The willingness to donate
So the goal of this research is to investigate if the liking of charities on social media could affect the willingness to donate of these likers. An often used variable to measure consumer behavior is consumers’ Willingness to Pay. Willingness to pay is defined by Krishna (1991) as the maximum amount of money a customer is willing to pay for a product or service. For current research this willingness to pay should be transferred in “Willingness to Donate”, the maximum amount of money a customer is willing to donate in total to charity and the maximum amount of money the customer is willing to donate to a specific charity.
12 Figure 1: Conceptual Model
Relation between Moral Self-Image and Liking on a Social Media Platform
Could “liking” the Facebook page of a charity be seen as an altruistic act? That is the main question in this relation, whether liking on a social network could be seen as a moral act. Whether people are more willing to like the page of a charity when they have an immoral self-image. Therefore, the hypothesis is:
Hypothesis 1: When people have a low MSI, their willingness to like the page of a charity is higher in comparison with people with a high MSI.
The research of Khan and Dhar (2006) shows that participants engaged in an altruistic act, display a change in self-concept, acting less altruistic subsequently. When liking could be seen as an altruistic act, the self-concept of respondents should increase due to the prior act (Monin & Jordan, 2009). The expectation is that people are more likely to like the page of a charity on Facebook when they perceive their self-image as immoral. The other way around, when they perceive their MSI as moral, they are less likely to engage in “liking” the page of the charity. This is shown by the research of Jordan et al. (2013), where the MSI is manipulated in order to find out how the MSI influences the moral behavior of the people. “Telling people that they achieved their ideal moral selves, led them to increase their MSI, where telling them that they were a ways away from achieving their ideal moral selves led them to decrease their MSI”(Jordan et al., 2013, p. 21).
Relation between Liking and Willingness to Donate
13 the egoistic dimension reflecting the benefits for one selves, making the individual feel better about him or herself. In line with the prior relation, making a donation in this research is about the egoistic dimension, to rebuild one’s own self-image to maintain a positive self-image. The research of Monin and Jordan (2009) shows that people are more eager to behave immoral when it succeeds a prior moral act. Khan and Dhar (2006) argue that people could select a charity as penalty for a driving violation, insinuating that donating to a charity could be seen as an altruistic act covering the prior egoistic act. Furthermore Khan and Dhar (2006) state that people who feel less altruistic donate less money to charity, especially when they performed an altruistic act before like helping the foreign student.
The assumption made in this research is that after people “liked” the page of a charity and they perceive this as an act of altruism, their willingness to donate would be lower due to their prior good deed (Khan & Dhar, 2006). When people perceive their self-image as immoral, their willingness to “like” the page of a charity would be higher, after engaging in this “liking” their self-image would be perceived as more moral, negatively influencing the amount they are willing to donate to the charity. Therefore, the hypothesis regarding this relation is:
Hypothesis 2: Liking the page of a charity on Facebook negatively affects the subsequent willingness to donate money to a charity.
Public Visibility of the “Like”
According to Goffman (1959), people are motivated to make a good impression on others and to present themselves well. People are aware of the fact that others make perceptions about them and therefore people are often trying to convey impressions of themselves that lead to desired outcomes (Leary & Allen, 2011). For example, people could make an impression by wearing clothes associated with a specific group or in contrary, avoiding products associated with dissociative groups (Berger and Heath, 2007). However, for making an impression, the action or behavior must be in public to be perceived by others.
14 demographical information is present and no pictures are displayed. The second one is a more revealing virtual presence, where pictures of real supporters or Facebook fans of the brand are displayed, showing some of the demographic characteristics of the fan (Naylor et al., 2012). In the method section, examples of ambiguous and more revealing displayed likes are shown.
According to White and Peloza (2009) people are more likely to engage in altruistic other-benefit action when the donation is in public. Also for firms the visibility is important. Consumers seem to evaluate a brand with identified supporters more positive in comparison with a brand whose supporters were not identified (Naylor et al., 2012), fostering companies to show more than only an ambiguous like. When the supporters of a brand are perceived similar to the consumer, the liking and willingness to interact with that brand increase; when the supporters are ambiguous the same effect happens (Naylor et al., 2012). The other way around, this means that when the liking is more revealing and public, but dissimilar to the target group, the liking and willingness to interact decreases. This fosters the importance of the public visibility of the “like” because the decision to reveal or disclose a brands’ fan base can be important in influencing consumer reactions to the brand (Naylor et al., 2012). In sum, the visibility of likes is important from a company perspective because potential supporters can identify the current supporters of the brand. For the “liker”, the visibility might boost their moral self-image by doing a good deed in public.
The visibility could have both a direct and an indirect effect of liking. The direct effect contains the relation between the visibility of the “like” and the willingness to like. According to White and Peloza (2009) people are more eager to donate when it is in a highly public area, and are less eager to donate when in a more private surrounding. This fosters the importance of the moderator visibility. When more people are aware of the donation, the donation could influence the self-image of the respondent.
Hypothesis 3: When the visibility of a “like” is high, people are more eager to like the page of a charity on Facebook than when the visibility of the “like” is low
Moderating effect of public visibility of the “like”
15 moral licensing. In sum, the presence of others or the visibility of the actual deed can influence the moral self-image of people. The hypothesis for this moderating effect is:
Hypothesis 4: the negative effect of MSI on liking (as hypothesized in hypothesis 1) becomes stronger in a highly visible situation compared to a low visible situation.
The assumption made is that when people engage in a highly visible “like”, their moral self-image would be higher in comparison when people engage in a more ambiguous like. The ambiguous like would have a weaker effect in comparison with the more revealing like. Therefore, testing the statement of Merrit, Effron and Monin (2010) regarding the moderating effect of the action is public or private.
Research Design
Design and participants
A 2 (MSI: low vs. high) x 2 (visibility of like: public vs. private) between-subjects factorial design was used to investigate the social media behavior of the participants and donations made by the participants. The main dependent variables were the number of points allocated by the recipients towards charities or egoistic causes. Respondents were attained by using social networks and personal networks. Participants were randomly assigned to the high MSI condition and the low MSI condition. The sample consisted of 31 (35,2%) men and 57 (64,8%) women, with an average age of 34 (Mage = 33,72, SD= 15,618). The division of the respondents among the four categories of conditions can be found in Appendix A.
16 Figure 2: Social Media Usage
Procedure
First the respondent received a bogus questionnaire, a questionnaire made up to convince the respondent he or she takes part in a real test, the so called “MID”. This bogus questionnaire was adopted from Carlo and Randall’s (2002) measurement of pro-social behavior. The bogus questionnaire consisted of five statements and one open question. Thereafter, the respondent received a message mentioning they just participated in a measurement called the “MID”, which shows their ideal moral self-image and how far they are away from that ideal self-image. This is the first manipulation. People got one out of two feedback options, randomly assigned to the participant. In the low MSI condition people learned that they are far away of their ideal moral self-image. In the high MSI condition, people are pretty close to their ideal moral self-image. The purpose of this is to manipulate people’s Moral Self-Image similar to how this was manipulated in the research of Jordan et al. (2013). When the participants had read the feedback, they received another questionnaire to measure their current moral self-image. This questionnaire is translated from the moral self-image measurement of Jordan et al. (2013). This questionnaire checks if the manipulation of the MSI due to the feedback has worked.
The next part of the survey consisted of an objective story regarding social media usage by charities and the KWF Kankerbestrijding (Dutch cancer combat charity). Then the second manipulation occurred. The participants were randomly assigned to two conditions. The participants were asked whether they want to like the Facebook page of the KWF Kankerbestrijding or not. They were randomly assigned into two conditions, Ambiguous or Visible. In the Ambiguous condition (figure 3), the name of the respondent is not displayed, while the name is displayed in the Visible condition (figure 4). 0 10 20 30 40 50 60 70 80
Facebook Other Social Media
17 Figure 3: Ambiguous condition
Figure 4: Visible condition
The third part of the survey consisted of the analysis of the willingness to donate. This was evaluated with a game. In this game, the participants could divide 100 points among nine different causes (eight charities and themselves). They had to divide all 100 points, but could also give the 100 points to themselves. Each point they donated to themselves equals one ticket to play in a real lottery, in order to give them the opportunity to behave egoistic. The willingness to donate can be evaluated using the amount of credits they assigned to the charities.
18 to be included. The questions used were adapted from the self-monitoring measurement of Snyder and Gangestad (1986).
The survey ended with a debriefing. The respondents should be debriefed because of the manipulation in their Moral Self-Image at the first step. This has been done by explaining the procedure and goal of the research before the respondent could finish the survey, using a mechanism that prevents that the respondent clicks forward without reading the debriefing. This could be done by displaying the forward or send button after 20 seconds for example; in a way the respondent has to wait at least 20 seconds before clicking the send button. It would include telling the respondents that the measure they took and the associated feedback was completely bogus (Jordan et al., 2013)
Plan of Analysis
Firstly the results of the MSI measurement of Jordan et al. (2013) were evaluated in order to check if the manipulation using the feedback on the bogus questionnaire has worked. This means evaluating if people have a high or low moral self-image before they start with the “liking manipulation”. Reliability analysis on the nine questions measuring Moral Self-Image showed that all nine questions together had a α = 0.798 (in research of Jordan et al., 2013, a α= 0.86 was found).
Secondly, the differences were checked within the two conditions of the “liking manipulation”. This in order to analyze if people are more willing to like the page of the KWF when they have a low MSI in comparison with the high MSI respondents. Also, the moderation could be checked with this data by comparing the amount of likes in the visible condition with the amount of likes in the ambiguous condition. In order to analyze the main effects (hypothesis 1 and hypothesis 3) and the interaction effect (hypothesis 4), an ANCOVA was used. The two manipulations were converted in dummy variables (0 vs. 1) and the outcomes of the MSI-scale were used as the continuous covariate that could influence the willingness to donate of the respondents.
Thirdly, the results from the donation game were linked to the respondent’s perceived moral self-image and their answer in the liking part. This in order to see if there are any differences in willingness to donate between high MSI respondents and low MSI respondents, and whether the likes are affecting these donations. A MANCOVA was used to analyze the effects on the willingness to donate, the donations for KWF and the donations to the self. The outcomes of the MSI-scale were used as the continuous covariate that could influence the willingness to donate of the respondents.
19 After deleting questions 2, 3, 4, 5 and 15, a α = 0.712 was found for the eleven remaining questions. The deleted questions were not removed from the data and another reliability analysis was used for the deleted questions .For question 3, 4, 5 and 15, α = 0.724 was found when question 2 was deleted. For analyzing the possible effect of self-monitoring three computed variables have to be used, the computed variable of the first reliability analysis with the 11 questions, the computed variable of the second reliability analysis with the 4 questions and last the results of the second question.
Results
Manipulation Check
In order to analyze whether or not the manipulation of the MSI was successful, a One-way ANOVA was performed of the dummy variable of the MSI manipulation on the MSI scale. This ANOVA was not significant, F(1,86) = 0,098, p = 0,755. The manipulation of the MSI did not significantly influence the MSI scale, thereby showing that the manipulation of the MSI failed. The two conditions (low MSI versus high MSI) do not differ for the MSI scale. Despite this finding, I did continue to analyze the rest of the data for exploratory reasons.
ANCOVA on Liking
To analyze the effects on liking behavior a 2 (MSI: high versus low) x 2 (Visibility: visible versus ambiguous) ANCOVA was performed with the MSI Scale as the covariate. The continuous MSI scale was included as covariate because the MSI scale influences the dependent variable, but is not part of the main experiment. The manipulation failed, but the MSI scale could be included to control for the moral self-image of the respondent during the survey.
20 F(1, 84)= 1.930, p = .169 > 0,05. The covariate does not significantly influence the liking behavior. The conclusion is that the neither differences in MSI, differences in visibility of the like nor the interaction between the MSI and the visibility does not have a significant effect on the liking behavior of the respondents.
MANCOVA of Liking on Donations
To analyze the effects of liking behavior on the donations, a MANCOVA was performed with three dependent variables. The overall Willingness to Donate, the amount of points allocated to the self and the amount of points allocated to the KWF charity. Since the Willingness to Donate (total amount of points minus points allocated to the self) and the points allocated to the self (total amount of points minus points donated to charities) are both sides of the same equation, the results will be reported once. Again the continuous MSI scale was included as covariate, because the moral self-image of the respondent could influence their willingness to donate.
The liking behavior seems to have no significant effect on the willingness to donate in general (F(1, 85)= .000, p = .998 > 0,05) and also seems to have no significant effect on the amount of points allocated to the self (F(1, 85)= .000, p = .998 > 0,05). However, the effect of liking behavior on the amount of points allocated to KWF charity appeared to be significant (F(1, 85)= 17.335, p < 0,001). When people liked the page of the KWF it leads to a higher amount of points allocated to KWF (M = 27.452) in comparison when people did not like the page of the KWF (M = 14,865). This significant effect contradicts the proposed hypothesis; therefore hypothesis 2 could not be supported and should be revised in order to fit the effect found. The covariate MSI scale, appears to be insignificant for Willingness to Donate and amount of points allocated to the self (F(1, 85)= .526, p = .470 > 0,05). Furthermore, the covariate is also insignificant for the amount of points allocated to the KWF charity (F(1, 85)= .019, p = .891 > 0,05). The covariate does not significantly influence the amount of the donation.
Impression Management
21 .01 < .05). The second variable of impression management computed out of four questions seems to have an insignificant effect on willingness to donate and amounts of points allocated to the self (F(1, 80)= .2.308, p = .133 > .05). The third variable of impression management, consisting of one question only, seems to have an marginally significant effect on the willingness to donate and the amounts of points allocated to the self (F(1, 80)= 3.474, p = .066 < .1). All three computed variables of impression management seem to have an insignificant effect on points donated to the KWF.
An additional regression analysis was performed to analyze the direction of the relation between the two (marginally) significant variables of impression management regressed on the willingness to donate. The regression analysis was significant (R² = 0,001, F(3,84) = 4.34, p = .007 < .05). The first variable of Impression Management, computed out of eleven questions, seems to influence the willingness to donate significantly (B = -4,753, t = -2,708, p = .008 < .05), therefore having a negative effect on the willingness to donate. The last variable of Impression Management, consisting of one question only, seems to be marginally significant (B = 1,376, t = 1,768, p = .081 < .1), therefore having a slight positive effect on the willingness to donate. The negative effect of the computed variable with the eleven questions (Beta = -,281) is bigger than the positive effect of the single question variable (Beta = -,188) and shows that the overall effect of impression management on the willingness to donate would be negative.
DISCUSSION
Discussion
The influence of Facebook behavior on the willingness to donate was studied using an experimental design with two manipulations. The first manipulation was needed in order to change the moral self-image of the respondents into a high moral self-self-image or a low moral self-self-image. The results of the manipulation checks suggest that the manipulation did not work. This finding shows that the manipulation of the moral self-image as proposed in the second study by Jordan et al. (2013) has not been replicated. The bogus questionnaire and the subsequent positive or negative feedback, did not succeed to affect the respondent’s moral self-image enough to be usable. The expectation that respondents MSI would increase after telling them they achieved their ideal moral selves (Jordan et al., 2013), and the other way around, was not supported. The failed manipulation resulted in two groups (Low MSI and High MSI) but both group means were equal. Despite this finding, the analysis of the rest of the data was continued for exploratory reasons. Due to the failed manipulation, the equal group means caused the absence of other effects during the rest of the analysis.
22 the high MSI condition. This finding does not support the findings of Khan and Dhar (2006) and Monin and Jordan (2009) regarding Moral Licensing or the change in the moral self-image. One big reason why the results do not support the findings of Khan and Dhar (2006) and Monin and Jordan (2009) is the failed manipulation of the moral self-image. The groups with a different MSI condition showed to have the same group mean, causing the absence of differences in the liking behavior between the two groups. Another possible reason could be that the respondents do not see the “like” of a charities Facebook page as an altruistic act as in comparison with the altruistic act of donating real money (Khan & Dhar, 2006) or behaving non-sexist (Monin & Miller, 2001). The findings cannot support the expectation that liking the social media page of a charity could be seen as an altruistic act, neither rejects it. So, the question remains, did they like the page to behave altruistic or just because they have a special interest in this particular charity? For future research, the manipulation should be improved in order to replicate the effect of Jordan et al. (2013). The manipulation should be improved in a way that the feedback used is less subtle and more explanatory regarding the outcome and consequences of the bogus feedback in order to influence the moral self-image of the respondents and therefore manipulating their moral self-image sufficiently. This is further elaborated in the limitations part.
23 agrees with a request, the need for consistency of the people encourages them to agree with the second request too (Cialdini, 2009). In this case it could be that someone likes the page of a charity on social media, subsequently his or her need for consistency encourages the person to agree with a subsequent request of the charity: a donation. Because of the need for consistency, the people that agreed to the first request are more likely to agree to the second, often bigger, request in order to remain consistent with the prior act the person committed to (Cialdini, 2009). So, in a way people liked the page of the KWF, they could donate to the KWF subsequently in order to remain consistent with their prior behavior.
For the third hypothesis, the expectation that when the like is in a more visible condition people are more willing to like the page of the KWF cannot be supported. This finding does not support the finding of White and Peloza (2009) that people are more eager to engage in an altruistic act when it is in a highly public area, and less eager to engage in an altruistic act when it is in a more private surrounding. Again, one possible reason for the found results could be that people do not see the “like” of a charities Facebook page as an altruistic act (Dhar & Khan, 2006) or as an act where others can benefit from (White & Peloza, 2009). A second reason could be that the passive exposure of a like in an online setting as described by Naylor et al. (2012) is not public enough to meet the findings of White and Peloza (2009). Online bystanders do not literally see the person act, but can only see a small outcome of the prior act, possibly making the act less public in comparison with an act performed in the “real” world. A third reason could be the manipulation of the visibility. The differences between the ambiguous condition and the visible condition are rather subtle, in a way that the manipulation possibly was not clear enough. The difference in the picture only consisted of the word “you” in the sentence “you and 2435 others like this” in comparison with the ambiguous condition stating “2435 people like this”. The introduction in the question explained the consequences of the “like”, but altogether the manipulation could have failed and therefore influenced the absence of an effect of visibility on the willingness to like the page of the KWF. The distinction between ambiguous and visible showed by Naylor et al. (2012) has not been replicated in this study and causes the group means of the ambiguous condition and the visible condition to be too equal. In future research this manipulation of visibility need to be more distinctive, in a way that the two different conditions significantly differ from each other in the eye of the respondents. One way could be using text as one condition and a picture as the other conditions as showed by Naylor et al. (2012). This is further elaborated in the limitations section.
24 the importance of the presence of others in moral licensing cannot be supported by the results of current study. One important reason that could cause the absence of a found moderating effect could be the absence of the two main effects (hypothesis 1 and hypothesis 3) as previously described. Due to a failed MSI manipulation and a possibly too subtle visibility manipulation, no main effects could be found. Therefore, it is also possible that no moderation effect could be found due to the subtle visibility manipulation. Another possible reason could be that people do not see the like as an act of altruism or a benefit for others (Dhar & Khan, 2006; Monin & Miller, 2001; White & Peloza, 2009). Altogether, the question if visibility of the act can moderate the willingness to like still needs to be investigated using improved manipulations of MSI and visibility.
An additional analysis was performed to measure the effect of Impression Management. The results suggest that the overall effect of impression management on the willingness to donate appears to be negative. This would mean that the more people are concerned with their image, the lower their willingness to donate would be. According to Goffman (1959), people are motivated to make a good impression on others and to present themselves well. In line with this statement, the expectation was that people that are concerned with their image would donate more make an impression on others. The findings contradict this expectation. According to Leary and Allen (2011), people convey impressions to others to accomplish the desired outcomes. The results of current study could suggest that people are not willing to donate in order to impress others. However, in order to make an impression on others, the action needs to be in public. The donation made in the survey was in private, suggesting that there was no one to impress. The importance of the presence of other people in moral licensing as described by Monin, Sawyer and Marquez (2008) could be a possible explanation for the outcome of this analysis. Due to the absence of others, there was no incentive for people with a high concern for impression management to do something good, because there was no one to impress.
25 base of a company or charity, the firms should use more advertising channels to promote their charity in order to be effective. According to Romaniuk (2011), the light users or the non-users of products play an important role for all companies to be beneficial. Adapting this for charities, the light-donators are people who only donate once or twice in a year, “just as the light-buyers buy only once or twice a year” (Romaniuk, 2011, p. 561). For charities to reach this light-donators or non-donators, other advertising media should be used instead of only advertising on social media, due to the fact that mostly heavy users are reached using social media advertising (Nelson-Field et al., 2012). Therefore charities in economic contraction should keep using multiple advertising media channels in order to communicate and attract the light donators and non-donators instead of exclusively using social media advertising to cut the costs.
The finding that people who liked the page of a charity on social media are willing to donate more to this particular charity could also lead to another practical implication based on the principles of the foot-in-the-door technique (Freedman & Fraser, 1966; Cialdini, 2009). The charities could try to encourage or even persuade people to like their social media page, therefore getting the consumers to comply with the first request. Is this the case, then according to Freedman and Fraser (1966) and Cialdini (2009), the desire to behave consistent fosters the compliance of the people with the second request, the donation to this particular charity. Therefore, charities should actively try to encourage people to comply with the request to like their page, in order to communicate the second request of donating real money to the people. However, the charity really has to make a request to people to become their Facebook fans and subsequently make a request for donating real money in order to replicate the effects found by Freedman and Fraser (1966).
Another new insight is the finding that people highly concerned with their image and impression management have a lower willingness to donate. A practical implication here is that charities should try to make the donations more public, in order to give people highly concerned with impression management the opportunity to impress their peers. According to Monin, Sawyer and Marquez (2008) and Goffman (1959) there need to be other people around to make an impression. Charities could give people the opportunity to make public donations in order to foster the willingness to donate of people that are highly concerned with their personal image. Further research is needed to study the effect of impression management both in real life donating situations and online conditions.
Limitations
26 of the research questions proposed in this study could not be addressed and therefore not be answered. This was not expected due to the example of Jordan et al. (2013) where this questionnaire and bogus feedback affects the MSI of the respondents. A possible explanation could be the small number of questions asked prior to the bogus feedback, this number could have been too small to give the respondent the idea that the feedback was really based on the answers they gave. Another possible explanation could be the length or the extensiveness of the feedback given to the respondent. The feedback scale showed was pretty straight forward, however the explanation assisting the scale was rather subtle. Altogether, these suggestions could possibly explain why the results of the procedure used by Jordan et al. (2013) could not be replicated in this study. In future research, the usage of bogus feedback should be less subtle and more evident for influencing the moral self-image of the respondents. The feedback and the accompanying consequences should be explained more thoroughly in order to significantly affect the moral self-image of the respondents. I suggest that the spot of both dots on the feedback scale should be described more in detail for the respondents to fully understand the outcome of the bogus questionnaire.
The second limitation concerns the possibly failed manipulation of visibility causing the group means of the ambiguous condition and the visible condition to be too equal. The distinction between ambiguous and visible showed in the research of Naylor et al. (2012) could not be replicated in this study. The subtle manipulation of visibility is the most probable explanation for this. In this study, the manipulation of visibility was only a small text within an image (see figure 3 and 4), with some additional explanation in text to support the image. While in the research of Naylor et al. (2012) this manipulation was made between a numerical condition representing the number of likes (figure 3) and a visible condition consisting of the Facebook picture of the respondent. The difference between a picture and a text is way more explicit in comparison with the distinction made in this study. Therefore in future research, the distinction between both conditions should be made more visible and explicit in order to replicate the findings of Naylor et al. (2012). This could be done by having one condition with text only while the other condition consist of pictures and images, therefore increasing the difference between the two conditions of visibility.
The third limitation is the deficit of respondents in the Moral x Visible condition. After removing false responses or unfinished surveys, only fifteen people remained that were randomly assigned to this condition. A number of fifteen respondents in one condition is not enough to meet the stipulation of a 2 x 2 design. In this design, each condition needs to have at least twenty respondents to suffice. The shortage of people in one condition could therefore have influenced the results found in this study.
27 Facebook”?. A common heard comment of respondents is the statement that they principally do not like anything at all on Facebook or other social media. Another possibility could be the fact that people only like things because they are getting something in return (e.g. a chance on winning a product or service). This is a common used technique on social media and in WOM marketing (Tuk, Verlegh, Smidts, & Wigboldus, 2009). The example in the introduction (Unilever and the World Food Program) could be seen as an example of rewarded WOM used by charities. This could have a positive effect for the number of like both Unilever and the World Food Program are receiving, however according to Tuk et al. (2009), this might also have the effect that the perceived sincerity of the referring Facebook fan liking the project of Unilever is lowered due to the fact that he or she is motivated by certain benefits. Both issues could have influenced the results found in this study and therefore possibly distorted the data. In future research, questions of these kinds need to be asked in order to be able to control for people that do not like anything at all or only like under special conditions as for example when the company or charity is offering extra benefits.
Future Directions
Future research could address the hypotheses proposed in this study by investigating them using proper manipulations for influencing the moral self-image and the visibility of the like. The most important question should be if liking could be seen as an act of altruism and therefore occur as a possible substitution for a donation to a charity. An important issue in investigating this should be the focus on light donators or non-donators to control for the effect of heavy users on the willingness to donate.
Another future direction could address both a practical implication as well as a limitation of this study, by investigating the conditions in which people are willing to like the social media page of a charity. Especially how charities could attract light donators or non-donators to like the social media page of a charity in order to communicate the message to them using social media or to encourage consistent behavior. These conditions could potentially be an important issue for charities to receive more donations using the social media platforms.
28 CONCLUSION
29 References
Alexandrov, A., Lilly, B., & Babakus, E. (2013). The effects of social-and self-motives on the intentions to share positive and negative word of mouth. Journal of the Academy of Marketing Science, 1-16. Aquino, K., & Reed II, A. (2002). The self-importance of moral identity. Journal of personality and social psychology, 83(6), 1423.
Argo, J. J., Dahl, D. W., & Manchanda, R. V. (2005). The influence of a mere social presence in a retail context. Journal of Consumer Research, 32(2), 207-212.
Argo, J. J., White, K., & Dahl, D. W. (2006). Social comparison theory and deception in the interpersonal exchange of consumption information. Journal of Consumer Research, 33, 99–108. Berger, J., & Heath, C. (2007). Where consumers diverge from others: Identity signaling and product domains. Journal of Consumer Research, 34(2), 121-134.
Carlo, G., & Randall, B. A. (2002). The development of a measure of prosocial behaviors for late adolescents. Journal of Youth and Adolescence, 31(1), 31-44.
Cialdini, R. B. (2009). Influence: Science and practice. Boston: Pearson Education
De Vries, L., Gensler, S., & Leeflang, P. H. (2012). Popularity of Brand Posts on Brand Fan Pages: An Investigation of the Effects of Social Media Marketing. Journal Of Interactive Marketing, 26(2), 83-91 Dunning, D. (2007). Self-image motives and consumer behavior: How sacrosanct self-beliefs sway preferences in the marketplace. Journal of Consumer Psychology, 17(4), 237-249.
Fitzsimons, G. J., & Lehmann, D. R. (2004). Reactance to recommendations: When unsolicited advice yields contrary responses. Marketing Science, 23(1), 82-94
Freedman, J. L., & Fraser, S. C. (1966). Compliance without pressure: the foot-in-the-door technique. Journal of personality and social psychology, 4(2), 195
Goffman, Erving (1959), The Presentation of Self in EverydayLife. Oxford: Doubleday.
Gruen, T. W., Osmonbekov, T., & Czaplewski, A. J. (2006). eWOM: The impact of customer-to-customer online know-how exchange on customer-to-customer value and loyalty. Journal of Business research, 59(4), 449-456.
Hung, Kineta, Stella Y. Li, and David K. Tse (2011), “Interpersonal Trust and Platform Credibility in a Chinese Multibrand Online Community”, Journal of Advertising, 40 (3),99–112.
Jordan, J., Gino, F., Tenbrunsel, A., Leliveld M.C. (2013). Moral self-image: Measuring the malleability of the moral self. Working Paper
Jordan, J., Mullen, E., & Murnighan, J. K. (2011). Striving for the moral self: The effects of recalling past moral actions on future moral behavior. Personality and Social Psychology Bulletin, 37(5), 701-713.
30 Koschate-Fischer, N., Stefan, I. V., & Hoyer, W. D. (2012). Willingness to Pay for Cause-Related
Marketing: The Impact of Donation Amount and Moderating Effects. Journal of Marketing Research, 49(6), 910-927.
Krishna, A. (1991). Effect of dealing patterns on consumer perceptions of deal frequency and willingness to pay. Journal of Marketing Research, 441-451.
Kumar, V. V., Bhaskaran, V., Mirchandani, R., & Shah, M. (2013). Creating a Measurable Social Media Marketing Strategy: Increasing the Value and ROI of Intangibles and Tangibles for Hokey Pokey. Marketing Science, 32(2), 194-212
Leary, R.M., & Allen, A. (2011). Self-Presentational Persona: Simultaneous Management of Multiple Impressions. Journal Of Personality & Social Psychology, 101(5), 1033-1049.
Leeflang, P.S.H., P.C. Verhoef, P. Dahlström and T. Freundt, 2013, Digitale marketing: uitdagingen en oplossingen, Holland Management Review, 148, pp. 41-49.
Lovett, M.J., Peres, R. & Shachar, R. (2013). On Brands and Word of Mouth. Journal of Marketing Research, 50(4): 427-444
Mazar, N., Amir, O., & Ariely, D. (2008). The dishonesty of honest people: A theory of self-concept maintenance. Journal of Marketing Research, 45(6), 633-644.
Merritt, A. C., Effron, D. A., & Monin, B. (2010). Moral Self‐Licensing: When Being Good Frees Us to Be Bad. Social and personality psychology compass, 4(5), 344-357.
Millward Brown, The Value of a Fan (22, Mar., 2011), http://www.slideshare.net/MillwardBrown /value-of-a-fan (Accessed: 6th January 2014)
Monin, B., & Jordan, A. H. (2009). The dynamic moral self: A social psychological perspective. In D. Narvaez & D.K. Lapsley (Eds.), Personality, identity, and character: Explorations in moral psychology, 341-354. New York: Cambride University Press.
Monin, B., & Miller, D. T. (2001). Moral credentials and the expression of prejudice. Journal of personality and social psychology, 81(1), 33.
Monin, B., Sawyer, P. J., & Marquez, M. J. (2008). The rejection of moral rebels: Resenting those who do the right thing. Journal of Personality and Social Psychology, 95(1), 76–93.
Naylor, R. W., Lamberton, C. P., & West, P. M. (2012). Beyond the “like” button: the impact of mere virtual presence on brand evaluations and purchase intentions in social media settings. Journal of Marketing, 76(6), 105-120.
Nelson-Field, K., Riebe, E., & Sharp, B. (2012). What’s Not to “Like?”. Journal of Advertising Research, 52(2): 262-269
Novemsky, N., & Dhar, R. (2005). Goal fulfillment and goal targets in sequential choice. Journal of Consumer Research, 32(3), 396-404.
31 Romaniuk, J. (2011). Are You Blinded by the Heavy (Buyer) . . . Or Are You Seeing the Light?. Journal Of Advertising Research, 51(4), 561-563
RTL Nieuws, Crisis Pijnigt Goede Doelen, (2, Jul., 2013), http://www.rtlnieuws.nl/nieuws/binnenland /crisis-pijnigt-goede-doelen (Accessed: 18th September 2013)
Sachdeva, S., Iliev, R., & Medin, D. L. (2009). Sinning Saints and Saintly Sinners The Paradox of Moral Self-Regulation. Psychological science, 20(4), 523-528.
Snyder, M. (1974). Self-monitoring of expressive behavior. Journal Of Personality And Social Psychology, 30(4), 526-537
Snyder, M., & Gangestad, S. (1986). On the nature of self-monitoring: matters of assessment, matters of validity. Journal of personality and social psychology, 51(1), 125.
Stephen, A. T., & Galak, J. (2012). The effects of traditional and social earned media on sales: A study of a microlending marketplace. Journal of Marketing Research, 49(5), 624-639.
Trusov, M., Bucklin, R., & Pauwels, K. (2009). Effects of Word-of-Mouth Versus Traditional Marketing: Findings from an Internet Social Networking Site. Journal Of Marketing, 73(5), 90-102
Tuk, M. A., Verlegh, P. J., Smidts, A., & Wigboldus, D. J. (2009). Sales and sincerity: The role of relational framing in word-of-mouth marketing. Journal Of Consumer Psychology, 19(1), 38-47. Van Heerde, H. J., Gijsenberg, M. J., Dekimpe, M. G., & Steenkamp, J. M. (2013). Price and Advertising Effectiveness over the Business Cycle. Journal Of Marketing Research (JMR), 50(2), 177-193.
32 APPENDIX A:
The number of respondents per category
Visible Ambiguous Total
Moral 15 24 39
Immoral 22 27 49
