THE ILLUSION OF CHOICE OVERLOAD:
THE MODERATING ROLE OF COMPLEXITY
AND CONSUMER INVOLVEMENT IN A
CAUSE-RELATED MARKETING CONTEXT
CYNTHIA BEER
University of Groningen Faculty of Economic and Business
TABLE OF CONTENT
1. INTRODUCTION ... 3
2. THEORETICAL BACKGROUND ... 5
2.1 THE ROLE OF CHOICE ... 6
3.2 CHOICE SET COMPLEXITY ... 10
3.2.1 The Influence of Similarity on Choice Set Attractiveness and Decision Making ... 10
3.2.2 The Influence of Similarity on Post-purchase Campaign Attractiveness ... 11
3.3 INFLUENCE OF VALUE TRAITS ON PREFERENCE UNCERTAINTY ... 12
3.3.1 The Influence of Value Traits on Choice Set Attractiveness ... 13
3.3.2 The Influence of Value Traits in CM Campaigns ... 14
3. METHODOLOGY ... 17
4. RESULTS ... 21
4.1 MANIPULATION CHECKS AND CONTROL VARIABLES ... 21
4.2 SET SIZE AND DISSIMILARITY ... 22
4.2.1THE ROLE OF SET SIZE AND DISSIMILARITY ON CHOICE ATTRACTIVENESS ... 22
4.2.2THE ROLE OF SET SIZE AND DISSIMILARITY ON CAMPAIGN ATTRACTIVENESS ... 23
4.3 THE IMPACT OF PERSONAL VALUE TRAITS... 24
4.3.1 The Role of Values on Choice and Campaign Attractiveness ... 24
4.3.2 The Moderating Role of Values for Choice Attractiveness as Dependent Variable .. 25
4.4 ADDITIONAL MEASURES ... 30
4.4.1 The Role of Set Size and Dissimilarity on other Dependent Measures ... 30
6. GENERAL DISCUSSION ... 31
6.1 FINDINGS ... 31
6.1.1 Theoretical Implications... 32
6.1.2 Practical Implications ... 36
6.2 LIMITATIONS AND FURTHER RESEARCH ... 37
7. REFERENCES ... 39
ABSTRACT
Cause-related marketing (CM) is increasingly used to fulfil the social responsibility of companies, to increase consumer loyalty, or to build reputation. This thesis looks at the role of having choice within a CM campaign to investigate whether providing the opportunity of a larger set of charities would increase the choice and campaign attractiveness, or if the freedom of choice is hindered by choice overload. To get a better understanding of the effect of choice set design I ran an experimental study with either a dissimilar or similar choice set (in terms of type of charities) and one of three sizes (2, 5, 10 options). Personal value traits were also taken into account to investigate the relationship of situational and personal factors on the choice effect. The thesis shows that larger choice sets of 5 or 10 charities (vs 2 charities) are perceived as more attractive, but no effect of choice overload could be observed. Dissimilar (vs similar) choice sets show a significant superior effect for campaign attractiveness, ease of choice, and perceived social responsibility. Consumers holding value traits congruent with those of charities perceive the choice set and the campaign as more attractive. The results suggest that CM campaigns should offer a rather large set of charities to choose for donation and indicate that charities should be selected properly with the aim to increase the variety within the set. The findings on consumer values put emphasis on the relevance of understanding personal value traits of the target. Note that further research is required to find the optimal balance between number of charities and costs associated with building long-lasting relationships to these organizations. Additionally, it is required to investigate further forms of dissimilarity between the charities and how for example an asymmetrical composition of charities characterized by different types of attributes influences the effect of increased choice.
In these days, companies are more and more expected to care about their social responsibility within society. In order to fulfil this demand, new campaigns are developed in which the company supports a partner charity when a specific product is purchased. For example, Procter & Gamble cooperates with UNICEF to enable vaccinations for new-borns in developing countries for every package of Pampers that is bought (“Pampers and UNICEF: A powerful partnership,” 2019). The cooperation with UNICEF enables a win-win-win scenario for the company, the charity, and the consumer (Adkins, 2003). Not surprisingly, companies increasingly realize the potential of those so called cause-related marketing (CM) campaigns, leading to continuous growth in sponsorship spending on causes (“The Most Active Sponsors Of Causes - IEG Sponsorship Report,” 2018). In 2018, only in the US 2.14 billion dollar were funded by brands like Coca-Cola, Google, Microsoft, or Shell (“The Most Active Sponsors Of Causes - IEG Sponsorship Report,” 2018). CM is a marketing technique, which can be defined as “ […] a commercial activity by which businesses and charities or good causes form a partnership with each other to market an image, product or service for mutual benefit” (Adkins, 2003, p. 670).
The increase in CM campaigns imposes new difficulties for the consumer. More campaigns also mean more options to help, which requires the consumer to decide whom to support. The role of choice builds the basis for the pareto optimum theory, which describes a hypothetical state in which social welfare can be optimized when every individual maximizes the own welfare (Burroughs & Rindfleisch, 2012). Achieving this state requires two conditions: 1) Everyone has the freedom of choice and 2) Individuals are fully capable of making informed choices. Following the idealistic concept, people would benefit from unlimited choice by leveraging their freedom of choice and optimizing societal welfare (Burroughs & Rindfleisch, 2012). The idea that choice always benefits welfare was refuted in the jam experiment on choice by Iyengar and Lepper (2000), where consumers exposed to a larger assortment of flavours were less motivated to engage in a purchase action compared to the condition in which just a limited number of options was available. Adults make on average about 35.000 conscious decisions during a day (Sahakian & LaBuzetta, 2013). The number of options can become overwhelming leading to a state of choice overload, where consumers avoid decisions in order to prevent themselves from making the wrong choice. The findings from the experiment show, that increased choice is not always desirable and can harm the consumers’ campaign and choice attractiveness.
leave the freedom to choose to the consumer. Hardly any research is known about how choice within the CM campaign affects decision making and if choice overload is similar to the one observed by Iyengar & Lepper (2000; see also Howie, Yang, Vitell, Bush, & Vorhies, 2018; Robinson, Irmak, & Jayachandran, 2012). This thesis aims to extend the knowledge of choice within CM campaigns by studying how the option to choose different charities influences the consumers’ campaign and choice set attractiveness and how the external structure of the assortment and internal, individual traits will influence their judgement.
The findings give indications about how a choice set of charities can increase consumer satisfaction and the way it needs to be designed to have an effect. Important influential variables are assessed, defining external conditions and internal personality traits that determine whether choice overload is likely to appear. Additionally, the thesis extends the literature on choice overload by providing a better understanding of key influencing factors and therefore forms the groundwork for further research.
The remaining thesis is structured as followed: The next section provides a review of previous literature and introduces the hypotheses investigated in this work. After defining the questionnaire used for the data collection, the effect of increased choice and the moderating role of internal personality traits and external choice set manipulation will be presented. Finally, a general discussion embedded in the too-much choice theory context will give implications for the management on how to improve cause-related marketing campaigns and presents limitations of this work to encourage further research on this topic.
2. THEORETICAL BACKGROUND
like brand-cause fit, product type, or donation type (Barone, Norman, & Miyazaki, 2007a; Bigné-Alcañiz, Currás-Pérez, Ruiz-Mafé, & Sanz-Blas, 2012; Chang, 2008, 2011; Christofi, Vrontis, & Leonidou, 2014; Grau & Folse, 2007; Gupta & Pirsch, 2006; Koschate-Fischer, Stefan, & Hoyer, 2012).
However, little is known about the role of choice within the campaign itself and whether choice overload might be a concern. The likelihood of the appearance of choice overload is, among other things, influenced by the amount of information which is processed to make a decision. Important characteristics that consumers use to evaluate charities can among others include the cause itself, the perceived efficiency of help provided by the charity, the geographical field in which the charity operates, or the amount of donations that is used to cover overhead costs (Baron & Szymanska, 2011; Barone, Norman, & Miyazaki, 2007b; Bigné, Currás‐Pérez, & Aldás‐Manzano, 2012; Gneezy, Keenan, & Gneezy, 2014a). Other factors relevant for choosing an option within a set of opportunities can be inter alia the individual processing motivation of evaluation criteria (Lin & Wu, 2006), the assortment size and attractiveness (Oppewal & Koelemeijer, 2005), display cues and product related information (Mogilner, Rudnick, & Iyengar, 2008), and time pressure (Haji, Krawczyk, Sylwestrzak, & Zawojska, 2019). This means that internal, consumer-related traits, as well as external, situational factors might moderate the consumers’ decision making and therefore the effect of choice (Chernev, Böckenholt, & Goodman, 2015).
2.1 The Role of Choice
two environmental charities). The effect of choice might be influenced by the composition of the choice set, more precisely the degree of similarity between the options. The role of choice set complexity as possible moderator will be discussed later. This thesis thus tries to extend the findings on the positive role of choice and aims to investigate whether a larger choice set is always more desirable or if there is a turning point where choice becomes demotivating.
Different reasons explain why larger assortments positively influence consumer behaviour. On the one hand, following the Self-perception theory of Bem (1972), a higher level of perceived freedom of choice strengthens the own responsibility perception of the consumer. This can be realized by adding distinct items to a choice set (Reibstein, Youngblood, & Fromkin, 1975). As a consequence, consumers are less likely to blame marketers, or the product itself, for their own bad decisions. On the other hand, larger assortments not just increase the perceived freedom of choice but are in general rated more positively. This effect even occurs, if a smaller assortment contains the preferred alternative which emphasizes the importance of good assortment management (Oppewal & Koelemeijer, 2005). In sum, larger assortments are evaluated more favourably for two main reasons. First, people are more attracted by variety as it enhances the probability to find what the consumer is searching for (Hoch, Bradlow, & Wansink, 1999). The chance for a better alternative explains why a larger assortment is preferred even when the smaller one holds the usually favoured item. Second, more alternatives can fulfil an educational function about market offers for variety-seeking consumers without, or with dynamic preferences which increases their flexibility and freedom of choice (Hoch et al., 1999). The role of consumer-orientation emphasizes individual differences in the perceived attractiveness of choice and provides the first indication for variances in the campaign and choice set attractiveness among consumer types. In section 3.3, the role of personality traits will be addressed more specifically.
require more comparison, more time to make the decision, evoke higher levels of regret and finally lead to less choice satisfaction. To find a maximizing solution, larger assortments are more appealing but those who choose from a smaller variety are eventually more satisfied (Schwartz, 2000). The paradox of being more attracted by larger assortments and yet unsatisfied with the choice, results from an imbalance between depleted resources for decision making and the final satisfaction, perceived as asymmetric cost-benefit trade-off (Dar-Nimrod, Rawn, Lehman, & Schwartz, 2009). Based on this one can reason that although increased freedom of choice may be attractive, it not necessarily needs to translate the positive impression in after-purchase satisfaction. In a CM context one can argue that although consumers are attracted by the opportunity to choose the charity, they will probably not translate this in their perceptions of the campaign. It can be assumed that as soon as the trade-off between required resources and satisfaction starts to be perceived as asymmetrical, a negative effect of choice occurs.
Another explanation why too much choice can have a negative effect states that high variety increases the risk of self-blame for the wrong choice (Arkin, Gleason, & Johnston, 1976). In this context, wrong choice is understood to be unsatisfied with the decision because of the feeling that a rejected option would probably have been more satisfactory. One can argue that this is also relevant for CM campaigns: consumers might experience feelings of guilt for choosing one charity and thus rejecting the help for others, triggering the perceived inefficiency of help. Indeed, pseudoinefficiency refers to a change in feelings due to the realization that supporting one charity means not supporting other causes and therefore mitigates the warm glow usually arising from charitable behaviour (Västfjäll, Slovic, & Mayorga, 2015). Negative feelings from this non-rational idea boost avoidance behaviour and therefore reduce the probability that consumers engage in the campaign (Västfjäll & Slovic, 2013). In sum, one can argue that increased feelings of inefficiency from choosing only one charity and an asymmetrical resource-satisfaction relation can decrease the campaign and choice satisfaction, suggesting that having more choice is a bad thing.
and is defined by the number of alternatives and the number of attributes on which options can be distinguished. The results imply that choice overload kicks in when six to nine choices are available. In line with her findings, one can assume that 10 charities are appropriate to study the effect of choice overload in the context of CM campaigns. Accordingly, the question arises how having 2 vs 5 vs 10 options will influence the perceived attractiveness of the choice set and of the overall campaign.
In sum, it can be hypothesized that a choice set of five causes distinguished by the mission and local focus of the charityis preferred over a choice set of two causes distinguished by the same attributes. Carrying on, a choice set of 10 causes will be less favourable than the five-causes-set due to negative effects of choice overload. This leads to a curvilinear relationship of the effect of increased choice in the cause-related marketing context and builds the hypothesis as followed:
H1: Increasing the number of options from 2 to 5 causes in CM, positively affects the
campaign and choice set attractiveness, while a further increase to 10 options has a reverse, negative effect.
3.2 Choice Set Complexity
Contextual variables can be described as choice set complexity and apply to the value of the set, essentially the overall attractiveness and similarity between the options (Payne, Bettman, & Johnson, 1993). Choice set complexity refers to characteristics of the provided assortment and is accordingly related to objective, situational factors (Chernev et al., 2015). (Dis)similarity within the set is an important characteristic influencing the perceived attractiveness of the choice set and thus decision making (Chernev, 2005). Additionally, this situational factor impacts after-purchase feelings of confidence and satisfaction in the consumers’ choice which are reflected in the perceived attractiveness of the whole campaign (van Dijk & van Knippenberg, 1998).
3.2.1 The Influence of Similarity on Choice Set Attractiveness and Decision Making
The self-determination theory of Schwartz (2000) states, that a self-determining self chooses the option which maximizes own preferences. Self-determined behaviour stems from intrinsic motivation and processes residing inside the consumer (Deci & Ryan, 1990). In a consumption related context, maximizing the own welfare results from finding the optimal cost-benefit solution. Being exposed to an overall attractive set of alternatives increases the preference for smaller assortments with less searching costs, while the attractiveness promises sufficient benefit (Chernev & Hamilton, 2009). But what if the choice set attractiveness is not directly obvious due to high similarity between the options requiring deeper processing to form evaluations?
shortcoming are eliminated as well. Therefore, the perception for multi-attribute options is biased, because the focus is isolated towards the shortcoming and ignores that the alternatives are composed with more than one feature (Kahneman & Tversky, 2013). For example, a consumer searching for a new laptop might be faced with a broad variety of options. A laptop can usually be described as a multi-attribute product characterized by lots of different features. To deal with the broad assortment an elimination-by-aspect process can be an efficient decision making method. This would mean that all laptops having for example less than 12" will be automatically eliminated from the consideration set, if the consumer is searching for a laptop with a big screen. This might be a very efficient process to make a choice but ignores alternatives characterized by features that are able to compensate the shortcoming of the display size. In sum, one can assume that high similarity among options can blind people for the best option.
3.2.2 The Influence of Similarity on Post-purchase Campaign Attractiveness
The role of choice set complexity impacts not just the decision process, but also determines post-purchase feelings like satisfaction or perceived endowment. In the cause-related marketing field one can assume that the selected charity is perceived as more favourable than the non-chosen target because the commitment for a cause is stronger than feelings for non-selected options. This effect can be explained by the endowment effect hypothesized by Thaler (1980). The effect states that people attach different weights to the impact of direct losses and gains by underweighting losses resulting from lost opportunities. The same amount of loss is perceived as stronger impact, compared to the same amount of gain (Kahneman & Tversky, 2013). Translated to the consumption context, products owned by the consumer are valued higher than those outside of their possession because giving up things from the own custody weights more than missing an opportunity. In sum, the endowment effect strengthens choice confidence and reduces the attractiveness of non-chosen subjects.
inhibits positive feelings from the endowment. In sum, it is required to find a way how to reduce perceived feelings of inefficiency associated with the donation and theory on the endowment effect provides insights on how to design the choice set to achieve this decline.
From the endowment effect is known that the preference for a selected item results from a value incomparability between dissimilar objects, meaning that uncertainty about the perceived value attached to the owned product lead to higher feelings of loss aversion (Dijk & Knippenberg, 1996). Conclusively, to reduce insecurities one need to increase the substitutability between objects emphasizing that loss aversion is stronger for more distinguishable objects (van Dijk & van Knippenberg, 1998). Transferring this to the problem of pseudoinefficiency of help, one can argue that the effect results from high uncertainty about perceived differences in the value attached to the donation for different causes. For example, a CM campaign could provide the opportunity to either support an environmental cause that focusses on a world-wide problem, or to support a humanitarian charity that builds schools in a specific geographical area. Being faced with making a decision between distinctly different causes might thus increase feelings of uncertainty about the effectiveness of the donation. According to the endowment effect, higher similarity between the options can reduce insecurities about the effectiveness of the donation. Providing a set in which all charities focus on the same issue makes it easier for consumers to estimate the value of their donation, because whichever charity they choose, the donation will contribute to solve the same problem. In sum, the findings from the endowment effect indicate that a higher substitutability within the choice set could reduce feelings of perceived inefficiency of help for the out of range target. In summary it can be hypothesized:
H2: The curvilinear effect of the increasing number of causes on the attractiveness of the
choice set and overall CM campaign is stronger for a similar rather than a dissimilar cause-set, in all three set sizes.
3.3 Influence of Value Traits on Preference Uncertainty
3.3.1 The Influence of Value Traits on Choice Set Attractiveness
Having the opportunity to choose between several options on various dimensions, does not necessarily imply that consumers have no idea what they enquire. Previous experiences and expectations retrieved from memory can form concreate ideas of the desired product, so called ideal points. Holding an ideal point not just influences the preference towards the assortment size, but also the way how information is processed as presented in the table below (Chernev, 2003a).
TABLE 1
Interrelation-Matrix: Ideal Point and Assortment Size
Small assortment Large assortment
Ideal point
Regret
Weak preference for selected item High regret for lost
opportunities Less attractive
Confidence
High confidence and preference
for the selected item More attractive No ideal point Convenience Easier to process more attractive more satisfied Struggling
Complex decision making required
Less attractive
Table 1: Interrelation Matrix: Ideal Point and Assortment size (adapted from Chernev, 2003b, 2003a; Slovic et al., 2002; Tversky & Kahneman, 1991)
Articulated ideal points can be described as ”[…] the degree to which [individuals] have readily established attribute preferences.” (Chernev, 2003a, p. 152). Clear requirements for the desired product enable elaborative processing in an experiential mode. Reactions towards incoming information are more intuitive and automatic, based on the valence placed towards the attribute constellation through previous experiences (Slovic et al., 2002). Therefore, larger assortments are more likely to encompass a selection that matches the ideal point, fostering
confidence in choice (Chernev, 2003b). Generally speaking, the ideal attribute-bundle is used
detail, instead of using a more efficient and less effortful elimination-by-aspect process. The consumer is not just required to eliminate the items holding less attractive attributes, but also needs to evaluate the importance of each function and create specific elimination criteria (Chernev, 2003a). Processing is less comprehensive and the number of subsequent value-weightings increase, as the assortment size increases (Chernev, 2003a).
In sum, consumers without ideal points and choosing from a large assortment struggle to find the optimal solution. In contrast, consumers without an ideal point choosing from a small assortment process information with more convenience and thus are more satisfied with their choice (Chernev, 2003a). The regret consumer is someone who has an ideal point and is searching for it in a smaller assortment, leading to low satisfaction with the selected product and feelings of lost opportunities. Finally, consumers with ideal points for their products and searching for them in large assortments are characterized by having the highest confidence in their choice.
All in all, it can be argued that larger assortments are more beneficial for those consumers with clear requirements whereas small assortments benefit those, who are more depending on information processing to decide. The question which becomes relevant for now is thus what determines whether people have an ideal point. I will focus on the role of consumers’ value traits within CM campaigns.
3.3.2 The Influence of Value Traits in CM Campaigns
In the CM field, ideal points are associated with the relative importance consumers attach to the charities’ goals and can thus more clearly be defined as personal value traits. Personal value traits describe the priority attached to individual goals unrelated to cultural or social norms prescribing behaviour (Kostelijk, 2016). According to Schwartz, (1992, 2006) six features define values inter alia that values provide guidance towards a desired goal, are used as standards/criteria facilitating evaluation and selection, and guide actions determined by the extend of relative importance. In sum, the value requirements highlight their ubiquity in determining actions, decisions and behaviours and depict the understanding of value traits for this paper.
system distinguishes between 5 different value categories, that are presented in the table below (Table 2). Each class can be characterized by its marker items, presenting the central focus of the value. Certain adjectives are used to more precisely define the characteristics of the value type (Kostelijk, 2016).
TABLE 2
Value Compass: Defined by (marker) items
Affiliation-Oriented values Increasing the future quality of life Hedonic Values Values aimed at making a difference with others Utilitarian Values
Care & affection, Intimacy, Honesty, Safety
Social
responsibility
Enjoying life Stimulation, Prestige, Beauty Functionality, Achievement Caring for someone Family life Friendship Harmony Cosiness Romance Sensuality Honesty Loyalty Safety Protection Feeling of security Being environment-friendly Providing for a better world Recycling Environmental protection Improving society Enjoying life ambition Excitement Being unique Fun Wisdom Pleasure Independence Vitality Adventure Being active Being sportive Courage Leadership Power Status Being successful Beauty Style Elegance Good-looking Efficiency Functionality Precision Reliability Creativity Innovation Usefulness Intellect Progress Smart solutions
Note: Italic items were used for the data collection
Table 2: Value Compass: Defined by (marker) items (adapted from Kostelijk, 2016)
Personal value traits of consumers are composed by the relevance attached to each value class. Therefore, it is important for a CM campaign to understand the target’s most important value traits to design the best suitable choice set. Transferring the value compass into a charitable context, one can argue that affiliation-oriented (AO) and increasing the future
quality of life (IFQL) are the categories best representing values included in mission and vision
the Children International,” 2019; Kostelijk, 2016). More environmentally oriented charities, like the World Wildlife Foundation (WWF), care about sustaining life quality for future generations and the world by following their social responsibility and encouraging environment-friendly behaviour (Kostelijk, 2016; “Values-Greenpeace International,” 2019.; “WWF’s Mission & Vision – World Wildlife Fund,” 2019).
In sum, it is likely that consumers scoring high in these “charity-defining” classes, have a better idea of what type of organization they prefer for donation, and thus have clear ideal
points. For example, consumers with strong relevance in value categories of increasing the
future quality of life, might be more attracted by environmental charities representing similar values like for example WWF. Accordingly, the donation decision is based in relation to the consumers’ ideal points, which is more likely to be concrete for someone with similar values like the ones represented by charities from the choice set. One can argue that this would imply that people with charitable values are more attracted by a higher variety of charities to find a match consistent with their own beliefs and are eventually more confident with their choice (Chernev, 2003a). For the increased choice effect this would imply that more options are even more beneficial when consumers have strong affiliation-oriented or increasing future quality of life values. Therefore, an increase from 2 options to 5 options would increase the perceived attractiveness even more for people who have charity-related values by strengthening the confidence in their choice and reducing feelings of regret for the selected option. Additionally, for the same consumer type, an increase from 5 to 10 options would dampen the negative choice effect. This leads to the following hypothesis:
H 3a: Having affiliation-oriented value traits and/or increasing the future quality of life value
traits, positively moderates the effect of choice options on the campaign and choice set attractiveness for an increase from 2 to 5 options and dampens the negative choice effect in the 10 options condition.
find a charity that matches their values. Thus, increased similarity is less desirable for those, who already have preferences for the donation decision. For consumers who lack in charitable values, an increase in choice set variety weakens the preference for the selected item, which strengthens the perceived inefficiency of help. In sum, it can be argued that the valence of the effect of dissimilarity depends on the extent to which consumers have already established preferences for their decisions. Putting in all in a nutshell, it can be hypothesized that:
H 3b: Having affiliation-oriented value traits and/or increasing the future quality of life value
traits increases the preference for dissimilar choice sets and thus reduces the moderating negative effect of set-dissimilarity on the effect of increased choice.
Summarizing the theory shows that choice is a complex construct with the potential to increase societal welfare. Achieving this requires the consideration of personal, internal as well as conditional, external effects. An overview of the conceptual model used in this thesis is provided below.
The following section presents the data collection used to investigate the role of similarity and value traits within different multi-option CM campaigns.
3. METHODOLOGY
from Amazon MTurk who completed the study in exchange for a small payment. The sample consisted of 233 (53.8 %) men and 200 (46.2 %) women, with a mean age of 40 years (Mage = 39.91, SD = 11.90).
First, participants answered questions to measure individual value traits. To assess the traits, I used Likert Scales (1 = Strongly disagree; 7 = Strongly agree) for asking in a random order about the personal relevance of marker items which are defining values, conceptualized by Kostelijk’s value compass (2016). Affiliation-oriented values were measured with five items (e.g. Honesty and loyalty are important characteristics; α = 0.73). Increasing the future
quality of life values were measured with two items (e.g. I think recycling is important; α =
0.81). Hedonic values were measured with two items (e.g. I am fun and excitement oriented; α = 0.68). Values aimed at making a difference with others were measured with four items (e.g. I care about my look; α = 0.77) and utilitarian values were measured with two items (e.g. I am seeking for smart solutions and innovations; α = 0.62). After measuring personal value traits, participants were asked some general demographic questions and an attention check was implemented to increase the reliability and statistical power of the data collected (Oppenheimer, Meyvis, & Davidenko, 2009). 460 participants finished the survey of whom 8 failed the first attention check and 19 failed to remember the number of charities within their set, which was asked at the end of the survey. In total, 27 participants were removed for the analysis.
Participants were then asked to “imagine that they are planning to buy cereal and that the cereal from company XY is thus an option to consider”.
TABLE 3
Description of charity organizations used in different conditions
Organisation Mission (for the full version, see
Appendix A)
Local focus
Condition Aids Healthcare
Foundation HIV prevention services and healthcare. Worldwide
10 - D 5 - D 2 - D
Rainforest Foundation Secure land rights for indigenous people by
providing technology, tools and training.
South America
10 - D 5 - D 2 - D
Art for the World
Building a bridge between art and society. Organization of events accessible for everyone.
Worldwide 10 - D
5 - D
Disability NGO’s Alliance
Support of disabled people. Increasing the
sensitivity in the country. India
10 - D 5 - D
The Water Project Providing access to clean and safe water. Africa 10 - D
5 - D
Church Urban Fund Bring people together and provide finaincial
support. England 10 - D
Franco-British Society
Encouraging closer relationships between France and GB. Responsible for events & exhibitions.
GB and
France 10 - D
Global Green Growth Institute
Ensuring balance of economic growth and
sustainability by providing green growth plans. Worldwide 10 - D
Poject Street Dogs Saving stray dogs & cats. Sterilization and
help with food and rehoming. Thailand 10 - D
Smile Foundation Finance and provide education and healthcare
for children. India 10 - D
Natural Resources Defense Council
Focus on water and ocean programs to protect
oceans and wetlands. Worldwide
10 - S 5 - S 2 - S
Ocean Conservancy Focus on protection of marine habitats.
Sustainable fishing and provide education. Worldwide
10 - S 5 - S 2 - S
Oceana Focus on preventing the collapse of fish
popluations and other damages of pollution. Worldwide
10 - S 5 - S
Surfrider Focus on protection of the oceans. Worldwide 10 - S
5 - S
The Ocean Cleanup Reduced pollution through new technologies to
clean the ocean. Worldwide
10 - S 5 - S
Marine Connection Focus on protection and conservation of
Oceanic Preservation Society
Protection of the ocean and the environment by
increasing awareness through documentaries. Worldwide 10 - S
Sea Shepherd
Conservation Society Fight against whale hunt and illegal fishing. Worldwide 10 - S Woods Hole
Oceanographic Institution
Marine science and engineering to solve
problems and provide education. Worldwide 10 - S
Environmental Defense Fund
Focus on global warming and ocean
protection. Supporting sustainable fishing. Worldwide 10 - S
10 - D = set of 10 dissimilar charities 10 - S = set of 10 similar charities
5 - D = set of 5 dissimilar charities 5 – S = set of 5 similar charities
2 - D = set of 2 dissimilar charities 2 – S = set of 2 similar charities
Subsequently, I measured the dependent variables. First, I measured the perceptions of the choice set with the ease of choice, set attractiveness, and perceived feelings from the choice. Ease of choice was measured with two items (e.g. It would be easy for me to choose the charity for the donation; α = 0.87). The set attractiveness was measured with two items (e.g. The choice of charities is satisfactory; α = 0.79). Perceived feelings from the choice was measured with two items (e.g. I believe purchasing cereal from company XY would make me feel positive; α = 0.77).
Next, the campaign perception was measured with attractiveness of the campaign, effectiveness of the campaign, and perceived social responsibility. Campaign attractiveness was measured with two items (e.g. Having the option to choose the charity makes products from company XY more attractive; α = 0.90). The effectiveness of the campaign was measured with two items (e.g. I believe this campaign will contribute to reach the goals of the charities; α = 0.83). For measuring the perceived social responsibility, two items were used (e.g. I believe the campaign of company XY aims to improve social welfare; α = 0.84).
After finishing these questions related to the protentional underlying processes, participants were asked whether they would buy a product from the company (yes vs. no), and if so, which charity they wanted to receive the donation.
much charities were provided to the participant, with the option to pick 2, 5, or 10. For a full overview of the experiment and dependent measures, see appendix A.
4. RESULTS
4.1 Manipulation Checks and Control VariablesTo check the manipulation of the similarity conditions, I performed a 2x3 ANOVA including the independent variables set size and dissimilarity (vs similarity) and the dependent measure perceived similarity. The test revealed a significant main effect of choice set size F (2, 427) = 11.19, p = .000, as well as for similarity, F (1, 427) = 853.33, p = .000. Consumers in the dissimilar choice set conditions perceived the choice set to be significantly less similar (M = 2.97, SD = 1.45) than those in the similar choice set conditions (M = 6.21, SD = 0.89).
The interaction effect was also significant, F (2, 427) = 4.93, p = .008, indicating statistical relevance of the interaction of set size and dissimilarity (vs similarity).
However, specific analysis on the choice set size did not support a statistically significant difference between the groups F (2, 432) = 0.52, p = .592. The differences between the 2 choice conditions (M = 4.47, SD = 2.19) and 5 choice conditions (M = 4.62, SD = 1.88), (t (430) = 0.64, p = .522), between the 2 choice conditions (M = 4.47, SD = 2.19) and 10 choice conditions (M = 4.71, SD = 1.96), (t (430) = 1.01, p = .313), and between the 5 choice conditions (M = 4.62, SD = 1.88) and 10 choice conditions (M = 4.71, SD = 1.96), (t (430) = 0.36, p = .718), were all non-significant.
Although the results indicated an interaction effect between the variables of set size and dissimilarity (vs similarity), this should not be an issue for this research due to the high difference between the perceived similarity of dissimilar (vs similar) sets which is crucial for this design.
men (M = 4.97, SD = 1.57) was significantly lower than the average campaign attractiveness of women (M = 5.40, SD = 1.47). The results reveal that female participants perceived the choice and campaign attractiveness on average as more attractive. To check whether age influences my results I checked my analysis with additional tests to control for the effect of gender. None of the tests did change my results significantly. Thus, I decided to not include gender as a variable in my model.
In order to analyse whether or not the age of consumers influenced their campaign attractiveness, I performed a regression analysis with age regressed to choice attractiveness. The regression analysis was non-significant, R² = 0.00, F (1, 431) = 1.33, p = .249. The age of consumers does not influence the choice attractiveness, B = -0.01, t (431) = -1.15, p = .249. The regression analysis was repeated with age regressed to campaign attractiveness. The regression analysis was non-significant, R² = 0.01, F (1, 431) = 2.04, p = .154. The age of consumers did not significantly influence the campaign attractiveness, B = 0.01, t (431) = -1.43, p = .154.
4.2 Set Size and Dissimilarity
This section of the analysis focusses first solely on the effect of choice set size and the degree of (dis)similarity between the charities on the perceived choice and campaign attractiveness, and thus does not take personal values into account yet. This makes interpretation of the effects easier. In chapter 4.3 I will test the complete model, including personal values, using PROCESS (Hayes, 2018; Hayes & Preacher, 2013). Unless stated otherwise all tests used a 2 (dissimilarity: Similar vs. dissimilar) x 3 (options: 2 vs 5 vs 10) between-subjects ANOVA.
4.2.1 The Role of Set Size and Dissimilarity on Choice Attractiveness
and 10 option set, t (430) = 0.03, p = .978. These results suggest, that participants were more attracted by a larger choice set (5 or 10) than by a smaller choice set of 2 options.
The effect of dissimilarity showed no significant difference between the choice set attractiveness of dissimilar (M = 5.16, SD = 0.10) and similar choice sets (M = 5.24, SD = 0.10), F (1, 427) = 0.36, p = .548. The interaction effect was also not significant, F (2, 427) = 0.21, p = .811.
To check whether the larger sets of 5 and 10 options can be summarized in one variable for the analysis in PROCESS, I tested Model 1 of PROCESS macro (5000 bootstraps, 95% CI; Hayes, 2013) including set size with either 2 options or more options (5 or 10 options) and dissimilar (vs similar) charities as moderator. The model revealed similar to the 2x3 ANOVA one significant main effect of set size, B = 0.41, t (429) = 2.76, p = .006. The more options conditions (5 or 10 options) (M = 5.33, SD = 1.46) were evaluated significantly higher in choice attractiveness than the 2 options conditions (M = 4.93, SD = 1.42). The main effect of dissimilarity (B = -0.09, t (429) = -0.63, p = .540) and the interaction effect (B = -0.02, t (429) = -0.05, p = .958), were both non-significant.1
4.2.2 The Role of Set Size and Dissimilarity on Campaign Attractiveness
The ANOVA yielded a significant main effect of choice set size F (2, 427) = 4.09, p = .017, as well as for similarity, F (1, 427) = 3.96, p = .047. Consumers in the dissimilar choice set conditions perceived the campaign to be more attractive (M = 5.32, SD = 0.10) than those in similar choice set conditions (M = 5.03, SD = 0.10).
Specific analysis on the choice set size supported a statistically significant difference between the groups F (2, 432) = 5.03, p = .007. Significant difference was found between the 2 choice conditions (M = 4.85, SD = 1.58) and 5 choice conditions (M = 5.40, SD = 1.48), t (430) = 3.02, p = .003. The groups of 2 choices (M = 4.85, SD = 1.58) and 10 choices (M = 5.27, SD = 1.51) also showed significant difference, t (430) = 2.33, p = .020. Comparing the 2-options set with the 5 and 10-2-options set revealed a significant difference with t (430) = 3.10, p = .002, whereas the contrast analysis showed no significant difference between the 5 options and 10 options set t (430) = -0.71, p = .478. These results suggest, that participants were more attracted by campaigns providing a larger choice set (5 or 10) than by a smaller choice set of 2 options. The interaction effect was not significant, F (2, 427) = 0.93, p = .394.
1 Testing a 2x2 ANOVA revealed, in line with PROCESS, the same significant main effect of set size, F (1, 429)
To check whether the larger sets of 5 and 10 options can be summarized in one variable for the analysis in PROCESS, I again tested Model 1 of PROCESS macro (5000 bootstraps, 95% CI; Hayes, 2013) including set size with either 2 options or more options (5 or 10 options) and dissimilar (vs similar) charities as moderator. The model revealed similar to the 2x3 ANOVA a significant main effect of set size, B = 0.44, t (429) = 2.84, p = .005. The more options conditions (5 or 10 options) (M = 5.33, SD = 1.50) was evaluated significantly higher in campaign attractiveness than the 2 options conditions (M = 4.85, SD = 1.58). The main effect of dissimilarity (vs similarity) was also significant, B = 0.29. t (429) = 2.01, p = .045. Campaigns including dissimilar choice sets (M = 5.33, SD = 1.48) were on average rated significantly higher than campaigns providing similar options (M = 5.00, SD = 1.58). The interaction effectB = -0.22, t (429) = -0.72, p = .471, was non-significant. 2
4.3 The Impact of Personal Value Traits
The following analysis takes into account the role of personal value traits to provide a complete picture of the choice effect.
4.3.1 The Role of Values on Choice and Campaign Attractiveness
A multiple regression was performed to assess the effect of personal values without taking into account the set size and dissimilarity, to check which values impact the choice attractiveness. The regression including all 5 value categories as predictors was significant, R² = 0.19, F (5,432) = 20.44, p = .000. The predictive paths of IFQL values, B = 0.38, t = 6.77, p = .000, and of AO values, B = 0.18, t = 2.00, p = .050, were significant. These results indicate that an increase of IFQL values results in an increase of the choice set attractiveness. Similarly, an increase of AO values increases the choice set attractiveness. The effects of hedonic values (B = 0.08, t = 1.19, p = .234), values aimed at making a difference (B = 0.05, t = 0.73, p = .463), and utilitarian values (B = 0.091, t = 1.15, p = .250), were all non-significant.
The same regression model was used to study the effects of values on campaign attractiveness. Again, the regression was significant, R² = 0.20, F (5,432) = 20.74, p = .000. The predictive paths of IFQL values, B = 0.28, t = 4.80, p = .000, of AO values, B = 0.32, t = 3.29, p = .001, and of values aimed at making a difference, B = 0.19, t = 2.90, p = .004, were significant. These results indicate that an increase of IFQL values results in an increase of the choice set attractiveness. Similarly, an increase of AO values and an increase in values aimed
2 Testing a 2x2 ANOVA revealed, in line with PROCESS, the same significant main effects of set size, F (1, 429)
at making a difference increases the choice set attractiveness. The effects of hedonic values (B = 0.02, t = 0.24, p = .811), and of utilitarian values (B = 0.09, t = 1.10, p = .271), were both non-significant.
4.3.2 The Moderating Role of Values for Choice Attractiveness as Dependent Variable
I used the PROCESS macro from Hayes for analysing the moderating role of values on choice attractiveness proposed in my conceptual model. Giving the findings of the ANOVA, set size was in accordance with the previous findings defined by either 2 options or more options (5 or 10 options), resulting in a variable with 2 conditions rather that a multicategorical variable. This made it possible to test my model in PROCESS. First, model 2 was tested with PROCESS macro (5000 bootstraps, 95% CI; Hayes, 2013) to include personal value traits as second moderator independent from degree of dissimilarity to test hypothesis 3a). Model 2 with IFQL as value category revealed to be significant R² = 0.17, F (5,427) = 18.07, p = .000. The model revealed only one significant main effect of IFQL values B = 0.49, t = 5.33, p = .000, 95% CI [0.31, 0.67]. The results indicate that an increase in IFQL values results in higher attractiveness of the choice set. All remaining main and interaction effects were non-significant as you can see in the table below (see Table 4).
Table4
Model 2: IFQL values as independent moderator on choice attractiveness
Variable B SE T P LLCI ULCI
SetSize 0.54 0.63 0.85 0.397 -0.71 1.78
Dissimilarity -0.23 0.22 -1.04 0.298 -0.67 0.21
SetSize*Dissimilarity 0.12 0.27 0.45 0.651 -0.41 0.66
IFQL 0.49 0.09 5.33 0.000 0.31 0.67
SetSize*IFQL -0.04 0.11 -0.31 0.753 -0.25 0.18
Table5
Model 2: AO values as independent moderator on choice attractiveness
Variable B SE T P LLCI ULCI
SetSize 0.31 1.04 0.30 0.767 -1.73 2.35
Dissimilarity -0.12 0.23 -0.53 0.594 -0.58 0.33
SetSize*Dissimilarity 0.01 0.29 0.20 0.983 -0.56 0.57
AO 0.46 0.15 3.19 0.002 0.18 0.75
SetSize*AO 0.02 0.18 0.10 0.918 -0.33 0.36
To test hypothesis 3b), if personal values moderate the first order moderator dissimilarity, Model 3 was tested with PROCESS macro (5000 bootstraps, 95% CI; Hayes, 2013). For the first test IFQL values were included and the model revealed to be significant R² = 0.19, F (7,425) = 14.27, p = .000. The results showed only one significant main effect of IFQL values B = 0.50, t = 4.64, p = .000, 95% CI [0.29, 0.72]. All remaining main and interaction effects showed insignificant results (see Table 6).
Table6
Model 3: IFQL values as second order moderator on choice attractiveness
Variable B SE T P LLCI ULCI
SetSize -0.59 0.81 -0.72 0.472 -2.19 1.02 Dissimilarity -0.00 1.15 -0.00 1.000 -2.27 2.27 SetSize*Dissimilarity 1.93 1.36 1.40 0.162 -0.77 4.58 IFQL 0.50 0.11 4.64 0.000 0.29 0.72 SetSize*IFQL 0.17 0.15 -0.21 0.251 -0.12 0.46 Dissimilarity* IFQL -0.04 0.20 -0.21 0.837 -044 0.36 SetSize*Dissimilarity*IFQL -0.32 0.24 -1.35 0.178 -0.80 0.15
effect of AO values remained to be significant B = 0.64, t = 3.32, p = .001, 95% CI [0.26, 1.03]. The results indicate an increase in choice attractiveness for higher AO values. All other main and interaction effects were non-significant (see Table 7).
Table7
Model 3: AO values as second order moderator on choice attractiveness
Variable B SE T P LLCI ULCI
SetSize 1.06 1.39 0.76 0.446 -1.67 3.78 Dissimilarity 2.27 1.73 1.32 0.188 -1.12 5.66 SetSize*Dissimilarity -1.76 2.10 -0.84 0.400 -5.87 2.35 AO 0.64 0.19 3.32 0.001 0.26 1.03 SetSize*AO -0.11 0.24 -0.47 0.639 -0.58 0.36 Dissimilarity* AO -0.41 0.29 -1.40 0.161 -0.99 0.16 SetSize*Dissimilarity*AO 0.30 0.36 0.85 0.396 -0.40 1.00
4.5 The Moderating Role of Values for Campaign Attractiveness as Dependent Variable
Table8
Model 2: IFQL values as independent moderator on campaign attractiveness
Variable B SE T P LLCI ULCI
SetSize -0.37 0.68 -0.54 0.589 -1.71 0.97
Dissimilarity 0.35 0.24 1.44 0.150 -0.13 0.82
SetSize*Dissimilarity -0.15 0.30 -0.50 0.616 -0.73 0.43
IFQL 0.29 0.10 2.94 0.003 0.10 0.49
SetSize*IFQL 0.16 0.12 1.32 0.188 -0.10 0.40
The test was repeated by using AO values and again the model was significant R² = 0.14, F (5,427) = 13.95, p = .000. The model revealed a significant main effect of AO values B = 0.34, t = 2.26, p = .025, 95% CI [0.04, 0.63]. The results indicate that an increase in AO values results in higher attractiveness of the campaign. The model additionally showed a significant interaction effect between the set size and AO values B = 0.38, t = 2.11, p = .035, 95% CI [0.03, 0.74]. The results indicate that stronger AO values positively influence the effect of increased choice on campaign attractiveness. All remaining main and interaction effects revealed non-significant (see Table 9).
Table9
Model 2: AO values as independent moderator on campaign attractiveness
Variable B SE T P LLCI ULCI
SetSize -1.66 1.07 -1.56 0.119 -3.76 0.43
Dissimilarity 0.41 0.24 1.70 0.091 -0.06 0.88
SetSize*Dissimilarity -0.23 0.29 -0.78 0.435 -0.81 0.35
AO 0.34 0.15 2.26 0.025 0.04 0.63
SetSize*AO 0.38 0.18 2.11 0.035 0.03 0.74
main effect of IFQL values B = 0.29, t = 2.43, p = .015, 95% CI [0.06, 0.52]. All remaining main and interaction effects showed insignificant results (see Table 10).
Table10
Model 3: IFQL values as second order moderator on campaign attractiveness
Variable B SE T P LLCI ULCI
SetSize -1.0 0.89 -1.12 0.263 -2.74 0.75 Dissimilarity 0.23 1.25 0.19 0.851 -2.23 2.70 SetSize*Dissimilarity 0.97 1.48 0.66 0.512 -1.93 3.87 IFQL 0.29 0.12 2.43 0.015 0.06 0.52 SetSize*IFQL 0.27 0.16 1.71 0.087 -0.04 0.58 Dissimilarity* IFQL 0.02 0.22 0.09 0.927 -0.41 0.45 SetSize*Dissimilarity*IFQL -0.20 0.26 -0.78 0.437 -0.71 0.31
Running the test again including AO values as second order moderator, the model revealed again to be significant R² = 0.14, F (7,425) = 10.16, p = .000. Within this model, all main and interaction effects were non-significant for α = 0.05 (see Table 11).
Table11
Model 3: AO values as second order moderator on campaign attractiveness
Variable B SE T P LLCI ULCI
For an overview of the significant findings within the conceptual model, see appendix B.
4.4 Additional Measures
The remaining results focus on additional measures outside of my conceptual model. I analysed these exploratory measures to get a better understanding of potential other relevant variables that are influenced by the choice set design.
4.4.1 The Role of Set Size and Dissimilarity on other Dependent Measures
Ease of choice To test whether my manipulation influenced ease of choice, I ran a 2x3
ANOVA. This test yielded only a significant main effect of the degree of dissimilarity, F (1, 427) = 8.50, p = .004. The results indicate that participants perceived their choice as being easier in dissimilar (M = 4.78) rather than similar (M = 4.32) conditions. The main effect of choice set size, F (2, 427) = 0.70, p = .497, nor the interaction effect, F (2, 427) = 1.86, p = .157, were significant.
Perceived social responsibility Testing the influence of the manipulation on perceived
social responsibility revealed a significant main effect of dissimilarity, F (1, 427) = 11.46, p = .001. The results indicate, that campaigns providing a dissimilar set are perceived as more socially responsible (M = 5.63) than campaigns providing similar charities (M = 5.21). The main effect of set size, F (2, 427) = 0.91, p = .404, nor the interaction effect, F (2, 427) = 0.19, p = .826, were significant.
Perceived feeling about making choice The test was repeated to assess the influence on
perceived feelings arising from the choice. No significant effect was observed. The main effect of set size was non-significant F (2, 427) = 0.67, p = .510 as well as the main effect of dissimilarity F (1, 427) = 0.29, p = .590. The interaction effect was also not significant F (2,427) = 0.39, p = .678. Participants in general perceived the same degree of positivity in their feelings about making a choice among all campaigns (M = 5.31 , SD = 1.30)
Campaign effectiveness Similarly, the test of choice set size and dissimilarity on
perceived campaign effectiveness revealed no significant effect. The main effect of set size was not significant F (2,427) = 2.20, p = 112 as well as the main effect of dissimilarity F (1,427) = 2.00, p = .160. The interaction was also not significant F (2,427) = 0.33, p = .718. Participants in general perceived the campaign to be equally effective (M = 5.31, SD = 1.26 ).
Brand-cause fit To rule out that the results are biased by the perceived brand-cause fit I
1.22, p = .269. The interaction effect was also not significant F (2,427) = 1.05, p = .352. Participants in general perceived the same degree of brand-cause fit among all campaigns (M = 5.11 , SD = 1.26)
6. GENERAL DISCUSSION
6.1 FindingsCause-related marketing campaigns become a popular tool to increase brand awareness or to fulfil the social responsibility of the company (Brønn & Vrioni, 2001; Varadarajan & Menon, 1988). Therefore, it is important to understand what factors drive the attractiveness of these campaigns. Prior work revealed that giving the opportunity to choose between a set of charities positively influences the willingness to participate in CM campaigns (Howie et al., 2018; Robinson et al., 2012). This thesis extends these findings by demonstrating how the effect of freedom of choice on the perceived attractiveness behaves for further increasing choice sets varying from 2, 5 and up to 10 charities. To get a better understanding of the composition of choice sets, I also studied whether the degree of similarity between the charities determines how attractive the choice set and campaign will be perceived. The theory suggested that dissimilarity (vs similarity) negatively moderates the positive effect of increased choice.
Results showed that larger choice sets (5 or 10 options) are perceived as more attractive than sets containing only a few charities (2 options) to choose for the donation. Additionally, campaigns that increase the freedom of choice are in general rated more positive, meaning that more options increase the overall attractiveness of the campaign. Contrary to the research on choice overload (Lepper & Iyengar, 2000), no negative “too much choice effect” was found on the set sizes provided within this research.
for campaign and company perceptions but does not moderate the impact of number of options on perceptions.
The thesis also tried to provide a better understanding of the effect of choice by looking at the role of personal value traits. I found that the attractiveness of the choice set and the campaign is generally increasing when people have stronger increasing future quality of life and/or affiliation-oriented values. Higher values aimed at making a difference with others also showed an increase in the overall campaign attractiveness but not for the provided set itself. Although the results provide evidence for the importance of consumer values, no moderation on the effect of (dis)similarity was found. Only AO values were found to positively moderate the effect of number of options on perceptions.
6.1.1 Theoretical Implications
These findings provide several insights for the theoretical understanding of the complexity of choice. The results are in line with the research of Howie et al. (2018) and Robinson et al. (2012), and show that increased choice within the number of options increases the attractiveness of the choice set and of the overall campaign. In general, the positivity of perceptions associated with larger choice sets are in line with choice-related and process-related benefits from increased choice. Choice-process-related benefits include a better option to find a product matching the consumers’ ideal points (Chernev, 2003a), more options to satisfy multiple or variety-seeking consumers (Broniarczyk, 2006) or the increased flexibility provided for uncertain choosers (Lehmann & Kahn, 1991). Complementary, process-related benefits increase the attractiveness of larger assortments through increased stimulation for consumers aiming to find optimal solutions (Trijp, Hoyer, & Inman, 1996) or for those who simply enjoy the shopping experience (Babin, Darden, & Griffin, 1994).
Donors often develop scepticism towards the expenditures of charities and for example avoid organizations investing high percentages in administration and funding (Gneezy, Keenan, & Gneezy, 2014b).
Considering risk and distrust associated with the donation decision, the heuristic-systematic model of information processing can help to understand the transcendence of larger choice sets (Boyd & Bahn, 2009). The heuristic-systematic model assumes that people try to balance their desire to make the right decision and the tendency to minimize processing efforts (Chen, Duckworth, & Chaiken, 1999). Consequently, in conditions of high certainty and low risk, people tend to focus on the reduction of processing effort leading to heuristic decision making (Boyd & Bahn, 2009). The opposite is the case for conditions characterized through high uncertainty, for example choosing the appropriate charity. Those situations evoke the need for more systematic processing which is enabled by providing more information through choice to create certainty (Boyd & Bahn, 2009).
The theory on the processing-benefit to reduced uncertainty also provides a possible explanation for the absence of the “too-much choice” effect that was hypothesized. The charities used for this research were chosen with the purpose to be less known through the media, to exclude biases based on possible previous experiences with the organizations. Thus, participants faced high uncertainty about the condition by having only the information provided in the survey. Following the theory of uncertainty reduction through increased processing ability, it can be argued that an increase in the choice set helps to form an evaluation and reduces risk associated with the decision. This would imply that further increases in the choice set would further positively increase perceptions as long as information processing provides certainty in the decision. In sum, the increase in attractiveness of larger sets can be explained by an increased fit in the expression of the moral self-image through the donation decision and reduced uncertainty about whom to support which is due to increased availability of information to process.
community link, reliability, commitment, congruence, benevolence, transparency, and broad impact. Linking these dimensions to my findings, one can argue that the dimension “reliability” increases for companies offering more diversified choice sets by signalling time and effort spend in forming partnerships with a high number of non-profit organizations. The number of charities willing to cooperate with the company might signal that charities indeed profit from the relationship which increases the reliability of the campaign (Joo et al., 2019).
The “perceived congruence” is another dimension of authenticity that could have been affected by the composition of the choice set. Two categories determine the degree of perceived fit in CSR perception: product dimensions, and personal involvement (de Jong & van der Meer, 2017). The consumers’ social involvement in the cause thus impacts the level of identification with the company and transfers the own behavioural intentions towards the campaign and the organization (de Jong & van der Meer, 2017). Therefore, diversified choice sets are more likely to match the personal social interest of the consumer, positively impacting the perceived congruence of the company.
Finally, the extent of help determines on how authentic CM campaigns and CSR are perceived. “Broad impact” describes the number of people that are affected by the support of the campaign which is perceived as more authentic the more people benefit from the help (Joo et al., 2019). Transferring this into the composition of the choice set, one might argue that more diversity within a choice set is associated with a broader target group for help. This implies that having different charities increases the authenticity of the campaign by helping more distinct people. In sum, the positive effect of dissimilarity on the campaign and organizational level can be explained by increasing authenticity through higher perceived reliability, congruence and a broader perceived impact of help associated with a more differentiated set.
between attributes in an asymmetrical assortment than forming evaluations within the same attribute category. In sum, the negative effect of cognitive cost associated with symmetrical assortments can outweigh the increased choice attractiveness through dissimilarity.
The thesis contributes the literature in not just focusing on external, situational factors but also including personal traits to draw a rather complete picture of the situation. The findings from this thesis show in line with previous literature that the attractiveness of the choice set and campaign is influenced by the relative importance attached to certain values (Galan Ladero, Galera Casquet, & Singh, 2015; Kropp, Holden, & Lavack, 1999). People who have strong affiliation-oriented values and/or increasing future quality of life values perceive the choice set and the campaign as more attractive compared to people scoring low on these values. Kropp, Holden and Lavack (1999) found that these findings hold for values related to the own role within society and the interaction with the social environment. In sum, this thesis confirms findings from earlier studies that high affiliation-oriented values increase the attractiveness of CM campaigns. Additionally, I provide evidence that the same holds for values of environmental protection associated with the consumers’ role in the biological environment and interaction with nature.
