What Difference Does a ‘Like’ Make?
An experiment on liking a charity’s Facebook page and subsequent donating behaviour, taking the role of moral self-image and issue involvement into account.
Christa Brouwer, 10181849 Master’s Thesis
Graduate School of Communication Master’s programme Communication Science
Supervisor: dr. M.H.C. Meijers February 2nd, 2018
Abstract
Since social media have almost become a primary necessity of life, this has led to a big increase of charities on Facebook and other social media platforms. Since it is so easy to express one’s appreciation of a charity, questions about meaningful support subsequent to a Facebook like arise. Does liking a charity’s Facebook page really add value to the
organization, in the sense that the liker will then also support the charity in a more tangible way by making a donation? Prior research has shown that an initial token of support towards a charity (such as liking its Facebook page) subsequently does not lead to more meaningful support of that charity (such as donating money to it). Moreover, a negative effect of liking on donating has been found. This slacktivism effect is a huge problem for charities, because one of the most important parts of their existence – receiving donations – might be jeopardized by liking. The present study therefore investigated the role of people's moral self-image and their issue involvement with the charity in a between-subjects experimental design, with 138 participants who were induced to either like (initial support) or view (no initial support) a charity’s Facebook page. No evidence was found for a direct effect of liking on donating, nor did a person’s moral self-image seem to play a mediating role in this relationship. Moreover, the level of involvement of people with the charity’s purposes did not influence how much they donate after liking the charity’s Facebook page, and their issue involvement also did not influence whether liking leads people to view themselves more or less morally, which again does not account for how much they donate. Limitations are discussed, and suggestions for future research are provided.
What Difference Does a ‘Like’ Make?
One of the most important activities for a charity – if not the most important one – is to raise money. In order to raise this money, charities need the public to become aware of what the organization does to make the world a better place. Only when they reach people with their message, they can get them to make a donation (Waters, 2010). Many charities try to reach their audience through several different channels, for example through a television campaign or by fundraising in the local supermarket (Webber, 2004). But television
commercials are very costly, and fundraising in supermarkets is time consuming, so a more attractive place for charities to reach people is online. In a world where the internet has almost become a primary necessity of life (Jackson, 2011), what better way to reach people than through the use of a social networking site?
There are many social networking sites where charities can stay in touch with their supporters. But since Facebook is used by 79% of internet users, and hence is one of the most popular ones, this study will focus on using Facebook as a channel to reach people
(Greenwood, Perrin, & Duggan, 2016). For charities especially, social networking sites are a good way to reach supporters, because they do not cost anything. Thus, the charity saves on publicity funds. This way of easily reaching many people with little effort and money, has led to a big increase in social media presence among charities (Nah & Saxton, 2013). With the introduction of the “like” button on Facebook in 2009, liking has become part of users’ daily Facebook routines, and for charities a way to engage with them. Every minute, more than 4 million posts, reactions or pages on Facebook are being liked (Carey-Simos, 2015). But what does all this liking on Facebook mean? Does it really add value to the organization behind that Facebook page, in a way that the liker views it more favourably? Or is liking just a way of expressing one’s fondness of the organization?
Assuming that liking a charity on Facebook means that a person becomes part of the charity’s online network, one may expect that liking a charity will have a positive effect on the
person’s behaviour towards that charity. This is because being a member or supporter of an organization (such as a charity) has been shown to lead for example, to positive word-of-mouth about that organization (Brown, Barry, Dacin & Gunst, 2005). However, research has also shown that liking a Facebook page does not necessarily lead to more donating behaviour among supporters (Kristofferson, White & Peloza, 2013).
On the contrary, previous research has demonstrated that an initial token of support (i.e., liking a Facebook page) does not lead to an increased, but rather a decreased willingness to provide more substantial contributions to the charity (e.g., donating; Kristofferson et al., 2013). A body of research suggests that this decrease in more substantial support after a smaller initial token of support towards the charity, is due to the fact that people feel they have already done something morally, and thus have done enough for now, which results in their feeling licensed to behave less morally (e.g., not donating; Monin & Miller, 2001; Khan & Dhar, 2006; Meijers, Verlegh, Noordewier & Smit, 2015). This is a serious problem for charities, because one of the most important parts of their existence – getting donations – might be jeopardized by liking. Therefore, it is very important to learn which factors play a role in the fact that people donate less money after liking a charity’s Facebook page, in order to try and find a way to make people donate more instead.
Previous research has looked into the reasons why people lack the will to enact meaningful change after having expressed an initial token of support. In five studies, Khan and Dhar (2006) have shown that prior expression of an intent that helps establish a moral self-image is subsequently more likely to liberate the person to make self-indulgent choices. The current study supports this idea by suggesting that a persons’ moral self-image (i.e., how moral a person considers him- or herself to be) mediates the relationship between a prior
expression of support (i.e., liking a charity’s Facebook page) and subsequent support (i.e., donating). The current study also adds to the existing literature by taking involvement with the charity into account as a moderator between liking the charity’s Facebook page and the persons’ moral self-image. Expecting that, due to a consistency effect (Gawronski & Strack, 2012), liking a charity’s Facebook page leads to a higher moral self-image and less donating than viewing a charity’s Facebook page, but only among people who are less involved with the charity’s issue, not among people who are more highly involved. The research question of this study therefore is: “Which role does a person’s moral self-image and their involvement with the issue of a charity play in the decision whether to take part in subsequent meaningful
support of that charity (i.e., donating), after either expressing initial support towards the
charity (i.e., liking its Facebook page) or no initial support (i.e., merely viewing its Facebook
page)?”
Theoretical framework Slacktivism
Previous research has demonstrated that liking a Facebook page does not lead to an increased willingness to provide more meaningful support, like making a donation to the charity. This demonstrates the existence of slacktivism, defined by Kristofferson, White and Peloza (2013) as “a willingness to perform a relatively costless, token display of support for a social cause, with an accompanying lack of willingness to devote significant effort to enact meaningful change” (p. 1149).
John et al. (2017) found that liking a brand (e.g., a charity) on Facebook has no positive effect on how favourable a person views this brand, but that liking in fact has a negative effect on people’s brand favourability, meaning that liking causes people to view a brand (e.g., a charity) more negatively. Which in case of a charity, logically would lead to less donating than if the favourability towards it was positive. In their experiment, John et al.
(2017) tested the effect of liking a brand’s Facebook page on how favourably people view that brand. They did this by conducting sixteen experiments, all testing for direct effects as well as for indirect ones, for example, whether a slacktivism effect only occurred after consumers had some time to experience and be influenced by the Facebook page (i.e., brand favourability was administered right after liking and after some delay). Results of the
experiments and meta-analysis suggest that brand Facebook pages (e.g., a charity) do not positively affect consumers’ attitudes or behaviours (e.g., donating to a charity), but that the act of liking actually has a negative effect on how favourably a person views the brand (John et al., 2017).
This slacktivism effect was also found by Krishna (2011), who showed that an initial token of support towards a charity led to less donating to that charity than showing no initial support. Directly donating money to a charity happened less if people had initially purchased a product where an amount of the price goes to the charity, than if the latter had not been the case (Krishna, 2011).
It can therefore be assumed that this slacktivism effect will hold, next to brands (John et al., 2017), also for non-profits like charities, which is in line with Krishna (2011). And thus, the assumption is that liking a charity’s Facebook page will lead to less donating behaviour than not liking, but merely viewing the Facebook page. Another body of research, to be discussed below, also supports the expectation that liking a charity’s Facebook page does not have a positive effect on subsequent moral behaviour like donating, explained by the
underlying processes of moral licensing.
Moral licensing
Previous studies have investigated the establishment of someone’s morality as a predictor of slacktivism, which is an underlying process of moral licensing. In the current study, liking a charity’s Facebook page can be regarded as a way to establish one’s morality,
because it is a form of supporting the charity (Kristofferson et al., 2013). The moral licensing effect entails that by performing moral behaviour, people establish their sense of morality, which licenses them to subsequently behave immorally (Monin & Miller, 2001). Sachdeva, Iliev and Medin (2009) studied what motivates people to act altruistically and they suggest that affirming a moral self-concept by acting morally, establishes moral credits which license people to subsequently behave less morally. This means that initially acting morally makes people feel like they have already done enough (i.e., established moral credits), and feel they momentarily do not need to behave morally again. This moral licensing effect can be related to the current study in the sense that liking a charity’s Facebook page is a way of acting morally, which establishes moral credits that in turn license people who have liked the
charity’s Facebook page to be immune to the urge to act morally towards the charity again by donating money to it.
For example, Khan and Dhar (2006) found that students who were first asked to help a foreign student with the course material donated significantly less money to a local charity due to moral licensing effects than students in the control condition, whose first task was to identify words that were misspelled in a passage of text (i.e., they were not asked to give initial support). This is also in line with research by Meijers et al. (2015), who showed that when people donated to a charity (i.e., initial support), they reported lower intentions to protect the environment than people who did not donate. This was due to the fact that donating (acting morally) gave them a sense of license that subsequently freed them from being less environmentally friendly (acting immorally).
Moreover, Monin and Miller (2001) found that people are more willing to express attitudes that could be viewed as prejudiced (i.e., acting immorally) when their past behaviour has established their morality. When people were given the opportunity to express
subsequently favour a man for a stereotypically male job in a cement manufacturing
company, than a woman (i.e., to act immorally). Moral licensing and its underlying process of establishing moral credits, can thus be seen as another alternative explanation, in addition to slacktivism, for why liking a charity’s Facebook page (i.e., acting morally) can lead to less donating behaviour (i.e., acting less morally or immorally).
Taking the above results of slacktivism and moral licensing together, it can be hypothesized that liking a charity on Facebook makes people likely to donate less money to the charity as compared to merely viewing the charity’s Facebook page.
H1 – Liking a charity’s Facebook page leads to less donating behaviour than merely viewing
the Facebook page.
The mediating role of the moral self-image
To what extent people view themselves as moral beings could be a possible explanation for the occurrence of a moral licensing effect. The moral credits that one
establishes with initial moral behaviour such as liking a charity’s Facebook page can be seen as a currency of moral behaviour (Merrit, Effron & Monin, 2010), where credits that allow for behaving immorally decrease with every immoral act one engages in. These moral credits thus bolster one’s sense of self-worth and therefore are a part of the extent to which people view themselves as moral beings, a.k.a. someone’s moral self-image; people with a higher moral self-image thus have more moral credits than people with a lower moral self-image. Therefore, liking a charity’s Facebook page is expected to lead to a higher moral self-image due to the establishment of moral credits, which results in the person’s donating less money to the charity as an effect of moral licensing.
Someone’s moral self-image is relatable to constructs like moral identity, self-esteem and moral disengagement. But, whereas the three constructs mentioned are somewhat static - at least in response to a single event - and are designed around a set of traits, the moral
self-image is a person’s dynamic and malleable moral self-concept related to those traits (Jordan, Leliveld, & Tenbrunsel, 2015). Examples of those traits, derived from Aquino and Reed’s (2002) construct of the moral identity, are caring, compassionate, helpful, hard-working, friendly, fair, generous, honest and kind. Jordan et al. (2015) found that the moral self-image is malleable and responds to an individual’s moral and immoral actions in the outside world, such as liking a charity’s Facebook page and donating money to it. The moral self-image thus suggests a malleable rather than static view of someone’s morality. Therefore, it is an
important variable to consider in the current study regarding moral behaviour.
The above reasoning is in line with previous research. According to Sachdeva et al. (2009), moral (or immoral) behaviour can result from an internal balancing of moral self-worth (i.e., a change in the moral self-concept) and the cost inherent in altruistic behaviour. Sachdeva et al. (2009) show that when people’s moral self-image was manipulated by having them write a story about themselves that includes positive traits, they donated less money than when people’s moral self-image was manipulated by using negatively worded traits. This shows that when people feel they have already proven that they are moral (e.g., by liking a charity’s Facebook page), because affirmation of their morality has given them moral credits (i.e., boosted their moral self-image), they will donate less (due to moral licensing effects). Thus, the establishment of moral credits as an underlying process of moral licensing is expected to influence someone’s moral self-image, resulting in slacktivism.
Moreover, Khan and Dhar (2006) studied the mediating effect of a changing self-concept (i.e., moral self-image) between a prior virtuous decision (i.e., acting morally) and the preference for a luxury option over a necessity (i.e., acting immorally). The results showed that an initial act of moral behaviour does indeed boost someone’s moral self-image (i.e., establishing moral credits), which in turn led to immoral behaviour (Khan and Dhar, 2006).
With the results of the studies described above taken together, it can be hypothesized that moral behaviour such as liking a charity’s Facebook page, increases the establishment of moral credits and thus boosts someone’s moral self-image, which leads to subsequent immoral behaviour like donating less money to charity as compared to merely viewing the Facebook page. Simply looking at the Facebook page will not establish moral credits because it is not seen as a form of support to the charity.
H2 – Liking a charity’s Facebook page leads to a higher moral self-image, which leads to less donating behaviour, than merely viewing a charity’s Facebook page.
The moderating role of issue involvement
When looking at how an initial act of moral support towards a charity (i.e., liking its Facebook page) predicts donating behaviour towards that charity, another factor comes in that plays a moderating role in this relationship. With the large numbers of types of charities that represent endless types of good deeds, people will feel more affinity with one charity than another. Consequently, people will show different donating behaviour towards different kinds of charities. For example, a person who recently lost a loved one to cancer may well feel more inclined to donate money to a cancer research charity than to one that focuses on helping people in war zones. This type of involvement with a charity is called issue involvement; it refers to the relevance someone perceives in a certain charity based on their own values, interests and needs (Cao & Jia, 2017).
This type of involvement has been proven to affect information processing and decision making. Many studies have found that information search, cognitive responses to advertising, and brand loyalty are all affected by how involved a consumer is with the product or topic (Bennet & Gabriel, 1999; Buchholz & Smith, 1991; Celsi & Olson, 1988). In the case of charities, if people are highly involved with the charity in question, they see the charity’s activities as more personally relevant, necessary and important (Bennett, Mousley, Kitchin, &
Ali-Choudhury, 2007), which leads to a higher tendency to donate compared to others who are less involved (Curtis, Cnaan, & Evans, 2014). Issue involvement thus plays an important role in people’s donating behaviour. Next to this direct effect of issue involvement, there is also research suggesting that issue involvement may be an important moderator with respect to liking and donating. This different body of research suggests that, next to moral licensing literature, initial support such as liking a charity’s Facebook page will only lead to less donating behaviour if the level of involvement one has with the charity is low.
That is, when someone has a high level of involvement with the charity, donating behaviour will logically be more. This is called a cognitive consistency effect and suggests that inconsistent cognitions (e.g., ‘I lost a loved one to cancer’, versus ‘I do not support a charity that focuses on cancer and will not donate money to it’) produce an aversive feeling of dissonance, which motivates people to reduce the underlying inconsistency and to maintain a state of consonance (i.e., ‘I lost a loved one to cancer’, and ‘I support the charity that focuses on cancer by donating money to it’; Festinger, 1957; Gawronski & Strack, 2012).
Kristofferson et al. (2013) studied the moderating role of issue involvement with the charity between liking and donating. This involvement was manipulated by having the
participants read a newspaper article about the recent activities undertaken by the organization that participants chose to support by liking its Facebook page. The described activities were ambiguous and could be interpreted as being either positive or negative. To divide the participants into a high involvement and low involvement condition, they were instructed to answer some manipulation-check questions. Participants who interpreted the actions
described in the newspaper article as positive were in the high involvement condition, while those who interpreted them as negative had misaligned values with the charity, and thus were in the low involvement condition. Participants who liked a charity’s Facebook page and were in the low involvement condition were less likely to help stuffing envelopes for a mailing
campaign of the charity (i.e., they acted less morally) than people who liked a charity’s Facebook page and had higher levels of issue involvement (Kristofferson et al., 2013). This suggests that after engaging in an initial moral act of support towards the charity (i.e., by liking its Facebook page), people will do less on a subsequent moral task for that charity (i.e., donate less money to it), but only when they have a lower issue involvement with that charity (slacktivism effect). In contrast, people who are more highly involved with the charity’s issue will also do more on a subsequent moral task for that charity (i.e., donate more money – consistency effect).
It can be hypothesized that in the current study, if people have a lower issue
involvement with the charity, they will donate less money after they have liked its Facebook page than if they have merely viewed the Facebook page. However, if people have higher issue involvement with the charity, they will not donate less money if they have liked its Facebook page than if they have merely viewed the Facebook page.
H3 – Liking a charity’s Facebook page leads to less donating behaviour than merely viewing
a charity’s Facebook page, but only for people who are less involved with the charity’s issue, not for people who are more highly involved with the issue.
Moreover, when taking the mediating role of the moral self-image into account again, issue involvement is also expected to moderate the relationship between liking a charity’s Facebook page and a person’s moral self-image. This means that only if a person has a low issue involvement, liking a charity’s Facebook page will lead to a higher moral self-image (and thus to less donating behaviour). Because, as discussed above, if a person already is highly involved with the charity’s issue, she is likely to donate more money to that charity after liking its Facebook page than less involved people (Kristofferson et al., 2013). The upshot is that she will not feel a boost in her moral self-image (i.e., will not establish moral credits) by liking the charity’s Facebook page, because this is normal behaviour in her eyes.
Less involved people on the other side will not regard the initial support as normal behaviour, but will consider it something better than normal. Therefore, they will feel a boost in their moral self-image when they like a charity’s Facebook page, establishing moral credits that license them to subsequently act less morally and thus donate less money to that charity than their more highly involved counterparts.
This reasoning is supported by Effron, Cameron and Monin (2008), who use the concept of racism to show a similar effect to the current study because expressing racism can be seen as a way of acting immorally. As initial support, participants were asked to endorse a black or a white candidate for president (Effron et al., 2008). As subsequent support,
participants were asked to indicate to which group they would donate money: either to people in a poor white neighbourhood, or to people in a poor black neighbourhood. Issue
involvement was assessed by measuring to what extent they were racist. The results showed that endorsing a black candidate for president rather than a white candidate (i.e., acting morally) established people’s morality, which led to more comfort in favouring whites over blacks (i.e., acting immorally; Effron et al., 2008). Acting morally (by endorsing a black candidate) thus led to subsequently acting less morally (by choosing to allocate funds to a poor white neighbourhood rather than a poor black neighbourhood), but this was only the case for people with pre-existing higher racist attitudes and thus a lower involvement with the black candidate (i.e., a lower issue involvement in the current study). These findings suggest that, with regard to charities and Facebook, liking a charity’s Facebook page establishes moral credits, leading to a higher moral self-image, which leads to donating less money. But this will only be the case for people with lower levels of issue involvement with the charity, not for people with higher levels of issue involvement with it (See Figure 1). On this basis, the following hypothesis is formulated:
behaviour than merely viewing a charity’s Facebook page, but only for people who are less involved with the charity’s issue, not for people who are more highly involved with the issue.
Figure 1. Conceptual model of the study
Methods Participants
The participants were 145 international Communication Science Bachelor students from the University of Amsterdam, who had given advance informed consent for participating in a study about charities and social media. They received 2 research credits for their
participation. The experiment was started by 148 participants, but three had to be removed. One did not complete the experiment, a second filled in the experiment twice and saw both conditions, and the third could not log in to Facebook due to security issues and therefore could not view or like the Facebook page. Therefore, three participants were excluded from the data for further analysis. Seven participants in the like condition indicated they did not actually like the Facebook page and were thus also excluded from analysis, so the remaining sample consisted of 138 participants. They were aged between 18 and 31, with an average age of 20.59 years (SD = 2.35). Of the respondents, 17.2 % were male and 82.8 % female.
Study design
independent variable and issue involvement as a continuous individual difference predictor. The independent variable initial support was manipulated in the experiment by having participants like (initial support) versus view (no initial support) a charity’s Facebook page and the dependent variable was donating behaviour. Of the 138 participants, 63 were in the like condition and 75 in the view condition.
Procedure
Upon arrival in the lab, participants first were instructed to read and sign a factsheet and informed consent before starting the questionnaire. They were escorted to a cubicle where they started the questionnaire on a computer. First, they were informed with a cover story which said that they were taking part in a study on Facebook pages from charities and that the study investigated how consumers would experience their social media pages. They were given a list of nine charities with a short description of their goals and then were asked to choose a charity whose Facebook page they would like to explore. These charities were the ALS Association, Amnesty International, the Cancer Research Institute, Doctors Without Borders, Greenpeace, the Salvation Army, Unicef, World Animal Protection, and the World Wildlife Fund. After choosing a charity, the participants were randomly assigned to either the like condition or to the view condition of that specific charity. At this point they were
instructed to click on the link to be redirected to the Facebook page and to either look at the page and then click the like button (i.e., initial support condition), or to merely look at the page (i.e., no initial support condition). Upon a participant’s mouse click, the Facebook page automatically opened in a new window on the screen.
Once they had returned to the questionnaire, the dependent variable of donating behaviour was measured by asking if they would (partly) donate the 40 Euros they could win to the charity of their choice. After that, some filler tasks were presented to distract their minds. Thereafter, moral self-image and issue involvement were measured and a like check
was done in order to see whether participants who had been assigned to the like condition actually did like the Facebook page or not. If they did not like the Facebook page, they were asked to briefly explain their reasons. Finally, the biographical items of gender and age were inventoried.
Moral self-image. To measure moral self-image, an existing scale created by Jordan
et al. (2015) was used. Participants were asked to indicate on a scale from 1 to 9 how much the statements of the nine traits from Aquino and Reeds’ (2002) moral identity applied to them: Compared to the caring/compassionate/
helpful/hard-working/friendly/fair/generous/honest and kind person I want to be, I am…, where 1 indicated much less than the person I want to be, 5 indicated exactly as the person I want to be, and 9
indicated much more than the person I want to be. See Appendix A for the full scale. A principal component analysis (PCA) shows that from the nine items, two components were extracted with an eigenvalue above one. For the first component, EV = 3.85; R2 = .43 and for the second component, EV = 1.06; R2 = .12. Since the scree plot shows a clear point of inflexion after component one and the eigenvalue of component two is only slightly above one, it can be assumed that the nine items form a uni-dimensional scale. All items correlate positively with the first component, the variable “Kind” has the strongest association (factor loading is .80) and the variable “Hard working” has the weakest
association (factor loading is .04). Reliability of the scale is good, Cronbach’s = .82. If the variable “Hard working” was deleted from the scale, Cronbach’s = .83. Because this is only a minor difference, no items will be deleted for further analysis. Therefore, it appears that the scale measures someone’s moral self-image accurately.
Issue involvement. To measure issue involvement, an existing scale by Bae (2008)
was used. Bae (2008) used it to measure people’s involvement with cornea donation by means of ten 7-point bipolar adjective items. For this study, the scale was adjusted for all of the nine
charities participants could choose from. Some examples of the ten items are (un)important, (un)exciting and (un)involving. For the World Wildlife Fund, for example, involvement was
measured by asking participants: For me, protection of plant and animal species on earth is unimportant – important, as this is the spearhead of the WWF. See Appendix B for the full
scale.
A principal component analysis (PCA) showed that from the ten items, two
components were extracted with an eigenvalue above one. For component one, EV = 4.87; R2 = .49 and for component two, EV = 1.36; R2 = .14. Since the scree plot shows a clear point of inflexion after component one, it can be assumed that the ten items form a uni-dimensional scale. All items correlate positively with the first component, the variable “Needed” has the strongest association (factor loading is .84) and the variable “Fascinating” has the weakest association (factor loading is .08). Reliability of the scale is good, Cronbach’s = .87. For none of the items Cronbach’s will be higher than .87 if they were deleted. Therefore, no items will be deleted for further analysis and it thus appears that the scale does indeed measure someone’s issue involvement with the charity.
Dependent variable
Donating behaviour. The dependent variable of this study was donating behaviour.
To measure this, a raffle was being held among all participants. Three bol.com vouchers of 40 Euros each were being raffled. To link this to donating behaviour, participants were told they could win the voucher, and if they did, they could donate it (or part of it) to the charity of their choice. They could indicate whether they would keep the 40 Euros completely for themselves if they won, whether they would donate a part of it to the charity and keep the remainder, or whether they wanted to donate the full 40 Euros to the charity. For analysis, a 9-point scale was used with 1 being keeping 40 Euros for myself, 5 being 20 Euros to the charity and 20 Euros to myself, and 9 being 40 Euros to the charity.
Results Randomization check
To check whether men and women were equally distributed between the two
conditions of initial support (yes; like vs. no; view), a Chi-square analysis was conducted. The test showed that there was no difference in participants’ gender across conditions, χ2
(1) = .452, p =.501. In the like condition, 82.5% were female and in the view condition 86.7% were female.
To check whether the mean age of participants was comparable over the conditions, a t-test was conducted. The t-test showed that age did not differ across the like (M = 20.59, SD = 2.61) and view conditions (M = 20.60, SD = 2.08), t(136) =.03, p = .975, 95% CI [-.79; .82]. This means that the participants’ ages were comparable over the conditions.
Main Analyses
To test hypothesis 1, that liking a charity’s Facebook page leads to less donating behaviour than merely viewing a charity’s Facebook page, an independent t-test was done with initial support (yes; like vs. no; view) as the independent variable, and donating behaviour as dependent variable. The t-test showed that donating behaviour did not differ among participants in the like (M = 5.95, SD = 3.00) and view conditions (M = 5.95, SD = 2.70), t(136) = -.01, p = .146, 95% CI [-.97; .95]. This means that hypothesis 1 was rejected: liking a charity’s Facebook page does not lead to less donating behaviour than viewing a charity’s Facebook page.
The role of moral self-image
To test hypothesis 2, that liking a charity’s Facebook page leads to a higher moral self-image which leads to less donating behaviour than merely viewing the Facebook page, a three-stage regression analysis was conducted, based on the approach of Baron and Kenny (1986). To establish mediation, the following conditions must hold: First, the independent
variable (initial support) must have an effect on the mediator variable (moral self-image) in the first model. Second, the independent variable (initial support) must have an effect on the outcome variable (donating behaviour) in the second model. Third, the mediator (moral self-image) must have an effect on the outcome variable (donating behaviour) when regressing the outcome variable (donating behaviour) on both the mediator and the independent variable in the third model. Lastly, the effect of the independent variable (initial support) in the third model must be weaker than when it is by itself in the second model. The data did not meet these criteria. First, moral self-image was not significantly higher in the like condition compared with the view condition, t(136) = -.35, b* =-.03, p =.728. Second, donating behaviour was not lower in the like condition than in the view condition, t(136) = .01, b* = .001, p = .991. Third, there was no effect of moral self-image on how much people donated, t(135) = -.16, b* = -.014, p = .876. And the effect of initial support on donating behaviour in
the third model was also insignificant, t(135) = .01, b* = .001, p = .994, F(2, 135) = .01, p = .988 (See Table 1).
This means that hypothesis 2 was rejected: liking a charity’s Facebook page does not lead to a higher moral self-image and less donating than viewing a charity’s Facebook page.
Table 1
Regression models to predict mediation of the moral self-image, from initial support (yes; like vs. no; view) on donating behaviour, hypothesis 2 (N = 138)
DV: Moral self-image Model 1 b* DV: Donating behaviour Model 2 Model 3 b* b* Constant Initial support Moral self-image 4.71* -.03 5.95* .001 6.14* .001 -.014 R2 < .01 < .01 < .01 F .12 < .01 .01
The role of issue involvement
To test the hypothesis that liking a charity’s Facebook page leads to less donating behaviour than merely viewing a charity’s Facebook page, but only for people who are less involved with the charity’s issue, not for people who are more highly involved with the issue, a simple moderation following Hayes’ (2015) Model 1 for PROCESS was done (using bootstrapping analyses with 5000 samples).
Results showed an insignificant effect of initial support (yes; like vs. no; view) on donating behaviour, t(134) = -1.48, b = -6.14, p = .140. This effect remains insignificant when looking at the interaction of initial support (yes; like vs. no; view) with issue involvement on donating behaviour, t(134) = 1.49, b = 1.02, p = .137. Also, the model as a whole was not significant, F(3, 134) = .750, p = .524. This means that hypothesis 3 was rejected: liking a charity’s Facebook page does not lead to less donating behaviour than viewing a charity’s Facebook page if people are less involved with the charity’s issue than more highly involved.
To test the hypothesis that liking a charity’s Facebook page leads to a higher moral self-image and less donating behaviour than merely viewing a charity’s Facebook page, but only for people who are less involved with the charity’s issue, not for people who are more highly involved with the issue, a moderated mediation following Hayes’ (2015) Model 7 for PROCESS was done (using bootstrapping analyses with 5000 samples). This allowed for testing the indirect effect of initial support (yes; like vs. no; view) on donating behaviour through moral self-image, moderated by issue involvement. Results are presented in Table 2 (See Appendix C) and b-values of the relationships in the conceptual model are shown in Figure 2.
Results showed an insignificant effect for initial support (yes; like vs. no; view) on donating behaviour, t(135) = .01, b < .01, p = .994, so the first condition for moderated mediation was not met. Also, an insignificant interaction term for initial support with issue
involvement in predicting moral self-image was found, t(134) = -.54, b = -.12, p = .592, violating the second condition for moderated mediation. Thus, for participants that are less involved, liking was not associated with less donating behaviour, as was expected under hypothesis 3. Also, the third condition was not supported; moral self-image did not
significantly affect donating behaviour, t(135) = -.16, b = -.04, p = .876. Lastly, the magnitude of the conditional indirect effect of initial support via moral self-image on donating behaviour was not different across high, average and low levels of issue involvement, where low
involvement was 1SD below average and high involvement 1SD above average. The results showed that for all three levels of issue involvement, there was no conditional indirect effect of initial support (yes; like vs. no; view) on donating behaviour through moral self-image, index of moderated mediation = .01, SE = .07, 95% confidence interval (CI) = [-.10 to .19]. The lower part of Table 2 (See Appendix C) presents the test of indirect effects for each level of issue involvement. Not in line with expectations, results demonstrated that the conditional indirect effect of initial support on donating behaviour through moral self-image was not significant for participants with lower nor higher levels of issue involvement, so hypothesis 4 was rejected.
Figure 2. B-values of the relationships between initial support and donating behaviour as mediated by moral self-image and moderated by issue involvement.
Conclusion & discussion
The aim of this study was to provide insight in which role a person’s moral self-image and their involvement with the issue of a charity play in the decision whether to take part in subsequent meaningful support of that charity (i.e., donating), after either expressing initial support towards the charity (i.e., liking its Facebook page) or no initial support (i.e., merely viewing its Facebook page). From the results in the analyses can firstly be concluded that there is no difference in how much people donate to charity depending on whether they like or merely view the charity’s Facebook page. Secondly, how moral people consider themselves to be at the time of liking or viewing the charity’s Facebook page does not account for how much they donate. Moreover, the level of involvement with the charity’s issue does not influence how much they donate after liking or viewing the charity’s Facebook page. Lastly, the level of involvement with the charity’s issue does not influence whether liking or viewing the charity’s Facebook page leads people to view themselves as more or less moral, which in turn also does not account for how much they donate.
There are some limitations to how the current research was set up, which could be a possible explanation for the insignificant results. For example, all participants participated voluntarily and knew they would receive research credits as a reward for taking part in the study, which creates a specific group of test subjects and influences the generalizability of the sample. Furthermore, because this group only included students, the sample is unlikely to be representative of the general population who may make their decisions based on other criteria (Bryman, 2012). Moreover, personal values and certain individual tendencies are proven to exert powerful influences on choosing which charity to donate money to (Bennet, 2003). Therefore, in the current study, participants’ involvement with the charity was likely to already be on the higher side of the scale, because people could choose for themselves which
of the nine charities’ Facebook pages they would focus on, which makes it difficult to find a distinction between different levels of involvement.
Although the findings do not support any of the hypotheses formulated, this could be explained by some differences between the literature they were based on and the current study. First of all, John et al. (2017) only found a marginal significant effect (p = .06) of liking a Facebook page on how favourably people viewed that brand (e.g., donating to
charity). Secondly, Khan and Dhar (2009) did find a mediating effect of the moral self-image between initial moral support and subsequent immoral behaviour, as mentioned in the
theoretical framework; however, to my knowledge this mediation effect has been found only once in research to date (by Khan and Dhar, 2009). Replication of this study was thus
important to see whether the moral self-image did indeed play a mediating role, which the results of the current study deny. For the future, more replications that study the direct effect of liking on donating and the mediating role of the moral self-image are needed in order to do a meta-analysis like John et al. (2017) which will give a better understanding of the effect of liking and the moral self-image on donating behaviour. Moreover, rather than manipulating issue involvement in the study like Kristofferson et al. (2013) did, in the current study the issue involvement was measured. This could be a possible explanation for the fact that findings of the current study are not in line with Kristofferson et al. (2013).
Lastly, Effron Cameron and Monin (2008) take the notion of moral credentials as an explanation for why initial moral support would lead to less subsequent moral support for people who are less involved with the subject of support. But in the current study, the establishment of moral credits due to liking a charity’s Facebook page was expected to increase a person’s moral self-image. The difference between moral credits and moral credentials is explained by Merrit, Effron and Monin (2010), where the former is seen as a moral bank-account that can become empty at a certain point and the latter as a model that
changes one’s morality as a whole. This means that moral credits (as the current study suggested) are a currency that is earned by initial moral behaviour, causing the moral self-image to increase (i.e., emphasizing the malleability of the moral self-self-image), whereas moral credentials function like a character witness on which one can call to testify that subsequent behaviour is not immoral (e.g., not donating to charity; Miller & Effron, 2010). Hence moral credentials provide an eternally lasting license to behave less morally (rather than the credits that only provide a license to behave less morally for a fixed amount of times). This
difference in the literature on moral licensing could explain the unexpected outcome that a person’s level of involvement with the charity’s issue does not influence whether liking or viewing the charity’s Facebook page leads people to view themselves as more or less moral, which in turn also does not explain how much they donate. Because if moral credentials were the explanation of the underlying process of moral licensing, rather than moral credits, this means that a person would always feel like she does not have to donate to charity, regardless of whether she has a low or higher issue involvement with the charity.
An interesting follow-up for this research could therefore be to see if there is a threshold for the amount of moral credits one needs to establish before a slacktivism effect occurs, for example, by taking into account different levels of effort put in the initial support. Liking a Facebook page is a very small action, performed in less than a second with just one mouse click. Future research could look at other types of initial support related to a charity’s Facebook page that require more effort. These types of support might be, for example, other types of social engagement on Facebook, such as commenting on the charity’s Facebook posts, or sharing them, accompanied by a personally written message. Aronson and Mills (1959) show that when a person has to exert herself more to gain membership of a discussion group, she will perceive the group as being more attractive than people who had to do little or nothing to become part of the group. This relates to engagement on a charity’ Facebook page
in that more effort (e.g., commenting on a charity’s Facebook post) will lead to perceiving the charity more favourably, which in turn could lead to more donating behaviour than when no initial support was given. Furthermore, Norton, Mochon and Ariely (2011) show the existence and magnitude of the so-called IKEA effect. This effect entails that people attach greater value to made products (e.g., sharing a charity’s Facebook post, accompanied by a self-written message) as opposed to ready-made products (e.g., merely sharing the ready-made post by the charity), meaning that the amount of effort put in to make the product (i.e., Facebook post), has increased its valuation. Thus, perhaps sharing a charity’s Facebook post and accompanying it by a self-written message would lead to more donating than merely liking the charity’s Facebook page because the greater effort increases how important a person finds the charity.
In sum, it would be a mistake to regard the outcomes of the current study as evidence for the complete lack of effectiveness of Facebook pages for charities. Instead, the act of liking a charity’s Facebook page could be valuable to the charity in the sense that it allows for building a relationship with supporters in the long run, providing them with knowledge and increasing awareness about the societal issues the charity intends to address. However, this does not mean that people will also donate more after they have liked the charity’s Facebook page.
References
Aronson, E., & Mills, J. (1959). The effect of severity of initiation on liking for a group. The Journal of Abnormal and Social Psychology, 59(2), 177.
Aquino, K., & Reed II, A. (2002). The self-importance of moral identity. Journal of personality and social psychology, 83(6), 1423.
Bae, H. S. (2008). Entertainment-education and recruitment of cornea donors:
The role of emotion and issue involvement. Journal of Health Communication, 13(1), 20-36.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of personality and social psychology, 51(6), 1173.
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.
Bennett, R., & Gabriel, H. (1999). Charity involvement and customer preference for charity brands. Journal of Brand Management, 7(1), 49-66.
Bennett, R., Mousley, W., Kitchin, P., & Ali-Choudhury, R. (2007). Motivations for participating in charity-affiliated sporting events. Journal of Customer
Behaviour, 6(2), 155-178.
Brown, T. J., Barry, T. E., Dacin, P. A., & Gunst, R. F. (2005). Spreading the word: Investigating antecedents of consumers’ positive word-of-mouth intentions and
behaviours in a retailing context. Journal of the Academy of Marketing Science, 33(2), 123-138.
Bryman, A. (2012). Social research methods. Oxford university press.
Cao, X., & Jia, L. (2017). The Effects of the Facial Expression of Beneficiaries in Charity Appeals and Psychological Involvement on Donation Intentions. Non-profit
Management and Leadership, 27(4), 457-473.
Carey-Simos, G. (August 19, 2015). How much data is generated every minute on social media? Retrieved November 27, 2017, from http://wersm.com/how-much-data-is-generated-every-minute-on-social-media/
Celsi, R. L., & J. C. Olson (1988). The Role of Involvement in Attention and Comprehension Processes. Journal of Consumer Research, 15(2), 210-244.
Effron, D. A., Cameron, J. S., & Monin, B. (2009). Endorsing Obama licenses favouring whites. Journal of experimental social psychology, 45(3), 590-593.
Festinger, L. (1957). A theory of cognitive dissonance. Evanston, IL: Row Peterson.
Gawronski, B., & Strack, F. (2012). Cognitive consistency: A fundamental principle in social cognition. New York, NY: Guilford Press.
Greenwood, S., Perrin, A., & Duggan, M. (2016). Social Media Update 2016: Facebook usage and engagement is on the rise, while adoption of other platforms holds steady.
Retrieved from Pew Research Centre website:
http://www.pewinternet.org/2016/11/11/social-media-update-2016/
Hayes, A. F. (2015). An index and test of linear moderated mediation. Multivariate Behavioral Research, 50(1), 1-22.
Jackson, N. (2011, June 3). United Nations declare internet access a basic human right. Retrieved November 27, 2017, from
https://www.theatlantic.com/technology/archive/2011/06/united-nations-declares-internet-access-a-basic-human-right/239911/
impact of joining a brand’s social network on marketing outcomes. Journal of Marketing Research, 54(1), 144-155.
Jordan, J., Leliveld, M. C., & Tenbrunsel, A. E. (2015). The moral self-image scale: measuring and understanding the malleability of the moral self. Frontiers in psychology, 6, 1878.
Khan, U., & Dhar, R. (2006). Licensing effect in consumer choice. Journal of marketing research, 43(2), 259-266.
Krishna, A. (2011). Can supporting a cause decrease donations and happiness? The cause marketing paradox. Journal of Consumer Psychology, 21(3), 338-345.
Kristofferson, K., White, K., & Peloza, J. (2013). The nature of slacktivism: How the social observability of an initial act of token support affects subsequent prosocial
action. Journal of Consumer Research, 40(6), 1149-1166.
Meijers, M. H., Verlegh, P. W., Noordewier, M. K., & Smit, E. G. (2015). The dark side of donating: how donating may license environmentally unfriendly behaviour. Social Influence, 10(4), 250-263.
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.
Miller, D. T., & Effron, D. A. (2010). Chapter three-psychological license: When it is needed and how it functions. Advances in experimental social psychology, 43, 115-155.
Monin, B., & Miller, D. T. (2001). Moral credentials and the expression of prejudice. Journal of personality and social psychology, 81(1), 33.
Nah, S., & Saxton, G. D. (2013). Modelling the adoption and use of social media by non- profit organizations. New Media & Society, 15(2), 294-313.
(Harvard Business School Marketing Unit Working Paper No. 11-091). Retrieved from Harvard University’s DASH repository:
http://nrs.harvard.edu/urn-3:HUL.InstRepos:12136084
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.
Waters, R. D. (2010). The use of social media by non-profit organizations: An examination from the diffusion of innovations perspective. In Social computing: Concepts, methodologies, tools, and applications (pp. 1420-1432). IGI Global.
Webber, D. (2004). Understanding charity fundraising events. International Journal of Non-profit and Voluntary Sector Marketing, 9(2), 122-134.
Appendix A Moral Self-image Scale
For all nine traits, a statement was presented to which participants had to indicate how much they applied to them on the 9-point existing scale from Jordan et al. (2015).
1. Compared to the caring person I want to be, I am:
2. Compared to the compassionate person I want to be, I am:
3. Compared to the fair person I want to be, I am:
4. Compared to the friendly person I want to be, I am:
5. Compared to the generous person I want to be, I am:
1 2 3 4 5 6 7 8 9 Much less caring than the person I want to be Exactly as caring as the person I want to be Much more caring than the person I want to be 1 2 3 4 5 6 7 8 9 Much less compassionate than the person I want to be Exactly as compassionate as the person I want to be Much more compassionate than the person I
want to be 1 2 3 4 5 6 7 8 9 Much less fair than the person I want to be Exactly as fair as the person I want to be Much fairer than the person I want to be 1 2 3 4 5 6 7 8 9 Much less friendly than the person I want to be Exactly as friendly as the person I want to be Much friendlier than the person I want to be 1 2 3 4 5 6 7 8 9
6. Compared to the hard-working person I want to be, I am:
7. Compared to the helpful person I want to be, I am:
8. Compared to the honest person I want to be, I am:
9. Compared to the kind person I want to be, I am: Much less generous than the person I want to be Exactly as generous as the person I want to be Much more generous than the person I want to be 1 2 3 4 5 6 7 8 9 Much less hard-working than the person I want to be Exactly as hard-working as the person I want to be Much more hard-working than the person I want to be 1 2 3 4 5 6 7 8 9 Much less helpful than the person I want to be Exactly as helpful as the person I want to be Much more helpful than the person I want to be 1 2 3 4 5 6 7 8 9 Much less honest than the person I want to be Exactly as honest as the person I want to be Much more honest than the person I want to be 1 2 3 4 5 6 7 8 9 Much less kind than the person I want to be Exactly as kind as the person I want to be Much more kind than the person I want to be
Appendix B Involvement scale
For all nine charities, a different statement was presented, to which participants had to indicate their involvement measured with ten adjective items on a 7-point bipolar scale adjusted from Bae (2008). Per charity, the core of their goal was detected and referred to in the statement. The statement was unobtrusive, meaning that the name of the charity was not mentioned so that participants would be less likely to connect this question to the charity. The items were (1) important–unimportant; (2) boring–interesting; (3) relevant–irrelevant; (4) exciting–unexciting; (5) means nothing–means a lot to me; (6) appealing–unappealing; (7) fascinating–mundane; (8) worth- less–valuable; (9) involving–uninvolving; and (10) not needed–needed.
The statements for each charity were framed as following, without mentioning the name of the charity: (ALS Association) To me, research on nerve and muscle diseases like ALS is…; (Amnesty International) To me, equal human rights worldwide is…; (Cancer Research Institute) To me, research on the combat of cancer is…; (Doctors Without Borders) To me, helping people in disaster areas is…; (Greenpeace) To me, solving large-scale
environmental issues is…; (Salvation Army) To me, helping the less fortunate of society is…; (Unicef) To me, equal rights for children worldwide is…; (World Animal Protection) To me, animal welfare is…; (World Wildlife Fund) To me, protection of plant and animal species is…
Appendix C Table 2
Regression results for moderated mediation, hypothesis 4 (N = 138)
Outcome: Moral self-image Outcome: Donating behaviour
Predictors Coefficient SE LLCI ULCI Coefficient SE LLCI ULCI
Constant 3.90* .93 2.07 5.74 6.14* 1.27 3.63 8.65
Initial support .67 1.37 -2.05 3.39 .003 .49 -.96 .96
Moral self-image -.04 .26 -.56 .47
Issue involvement .13 .15 -.17 .44
Initial support x issue involvement -.12 .23 -.57 .33
R2 .01 .00
Issue involvement Indirect effects Boot SE Boot LLCI Boot ULCI
5.29 (low) < -.01 .06 -.16 .10
6.00 (average) < .01 .04 -.09 .09
6.71 (high) .01 .07 -.11 .17
Note. Unstandardized regression coefficients are reported. Bootstrap sample size: 5.000. LLCI = lower level of the 95% bootstrap percentile confidence interval; SE = standard error; ULCI = upper level of the 95% bootstrap percentile confidence interval. *p < .001.
