MASTER MARKETING
MANAGEMENT THESIS
T
HE POWER OF WORDS IN CHARITY ADVERTISEMENTS
:
S
TUDYING THE EFFECTS
OF
“
HELP VERSUS SUPPORT
”
1 COLOPHON
Rapport data
Title: Master Marketing Management Thesis
Subtitle: The power of words in charity advertisements: Studying the effects of “help versus support”
Author data
Name: Sharen Baijense
Date of birth: 25/08/1994 Address: Fongersplaats 121
9725LE Groningen The Netherlands
Phone: +316-11140086
Education: MSc Marketing Management Email: s.r.baijense@student.rug.nl
Student number: 3195198
University data:
Instance: University of Groningen Department Economics and Business Education: MSc Marketing Management Qualification: Master thesis
Language: English
Address: Grote Kruisstraat 2/1 9712 TS Groningen The Netherlands Supervisor: Marijke Leliveld
Email: m.c.leliveld@rug.nl
2 MANAGEMENT SUMMARY
Charities have new ways of acquiring donations and members. They try to activate people by using the popular social media platform Facebook. Here they try to generate so-called leads by using Facebook posts, which contains a link to an online petition. When people sign this petition with their name and e-mail address a lead is created and used internally for the mission of the charity. Little is known about what makes an ad on Facebook effective in generating leads. In this study I will focus on the words that are used in the Facebook post and see whether there is a difference in the effectiveness of the word ‘help’ versus ‘support’ when trying to persuade people to sign the petition. This study has been performed in Dutch, so the words used are ‘help’ and ‘steun’. These two words are the most frequently used words in charity’s advertisements and are used interchangeably. However, there might be a difference in the effectiveness of these words.
Previous literature shows that some words look interchangeable, but in practise they sometimes are not. With ‘help’ and ‘support’ I found a difference in the definitions of these words when looking at the activeness of the question when one of these words is used. I tested this via a pre-test and found out that ‘help’ is perceived as more active and direct than the word ‘support’, which was seen as a more passive way of asking someone to aid you. Therefore, I hypothesised that help would be more effective in generating leads and engagement on Facebook than support, because when someone asks you more actively you would expect that people are more willing to help you. In this research I also looked at three possible indicators that could have an influence on the effectiveness of ‘help’ versus ‘support’. Namely, how good people feel after helping/supporting, how effective they think their help/support is and how committed the people are to the organisation and its mission. With these three indicators I hypothesised that ‘help’ has a stronger impact on the indicators, which leads to more effectiveness of the Facebook post when looking at engagement. The last thing I looked at was the dynamics between the likes, comments and shares of the Facebook post with the leads to investigate if there would be results in line with moral consistency (I behaved good once, so I will continue to do so) or moral licensing (I did good thing once so now I can do a bad/less good thing).
3
‘support’ condition. The sample was generalizable because this study covers all demographic segment groups, however it is not known if the results are also generalizable in other languages like English. For this survey, I used a real Facebook post to make this study as close to reality as possible.
In the end I had five effectiveness measurements: 1) likes, 2) comments, 3) shares, 4) leads and 5) engagement (combination of the other four). Even though support scored higher on almost all these effectiveness measurements (except share), none of these differences was significantly different. Indicating that ‘help’ or ‘support’ did not influence the effectiveness and engagement of the Facebook post differently. The same result was found with the indicators. The word conditions did not influence the three indicators. The positive feeling people get from donating did, however, have a positive effect on all the effectiveness measures and all the indicators did have a positive effect on the leads (not significantly on other effectiveness measures), except for commitment to the organisation, which had a negative effect. This was unexpected, especially because I found a positive relationship between past behaviour towards the charity (via donations) and engagement with the Facebook post. Lastly, I found evidence of moral consistency as most people who ‘liked’, shared and/or commented the Facebook post also generated leads.
4 PREFACE
Hello and welcome to my master thesis, which I wrote for the Master “Marketing management” at the University of Groningen. My name is Sharen Baijense and I am the author of this thesis.
After a guest lecture in December 2017 my interest in charity organisations and their way of acquiring donations was triggered. That is why I decided to do a study in this field. Along the way I specified my study on Facebook, which was very interesting because there is not much research on the effectiveness of this very popular social media platform. I enjoyed my time writing this report and I hope you enjoy reading it.
I want to thank a few people that helped me along the way in writing this master thesis. My first and biggest thanks goes to my supervisor Marijke Leliveld for all the help, support, tips, time and effort she took to help me get to a next level in writing a master thesis. Next up I want to thank Bart van Langen and Floyd van den Berg for helping me with optimizing the English spelling and grammar in this report.
5 TABLE OF CONTENTS COLOPHON ... 1 MANAGEMENT SUMMARY ... 2 PREFACE ... 4 INTRODUCTION ... 6 LITERATURE REVIEW ... 8
Perceived word differences ... 8
Warm glow ... 9
Perceived effectiveness ... 10
Commitment ... 11
The dynamics of likes, shares, comments and leads ... 11
CONCEPTUAL MODEL ... 13
METHODOLOGY ... 14
Participants and Design ... 14
Independent variable ... 14
Behavioural intentions ... 15
Mediators: expected warm glow, perceived effectiveness and commitment... 15
RESULTS ... 17
Independent variable on the dependent variable: ... 17
Independent variable on the mediators: ... 17
Mediators on the dependent variable: ... 18
6 INTRODUCTION
Charities use many advertisements to persuade people to donate money, e.g. via one-time donations or by getting new members to the charity, time or products depending on the charity in question. For these charities it is critical that the advertisements are effective to generate as much donations as they need to help their cause. Recently, charities have implemented new tools for their advertisements with the growth of social media and developed new forms of supporting the charity, e.g. on Facebook, charities can use advertisements to receive “likes”, as this increases the spread of their ads, the reach of the ad and thereby the ad and brand awareness.
Another way to activate people is that charities try to generate so-called ‘leads’. To create such ‘leads’ charities can post advertisements on Facebook. This post contains a link, which brings people to a website where they can sign a petition for the cause advertised by the charity. When someone indeed leaves their contact information on such a site, this is called a lead. Generating such leads is growing in popularity. According to Marketing facts (2014), we live in a lead
generation. Getting information is more important than ever to reach people and to persuade them
with personalized ads. This area of charity campaigns on social media is a rather new topic and not many studies researched the effectiveness of charities on social media yet. One study that did investigate this effectiveness is the study of Phethean, Tiropanis and Harris (2015), which found that posts using pictures, compared to no pictures (type not stated), receive more responses on Twitter and Facebook.
7
Charity Steun (Support)
Help (Help)
Other words used to get donations
WNF X X Kom in actie, doe mee (come in action, act with us) KWF Kankerbestrijding X Sta op tegen (stand up against)
War Child X Doe mee (act with us)
Amnesty X X
Unicef X Geef (give)
Greenpeace X X Kom in actie (come in action) Artsen zonder grenzen X X Kom in actie (come in action) Dierenbescherming X X Word lid (become a member) Cliniclowns X X Kom in actie (come in action) Hartstichting X X
Table 1. Words used in Dutch charity organisations advertisements: help, support and other words. From January till March 2018 on their websites and in their advertisements.
However, it is not known which of these words are most effective in advertisements for generating the likes and leads on Facebook, especially for charity organisations. In a study on Twitter, Zarrella (2009) showed that the word ‘help’ (and not support) is in the top 20 words used in tweets that gets “retweeted” (shared by the readers with their network), suggesting that ‘help’ is more effective than ‘support’ as that word was not in the top 20. However, we do not know how effective the word ‘help’ or ‘support’ is in generating leads via Facebook. This study fills that gap by systematically study the differences and underlying processes of these two words. Furthermore, this study focusses on likes, leads and engagement on Facebook instead of money or time donations, which is also an area that has not been studied before within the charity industry.
8 LITERATURE REVIEW
Perceived word differences
As indicated in my introduction, in this thesis I will focus on the difference between the words ‘help’ and ‘support’ in charity advertisements posts on Facebook. One reason why words can have different effects on people is because people can interpret words differently. Multiple studies show that interchangeably-looking words sometimes differ when we look at the interpretation or consequences of these words. For example, Boholm (2012) shows that the words ‘risk’ and ‘danger’ differ in interpretation. Both words are used for situations that can have negative outcomes (e.g. physical injuries), but they differ in the situation where they should be used. ‘Risk’ is used when the person affected by the action is also the one causing it, while ‘danger’ is used when the one that is affected is not the one in action. However, people still use them interchangeably (Boholm, 2012). Besides that, Wilkinson (2008) shows the difference between the words ‘unwanted sex’ and ‘rape’. Even though these words look interchangeable, people feel they are different. When journalists use the words ‘unwanted sex’ people blame the victims more and give less severe punishments for the perpetrator than when they use the word ‘rape’ (Wilkinson, 2008). Thus, words can have different meanings and activate people in different ways even though the words look interchangeable.
9
To test if people indeed interpret the words ‘help’ and ‘support’ differently in terms of active-passive, I conducted a pre-test (see appendix 1 for full details). In this pre-test 64 participants answered different questions measuring their interpretation of the words ‘help’ and ‘support’ on a scale from 1 (absolutely passive) to 9 (absolutely active). Results show that the word ‘help’ (M = 6.95, SD = 1.70) is seen as a more active word than ‘support’ (M = 5.84, SD = 2.02) (t (63) = 3.01,
p = .004). Moreover, when participants were asked to classify the words as active versus passive,
82.8% of the participants classified ‘help’ as an active word, while only 53.1% classified ‘support’ as an active word. The Chi-Square test shows that this difference is significant (2 (1, N = 64) = 7.61, p = .006). These findings suggest that helping is more active than supporting. That said, asking more actively for aid could also lead to more active help in return as well, because people feel more the need to help when actively asked for it instead of passively. Thus, it is expected that the word ‘help’ will be more effective in the end than the word ‘support’. Put formally, H1: The
word ‘help’ (vs ‘support’) makes the Facebook post more effective in terms of generating leads and engagement.
To explain this argumentation in more detail, I will also study the potential drivers of this effect. That is, the idea that the two words differ in active- versus passive perceptions, might trigger various other anticipated emotions or cognitions, which I will discuss below.
Warm glow
10
states: “your donation will help (insert child name)” instead of “3000 people got injured because of this disaster, please help and donate here”. This feeling of not being able to help everyone, reduces the warm glow and therefore the intention to donate (Västfjäll, 2015).
In this research it is expected that the use of the word ‘help’, compared to ‘support’, will increase the expected warm glow due to the perceived activeness of the word. It can be expected that when people feel like they help more actively they also feel happier and have a higher warm glow effect. Therefore, the following hypothesis is generated: H2: The word ‘help’ (vs support) gives people
higher expected warm glow which results in more engagement.
Perceived effectiveness
As introduced above, a second potential mediator for the word effect can be the perceived effectiveness of the charity (the impact of the donations). Leliveld and Bolderdijk (2018) show that charities which are more effective in turning the donations to action are more likely to get donations than ineffective charities. This, however, is only the case when the action is defined in terms of number of lives saved and not when the action is defined in monetary terms (number of euro raised). This is understandable as people only want to donate if their donation matters for the moral cause and people feel betrayed when the charities turn out to be ineffective (Hager, 2004). Gneezy, Gneezy and Keenan (2014) found that donators have the tendency to avoid non-profit organisations/charities that use a high percentage of their income to fundraising and administrative costs. This, in turn, limits the effectiveness of the charity and lowers the incoming donations (Gneezy, et al., 2014). This perceived effectiveness can also influence the perceived warm glow that I mentioned before: The more effective the donation the warmer glow the people perceive, because when they have the feeling they can help more effectively they would feel better doing it. Therefore, I suggest that: H3: the word ‘help’ (vs support) gives people a higher perceived
effectiveness of the charity which results in higher expected warm glow.
11
(vs support) gives people a higher perceived effectiveness of the charity, which results in more engagement on the Facebook post.
Commitment
The last mediator that I will look at during this study is the commitment towards charity organisations of the Dutch citizens. The commitment of employees has shown to have a positive influence on the success of a company. However, it is also shown that customers have an influence on the success of the organisation especially when they are committed. The more people are committed to your organisation the more they are willing to donate and help your cause (Valéau et al., 2013). Commitment is one important way to link volunteers and donators to your company and gain income, as they do not benefit from donating except for earlier mentioned warm glow and satisfaction. Committing them to your organisation makes them more willing to donate and less willing to switch to another charity (Valéau et al., 2013).
Jak and Evers (2010) mention that there are different ways of organizational commitment: Affective commitment, continuity commitment and normative commitment. Affective commitment is the personal meaning people feel towards the organisation. People are continuity committed when they feel like they have no other choice than staying with the organisation, while people are normative committed when they find the organisations mission important (Jak, Evers, 2010). In charities continuity commitment is not relevant as members and donators are never committed to stay with the organisation if they do not want to, something which is relevant with their employer. Thus, in this study I will only focus on affective and normative commitment.
The dynamics of likes, shares, comments and leads
12
consistency, one in line with moral licensing and one in line with moral do gooder derogation theory (cf. Leliveld & Risselada, 2017).
Moral consistency implies that people have a relative stable ethical behaviour. When they behaved
ethically in the past it is likely they will act ethically in the future as well to signal to themselves how moral they are. Consequently, their past behaviour is a signal of their identity. This would imply that in this study when participants choose for example to “like” the post they are also more likely to participate in the lead, and vice versa (Joosten, Dijke, Hiel, 2014; Mullen, 2016; Leliveld & Risselada, 2017). Moral licensing, on the other hand, is when people balance their ethical behaviour and unethical behaviour. First people behaved ethically and then they behave unethically after and vice versa. After all they keep their ideal self-image, because they balance their ethically and unethically acts. The past behaviour determines which behaviour they need to balance their self-image out again. In this research that would mean that people after they for example share the Facebook post feel that they acted good enough and do not need to do more good: in other words, there is no need to sign the petition after a share (Merrit, 2010; Mullen, 2016; Leliveld & Risselada, 2017).
13 CONCEPTUAL MODEL
In Figure 1, the setup of this thesis is visualized:
Figure 1. Conceptual model.
14 METHODOLOGY
Participants and Design
To measure the earlier stated hypothesis, I have conducted a survey in Dutch and spread it via the online survey platform Qualtricsto the Dutch citizens. 204participants have filled in this survey during the first week of May 2018. Participants were randomly assigned to a one-factor (Word: help vs support) between-subject design. 51% saw the support factor and 49% the help factor. The sample consisted of 104 (51%) men and 100 (49%) women, with a mean age of 46 years (M = 45.84, SD = 17.17). In table 2 and 3 one can see that all the school levels and employment statuses participated in this study. All these demographics suggest that the sample is generalizable as this study covers all the demographic segment groups.
Highest school level Frequency
Not graduate high school 2.0% High school 26.0% Secondary vocational education (MBO) 29.4% Higher professional education (HBO) 28.9% University (WO) 13.7% Table 2. Highest graduated school level (N=204).
Table 3. Employment status level (N=204).
Relevant for the current study is that out of these 204 participants only 24 (11.76%) participants were already member of WWF and 48 (23.53%) gave money to WWF incidental last year (via text or via collection at the door). This means that 132 participants (64.71%) did not donate to WWF at all last year. This is useful for this study, because these 132 participants are the target group that needs to be reached via these Facebook posts and were the leads are the most important.
Independent variable
The independent variable was the word ‘help’ vs ‘support’ (in Dutch ‘help’ and ‘steun’). To manipulate this variable two ads were made which were identical to each other except for these Employment status Frequency
Student 10.8% Full-time 31.4% Part-time 12.7% Freelancer/ZZP 10.3% Out of work, but looking 5.4%
15
words. To make it more realistic I used a real Facebook ad with all the buttons as like, comment and share, as can be seen in the picture below. I used the post from WWF, which originally contained the word ‘support’ (steun) and used Photoshop to replace it with ‘help’.
Figure 2. Picture used in the survey with “support” (steun in Dutch).
Behavioural intentions
To measure the effectiveness of the Facebook post, participants were provided the option to 1) like the ad, 2) Comment on the ad, 3) share the ad and/or 4) sign the petition.
Mediators: expected warm glow, perceived effectiveness and commitment To measure the three mediator’s different theories were used:
16
effectively worldwide in three different items (see appendix 2). They also indicated on a nine items Likert scale how likely their help/support would make a difference (Vastfjäll, 2015). The four questions were combined in one sum variable after conducting a successful Cronbach’s alpha tests (0.857).
2. To measure the warm glow, I used the methodology of Vastfjäll et al. (2015). They measured the anticipated warm glow in a scale from 0 (no warm glow) to 100 (very strong warm glow). In this study I used the same scale.
17 RESULTS
Independent variable on the dependent variable:
In this chapter I will first look at the main effect of H1: The word ‘help’ (vs ‘support’) makes the
Facebook post more effective in terms of generating likes and engagement. I created one variable,
which is engagement by combining the four reactions options to 0 (they did not click yes on anything) and 1 (they clicked at least one time yes).
In Table 4, one can see that likes and leads are the most popular reactions on the advertisement. When looking at the difference between support and help, there is no significant difference between the percentages of the two words. So even though support scores slightly better on like and comments and a lot better on leads, which is in contradiction with the hypothesis H1, this difference is not big enough to be significantly different.
Overall percentage Support condition Help condition X2 (1, N = 204) Like 63.2% 65.4% 61.0% 0.42, p = .516) Comment 23.5% 25.0% 22.0% 0.25, p = .614) Share 40.2% 37.5% 43.0% 0.64, p = .423)
Go to petition and sign 68.1% 72.1% 64.0% 1.55, p = .214)
Engagement 78.9% 82.7% 75.0% 1.81, p = .178)
Table 4. Facebook reactions on the post overall, with support and with help (N=204).
Independent variable on the mediators: I then looked at three possible mediators:
Warm glow. The mean of peoples expected warm glow is 75.00 (SD = 20.86), which means that
people do indeed expect to feel better after helping or supporting the cause. However, the expected warm glow for people in the support condition was not significantly different (M = 76.59, SD = 19.68) from those in the help condition (M = 73.35, SD = 22.01) (t (204) = 1.108, p = .269). This does not support hypothesis 2.
Perceived effectiveness. The overall mean of effectiveness was 4.90 (SD = 1.40). However, the
18
effectiveness on the mediator warm glow, as I predicted in the literature review and hypothesis 3. Meaning that the perceived effectiveness has an impact on the warm glow.
Commitment. Overall, the mean of affective commitment is 4.27 (SD = 1.44) and normative
commitment 4.35 (SD = 1.50). In line with the previous results the support condition has a slightly higher mean in the affective commitment, but this support condition was not significantly different (M = 4.32, SD = 1.44) from those in the help condition (M = 4.21, SD = 1.45), (t (204) = 0.527, p = .598). The same results can be seen with the normative commitment. The support condition has a higher mean (M = 4.42, SD = 1.56) than the help condition (M = 4.29, SD = 1.44), however this difference is not significant (t (204) = 0.602, p = .548).
Mediators on the dependent variable:
I tested the effect of the mediators on all five dependent variables. For all these analyses I used binary regression analysis. In table 5, I show the B values and the significance of the different mediators. In this analysis 0 meant not doing anything and 1 meant clicking on the button.
Engagement Likes Comments Shares Leads
Warm glow .054 (p .000) .029 (p .004) .033 (p .021) .050 (p .000) .041 (p .000) Perceived effectiveness .778 (p .002) .198 (p .285) .130 (p .586) -.096 (p .641) .534 (p .010) Affective commitment -.417 (p .174) .138 (p .525) .084 (p .728) .000 (p 1.000) -.832 (p .001) Normative commitment .449 (p .077) .149 (p .430) .397 (p .098) .476 (p .024) .763 (p .001)
Table 5. Mediator effect on dependent variables: engagement, likes, comments, shares and leads (N=204).
19 Additional analysis
For additional analysis I looked at other possible independent variables (like gender, past behaviour and age) that could influence the dependent variables and looked at the relationship between the dependent variables (likes, comments, shares and leads). I only report the significant test result here.
First, I found an effect between the past relationship/commitment with WWF (they either donated last year or not) and the engagement with the Facebook post. From the donated group 90.27% were engaged in the Facebook group and only 72.72% from the not donated group were engaged, which is a big difference in the engagement of the two groups. This is indeed a significant difference (2 (1, N = 204) = 8.626, p = .003). Thus, this indicates that engagement is indeed related to the earlier connection of the charity with the donators.
20 GENERAL DISCUSSION
Conclusion
I started this study to investigate a new area in research: The effectiveness of advertisements on social media. I especially looked at the effectiveness of Facebook posts for charity organisations. In these posts I manipulated the first word, namely ‘support’ or ‘help’ to see which word was more effective in the advertisement when we look at the engagement of the Facebook users with the advertised post. Via a pre-test I discovered that people perceive the word ‘help’ as more active than the word ‘support’. Therefore, I expected that the word help would be more effective than the word support in creating engagement. I researched this word effect on behavioural intentions but also investigated three possible mediators: expected warm glow, perceived effectiveness and commitment. However, the results suggest that there is no noticeable difference between the two words on Facebook ad effectiveness. This holds for the behavioural intentions as well as the proposed mediators.
However, I found that likes and leads have the highest percentages of behavioural intentions, which is an interesting discovery. I discussed earlier that people might not want to show their friends that they support ‘World Wide Fund for Nature’ which can explain the lower comment and share, however likes are also seen by Facebook friends. Likes on the other hand are very easy and cost the least time and effort compared to comment, shares and leads, yet leads have the highest percentage, which cost the most time and effort. The reason for these outcomes remain unknown.
21 Limitations/future research
This study has some limitations which can be considered for future research. As there is only limited research on engagement of charities on social media it would be interesting to conduct future research to gain more knowledge about this social media platform.
Firstly, this research did not measure real behaviour, only the behavioural intentions. I tried to copy reality as close as possible, but participants still had to answer “would you click the like (or comment or share) buttons” instead of being able to click it. Someone also mentioned in the feedback that it depends on how busy they are that day if they look at the ad or not. It also would be interesting to see if there is a difference in behaviour when people use the so-called system one versus system two (Kahneman, 2012). System one stands for the “automatic mode” where people think fast and rely on emotions and stereotypes, system two is the “reflective mode” and people rely on logical thinking and conscious reasoning (Kahneman, 2012). These systems might influence the behaviour on Facebook because when people are busy they might rely on system one, whereas when people are in a relaxing state they might rely more on system two, which can change their behaviour. In future research it would therefore be interesting if this study can be performed on Facebook itself, instead of via a survey to see the real engagement and behaviour of participants.
22
Thirdly, in this study I measured all the mediators before I measured the main dependent variables (behavioural intention), which is why I asked the participants to indicate their expected warm glow instead of perceived warm glow of helping/supporting. However, this is a difficult measurement for people as you never know for sure what you are going to feel in the future. In a future research it would be interesting to ask the perceived warm glow after answering the dependent variable, behavioural intention, to see how people feel after helping, because as Andreoni (1990) stated: people feel happier when they helped others (Andreoni, 1990).
Fourthly, it would be interesting to see if the results hold for other languages as well, like English. The word effect for ‘help’ versus ‘support’ might be different than the literal Dutch translation of ‘help’ and ‘steun’.
Lastly, in this study I did not find any significant difference in the word effect of support and help on Facebook ad effectiveness. However, it would be interesting to study the word effect on actual donations like money or time. As money is more passive and donating time is more active the word condition could in this case make a difference (compared to online), because even though leads are more active than likes, they are still both more passive as people can do it from phone or laptop within a few minutes. With donating money or time this is not the case. Bauer, Bredtmann and Schmidt (2013) found that as soon as people do not have time for it because, e.g. due to the growing demand in the working market, people substitute their time donations to money donations. This indicates that donating time needs actively commitment, while money donating does not cost much effort and therefore can be done more passively (Bauer et al., 2013).
23 REFERENCE LIST
Andreoni J. (1990). “Impure Altruism and Donations to Public Goods: A Theory of Warm-Glow Giving”. Economic journal, 100: 464-477.
Bauer T.K, Bredtmann J, Schimdt C.M. (2013). “Time vs. money — The supply of voluntary labour and charitable donations across Europe”. European Journal of Political Economy, 32: 80-94.
Boholm M. (2012). “The Semantic Distinction Between “Risk” and “Danger”: A Linguistic Analysis”.
Risk Analysis, 32: 281-293.
Cambridge dictionary. (2018). “Help” (Accessed on 29/05/2018) [available on: https://dictionary.cambridge.org/dictionary/english/help]
Cambridge dictionary. (2018). “support” (Accessed on 29/05/2018) [available on: https://dictionary.cambridge.org/dictionary/english/support]
Dunn E.W, Aknin L.B, Norton M.I. (2008). “Spending money on others promotes happiness”.
American Association for the Advancement of Science, 319: 1687-8.
Encyclo. (2018). “Help” (Accessed on 09/03/2018) [available on: http://www.encyclo.nl/begrip/help] Encyclo. (2018). “Steun” (Accessed on 09/03/2018) [available on: http://www.encyclo.nl/begrip/steun] Gneezy U, Gneezy A, Keenan E.A. (2014). “Behavioural economics. Avoiding overhead aversion in charity”. American Association for the Advancement of Science, 346: 632-5.
Hager M.A. (2004). “Public Trust in the Public Face of Charities”. (Accessed on 04/04/2018) [available on: http://webarchive.urban.org/UploadedPDF/1000630_public_trust.pdf]
Jak S, Evers A.V.A.M. (2010). “Onderzoeksnotitie: Een vernieuwd meetinstrument voor organizational commitment” Gedrag en Organisatie, 23: 158-171.
Joosten A, Dijke M, Hiel A. (2014). “Feel Good, Do-Good!? On Consistency and Compensation in Moral Self-Regulation”. Journal of Business Ethics, 123: 71-84.
Kahneman D. (2012). Thinking, Fast and Slow. Penguin Books Ltd. 1st edition.
Leliveld, M.C., & Bolderdijk, J.W. (2018). (In)Effective Altruism: Charities Using Their Head Alienate Consumers Who Use Their Heart. Working paper
Leliveld M.C, Risselada H. (2017). “Dynamics in charity donation decisions: Insights from a large longitudinal data set”. Science advances: 1-8.
Marketing facts. (2014). “Lead generation”. (Accessed on 09/03/2018) [available on: https://www.marketingfacts.nl/topic/lead-generation]
Merritt A.C. (2010). “Moral self-licensing: When being good frees us to be bad”. Social and Personality
Psychology Compass, 4: 344-357.
Monin B, Sawyer P, Marquex M.J. (2008). “The Rejection of Moral Rebels: Resenting Those Who Do the Right Thing”. Journal of Personality and Social Psychology, 95: 76-93.
Mullen E. (2016). “Consistency versus licensing effects of past moral behaviour”. Annual Review of
Psychology, 67: 363-385.
Phethean C, Tiropanis T, Harris L. (2015). “Engaging with Charities on Social Media: Comparing Interaction on Facebook and Twitter”. Springer International Publishing, 9089: 15-29.
Touré-Tillery M, Fishbach A. (2017). “Too Far to Help: The Effect of Perceived Distance on the Expected Impact and Likelihood of Charitable Action”. Journal of Personality and Social Psychology, 112: 860-876.
Valéau P, Mignonac K, Vandenberghe C, Gatignon Turnau A. (2013). “Study of the Relationships Between Volunteers’ Commitments to Organizations and Beneficiaries and Turnover Intentions”
24 Västfjäll D, Solvic P, Mayorga M. (2015). “Pseudo inefficacy: negative feelings from children who cannot be helped reduce warm glow for children who can be helped.”. Frontiers in psychology, 6: 1-12. Wilkinson C. (2008). “Unwanted sex versus rape: How the language used to describe sexual assault impacts perceptions of perpetrator guilt, victim blame and reporting.”. Dissertation Abstracts
international, Section B: The Sciences and Engineering, 69: 4450
25 APPENDIX 1: PRE-TEST
To see if people perceived the words help and support differently in terms of activeness, I conducted a pre-test. This pre-test was spread through my network and at the university of Groningen (age ranged from 19 to 85, Mean age: 36.2; 59.4% female and 40.6% male). A total of 64 participants answered the next questions (which were asked in Dutch, but translated to English for this rapport):
Q1: For my research I would like to know how you perceive certain verbs based on the active behaviour or passive behaviour. Give your answer by clicking on one of the choices below.
1: Absolutely passive 2 3 4 5 6 7 8 9: Absolutely active Help Support
Q2: If you must choose, do you perceive these two words as passive or active?
Passive Active
Help Support
Q3: What is your gender? o Male
o Female
Q4: What is your age? (Open question)
I used a repeated measures ANOVA to see if there was a perceived difference within subjects as well.
The results are as follows (N=64):
26 APPENDIX 2: SURVEY
The survey is made on Qualtrics and the questions are asked in Dutch. For this report I translated the questions and possible answers in English.
Introduction:
Q1: I asked the participants if they are above 18 and willing to take part in this survey, if they answered no, they were skipped to the end of the survey.
Q2: They were asked if they have active Facebook account (been online at least once this month), if they answered no, they were skipped to the end of the survey.
Demographics:
Q3: What is your gender? o Male
o Female
Q4: What is your age?
(Open question) (If they put less than 18: skipped to the end.
Q5: What is your highest finished education level or your current education level? o Less than secondary school
o Secondary school or comparable o MBO
o HBO (bachelor or master) o WO (bachelor, master or PhD)
Q6: Which of the following categories describes your work status best? o Student
o Full-time worker (more than 32 hours per week) o Part-time worker (less than 32 hours per week) o Freelancer/ZZP
o Out of work, but looking for work o Unpaid work (like homemaker)
o Unable to work because of health issues o Retired
27
Independent variable:
Q7: Down here you see the advertisement of WNF like you can see it on Facebook. Look at this advertisement attentively.
Q8: people either saw this picture or the one where the words ‘help’ were replaced by ‘steun’ (support). In this question I could see how long people spend on this page.
Mediators: in these questions people either saw the words help or support depending on the ad
they saw. In this appendix I will state the questions with help.
The next questions are about what you think and feel when you would help the elephants and rhinos by signing the petition.
Q9: If I help the elephants and rhinos, then that would give me:
Slider from (0: not a good feeling at all – 100: a very strong good feeling)
Q10: How likely is it that my help will make a difference? 9 items Likert scale, 1= very unlikely – 9= very likely
Q11: Indicate how much you agree with the next theses: 1= absolutely disagree, 7= absolutely agree
If I help the elephants and rhinos:
1 2 3 4 5 6 7
Then WNF means a lot to me
Then if WNF has a problem, I personally accept it Then I feel like I am a part of the WNF family
28
Q12: Indicate how much you agree with the next theses: 1= absolutely disagree, 7= absolutely agree
If I help the elephants and rhinos:
1 2 3 4 5 6 7
Then I think that WNF helps effectively in helping the animals worldwide
Then I think WNF uses their donations effectively for their case Then I think that my contribution will make a difference in helping the animals
Dependent variable:
Q13: here you see the advertisement again. We ask you to react the same way as you would do when you would see this add on Facebook. Like on Facebook you can do multiple things or do nothing at all.
Q14: advertisement picture again. Same one as before. Q15: I would:
Yes, I would to that No, I would not do that Like the ad (by clicking on the thump)
Place a comment Share the advertisement
Go to the petition by clicking on the ad
Q16: when they clicked I would do that at place a comment then those participants get an extra open question where they can say which comment it would be.
Anything else:
Q17: have you donated to WNF in the past year? o No, I have not donated to WNF in the past year o I have incidentally donated to WNF in the past year
o I am a member of WNF and donate monthly via this subscription
Q18: If you have any questions or feedback after this survey you can fill it in here (you are also allowed to leave this one empty):
