How Egoistic, Altruistic and Social Motivators and Personalization Influence Online Reviewer Engagement

(1)

How Egoistic, Altruistic and Social Motivators

and Personalization Influence Online Reviewer

Engagement

By

Arjan te Brake

University of Groningen Faculty of Economics and Business

Msc Marketing Management track January 2019/20 Rabenhauptstraat 5 9725CA Groningen 0650218522 Arjantebrake92@hotmail.com Studentnummer: S2372215 Completion date: 13-01-2020

Supervisor: dr. J.A. (Liane) Voerman

(2)

Abstract

In recent years, eWOM communication has become fundamental in the online

communication for both businesses and consumers. This master thesis focuses

on how companies can induce consumers to write reviews through the use of

(personalized) requests. This study provides several insights on the motivations

that drive eWOM communication and looks at the effect of these motivations

and the effects of personalization on the willingness to write a request.

Additionally, consumer characteristics, privacy concerns and

individualism/collectivism, that could influence personalization and the three

motivators are examined. From the theory a conceptual framework is build and

tested in an experimental setting. The results of this study show a significant

effect of the egoistic and altruistic motivations on the willingness to write a

review. Additionally, the results of this study show an unexpected that

personalization has a negative effect on the effect of the altruistic motivator on

the willingness to write a review. The implications demonstrate that managers

should use either an egoistic or altruistic motivator in a request to increase OCR

creation. Additionally, if it uses an altruistic motivator in the request it should do

so without personalization. The findings show no significant direct impact of

personalization on the willingness to write.

(3)

Chapter1 Introduction

- “Nothing influences people more than a recommendation from a trusted friend” – Mark Zuckerberg

- “Do what you do so well, that people can‟t resist telling others about you” – Walt Disney

1.1 Introduction/ background problem

Since the beginning of humanity, people have relied on social proof to make informed decisions for a given situation, i.e., the phenomenon where people rely on other people when making decisions (Acquisti, John and Loewenstein, 2012). In the fast-moving world of today, where people have to interpret more and more impulses, social proof is more prominent than ever. Be it in the "most chosen menu" in a restaurant, canned laugh tracks in a comedy, or looking at review websites like Yelp, Facebook, or Amazon when buying online products. These reviews sites are examples of electronic Word-of-Mouth communication (eWOM). With more than half of the world using the internet (Union, 2018). The online world is becoming an integral part of our day-to-day life, and electronic Word-of-Mouth communication is increasingly important in the decision making of people (Kim, S., Kandampully, J., & Bilgihan, 2018). This thesis defines eWOM communication as: "any positive or negative statement made by potential, actual, or former customers about a product or company, which is made available to a multitude of people and institutions via the

Internet.”(Hennig-Thurau et al., 2004). The business side can also mostly gain from using eWOM communication in their marketing strategy. According to Floyd et al. (2014), several main advantages have been noted about the electronic counterpart of offline Word-of-Mouth. First, eWOM is more accessible; within seconds, you can reach an audience of millions all over the world, making it a powerful marketing tool. Additionally, there is more transparent and balanced information since it gives a platform to anyone on the planet with access to the internet. Lastly, the information on the internet is more comfortable to analyze, and

companies can use online feedback to gain helpful insights.

(5)

consumer concerning his experience about a particular product or service. An OCR can be positive, neutral or negative feedback on an online platform. Studies estimate that product and service reviews are the primary factor behind 20–50% of all purchase decisions. (Bughin, Jacques, Jonathan Doogan, 2010). Yet, not all consumers engage in writing these reviews.

Given the previous, the importance of OCRs is fundamental in the purchases in today's online environment. Consequently, to influence purchasing behavior as a company, it is essential to increase OCRs. But for a company to make use of the opportunity that OCRs present by creating involvement and boosting purchase intention, it is critical to get

consumers to write OCRs in the first place. This means that companies have to find a way to stimulate people to write reviews for their products, a form of ORE. Hence the problem at hand is how companies can to trigger people to write OCRs.

1.2 Online reviewer engagement (ORE)

(6)

1.3 Personalization

E-mail marketing is among the top 3 most used digital marketing channels, and is 69% of the companies use it to reach customers (Herhold, 2018). For this reason, consumers get their mailboxes filled with similar messages, advertisement, and spam, e.g., "spam messages accounted for 54.68 percent of e-mail traffic in September 2019 (j. clement, 2019).

Consumers are in over their heads and have an overload of choice. This overload means they cannot distinguish between all messages received and treat each one of them the same. Personalization helps the request to stand out among the other messages and grab the readers' attention (Howard and Kerin, 2004). Therefore, in this thesis, the effect of personalization on the ORE and thus the willingness in the request to write an OCR is researched.

According to Tam and Ho (2006), Personalization is the use of content specifically targeted towards a person, be it in the name or experience of the user. According to them, "It is common to include cues (e.g., the user's name in a greeting message) that connect to personally relevant concepts. A greeting message such as "Dear John, welcome back to giftshop.com. I hope you enjoyed the latest release of Harry Potter in your last purchase," is an example of a simple personalization. It greets the user by his name and refers to a previous experience."

Scientific papers widely regard personalization as one of the tools that can establish a higher response, have a significant impact on satisfaction (Devaraj, Fan and Kohli, 2006)and can increase both advertising and sales revenues(Ho and Bodoff, 2014). Though recent literature has examined personalization in a wide variety of contexts, there is still a lot of debate on the effectiveness since there are also insignificant effects on satisfaction reported (Kim, Kim and Kandampully, 2009). Additionally, privacy is widely debated since highly tailored

advertisements increase the feelings of intrusiveness (van Doorn and Hoekstra, 2013).

1.4 Willingness to write OCRs

(7)

The scientific literature on eWOM focuses mainly on the effect of eWOM on purchase behavior (Howard and Kerin, 2004)or the motivations driving eWOM behavior. (Hennig-Thurau et al., 2004; Mathwick and Mosteller, 2017). When researching the drivers behind eWOM behavior, most research focuses on the motivations to write OCRs in an online community but did not include responses from individual consumers who write reviews online. i.e., the study of (Hennig-Thurau et al., 2004) focuses on the motives for writing online articulations and drew their sample from an online community who engage in eWOM communications on a web-based opinion platform. Mathwick and Mosteller (2017) research the psychological needs and when a consumer is triggered to write a review. They also base their research solely on a reviewer's community (Amazon) and focus mainly on the reviewers who have posted quite some content. Additionally, they state that the applicability of using their theory and testing it in a practical setting warrant further investigation. Furthermore, no papers were found about the business applications of drivers of eWOM and how to best use motivators to induce eWOM behavior. Therefore, this research focuses on types of

motivations used in the request to convince consumers to write a request, and whether it should use personalization in the request or not.

This thesis focuses on finding out what motivates people to write OCRs and which of these motivators a company can best employ to encourage customers to write OCRs. What is the effect of different motivators with/without personalization in the setup of the request on OCR creation?, and finally what are the consumer characteristics impact the effects on the

willingness to write?

(8)

Main questions:

How can companies induce consumers to write OCRs through the use of (personalized) requests?

Sub-questions:

What are the primary motivators for consumers to write OCRs?

How can we use these motivators to induce consumers to write OCRs about the product they bought?

How does personalization in a request to the consumer impact the willingness to write a review?

To what extent do the motivators, egoistic, altruistic, and social impact the targeted consumer In their willingness to write a request? And which one has the most impact on the willingness to write a review?

How does personalization impact the effect of the three motivators, egoistic, altruistic and social, on the willingness to write OCRs

What other consumer characteristics have an impact on the effect of motivators and personalization?

1.4 Structure rest of the thesis

The rest of the paper is organized as follows. In §2, this thesis first discusses the literature surrounding the motivations. Next, this thesis examines personalization, the effect it has on the impact of the motivators on willingness to write a request. The consumer characteristics, individualism/collectivism, and privacy concerns will be introduced in the fourth paragraph. At the end of chapter two, the conceptual model will be graphically presented. In chapter four, the results will be presented, which will be discussed in section five. To conclude the

(9)

The methodology used is an extensive review of articles to comprehend and establish clear definitions as well as find major gaps. databases like Science Direct, JSTOR and EBSCOhost which gave access to a majority of the journals worldwide. First top tier journals regarding marketing were used like Journal of Marketing Research and Journal of Marketing to base the ideas on. When the idea was established these articles were used to find other articles specific to the topic. Furthermore, to improve triangulation this thesis tried to find multiple sources of support. First an exploratory research was done in order to find articles related to eWOM and OCR‟s. Search terms included “eWOM” “Online reviewers engagement” etc. were used to first get an overall view on the topic eWOM. Once a few articles were found each article was reviewed and an emerging pattern was discovered. Consequently, the search could be narrowed down to Motivations to eWOM and how to induce this behavior. References in relevant articles were used to find additional articles, papers and books. The selection of articles included was based on the relatedness of the subject and each article was summarized and shortly reviewed to find an overarching theme. Furthermore, by triangulation of time-span, writers of multiple nationalities and journals from cross-sectional international business topics this article tries to ensure reliability. The included papers used explanatory, exploratory, cross-sectional, and longitudinal research. The papers were checked for accuracy by checking facts and dive into the sources as well verify the documents against other papers. The authority of the writers was checked and papers with a lot of citations, credentials and top journals were assigned a heavier weight. Objectivity was checked by using multiple articlescovering the same topics.)

Chapter2 Literature review

Theoretical framework

H2.1 Overview

The main question of this thesis is how to induce consumers, which have bought a product, to write an OCR through the use of a request. Firstly, this chapter examines the different

perspectives on how people are motivated to write OCRs online, the so-called ORE. This chapter derives three categories of motivators to write OCRs based on several scientific articles: egoistic, altruistic, and social. Secondly, this chapter delves into the literature about the supposed effects of personalizing this request. Thirdly, this thesis discusses the interaction effect between personalization and the three motivators. Lastly, the implications of the

consumer characteristics individualism/collectivism and privacy concerns on the effects of the motivators and personalization are discussed. Concluding, from these theories, the conceptual framework is derived and graphically presented.

2.1.1 Motivators to write OCRs

(10)

(Mathwick and Mosteller, 2017) take into account that writing a review and sharing it online is the process of the market helping behavior since OCRs help people in the decision-making in the process of buying. Market helping behavior is characterized by both altruistic and egoistic motives (Bendapudi, Neeli, Surendra N. Singh, 1996). Altruistic behaviors are unselfish concerns by investing their own tangible and intangible assets (i.e., capital, labor, or time) to the benefit of others without expecting something in return. (Simon, 1993) On

contrast, a motivation to write a review can be for selfish purposes. Egoistic behaviors are selfish behaviors to enhance one‟s interests, be it economical, hedonic, or self-gratifying (Wasko and Faraj, 2005).

Both Mathwick and Mosteller (2017) and Hennig-Thurau et al. (2004) conclude in their research that egoistic and altruistic motivations drive the creation of eWOM communication. On the one hand, both papers found that for manyreviewers, egoistic motivators play a crucial role in the reviewing experience. Mathwick and Mosteller (2017) found that personal achievement, professional benefits, i.e., remunerations and enjoyment, drives the behavior of writing OCRs. Hennig-Thurrau et al. (2004) found that economic incentives and the potential to enhance their self-worth were both motivations to articulate oneself online.

On the other, the reviewing process engages consumers to act in altruistic behavior, helping other consumers in the process of buying. Mathwick and Mosteller (2017) find three distinct categories, which al had a different reason of why they engaged in altruistic market helping behavior. Reviewers, who were motivated by altruistic motivators, saw it as their duty to prevent other people from buying junk and wanted to help people make an informed decision. However, they also argue a third category of people who saw altruistic market behavior as either something incidental to their opinions or instrumental to their personal goals of advancing in rank. Also, Hennig-Thurrau et al. (2004) find that the main reason reviewers engaged in eWOM behavior are because of concern for other consumers or to help the company.

Another component that plays a crucial role in both papers is the social component. Mathwick and Mosteller (2017) state that “The significance of CC reviewers’ social connectedness suggest that reviewing is a socially embedded experience for them

(11)

Therefore, it is argued from these papers that the motivators to write OCRs fall into three distinct categories: Egoistic, altruistic, and social.The next paragraph explains these

motivators separately before the discussion of the effect of the motivators on the willingness to write a review.

2.1.2 Egoistic motivator

Hennig-Thurrau et al. (2004) identify two concrete motives that are egoistic: egoistic self-enhancement and economic incentives. Mathwick and Mosteller (2017) identify multiple egoistic motivators like competence, reputational benefits, and monetary incentives. First, self-enhancement, reputational benefits, and competence are discussed and then the economic incentives. Maintaining positive self-esteem is the driving motivation of self-enhancement. Self-enhancement is both an egoistic as a social motivator. The primary driving motivation for self-enhancement is to create the image of being competent. According to Mathwick and Mosteller (2017) competence refers to the need to be recognized in one‟s individual

capacities and capabilities. self-enhancement is very closely related to the need for

competence which emphasizes “demonstrated effectiveness resulting from persistent effort to hone one’s capabilities and the sense of confidence derived from “feelings of efficacy”.” (Mathwick and Mosteller, 2017).

This study finds explanation of the egoistic motives self-enhancement and the need for competence in the theory of Ryan and Deci, (2002). They explain that rewards or

positive/negative feedback can intrinsically motivate people to engage in voluntary behavior (i.e., writing an OCR). The fulfillment of mastering a capability gives a sense of purpose and confidence; for example, everyone can remember the joy they experienced when they cycled on their own for the first time. In the context of eWOM communication, an egoistic motive of competencefor writing an OCR can be, for example, when someone is trying to improve their writing. E.g. in the research of Mathwick and Mosteller (2017) a subject replied: “greatly aided my writing skills…[T]he Amazon reviewing system is a major contributing factor for my love of writing.”

Both Hennig-Thurrau et al. (2004) and Mathwick and Mosteller (2017) discuss

economic benefits as a motivator that can drive OCR creation. Financial incentives are said to exist where an agent can expect some form of material reward –, especially money – in exchange for acting in a particular way (Dalkir, 2013). Economic rewards for eWOM

(12)

payment in this case. Companies employ a multitude of economic rewards besides monetary payment. Free memberships, credits, and discounts are a few of the many marketing tactics in practice to create an incentive to write online content.

This thesis defines egoistic purposes as the desire to advance or maintain one‟s welfare by writing a review. This self-enhancement can stem from either receiving a

monetary reward or gratification incentive. In this thesis, we examine both financial/economic incentives and self-enhancement.

So, egoistic motivators are motivators that are a result of acting in one‟s self-interest. I.e., the desire to advance or maintain one‟s welfare and are related to personal gratification like the need for competence or self-enhancement or rewards like economic incentives. Due to

intrinsic and extrinsic motivations, people can be motivated to write OCRs. An example of an extrinsic egoistic motivator to write a review would be offering a free coupon with a 50% discount; an intrinsic egoistic motivator could be for someone to enjoy writing or becoming a better writer.

2.1.3Altruistic motivator

Anthony, Smith, and Williamson (2009) gave substantial evidence for an altruistic motivator for OCRs. In their research, they examined the open-source online encyclopedia Wikipedia. In their findings, they prove that registered members, motivated by reputation, make regular contributions. However, a substantial quantity of articles of high quality came from anonymous patrons. From these findings, they conclude altruistic motivators drive the authors to contribute. They argue that these authors are not motivated by social recognition since they contribute anonymously.

(13)

and his or her subsequent desire to help the company (Sundararam and Kaushik, 1998). According to Hennig-Thurrau (2004), “The customer is motivated to engage in eWOM communication to give the company “something in return”for a good experience. The intended effect of his or her communicative activities is that the company will become or remain successful.” On the other hand, the motive to engage in eWOM communication can be to ventilate negative experiences with the product and prevent other consumers from making the same mistakes. Mathwick and Mosteller (2017) define: “altruistic helping motives as the desire to help consumers make informed buying decisions.” In their paper, there are two examples of helping other consumers. First, respondents were helping other people to make informed buying decisions by giving them honest views about the products. Second,

respondents saw it as their duty to prevent people from buying junk, i.e., their main reason to engage in altruistic behavior is to avert potential buyers from making a mistake.

So, altruistic motivators are motivators that driven by acts in the interests of helping others and are related to helping other consumers in the buying-process of a product or

helping the company succeed. When writing an OCR, an altruistic motivator would be to help other people or the company.

2.1.4 Social motivator

In this thesis, the social motivator is defined as the motive to write OCR‟s to become part of a community and gain social benefits and fulfills the social needs to be integrated into a social network, as will be explained below.

Thriving online communities are the evidence of a sharp increase in the „social‟ aspect of OCRs (Chen and Xie, 2008). Sun, Dong and McIntyre (2017) explain that: “The increase in social connectedness is understandable since social connections tend to be more attractive than "anonymous" reviews because of the high level of trust and personal knowledge that make such recommendations more relevant.” The two before mentioned articles argue that a social component is important, if not crucial, for ORE. Ryan and Deci (2002) also say that a sense of belongingness can motivate a person intrinsically. Mathwick and Mosteller (2017) also give substantial evidence that reviewing is a socially embedded experience, and social interaction was the main reason they were motivated to write OCRs. Next, in the paper of Hennig-Thurrau et al. (2014) the motive factor with the most substantial impact in their model is social benefits. According to Hennig-Thurrau et al. (2014), social benefit and

(14)

virtual community can represent a social benefit to a consumer for reasons of identification and social integration. Social self-enhancement: is driven by one‟s desire for positive recognition from others, underlying behavior that can be gratified only through social interaction. Aforementioned under the egoistic motivator, this concept can be both a social and egoistic driver. In this thesis, self-enhancement is shared under the social motivator when self-enhancement is focused on getting social recognition. eWOM communication that is read by others gives writers a degree of social reputation that can become valuable to one‟s self-concept. Adding value to the community through the writing of OCRs is initiated because the goal of the individual is to gain social benefits, i.e. to be seen as a consumption expert or intelligent shopper by the rest of the community. This reputational enhancement can be a motivator to write OCR‟s. This paper defines the social component according to Hennig-Thurrau et al. (2014) and Mathwick and Mosteller (2017). The definition is to be socially integrated into a network and engage in interaction with other people.

So, social motivators are motivators driven by a desire to be integrated into a network and are related to social interaction, integration, social benefit, and self-enhancement. An example of the social motivator would be the writing of OCRs to establish a social connection with other consumers.

2.1.5 The Effect egoistic, social and altruistic motivator

(15)

benefits or reputational benefits. Sun et al. (2017) argue that “monetary rewards may actually suppress intrinsic motives, and consequently, become ineffective or even counterproductive.” I.e., monetary rewards could decrease the willingness for socially integrated members because being paid for reviewing could damage their reputation. As helping is a behavior that is highly regarded in most social contexts, when helping in public, helpers usually receive social

benefits in return (social recognition, praise, and gratitude) (Bénabou and Tirole, 2006). The community could see payment as the primary reason to engage in reviewing instead of the social aspect (Heyman and Ariely, 2004). When looking in the context of sending a request to write a review, targeted customers are likely to be uninvolved on the reviewing platform. Therefore, social benefits, reputation, or reputational benefits do not play a role.

Additionally, when social benefits motivate reviewers the recognition and reputational rewards received likely increase with the size of the audience (e.g., Zhu and Zhang, 2010; Toubia and Stephen, 2013). However, in the context of sending a request to an individual, the perceived audience is dependent on the perception of the customer of the online review platform on which the review will be posted. Assuming that most customers targeted with a request are not on a platform yet, the perceived social benefits will probably be low. Bénabou and Tirole (2006) also state that if the visibility of the actions plays a role in

motivations to act egoistic or altruistic. When actions are visible in a community, people tend to act more generous than in a private situation. This behavior has probably to do with

maintaining a positive self-image and anticipating a negative bias towards egoistic behavior. (Siem and Stürmer, 2019) White, Peloza, and Emerson (2009) supported this theory:

“collectivist appeals were found to be less effective than individualist appeals when responses were private rather than public, because people could not be held accountable for not

engaging in socially desirable actions”. In the context of sending a request to a customer the visibility is limited to the individual and therefore customers do not experience social pressure to act more altruistic.

(16)

This leads us to the following hypothesis:

Hypothesis 1: Compared to no motivator, it is expected that the egoistic, social, and altruistic motivators have a positive impact in varying degrees. Relatively speaking, the egoistic

motivator will have the highest positive impact, the altruistic motivator the lowest positive impact, with the social motivator having a positive effect somewhere between egoistic and altruistic.

2.2 Personalization of the request to write an OCR

In recent years, personalization has been widely researched, both in the field of marketing and psychology. In the majority of studies, personalization was found to have a significant

positive effect on the satisfaction of consumers (Devaraj, Fan and Kohli, 2006) and a positive impact on consumer loyalty (Zhu and Zhang, 2010). According to Salonen and Karjaluoto (2016), satisfaction and loyalty are vital concepts to assess successful personalization since they are key drivers of a firm‟s performance. In their overview of articles over the last ten years, they find that most studies found a significant positive relationship of personalization on satisfaction. Furthermore, Howard and Kerin (2004) find that messages targeted at the individual that use personalization or personalized product offerings boost responsiveness. They argue that personalization has higher relevance to their personal needs and preferences, thus increasing purchase intentions. Although the majority of articles find a positive effect of personalization on satisfaction, Wattal et al. (2012) find that: “consumers respond negatively when firms are explicit in their use of personally identifiable information (i.e., a personalized greeting).” Although personalization based on product history, i.e., purchase of a product, elicits a positive response.

The underlying psychological mechanic is the self-reference effect, which refers to people‟s tendency to better store information and gives attention to messages that involve themselves (Rogers et al., 1977). According to Tam and Ho (2006), self-reference in web personalization is the use of content specifically targeted towards a person, be it linked to the name or experience of the user. They found that the use of self-referent web content creates a larger attention span; users recall the material faster and spend less time on decision making than non-self-referent web content. Additionally, people are more inclined to accept offers targeted explicitly to the self, were more likely to create preferences. The goal of

(17)

engaged in a favorable response to the company and additionally tries to establish a positive connection with the consumer hoping to influence future buying decisions. Sending a personalized request makes it more convenient for the customer to write an OCR since it reduces the amount of search effort to post an OCR, therefore speeding up the decision-making process of the customer. According to Sahni, Wheeler, and Chintagunta (2018), personalization has a positive effect even when the message has no link to the company in any way. Adding the recipient‟s name to the subject line increased the probability of the recipient opening it by 20%. Furthermore, When personalizing a request, the personalization must have some personal relevance, e.g., a name or situation, because a poor fit often results in

annoyance or even an adverse purchase reaction (Thota and Biswas, 2009).

Therefore, we derive the following hypothesis

Hypothesis 2: A personalization of the request to write a review will increase the willingness to write a online customer review.

2.3 Interaction personalization on egoistic, altruistic and social

Much research is done on both personalization and the different motivators. However, research on the effect of personalization on the different motivators is limited. According to Salonen and Karjaluoto (2016) , who did a literature review of the top 20 marketing and information system journals from 2005 to 2015, show that there is substantial research on the effect of personalization. A multitude of items has been researched, e.g., satisfaction/loyalty, implementation, and privacy and trust. Additionally, they argue that the research of contextual factors is a trend that has recently emerged. Contextual factors, in this case, the presence of egoistic, social, or altruistic motivations in the request, can be used in assessing the contextual effects of personalization. Some researchers studied related effects. For example Steenkamp and Geyskens (2006) found the effect of customization on the perceived value of websites to be higher in countries where national-cultural individualism is higher. However, there is, to my knowledge, not a study on the effect of personalization on egoistic, social, or altruistic motivators. Therefore, the reasoning on the interaction effects is mostly based on

assumptions. First, for the egoistic motivator, personalization of the request could affect the self-enhancement of the person in question. Personalization could trigger a sense of

(18)

personalization had a larger effect. They found that: “people who lived in more individualistic countries give more weight to pleasure, to privacy/security protection, and to customization than people from collectivist countries.” Since pleasure is an egoistic motivator this research argues that individualistic cultures are more egoistically oriented. Thus, for people of

individualistic countries the egoistic motivator would be most affected by the egoistic

motivator. According to Steenkamp and Geyskens (2006) individualistic cultures are positive about customization. However, if people find the personalization to be intrusive negative privacy concerns are triggered. Therefore, it is possible to hypothesize a positive and negative effect. In the context of this research, this thesis expects a more significant effect of

personalization than privacy concerns; thus, it is argued that there is a positive impact on the effect of the egoistic motivator on the willingness to write a review.

Therefore, we come up with the following hypothesis:

Hypothesis 3a: There is a positive interaction effect of personalization on the positive effect of the egoistic motivator on the willingness to write OCRs

Based on the previous logic, when looking at collectivist countries, which are more socially embedded, it would be expected that there is a smaller effect of personalization on the effect of the social motivator on the willingness to write a review. However, the request is sent to individuals and is not demonstrated within a group, in which social dynamics have a more prominent role. Therefore, this thesis argues that the social motivator and personalization have a smaller or no interaction effect than the interaction effect between the egoistic motivator and personalization.

Hypothesis 3b: There is a smaller effect than H3a/no significant positive effect of

personalization on the positive effect of the social motivator on the willingness to write OCRs Altruism in itself as a concept has nothing to do with personal motivations and is targeted towards allocating resources like time, effort, or money towards the benefit of others.

(Mathwick and Mosteller, 2017) It is argued that there is no effect of personalization on the altruistic motivator on the willingness to write a review other than the direct effects of

personalization on the willingness to write a request. The direct effects of personalization include a larger attention span, faster recall, and better persuasion.

(19)

2.4 Other variables

Besides the effect of the motivators, egoistic, altruist, and social, and the personalization and the interaction of personalization on the ORE, there are two expected moderating effects that also impact the motivations to write OCRs: Individualism/collectivism which moderate the motivators and privacy concerns which moderate personalization.

2.4.1 Individualism/ Collectivism

Collectivist societies subordinate individual goals to the goals of a few large in-groups (Triandis and Gelfand, 1998). In individualistic societies, social fabric and group norms are much looser. Central to individualism are individual goals, independence and personal gratification. According to Luo et al. (2014) “people with individualistic culture (high ICO) will tend to be more independent, they can be regarded as “private self”, their own goals, values and beliefs are the dominant impact factors to affect their cognitions ,they will utilize their internal cognition to evaluate the target issues regardless of others’ opinions; whereas, persons with collectivistic cultural orientation (low ICO) can be regarded as “collective self” or “public self”, they are more likely to follow the social/group norms instead of their own beliefs.” According to this reasoning, when an individualistic person receives a request to write a review, the egoistic motive will probably have a more significant impact. On the other hand, an altruistic or social motivation will persuade a consumer with a high collectivistic orientation more likely.

All this leads to the following hypothesis:

Hypothesis 4: The more individualistic (collectivistic) oriented someone is the more likely they respond to egoistic (altruistic and social) motivators in the willingness to respond to the request to write an OCR.

2.4.2 Privacy regards

(20)

higher degree of personalization, by, for example, adding personal identification, the better fit the ad has to the customer. On the other hand, the more personalization an ad has, the likelier the chance that the customer experiences feelings of intrusiveness, which negatively affect their purchase intentions. The amount of intrusiveness experienced by the consumer is also dependent on the expected potential benefits gained. I.e., people were more willing to give up information to be profiled on websites than online ads because they were hoping to gain more benefits from these websites (Awad and Krishnan, 2006). According to Rathore and Panwar (2015), “online anonymity and privacy are among the many characteristics that are unique to eWOM” when they compare eWOM to WOM communication. Therefore, privacy concerns could also have a direct impact on the willingness to write a review. If online anonymity and privacy are not guaranteed, and people have privacy concerns, they are less likely to engage in the writing of OCRs

Therefore, we come up with the following hypotheses:

Hypothesis 5a: A greater personalization of the request to write a review will decrease the willingness to write an online customer review.

Hypothesis 5b: a greater concern about privacy will decrease the willingness to write a online customer review.

2.6 Conceptual model + hypothesis

To give a clear picture, the conceptual model is hypothesized from the literature discussed in this chapter is represented in see figure 1. The main EVs are the motivators, egoistic,

altruistic, and social (H1), which all have a speculated positive effect on the willingness to write a review. However, the main assumption is that their impact on the DV differs from each other. The egoistic motivator is argued to have the relative highest positive effect (shown by a +) on the willingness to write, the altruistic motivator the lowest positive effect

(21)

Figure 1 Conceptual model

Hypotheses

- Hypothesis 1: Compared to no motivator it is expected that the egoistic, social and altruistic motivators have a positive impact in varying degree. Relatively speaking, the egoistic motivator will have the highest positive impact, the altruistic motivator the lowest positive impact, with the social motivator having a positive impact somewhere between egoistic and altruistic.

- Hypothesis 2: A personalization of the request to write a review will increase the willingness to write a online customer review.

- Hypothesis 3a: There is a positive interaction effect of personalization on the positive effect of the egoistic motivator on the willingness to write OCRs

- Hypothesis 3b: There is a smaller effect than H3a/no significant positive interaction effect on personalization and the social motivator on the willingness to write OCRs

- 3c: There is no expected significant effect of personalization on the positive effect of altruistic motivator on the willingness to write OCRs

- Hypothesis 4: The more individualistic (collectivistic) oriented someone is the more likely they respond to egoistic (altruistic and social) motivators in the willingness to respond to the request to write an OCR.

- Hypothesis 5: A greater personalization of the request to write a review will decrease the willingness to write a online customer review.

(22)

Chapter 3 Research design

3.1 Type of research

To test the hypotheses in the conceptual model, experimental research was conducted to see if a change in personalization or the three motivators have a significant effect on the DV

(Willingness to write a review). From the theory, the conceptual model has derived three distinct motivators, egoistic, altruistic, and social, that influence the willingness to write an OCR. Furthermore, the literature review gave reason to research effect of personalization of the request on the willingness to write a review. Additionally, this thesis looks at the

moderating effect of both privacy concerns and individualism/collectivism. This thesis conducts research to empirically make an assessment of the relevance of the effect of the personalization and motivators used in a request. The goal of this experimental research is to test the hypothesis through deliberate manipulation of the independent variables.

3.2 Participants and design

253 participants responded to the online questionnaire that was developed and put online. A total of 253 respondents started the questionnaire, of which 42 contained missing data. Of the completed 210 Participants the average time to complete the survey was 280 seconds (4.65 min). Of the 210 participants ,118 were male and 92 females between the age of 18 and 74 (M-27,04, SD=9,842). (see table 1) Most of the respondents were in their 20‟s. (Table 1) For completion of the survey, the participants could win either a gift card for 15€ or a gift box of craft beers. There was no significant difference in all eight conditions for gender (χ2 (17) =16,020 p =,522), though there was for age (χ2(442) = 552,176 p = ,000). (See appendix 1.1 & 1.2)

Table 1

Variable Categories Percent of sample N

(23)

The participants were randomly assigned to the conditions of a counterbalanced 2 (no personalization vs. personalization (self-referent) ) x 4 (No motivator vs. Egoistic vs.

Altruistic vs. Social.) between-subject design (see table 2). For this research, the respondents were given one of the eight conditions which were either personalized or non-personalized in combination with no motivator or an egoistic, altruistic or social motivator.

Table 2

Conditions No personalization Personalization

No motivator 1 2

Egoistic motivator 3 4

Altruistic motivator 5 6

Social motivator 7 8

3.3 procedure

(24)

3.4 Experimental variables

This paragraph gives a detailed description of the IVs and show the manipulation and measurement of the IVs.

3.4.1 Manipulation EV1 and EV2

The scenarios were built up around the fictional situation in which the respondents had bought a new mobile phone with the following scenario.

Imagine the following situation: As your phone broke down, you've decided to buy a new mobile phone. Your main criteria are that the phone should be easy in use and it has to have an excellent camera. After a long time of searching and comparing several phones you've decided to buy the Nona 8408-X, which perfectly fits your criteria. One month after the purchase, you receive the following message (see next page).

This research chooses a mobile phone because it is a high involvement good. According to (Tam and Ho, 2006), high involvement goods are products that are more expensive. Thus, there is more risk associated when buying the product. Therefore, the consumer is more willing to spend time and resources to reduce uncertainty in decision making. Because they are eager to spend more time and resources, there is a higher chance they will write a review on the product. Additionally, OCRs for high involvement goods is thought-out to be more beneficial to the firm (Kim et al., 2012). Subsequently, the respondents were shown one of the eight. This is where the IVs are manipulated in the following ways. (See appendix 2.2)

(25)

Thirdly, a small sentence was added at the bottom of the request with the motivator in it. For egoistic motivator: “8/10 consumers cashed their discount coupon, earn yours now!”

Altruistic motivator: “8/10 consumers use reviews to make a decision, help others make their decision!” and for the Social motivator: “8/10 consumers who become a member become an active member! In figure 2 below you see on the left no personalization with no motivator and personalization with an egoistic motivator.

Figure 2

The manipulation of the conditions with motivators is checked after the scale for willingness to write a review with the question: “The request to write a review was aimed at helping others” (see survey Q6) on a 7 Likert scale (1 being “strongly agree”, 7 being “strongly disagree”). Expected is that subjects who have a condition with an egoistic motivation score a high score and the condition with an altruistic motivator a low score. The expected result from a condition with social motivation is to fall somewhere in the middle.

3.4.2 Manipulation check

(26)

motivator had the highest mean (Appendix 2.3.1) so these subjects agreed the least (4,77). Like expected, the altruistic scenario had the highest agreement (2,74) and the egoistic scenario the least (4,24). The social scenario was between the egoistic and altruistic, with a mean of 3,66. (F(3,206)=16,370, p<0.00 (Appendix 2.3.2). A Tukey HSD test was conducted (Appendix 2.3.3) and indicated that there was a significant difference between the no

motivator and the altruistic and social motivator. The altruistic motivator has a significant difference with all the other motivators. The egoistic motivator has a difference with the social motivator. Concluding, support was found for our manipulation of the motivators. IV2 (Personalization): Comprised of no personalization and personalization

Manipulation in the conditions where personalization of the request was present was done in two ways. First, the name, which was asked as a first question in the survey, was filled in. So instead of Hi Customer (No personalization) the message begins with Hi {filled in name} (personalization). Additionally, in the personalized scenario, the specific buying situation is referred to twice. The first manipulation is in the second sentence, where it reads, “Thank you for buying our Nona 8408-X!” instead of “Thank you for buying our product!”. The second manipulation is on the blue button, which reads “Write a review for the Nona 8408-x” instead of “Write a review.” The manipulation of the personalized condition is checked after the scale for willingness to write a review with the question: “The request to write a review was very personalized” (see survey Q6) on a 7 Likert scale (1 being “strongly agree”, 7 being “strongly disagree”). Expected is that subjects who have an impersonalized condition score a high score and condition with personalization score a low score.

Results showed that the scenario with no personalization had the highest mean (Appendix 2.4.1), so these subjects agreed the least (5,11). Like expected, the scenarios with

personalization in it had a higher agreement (4,36). (F(1,208)=11,146, p<0.01. (Appendix 2.4.2)) Concluding, support was found for our manipulation of personalization.

3.5 Operationalization of variables

(27)

Table 3

3.5.1 Measured scales

The survey was comprised of three parts: willingness to write a review, privacy concerns, and individualism/collectivism (see table 3). The dependent variable is ORE in the form of the willingness to write a review. The dependent variable was immediately asked after one of the eight conditions when participants were asked to signify their willingness to write a review. To measure the willingness to write a review, a scale of Wu, Hu, and Wu (2010) was adopted and transformed. This scale had a reliability with a Cronbach‟s alpha of ,947. The

measurement of the questions is done at an interval level on a seven-point Likert scale ranging from (1) “Extremly unlike” to (7) “Extremely likely”. The items of the intention to write a review had to be transformed with the second item being deleted, which resulted in a 3 item scale which had significant internal consistency with a Cronbach‟s alpha of .917.

Additionally, a slider is used to test the willingness to write a review. The slider is used to make sure the internal validity rises. The question used: is what is the probability that you will write an review with a slider ranging from 0-100. To measure privacy concerns a scale of

Intention to write a review Purchase intention

,947 / ,841 2.572 85,722 Wu, Hu, Wu and Zheng (2010)

1..My intention to write a review for this company is …

2. The likelihood that I would not write a review for this company is .. ( R )

3. The probability that I would consider writing a review for this company is

4. My willingness to write a review for this company is.

Privacy concerns Privacy concerns

,90 / ,841 2,974 74,358 Okazaki, Li, and Hirose (2009)

1. It usually bothers me when __________ ask me for personal information.

2. When __________ ask me for personal information, I sometimes think twice before providing it.

3. It bothers me to give personal information to so many __________.

4. I‟m concerned that __________ are collecting too much personal information about me.

Individualism/collectivsm Individualism / Collectivism ,735 / ,548 1,426 28,516 Erdem, Swait, and Valenzuela (2006)

1. I like sharing little things with my neighbors 2. Being a unique individual is important to me. (r) 3. Decisions reached in groups are better than those reached by single individuals.

4. I usually sacrifice my self-interest for the benefit of my group.

(28)

Okazaki, Li, and Hirose (2009) was adopted where it had a Cronbach‟s alpha of ,90. The measurement of the questions is done at an interval level on a seven-point Likert scale ranging from (1) “Extremly unlike” to (7) “Extremely likely”. The 4 item scale had a high reliability with a Cronbach‟s alpha of ,882. To measure the individualistic or collectivistic orientation of a person a scale with the reliability with a Cronbach‟s alpha of ,735 of Erdem, Swait, and Valenzuela (2006) was adopted and transformed. The measurement of the questions is done at an interval level on a seven-point Likert scale ranging from (1) “Strongly disagree” to (7) “Strongly Agree”. In this thesis it is transformed to a two item scale comprised of question 2 and 5, which had a Cronbach‟s alpha of ,458.

3.5.2 Factor analysis and reliability analysis

First the scales “willingness to write a review”, “privacy concerns‟ and

“Individualism/collectivism” are analyzed through a factor analysis, principal component analysis (PCA) to find the underlying dimensions and tested for their internal consistency with a reliability analysis (Malhotra, 2018).

Willingness to write a review

“Willingness to write a review” consists of 4 item scale which was measured on a 7 point Likert scale. Several tests are used to check for the appropriateness for the factor

analysis. Firstly, the KMO measure of sampling adequacy, which predicts if data are likely to factor well, based on correlation and partial correlation (Malhotra, 2018). If KMO is below 0,5 it is common to drop the variable with the lowest individual KMO Statistic(Malhotra, 2018). KMO is higher than 0.5 (,763). (see appendix 2.4.1)

Bartlett‟s Test of Sphericity show a p-value smaller than 0.05. (P = 0,00, appendix 2.4.1) Therefore, the null hypothesis that the correlation matrix is an identity matrix is rejected and significant correlations exist which gives us more evidence that factor analysis is applicable. (Malhotra, 2018) Lastly, as communalities measure the percent of variance in a given variable explained by all the extracted factors, we see that in our case 3 out of the 4 extracted

communalities are bigger than 0.4. (see appendix 2.4.2) However, the second question “The likelihood that I would not write a review for this company is .. ( R )” Has a communality of below 0.4 (,305). From the data one could assume that the scale can be reduced to 1 factor. The eigen values is >1, total variance >60% and the factors explain > 5% each. Additionally, the scree plot also points at 1 factor. (appendix 2.4.3&4). To extract the components, we use the method of Principal Component Analysis. This method is commonly used among

(29)

by the same underlying dimension (or factor) and are highly correlated with each other (Malhotra, 2018). Varimax is used to rotate the factor matrix to minimize the number of variables which have high loadings on each given factor cannot be used since there is only one factor. All loadings where >0.5 (appendix 2.4.5).

Next we do a reliability analysis which tests the Cronbach‟s alph to measure internal consistency of the scale. (Malhotra, 2018)Reliability Analysis is in a sense the measurement of strength of these dimensions and how reliable the scale is. If this statistic is higher than 0.6 the factors are strong enough. The factor representing “Willingness to write a review has a value of ,841 (appendix 2.5.1). But looking at the appendix 2.5.2 it can be seen that if question 2 is extracted the cronbach‟s alpha rises to ,917. Question 2 is a reversed scale so there is reason to believe that it is not as correlated to the rest of the scale due to a couple of reasons. It could be that people missed that it was a reversed scale and did not notice the word “not”. Another reason could be that the intention of liking and not liking is not evenly

distributed and people are drawn to be more positive and avoid being negative. Question 2 is dropped from the scale and a new factor scores are computed for the willingness to write a review using question 1, 3 and 4.

Privacy regards

“privacy regards” is also a 4 item scale on a 7 likert-scale The same tests are applied to the scales. For the Factor analysis the KMO is higher than 0,5 (,828) and Bartlett‟s test of sphericity is significant (,000) (Appendix 2.6.1) The communalities are all above 0.5

(Appendix 2.6.2). And the eigenvalue of 1 factor is above one, from which more than 60% of the variance is explained (74.358%) (Appendix 2.6.3). The PCA loads all on one factor so there is a high support that the scale used is measured on one dimension. The internal consistency has a cronbach‟s alpha of ,882 and deletion of items of the scale do not further improve the alpha. (Appendix 2.6..5&6)

Collectivism/individualism

“Collectivism/individualism” is on a 6 item scale on a 7 likert-scale The same tests are applied to the scales. For the Factor analysis the KMO is higher than 0,5 (,557) and Bartlett‟s test of sphericity is significant (,000) (Appendix 2.7.1) 3 of the 6 communalities are above 0.5 (Appendix 2.7.2). The communality of “I like sharing little things with my neighbours” is way to low so it is deleted. The new factor analysis of 5 items gives a KMO of (,558) ) and

(30)

2.7.5). In the PCA. Varimax is used to rotate the factor matrix to minimize the number of variables which give high loadings on 2 factors. (2.7.6). This is explainable because the two items (2 and 5) are focused on more of the individualistic scale and the items more towards collectivism. However, being individualistic does not necessarily mean that you cannot be sharing things or think group consensus is bad. A reliability analysis is done for the two factors which give an a cronbach‟s alpha of ,458 for the individualistic one and a cronbach‟s alpha of ,423 and deletion does not lead to an increase in the internal consistency. (Appendix 2.7.7&8). Because the alpha‟s are under .6 the internal consistency is unreliable. However, for the sake of the analysis the individiualistic scale is used in further analysis. The questions are pretty straightforward measuring individualism which best fits into the conceptualization of the model. .

3.6 Plan of analysis.

To do the linear regression and build the model several things first will have to be checked. In this paragraph the way of analysis is explained step by step.

In order to proceed with a regression analysis, the data has to be ordered and categorized. The output from Qualtrics was imported into SPSS. Missing data was deleted and different

dummies were created.

3.6.1 Dummies and other variables

First, a variable was created for all the conditions, which had a value corresponding with table 2. From this variable 1 dummy (“self-reference”) for personalization is created. A dummy variable is a numerical variable, which can take the value of 0 or 1, used in regression analysis to represent subgroups of the sample in your study (Malhotra, 2018). Another three dummies were created for the different motivators (“egoistic”, “altruistic”, and “social”).

Second, the items on the scales resulting from the factor analysis are computed. The DV = “Willingness to write a review” is computed from question 4.1, 4.3, and 4.4 (Appendix 4.2) by the formula: (4.1+4.3+4.4)/3. The items “privacy concerns” and

(31)

The slider is checked (appendix 2.8) against the DV=willingness to write a review. There is a high correlation (,848, p=,000). Which means that the DV slider can also be used for

triangulation of the data.

Next, the interaction effects are computed by multiplying the dummy “self-reference” by the motivator dummies. As a result, the variables “interaction_egoistic”, “interaction_altruism” and “interaction_social” are created.

Additionally, the interaction effects for the moderators are created. Privacyconcerns*self-reference and Indivualistic/collectivistic*motivators.

3.6.2 Correlations and multicollinearity

Before doing the regression, excessive correlations between the variables should be examined. Multicollinearity seems to be a problem once the moderators are added to the equation. The VIF scores rise far above the accepted level (VIF>10, see appendix 2.8) In appendix 2.8 you can also see that there are multiple items that have a high correlation (>,9). To interpret the interaction effects the moderator scales are mean-centered. This means that the mean of the scale is deducted from the scores on the scales. This means that for the variables Privacy concerns and Individualism are transformed in Mcprivacyconcerns and MCind and the new interaction variables are computed. The new model have normal VIF scores and no

correlations above ,90 (see appendix 2.9.1 full model)

3.6.3 Anova

(32)

3.6.4 Regression

The regression model is built by sequentially adding variables till the full model.

Y = b0 + b1Xegoistic + b2Xaltruistic + b3Xsocial + b4Xpersonalization + b5Xp*e + b6Xp*a + b7Xp*s + b10ind/col + b10mot*ind/col + b8privacy + b9p*privacy + b11age + b12gender + b13freqwrit + error (formula)

The model is built on the assumptions that the dummies of the different motivators are put in compared to a situation with no motivator. The model will be tested on two DV‟s. “Willingess to write” and the DV “slider Willingness to write”

Model 1: Y = b0 + b1Xegoistic + b2Xaltruistic + b3Xsocial + b4Xpersonalization + error

(formula)

Model 2: Y = b0 + b1Xegoistic + b2Xaltruistic + b3Xsocial + b4Xpersonalization + b5Xp*e

+ b6Xp*a + b7Xp*s + error (formula)

+ b6Xp*a + b7Xp*s + b10ind/col + b10mot*ind/col + error (formula)

Model 4 Y = b0 + b1Xegoistic + b2Xaltruistic + b3Xsocial + b4Xpersonalization + b5Xp*e +

b6Xp*a + b7Xp*s+ b8privacy + b9p*privacy + error (formula)

+ b6Xp*a + b7Xp*s + b11age + b12gender + b13freqwrit + error (formula)

(33)

(34)

Chapter 4 Results

In this chapter the results of the anova and regression will be presented and discussed. First, the tables of the means of the different conditions are shown. Next, the Anova and results will be interpreted. Then, the regression model will be sequentially be shown till the full model is presented. Finally, the results will be interpreted and discussed.

Table 4 Means (SD) No motivator M = 3,307 (SD = 1,607) Egoistic motivator M = 4,0261 (SD = 1,387) Altruistic motivator M = 3,7284 (SD=1,824) Social motivator M = 3,094 (SD = 1,670) No personalization M = 3,470 (SD = 1,665) M = 3,000 (SD = 1,497) M = 3,905 (SD = 1,390) M = 3,8452 (SD = 1,900) M = 3,048 (SD = 1672) Personalization M = 3,6078 (SD = 1,680) M = 3,560 (SD = 1,675) M = 4,174 (SD = 1,399) M = 3,603 (SD = 1.766) M = 3,147 (SD = 1,775)

Means DV = Willingness to write

Means (SD) _{No motivator} M = 29,56 (SD = 23,15) Egoistic motivator M = 41,16 (SD = 23,51) Altruistic motivator M = 37,75 (SD=25,42) Social motivator M = 29,88 (SD = 21,89) No personalization M = 33,71 (SD = 23,13) M = 26,167 (SD =20,66) M = 36,179 (SD = 22,02) M = 40,889 (SD = 25,99) M = 30,786 (SD = 21,98) Personalization M = 34,46 (SD = 24,75) M = 32,464 (SD = 25,09) M = 47,217 (SD = 24,30) M = 34,360 (SD = 24,86) M = 28,833 (SD = 22,20) Means DV = DV slider

4.1 Anova

4.1.1 DV “Willingness to write a review”

To compare the effect the different motivators had on the willingness to write a review a one-way between subjects ANOVA was conducted. There was a significant effect of IV

Motivators on DV willingness to write a review for the 4 conditions at the p<.05 level.

(F(3,205) =3,370,p = 0,19). A one-way ANOVA tells that the conditions differ, but it does not tell which conditions differ. Therefore, a post-hoc test is conducted to see which groups are statistically different. A Tukey HSD test was conducted and indicates that the egoistic motivator ( M = 4,026, SD = 1,387) and social motivator ( M = 3,094, SD = 1,70517) differ significantly. However, the egoistic and social motivator were the only ones significantly different. No motivator ( M = 3,307, SD = 1,607) did not significantly differ from egoistic, altruistic and social motivators. Also the Altruistic motivator ( M = 3,728, SD = 1,824) had insignificant difference with the no motivator, egoistic and social motivator. These results mean that the egoistic motivator has a positive significant difference on the willingness to write a review.

To compare the effect of personalization of the request on the willingness to write a review a one-way between subjects ANOVA was conducted. There was no significant effect of IV

(35)

no personalization) at the p<.05 level. (F(1,207) = ,343,p = ,553). This means that there was no significant difference between the non personalized condition ( M = 3,470, SD = 1,665) and the personalized condition. ( M = 3,608, SD = 1,680)

To compare the interaction effect of personalization of the request and the motivators on the willingness to write a review a two-way between subjects ANOVA was conducted. There was no significant effect of the interaction effect IV1 motivators to write a request and IV2

Personalization on DV willingness to write a review at the p<.05 level. (F(3,209) = ,539,p = ,656).

Figure 3

4.1.2 “DV Slider Willingness to write a review”

To compare the effect the different motivators had on the Slider willingness to write a review a one-way between subjects ANOVA was conducted. There was a significant effect of IV Motivators on DV willingness to write a review for the 4 conditions at the p<.05 level.

(36)

had insignificant difference with the no motivator, egoistic and social motivator. These results mean that the egoistic motivator has a positive significant difference on the willingness to write a review when compared to the no motivator and social motivator.

To compare the effect of personalization of the request on the willingness to write a review a one-way between subjects ANOVA was conducted. There was no significant effect of IV

Personalization on DV willingness to write a review for the 2 conditions (personalization vs no personalization) at the p<.10 level. (F(1,205) = ,276,p = ,600). This means that there was

no significant difference between the non personalized condition ( M = 33,71, SD = 23,13) and the personalized condition. ( M = 35,46, SD = 24,75)

To compare the interaction effect of personalization of the request and the motivators on the willingness to write a review a two-way between subjects ANOVA was conducted. There was no significant effect of the interaction effect IV1 motivators to write a request and IV2

Personalization on DV willingness to write a review at the p<.05 level. (F(3,207) = 1,461,p = ,226).

4.2.1 Regression model DV “Willingness to write a review”

Table 5 shows the regression table which covers the main effects, the interaction effects, moderators and control variables. The regression is comprised of six models. Model 1 is based on only both IV‟s. In model 2 the interaction effects of personalization on the effect of the moderators are added. Model 3, 4 and 5 sequentially includes the two moderators, privacy concerns and individualism/collectivism, and the control variables, age, gender and frequency of writing. Model 6 is the full model which incorporates all the variables and effects.

(37)

In Model 1, EV1 (Motivators) and EV2 (Personalization) explain 0.049 of the variance. The model is significant ( P = ,034), which seems to be mostly explained by the egoistic motivator which has a beta of ,735** ( p=,025). The other motivators and personalization have no significant beta‟s.

(38)

egoistic (β=,905,p<0.1) and altruistic motivator (β=,845,p<0.1) have a significant effect with pretty high betas. The beta of the egoistic motivator is higher than that of the altruistic beta.

In Model 3, EV1 (Motivators) and EV2 (Personalization), the interaction effects and individualism/collectivism explain 0,064 of the variance. The model itself becomes

insignificant. Individualism/collectivism has a positive beta though it is insignificant and the same pattern emerges for the rest of the variables as model 2. Egoistic and altruistic

motivators stay significant.

In Model 4, EV1 (Motivators) and EV2 (Personalization), the interaction effects and Privacy concerns explain 0,065 of the variance, the model is insignificant. Privacy concerns has a negative beta but is insignificant. Egoistic and altruistic motivators stay significant.

In Model 5, EV1 (Motivators) and EV2 (Personalization), the interaction effects and the control variables explain 0,171 of the variance. This is a significant increase when comparing to the other models. the model is highly significant ( p<,01). Frequency of writing has high beta ((β=1,166)and is highly significant (P < ,01). Age and gender have a negative beta but are insignificant. Egoistic (β=,928, p<0.05) and altruistic (β=,993, p<0.05) both increase in significance but the altruistic has a higher beta. Additionally the effect of personalization on the effect of the altruistic motivator on the willingness to write a review becomes significant (β=-1,,1044 p<0.1).

(39)

4.2.2 DV Slider “Willingness to write a review”

In Model 1, EV1 (Motivators) and EV2 (Personalization) explain 0.046 of the variance. The model is significant ( P = ,048), which seems to be mostly explained by the egoistic motivator which has a beta of 11,78** ( p<,05) and the altruistic motivator with a beta of 8,317 (p<0,1). The other motivators and personalization have no significant beta‟s (see table 6).

In model 2, EV1 (Motivators) and EV2 (Personalization) and the interaction effects explain 0.067 of the variance. The model is significant ( p = ,067)The interaction effects seem to all have a negative beta although all insignificant. Personalization and the social motivator both also have an insignificant effect on the willingness to write. The main difference is that the effect of personalization on the effect of the egoistic motivator is positive when comparing to the regression with the DV willingness to write. The altruistic motivator (β=14,72, p<0.05) has a significant effect with pretty high betas.

Model based on second DV (slider willingness to write a

review)

Model 1 Model 2 Model 3 Model

4 Model 5 Model 6 Main effects Constant Ev1 Motivators - Social - Egoistic - Altruistic Ev2 Personalization Interaction effects Personalization x Ego Perosnalization x Alt Personalization x Soc Moderators Privacy concern

Privacy concern x Selfreference Individualism

Ind x Motivators

Age Gender

How Egoistic, Altruistic and Social Motivators and Personalization Influence Online Reviewer Engagement