Benefits of Information Disclosure:
How Trust Influences the Effect of
Reciprocity and Relevant Advertising
on the Acceptance of Disclosing Information
Master Thesis
University of Groningen Faculty of Economics and Business
M.Sc. Marketing Management
20th of June 2016
Frist Supervisor: Prof. Dr. Peter C. Verhoef Second Supervisor: Frank T. Beke
Felix Böhm
Winschoterdiep 46, 9723AC Groningen Studentnumber: S2948427
Phone: +49 170 3428788 Email: felix.boehm@audivo.com
III Management Summary
Internet advertising gives online retailers the opportunity to develop certain advantages. An increased fit of advertisements leads to a rise of purchase intention, which concludes in higher profits. Personalized advertisements are only possible if the consumer is willing to disclose information. The consumer only accepts the disclosure, if the rewards outweigh the costs. Therefore, companies have to show the benefits of information disclosure to consumers. Two ways of demonstrating these benefits are the reciprocity and the relevance approach. Reciprocity lets people feel indebtedness, which online retailers can use to convince consumers to disclose information. Showing the consumer the benefits of relevant advertisements, like time savings because of reduced search costs, can be used to derive personal information as well. In order to give managers useful insights, this study investigates the effect of both benefits on the acceptance of information disclosure. As trust plays an important role in e-commerce and in terms of information disclosure, it was furthermore investigated which effect this component has on the relationship between the benefit arguments and disclosure. Eight different online scenarios were therefore implemented in a survey and presented to 208 respondents. The reciprocity and relevance argument were embedded in the website of either a high-trusted or low-trusted online retailer. The responses were analyzed with the help of two multiple regressions. None of the benefit approaches had a significant influence on the acceptance of information disclosure. Furthermore, trust had no effect on the relationship of the benefit arguments and disclosure. These findings show that companies should not use the argument of relevant advertising as main argument convincing consumers to disclose information. As the reciprocity approach has a significant positive effect in a free web service context (Schumann, Wangenheim, and Groene 2014), but not in an e-commerce context, this approach needs further distinctions.
Keywords: e-commerce, online advertising, reciprocity, relevant advertising, trust,
IV Preface
During my bachelor studies at the University of Erlangen-Nuremberg I invested all my time to learn as much as possible about the field of marketing. After I had the chance to do an internship in online marketing, I decided to fully focus on the communication between companies and their consumer audience in the online world. My bachelor thesis about consumer’s expectations on e-commerce gave me further insights into the possibilities of online marketing. I am glad that I had the opportunity at the University of Groningen to deepen my knowledge about this subject in combination with the important factor of privacy.
I would like to thank my supervisors Dr. Peter Verhoef and Frank Beke for their help and useful recommendations during my research. Furthermore I would like to thank my family and friends for their support.
V Table of Contents
Management Summary ... III Preface ... IV
1. Introduction ... 6
2. Literature Review ... 8
2.1 Relevance Argument ... 8
2.2 Reciprocal Argument ... 8
2.3 Reciprocal Argument versus Relevance Argument ... 9
2.4 Trust ... 11
3. Conceptual Model ... 13
4. Methodology and Data Collection Plan ... 13
5. Results ... 16
5.1 Descriptive Statistics ... 16
5.2 Disclosure Through Tracking ... 16
5.3 Model 1: Reciprocity versus Relevance ... 17
5.3 Model 2: Trust ... 19
6. Discussion ... 21
6.1 Theoretical Implications ... 21
6.2 Managerial Implications ... 23
6 1. Introduction
In 2015, a new peak of internet advertising was reached with $15 million spent by US-American companies in the third quarter of the year (Interactive Advertising Bureau 2015). The spending on digital display advertising will surpass search advertising spending within this year, with banner ads as the most important category member (eMarketer 2016). Personalization of advertisements became an important factor in the advertising industry and thus affects digital banner ads as well. 85% of companies in France, Germany, United Kingdom and the United States claim to use at least a basic form of personalization (Forrester Research 2014). “Targeted online advertising refers to any form of online advertising that is based on information the advertiser has about the advertising recipient, such as demographics, current or past browsing or purchase behavior, information from preference surveys and geographic information” (Schumann, Wangenheim, and Groene 2014, p. 59). The benefit of targeted advertising is the increased purchase intentions of consumers due to a higher fit of their preferences (Goldfarb and Tucker 2011). This, in turn, increases the potential revenues and becomes therefore an important benefit for profit-driven firms. In order to take advantage of this possible increase in profit, companies actively ask consumers to disclose personal information (Alpert et al. 2003). But asking consumers to provide an online retailer with personal information is delicate and depends on several components. The sensitivity of information (Mothersbaugh et al. 2012), the amount of information requested (Hui, Teo, and Lee 2007), the depth of the relationship between both parties (White 2004), and privacy concerns (Smith, Milberg, and Burke 1996) are only some of the issues that influence the consumer’s decision to disclose information. Consumers evaluate the risks and benefits of these issues before providing the online retailer with personal information, known as the privacy calculus (Laufer and Wolfe 1977) and only disclose information if benefits outweigh risks or if both are balanced (Culnan and Bies 2003). It is therefore in the interest of the company to show consumers explicit benefits when requesting the voluntary disclosure of information (Hui, Tan, and Goh 2006; Zimmer et al. 2010). This study focuses on the implementation of specific and relevant benefits when asking consumers for disclosing information because it is crucial for companies to convince consumers to provide personal information in order to increase revenues while at the same time adding value to the consumer, resulting in a win-win situation.
7
study on a free web service context, using these two benefit approaches, besides the control of data usage. They found that consumers are more likely to give websites, which are offering free services, access to their personal information when they are reminded of a reciprocal appeal. In their study, this implies the provision of free content, which was more effective than showing visitors benefits of more relevant, personalized advertisements. However, they also state this only works if the website provides a high utility and quality. For websites using website characteristics which support the relevance argument, the outcomes might be different (Schumann, Wangenheim, and Groene 2014). The results provide further evidence that the reciprocity argument does work in a public situation offline (Alpizar, Carlsson, and Johansson-Stenman 2008), as well as in an online situation, even though the important force of observation is absent.
Their results give companies important insights on how to increase the acceptance of information disclosure and targeted online advertising. However, these findings might only hold for companies offering free web services. For online concepts without free services, like e-commerce, the reasons for accepting the disclosure of personal information might be different. In order to investigate this, the study follows the design and measurement of Schumann and colleagues’ first study (Study 1) to test the acceptance of online advertising in an e-commerce environment. According to Premazzi et al. (2010), the involvement with a service category is an antecedent for the willingness to share information, which supports the assumption that free web services might influence information disclosure in a different valence or strength than e-commerce interactions. A transaction between two parties includes money, which might result in a different involvement of consumers compared to websites without a monetary transaction (like free web services). Considering this possible differences, the following question arises:
Are the effects of the reciprocity argument and the relevance argument in an e-commerce context the same as in a free web-service context?
8
Choudhury, and Kacmar 2002, p. 334). As both, trust (Gefen, Karahanna, and Straub 2003) and the two benefit approaches (Schumann, Wangenheim, and Groene 2014) influence the disclosure of information, it is tested if trust serves as a moderator on the effect of the benefits and the disclosure. Since trust is especially important in an e-commerce context because of a limited web interface (Reichheld and Schefter 2000), there emerges the question of its importance when using benefit arguments to increase the acceptance of information disclosure:
Does trust influence the acceptance of information disclosure when using the reciprocity argument and the relevance argument in an e-commerce context?
2. Literature Review
2.1 Relevance Argument
Relevant advertising is defined as interesting and useful advertisement of which consumers think it is worth to pay attention to (Laczniak and Muehling 1993; Schumann, Wangenheim, and Groene 2014). The probability of paying attention to an advertisement depends on its fit of the consumer’s needs and wants. If the consumer is paying attention to an advertisement, he has to evaluate if it is worth reacting to the advertisement. According to the social exchange theory people evaluate exchanges (Thibaut and Kelley 1959), in terms of costs and rewards. Relevant advertising is thus a self-benefit appeal that focuses on the benefits of prosocial actions to the individual self (Nolan et al. 2008; White and Peloza 2009; White and Simpson 2013). This implies that consumers accept the exchange if the rewards are bigger than the costs. Concerning the disclosure of information this leads to the decision of consumers whether they are willing to disclose information, known as the privacy calculus (Laufer and Wolfe 1977). Chellappa and Sin found that “[…] a consumer is willing to share preference information in exchange for apparent benefits, such as convenience, from using personalized products or services” (2005, p. 186). This convenience could, for example, be a reduced search cost, which saves the consumer’s time (Bhatnagar, Misra, and Rao 2000) and was therefore used in this study as a benefit to convince the consumer to disclose information.
2.2 Reciprocal Argument
9
friend, this favor is likely to be returned since this behavior can be seen as social norm, which people automatically obey. But there are more reasons why people act reciprocally, for example to be regarded positively by others or to have a positive view of themselves (Alpizar, Carlsson, and Johansson-Stenman 2008). The feeling of indebtedness is the underlying phenomenon of returning a favor (Greenberg 1980), which holds not only for an interaction between two people, but also between a person and a firm or other organization. Falk (2007) showed, that people are more likely to donate money when charity organizations take advantage of the reciprocal approach by including small gifts when asking for a donation.
In this study the visitor of an online retailer website is reminded about the money he saves when purchasing products due to low prices, in order to persuade the visitor to disclose personal information with the reciprocity appeal afterwards. The retailer admits that low prices are enabled by reducing costs, which are achieved by the use of advertisements. This fact is stated by the online retailer to the consumer when visiting the retailer’s website and thus reminds the visitor of his “debt” which requires to return the favor by disclosing personal information. The reciprocity argument involves therefore not direct information disclosure, but rather uses the low prices argument as an excuse in order to obtain personal information through mentioning the use of advertisements. Low prices result in monetary savings for the consumer, which include price reductions such as discounts in any form (Chandon, Wansink, and Laurent 2000; Schindler 1998). Consumer’s choice of alternatives is significantly influenced by monetary incentives (e.g. price) when forming preferences (McFadden 2001). Thus, saving money should be a motivating factor behind consumer disclosure in an online context as well (Hui, Tan, and Goh 2006). The monetary savings used here, are intentionally not including any compensation for information disclosure, such as vouchers or gifts, because research showed that monetary compensations decrease the consumer’s willingness to disclose information when the reputation of the company or the trust towards the company is high (Premazzi et al. 2010; Xie, Teo, and Wan 2006). Instead, price reductions referring to low prices are used to signalize lower costs when comparing prices to competitors.
2.3 Reciprocal Argument versus Relevance Argument
10
(Alreck and Settle 2007). Thus, relevant advertising, enabled by tracking, is a positive attribute for the consumer. However, tracking is a form of information disclosure and disclosing information depends, as mentioned in the introduction above, on different components like the amount of information requested (Hui, Teo, and Lee 2007). Some of these components might be a negative attribute for the consumer, which might influence the response of the relevance argument negatively as well.
The reciprocity argument is especially high if it takes place in public rather than in private situations (Alpizar, Carlsson, and Johansson-Stenman 2008). Such public situations include the use of the internet as well, even though the possibility of being observed by others is not present. However, as the online world is thus not public in a traditional definition, the reciprocity argument might be less effective. On the other hand, Greenberg, Block, and Silverman (1971) found that cooperative behavior is determined by the extent to which a person already benefited from a counterpart. Consumers who visited the website of the online retailer before might have already benefited from the low prices offered (reciprocity argument) but not from the personalized and thus better advertising in the future (relevance argument), which implies that the reciprocity argument in this study might be more effective.
Schumann, Wagenheim, and Groene (2014) argued that the reciprocity argument is more effective because it uses the free service of websites to convince the user that is only fair if they provide information in exchange. As they have shown, this holds in a free web service context, but it is not assumed to hold in an e-commerce context, simply because the service an online retailer provides is not for free. Consumers pay money in exchange for products and consequently there is no reason for the company to ask for a favor. The “low price” argument used in this study is therefore expected to be evaluated as a weak argument by the consumer and thus not convincing him to disclose personal information. As described above, the reciprocal approach of low prices has no direct link to information disclosure, which weakens the argument additionally. The relevance argument is therefore assumed to dominate the reciprocity argument in an e-commerce context.
In combination, the reasons for the relevance argument outweigh the reciprocity argument in an e-commerce context. Compared to a neutral condition, which includes neither the reciprocity nor the relevance argument, it is expected that:
H1a: The reciprocity argument has a positive effect on information disclosure.
11
H1c: The relevance argument has a stronger effect on information disclosure than the reciprocity argument.
2.4 Trust
Customers engage in trust-related behaviors with online retailer such as sharing information or making purchases because trust reduces uncertainty and risk (McKnight, Choudhury, and Kacmar 2002). Especially in an e-commerce context trust is crucial because of the limited web interface (e.g. no face-to-face interaction, no touch and feel of products), which makes it hard for the consumers to judge the retailer’s trustworthiness (Reichheld and Schefter 2000). Referring to existing research, no uniform definition of trust does exist, even though “confident expectations” and “willingness to be vulnerable” are components that constantly define trust (Rousseau et al. 1998). Mayer, Davis, and Schoorman define trust as “[…] the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party” (1995, p. 712).
Trust is mostly viewed as a multidimensional construct (Gefen, Karahanna, and Straub 2003; Mayer, Davis, and Schoorman 1995; McKnight, Choudhury, and Kacmar 2002; Rousseau et al. 1998). Even though there are different conceptualizations of trust in the literature, this study focuses on the concept of specific beliefs because actual behavioral intentions (e.g. disclosure of information) are a proxy of trust (Gefen, Karahanna, and Straub 2003). McKnight, Choudhury, and Kacmar (2002) provide a broad overview of the specific beliefs (or trusting beliefs) researched in literature and emphasize the following three most important beliefs, which are also proposed by Mayer, Davis, and Schoorman (1995): The first is competence, which is the ability of the online retailer (trustee) to do what the consumer (trustor) needs, for instance showing the consumer personalized advertisements due to the consumer’s disclosure of information. The second is benevolence, defined as the trustee’s motivation to act in the trustor’s interest, as showing the customer the best recommendations possible. The third component is integrity, the trustee’s honesty and promise keeping (i.e. not to give the disclosed information to third parties) (McKnight, Choudhury, and Kacmar 2002).
12
approaches nor information disclosure, it shows that as a moderator, trust plays an important role in e-commerce. Trust is, together with privacy concerns, the main component of consumer’s decisions to disclose information (Gefen, Karahanna, and Straub 2003; Pavlou and Gefen 2004). Gefen, Karahanna, and Straub (2003) found a main effect of trusting beliefs on a consumer’s intention to disclose information. More specifically, trust in a specific online retailer increases the willingness of a consumer to disclose information to that firm (Premazzi et al. 2010). It is therefore assumed that trust has a positive moderating effect on the relationship between both benefit approaches and the acceptance of information disclosure. Consumers should have a higher acceptance for information disclosure if a retailer is trustworthy. On the other hand, the acceptance should be lower if the consumers holds a low level of trust towards a retailer. This assumption should hold for the reciprocity argument, because people are more likely to act reciprocally if they trust their counterpart (Berg, Dickhaut, and McCabe 1995). As shown in the literature review of relevance, the argument is closely linked to outcomes of the privacy calculus, which depend on the trust component (Dinev and Hart 2006). It is therefore predicted that both benefit approaches are influenced by trust equally. Thus:
H2a: High Trust in the online retailer strengthens the effect of the reciprocity argument on information disclosure.
13 3. Conceptual Model
Figure 1 shows the conceptual model and the hypotheses, which were explained in the previous section.
4. Methodology and Data Collection Plan
To test the two hypotheses, a study was conducted with the acceptance of information disclosure as the dependent variable and the benefit approaches as well as trust as independent variable. The disclosure of information was presented to the respondents in the form of behavioral targeting. This technique refers to the consumer’s search and browsing behavior as gathered by cookies, which are placed in his browser (Schumann, Wangenheim, and Groene 2014). The participants were asked to rate the extent of their acceptance concerning the use of the data gathered by these cookies. An online survey was used to collect information about the different responses of the participants confronted with the reciprocity or relevance argument, as well as trust. In order to compare results with the paper of Schumann and his colleagues, a similar study design was used. The online study consisted of a 2 (relevance vs. neutral) x 2 (reciprocity vs. neutral) x 2 (high vs. low trust) between-subjects experimental design. Table 1 gives an overview of all eight different conditions. In order to check the manipulations, reciprocity, relevance, and trust were measured in the survey as well. Reciprocity was measured with the help of distributive justice, which refers in the context of information disclosure to the perceived appropriate value of a return (Wirtz and Lwin 2009). Relevance was measured with questions relating to the anticipation of relevance (Laczniak and Muehling 1993; Schumann,
Control Variables:
• Internet affinity
• General concern of privacy • General attitude towards advertising
Reciprocity Argument Relevance Argument Neutral Acceptance of Information Disclosure Trust • Competence • Benevolence • Integrity
14
Wangenheim, and Groene 2014). Trust was measured with the three trusting beliefs benevolence, integrity and competence (McKnight, Choudhury, and Kacmar 2002).
Table 1: Overview of the conditions
Schumann and colleagues included control of information in their study as well, which refers to the consumer’s possibility to see, edit or delete the disclosed information. However, this variable was not implemented in the research design. The Directive 2009/136/EC of the European Parliament (the “EU Cookie Law”) allows website owners to use behavioral tracking with the help of cookies only if “[…] the subscriber or user concerned has given his or her consent, having been provided with clear and comprehensive information […] about the purposes of the processing” (European Union 2009, p. 30). This consent give consumers only a minimal control of their disclosed information, because the only way to refuse behavioral tracking is to not use the website at all. Hence, website owners just inform visitors on their website about cookies, but do not further explain how to edit or delete personal information, as added by Schumann, Wangenheim, and Groene. In order to conduct a more realistic scenario, the control condition was not considered in the online survey.
In order to use an appropriate high and low trust variation a pretest was conducted. 18 participants rated the trustworthiness (McKnight, Choudhury, and Kacmar 2002) of five out of ten online retailers, which were all presented with the logo and a short description about each company to the respondents. All retailers belonged to the 20 largest online retailers in the world by revenue in 2014 (Statista 2016). The highest (Zalando; M = 5.78; SD = 0.577) and lowest (JD.com; M = 3.11; SD = 1.18) trust-rated online retailers were used for the online survey as conditions (Appendix 1).
The 208 participants of this study were randomly assigned to one of the eight experimental conditions. Most of the participants were students, living mainly in the Netherlands and Germany. In the beginning of the study a screenshot was exposed to the participants which showed the English homepage of either the high trusted online retailer or the low trusted online retailer. In a second step, the participants were confronted with the experimental treatment in the form of a flash layer which overlapped the homepage (Appendix 2). The message started
Reciprocity Relevance Reciprocity &
Relevance Neutral
High Trust High Trust, Reciprocity High Trust, Relevance High Trust, Reciprocity & Relevance High Trust, Neutral
Low Trust Low Trust,
15
with a short greeting (“Dear visitor …”) followed by one combination of the four experimental benefit conditions shown in Table 2. These short texts were adapted from Schumann, Wagenknecht and Groene (2014) and all ended with the sentence “We would like to hear your opinion on this. Please click ‘continue’”. The third step consisted of a survey, asking the participants about their attitudes towards the willingness to disclose information, as well as manipulation check, control, and demographic questions.
Table 2: Description of Treatments
Appendix 3 lists the constructs measured in this study, the measurement items used to measure these constructs, as well as the sources from where these items are derived. These constructs include the disclosure through tracking, measurements for the manipulation checks (relevance anticipation, distributive justice, and trusting beliefs), as well as the control variables (internet affinity, general concern for privacy, and attitude towards advertising). All items were measured on a 7-point Likert scale ranging from “strongly disagree” (1 point) to “strongly agree” (7 points). Besides the measurement of trust all items are equal to Schumann et al. (2014), i.e. the control variables. Demographic questions included gender, age, highest obtained degree and income, which were measured with the help of categories.
H1 was tested with an ANOVA and two multiple linear regressions. The ANOVA was used to show differences in the means of the four benefit conditions. The first multiple regression supported the findings of the ANOVA. The second regression was used to check if the benefit approaches were manipulated successfully. The dependent variable in both regressions was the acceptance of information disclosure. A potential indirect effect of the benefits conditions on information disclosure through the measured variables of ‘Distributive Justice’ and ‘Relevance Anticipation’ was analyzed with two additional multiple regressions. To test an indirect effect of reciprocity, distributive justice was used as dependent variable. The anticipation of relevance was used as dependent variable in order to test an indirect effect of relevance (Model 1).
Constructs Treatment Neutral Condition
Reciprocity We offer you products at the lowest
price. This is possible because we show you advertisements. Effective advertisements save us money. Only this way we can keep our prices low.
Thank you for visiting our website. We offer you products at the lowest price. Besides, we display advertisements to you.
Relevance We would like to show you
advertisements you are interested in. This way you only see relevant advertisements which are interesting for you and which reduce your time searching for products.
We would like to show you
16
H2 was tested with an ANOVA and two multiple regressions as well. The ANOVA tested if any difference in the means of the eight conditions occurred. A multiple regression with all eight conditions was used to confirm the findings of the ANOVA. A second multiple regression was used to check if there has been a successful manipulation of trust. In both multiple regressions acceptance of information disclosure was the dependent variable (Model 2).
5. Results
5.1 Descriptive Statistics
223 voluntary participants finished the survey. 15 of them were excluded from the analysis because they failed to answer the control question, which made sure that the participants were actually reading the questions. Each of the eight conditions consisted therefore of exactly of 26 respondents. The gender ratio was almost equal, with 57.7% of male and 42.3% of female representatives. 96.6% of the respondents were between 15 and 34 years old, which is realistic considering that I mainly asked friends and colleagues to participate.
Roughly the same distribution can be seen when looking at the highest degree obtained and the monthly net income. 79.4% of all respondents had a bachelor’s or master’s degree. 64.4% of the participants had a monthly net income of less than 1000€. Again, considering that most of the respondents were students, this is not a surprise. The descriptive statistics are summarized in Appendix 4. The participants of the survey rated themselves as internet affine (M = 5.79; SD = 0.89) and have concerns about their privacy when using the internet (M = 5.50; SD = 1.24). In contrast to that, their attitude towards advertising (M = 3.99; SD = 1.58) was two-sided. 77 respondents (37.00%) “agreed” or “strongly agreed” to the statement that advertising is favorable, whereas 72 people (34.62%) “disagreed” or even “strongly disagreed” with the statement that advertising is a good thing.
5.2 Disclosure Through Tracking
17
Figure 2: Frequencies Disclosure Through Tracking
5.3 Model 1: Reciprocity versus Relevance
The use of the different conditions and the 7-point Likert scale, with which the variables were measured made it not necessary to exclude outliers. The two variables ‘Distributive Justice’ (manipulation check for reciprocity) and ‘Relevance Anticipation’ (manipulation check for relevance) were used in the second multiple regression. There was no heteroscedasticity present in the regression for manipulation check (Appendix 5a). The normal probability plots of the independent variables and the dependent variable show a normal distribution of errors (Appendix 5b). The Lavene Statistic was not significant in the ANOVA (p > 0.1), thus a homogeneity of variances occurred (Appendix 6a). Thereby assumptions of ANOVA and linear regression are met and the results are proved to be valid.
In order to test H1, a one-way ANOVA was used to show if there is a significant difference in the means of the conditions (reciprocity, relevance, reciprocity and relevance, and neutral condition). As shown in Appendix 6a, there are no significant differences in the group means (F(3,204) = 0.712, p > 0.1). Because of this insignificance there was no ad-hoc test conducted. This insight can be confirmed when looking at the distribution of the conditions on the 7-point Likert scale (Figure 3). Each of the four conditions seems to be relatively similar distributed as the average acceptance of disclosure through tracking (Figure 2). There is a significant peak at ‘Strongly disagree’ as well as ‘Disagree’ in each of the four conditions.
However, these findings give only limited insights since they only show whether an effect occurs or not. In order to assess the valence and directions of the four conditions a multiple regression was used to get detailed findings about the relationship and valence of the four conditions and the acceptance of disclosure through tracking. The model was highly significant (F(6,201) = 12.890, p < 0.05) and showed that the conditions explain 27.8% of the variance of
0 5 10 15 20 25 30 35 40 45 Strongly disagree Disagree Somewhat disagree Neither agree nor disagree Somewhat agree Agree Strongly agree Fre q u e n c y
18
the dependent variable (Appendix 6b). Against expectations the reciprocity argument and the relevance argument had no significant influence on information disclosure (p > 0.1), which confirms the first findings of the ANOVA and the illustration.
Figure 3: Frequencies Acceptance of Disclosure Through Tracking in each benefit condition
A second multiple regression was conducted in order to test whether both benefit approaches were successfully manipulated. This regression used ‘Distributive Justice’ and ‘Relevance Anticipation’, which were both measured in the survey on a 7-point Likert scale (from 1 = strongly disagree to 7 = strongly agree) as independent variables. Their interaction effects with the trust conditions were included as well to check for manipulation. The dependent variable was the acceptance of information disclosure. As it can be seen in Appendix 6c, the model was significant and the independent variables explained 46.4% of the variance on the dependent variable (F(8,199) = 21.493, p > 0.05). The main effect of ‘Distributive Justice’ and ‘Relevance Anticipation’ was statistically significant (p < 0.05), while their interaction effect had no significant influence (p > 0.1). This shows that the benefit approaches were successfully manipulated, but they had against expectations no influence on the acceptance of information disclosure.
The manipulation check of reciprocity and relevance showed additionally that the there is no mediation effect of the two benefit approaches on acceptance of information disclosure through the two variables ‘Distributive Justice’ and ‘Relevance Anticipation’ (Figure 4). This insignificance shows therefore that there is no indirect effect of reciprocity and relevance on information disclosure. However, there might be a significant effect of the benefit manipulations on these two manipulation check variables, which represents the first part of the mediation. In order to test this, a third and fourth multiple regression were carried out. For the third regression ‘Distributive Justice’ was used as dependent variable and the four benefit
0 1 2 3 4 5 6 7 8 9 10 11 12 13 Strongly disagree Disagree Somewhat disagree Neither agree nor disagree Somewhat agree Agree Strongly agree Fre q u e n c y Treatment Conditions:
Acceptance of Disclosure Through Tracking
19
conditions were used as independent variable. The fourth regression used ‘Relevance Anticipation’ as dependent variable and the four benefit conditions as independent variables. Since in both cases the independent variables are only conditions and no measurements, all assumptions of a multiple regression are given. As can be seen in Appendix 6d, reciprocity condition had no significant effect on ‘Distributive Justice’. Moreover, the relevance condition had no significant effect on ‘Relevance Anticipation’ as well (Appendix 6e). Thus, the benefit approaches do not affect their manipulation checks and a mediation effect is not even partly existent.
Figure 4: Mediation
Overall, the reciprocity as well as the relevance approach had no influence on the acceptance of information disclosure and therefore both parts of Hypothesis 1, H1a: The reciprocity
argument has a positive effect on information disclosure as well as H1b: The relevance argument has a positive effect on information disclosure can be rejected. Furthermore, the third
part of Hypothesis 1, H1c: The relevance argument has a stronger effect on information
disclosure than the reciprocity argument be rejected as well, because both arguments are not
statistically significant. 5.3 Model 2: Trust
No outliers had to be deleted because it were only used the different conditions as well as a Likert scale to measure trust. The variable trust (trusting beliefs) was used as a manipulation check in the second multiple regression. As it can be seen in Appendix 7a there is no heteroscedasticity present. The normal probability plots of ‘Disclosure Through Tracking’ and
20
the measurement of ‘Trust’ show a normal distribution of errors (Appendix 7b). Because the VIF-scores were very high in the first run, but none of the correlations where higher than 0.8, trust was mean-centered. The Lavene statistic in the ANOVA was not significant (p > 0.1), which means that there is a homogeneity of variances (Appendix 8a). As all assumptions for an ANOVA and a linear regression were met, the analysis provides valid results.
Similar to the first analysis, a one-way ANOVA was chosen to give first insights about the influence of trust on the relationship of the four conditions and the acceptance of information disclosure. The means of all eight conditions (four treatments x high/low trust) were therefore compared. The ANOVA shows that none of the conditions has a statistically different mean (F(7,200) = 0.399; p > 0.1) (Appendix 8a). Again, these findings become clearer when looking at the distribution of the conditions on the Likert-scale (Figure 5). None of the condition in the illustration seems to be significantly different and thus there seems to be no moderation effect of trust.
As an ANOVA analysis does not provide details about the relationship, a multiple regression was used to show the moderating effect of trust. The eight conditions and the control variables were used as independent variables, whereas the dependent variable was the acceptance of information disclosure through tracking. The moderation of trust in this regression was measured with the responses of the eight different conditions. The regression was statistically significant (F(10,197) = 7.825, p < 0.05) and the independent variables explained 28.4% of the variance in the dependent variable (Appendix 8b). Furthermore, the analysis revealed that none of the conditions had a statistical significant influence on the dependent variable.
For a second regression the measurement of trust (trusting beliefs; measured on a 7-point Likert scale), instead of the trust conditions, was used in order to check if there has been a successful manipulation of trust. The interaction effect between the trust measurement and the four benefit conditions shows if the condition trust was successful manipulated. The model was significant (F(10.197), p < 0.05) and the independent variables explained 37.8% of the variance in the dependent variable (Appendix 8c). None of the interaction effects of trust and the benefit approaches had a statistical significant influence, whereas trust was highly statistically significant (p < 0.05). The outcome of these regressions shows that trust was indeed successfully manipulated, but had no effect on the acceptance of disclosure through tracking. The findings of the ANOVA and the multiple regression shows that the moderating effect of trust does therefore not exist and both hypotheses H2a: High Trust in the online retailer
21
Trust in the online retailer strengthens the effect of the relevance argument on information disclosure have to be rejected.
Figure 5: Distribution of Treatment and Trust Conditions on Information Disclosure
6. Discussion
6.1 Theoretical Implications
As shown at the beginning of the results section, the acceptance of information disclosure through tracking was low in general. People don’t want to provide retailers with information, even though they were confronted with benefits of disclosure. Table 3 shows an overview of the hypotheses and the final results.
Hypotheses Result
H1a: The reciprocity argument has a positive effect on information disclosure rejected
H1b: The relevance argument has a positive effect on information disclosure rejected
H1c: The relevance argument has a stronger effect on information disclosure than
the reciprocity argument rejected
H2a: High Trust in the online retailer strengthens the effect of the reciprocity
argument on information disclosure rejected
H2b: High Trust in the online retailer strengthens the effect of the relevance
argument on information disclosure rejected
Table 3: Summary of hypotheses testing
Based on the findings, in the context of e-commerce the acceptance of behavioral tracking is neither directly nor indirectly influenced by the argument of relevant advertisements. As shown
0 1 2 3 4 5 6 7 8 9 Strongly disagree Disagree Somewhat disagree Neither agree nor disagree Somewhat agree Agree Strongly agree Fre q u e n c y
Treatment and Trust Conditions: Acceptance of Information Disclosure
High Trust, Reciprocity&Relevance Low Trust, Reciprocity&Relevance
High Trust, Reciprocity Low Trust, Reciprocity
High Trust, Relevance Low Trust, Relevance
22
in the study, privacy concerns are rated high by the respondents (M = 5.50; SD = 1.24) and negatively influence the acceptance of tracking (βPrivacy Concerns = -0,441; Appendix 5b). These concerns are seen by the consumer as costs which cannot be outweighed by the benefit of relevant advertising and its advantage of a reduced search time. This might be the case because the perceived benefits of the privacy calculus are lowered by the period of time until the benefit actually occur to the consumer. Targeted advertisements will be approached in the future and have therefore less value at the point of time in which the consumers discloses personal information. Consumers reduce the value of a reward, to which they will have access at a later point in time. This is known as time or hyperbolic discounting (Zauberman et al. 2009) and influences the decision making about the acceptance of disclosure through tracking. The consumer has not benefitted from its counterpart yet (Greenberg, Block, and Silverman 1971) and decides therefore not to accept the disclosure. This result gives a good explanation why the relevance argument does not have an influence in an e-commerce context.
23
content-based and free web services as well, as proven by Schumann and colleagues (2014), even though the internet is not a public sphere in a traditional, offline way. Although this argument is less strong because of the findings of Schumann, the public effect might not be present in an e-commerce context, because the consumer perceives it as a private transaction-based situation between him and the retailer.
Considering the benefit treatments used in the study there might be another valid reason why both approaches are insignificant. The reciprocity and the relevance treatment mention only the use of advertisements, but not the need to disclose information. Even though respondents rated themselves as internet affine (M = 5.79; SD = 0.89), this might lead to a missing link between showing advertisements and disclosing information by the consumers, which weakens the acceptance of information disclosure in both approaches. Especially when looking at participant’s response to their attitude towards advertising (M = 3.99; SD = 1.58) this missing link might be influential. Participants who have a negative view of advertising might have responded differently in the survey when there would have been included a clear link to information disclosure. These participants might connect other benefits to information disclosure, e.g. a faster checkout or better recommendations.
High trust as well as low trust have no significant influence on the relationship of the benefit approaches and the acceptance of disclosure through tracking. However, this finding does not dispute the importance of trust in an online context because, as shown in the analysis, trust itself had a significant main effect on the acceptance of information disclosure. Consequently, the findings do not weaken the well proven role of trust in an e-commerce context (Reichheld and Schefter 2000) and online information disclosure (Gefen, Karahanna, and Straub 2003). The relevance and reciprocity approach might have influenced the participants’ response to the acceptance in greater way than trust. Trust was, compared to the two approaches, not explicitly shown to them in form of words, only shown visually. It approached the respondents indirectly through the screenshot of the retailer’s website and thus might have influenced people less when answering the questions.
6.2 Managerial Implications
24
collection of data but leads to a decrease in acceptance of tracking and use of personal information.
The usage of a reciprocity approach, on the other hand, depends on the business model and the type of the company. The return of a favor is effective for companies which offer a free web service, but not for online retailer. As there is no favor to ask for, due to an exchange of products and money, consumers do not accept information disclosure. It remains unclear for which forms of companies this approach holds and for which it does not. A possible distinction between the effective use of reciprocity as a consumer-convincing tool, may lie in whether the company is offering goods and services. However, considering that free web service websites can be profit-driven (e.g. Facebook) or nonprofit-profit-driven (e.g. Wikipedia) this distinction needs further research. Managers should therefore not only rely on research paper, but evaluate both approaches with the help of, for example, internal A/B tests in order to achieve the highest possible effectiveness in consumer data collection.
7. Limitations and Future Research
25
advertisements. This might have biased the respondents in the reciprocal condition and might have led to different findings than the relevance argument.
V References
Alpert, Sherman R., Karat, J., Karat, C., Brodie, C., and Vergo, J. G. (2003), "User Attitudes regarding a User-Adaptive eCommerce Web Site," User Modeling and User-Adapted
Interaction: The Journal of Personalization Research, 13 (4), 373-396.
Alpizar, Francisco, Carlsson, F., and Johansson-Stenman, O. (2008), "Anonymity,
Reciprocity, and Conformity: Evidence from Voluntary Contributions to a National Park in Costa Rica," Journal of Public Economics, 92 (5), 1047-1060.
Alreck, P. L. and Settle, R. B. (2007), "Consumer Reactions to Online Behavioural Tracking and Targeting," Journal of Database Marketing and Customer Strategy Management, 15 (1), 11-23.
Anderson, Rolph E. and Srinivasan, S. S. (2003), "E-Satisfaction and E-Loyalty: A Contingency Framework," Psychology & Marketing, 20 (2), 123-138.
Berg, Dickhaut, J., and McCabe, J. (1995), "Trust, Reciprocity, and Social History," Games
and Economic Behavior, 10 (1), 122-142.
Bhatnagar, A., Misra, S., and Rao, H. R. (2000), "On Risk, Convenience, and Internet Shopping Behavior," Communications of the ACM, 43 (11), 98-106.
Carter, Michelle, Wright, R., Thatcher, J. B., and Klein, R. (2014), "Understanding Online Customers’ Ties to Merchants: The Moderating Influence of Trust on the Relationship between Switching Costs and E-Loyalty," European Journal of Information Systems, 23 (2), 185-204.
Chandon, Pierre, Wansink, B., and Laurent, G. (2000), "A Benefit Congruency Framework of Sales Promotion Effectiveness," Journal of Marketing, 64 (4), 65-81.
Chellappa, Ramnath K. and Sin, R. G. (2005), "Personalization Versus Privacy: An Empirical Examination of the Online Consumer’s Dilemma," Information Technology and
Management, 6 (2-3), 181-202.
Culnan, Mary J. and Bies, R. J. (2003), "Consumer Privacy: Balancing Economic and Justice Considerations," Journal of Social Issues, 59 (2), 323-342.
Dinev, Tamara and Hart, P. (2006), "An Extended Privacy Calculus Model for E-Commerce Transactions," Information Systems Research, 17 (1), 61-80.
eMarketer (2016), "US Digital Display Ad Spending to Surpass Search Ad Spending in 2016," [available at http://www.emarketer.com/Article/US-Digital-Display-Ad-Spending-Surpass-Search-Ad-Spending-2016/1013442].
European Union (2009), "Directive 2009/136/EC of the European Parliament and the Council," [available at
VI
Forrester Research (2014), "Unlock the Promise of Customer Data - how Comprehensive Profiles Hold the Key to Effective Personalization," [available at
http://www1.janrain.com/rs/janrain/images/Industry-Research-Unlock-Customer-Data.pdf].
Gefen, David, Karahanna, E., and Straub, D. W. (2003), "Trust and TAM in Online Shopping: An Integrated Model," MIS Quarterly, 27 (1), 51-90.
Goldfarb, A. and Tucker, C. (2011), "Online Display Advertising: Targeting and Obtrusiveness," Marketing Science, 30 (3), 389-404.
Gouldner, Alvin W. (1960), "The Norm of Reciprocity: A Preliminary Statement," American
Sociological Review, 25 (2), 161-178.
Greenberg, Martin S. (1980), "A Theory of Indebtedness," in Kenneth J. Gergen, Martin S. Greenberg, and Richard H. Willis, eds. Boston, MA: Springer US, 3-26.
Greenberg, Martin S., Block, M. W., and Silverman, M. A. (1971), "Determinants of Helping Behavior: Person's Rewards Versus Other's Costs," Journal of Personality, 39 (1), 79-93. Hui, Kai-Lung, Tan, B., and Goh, C. (2006), "Online Information Disclosure: Motivators and
Measurements," ACM Transactions on Internet Technology (TOIT), 6 (4), 415-441. Hui, Kai-Lung, Teo, H. H., and Lee, S. T. (2007), "The Value of Privacy Assurance: An
Exploratory Field Experiment," MIS Quarterly, 31 (1), 19-33.
Interactive Advertising Bureau (2015), "Q3 2015 Digital Ad Revenues Climb to $15 Billion, Marking all-Time Quarterly High," [available at
http://www.iab.com/news/q3adrevenue/].
Kallgren, Carl A., Reno, R. R., and Cialdini, R. B. (2000), "A Focus Theory of Normative Conduct: When Norms do and do Not Affect Behavior," Personality and Social
Psychology Bulletin, 26 (8), 1002-1012.
Laczniak, Russell N. and Muehling, D. D. (1993), "The Relationship between Experimental Manipulations and Tests of Theory in an Advertising Message Involvement Context,"
Journal of Advertising, 22 (3), 59-74.
Laufer, Robert S. and Wolfe, M. (1977), "Privacy as a Concept and a Social Issue: A Multidimensional Developmental Theory," Journal of Social Issues, 33 (3), 22-42. Malhotra, Naresh K., Kim, S. S., and Agarwal, J. (2004), "Internet Users' Information Privacy
Concerns (IUIPC): The Construct, the Scale, and a Causal Model," Information Systems
Research, 15 (4), 336-355.
Mayer, Roger C., Davis, J. H., and Schoorman, F. D. (1995), "An Integrative Model of Organizational Trust," The Academy of Management Review, 20 (3), 709-734. McFadden, Daniel (2001), "Economic Choices," The American Economic Review, 91 (3),
VII
McKnight, D. H., Choudhury, V., and Kacmar, C. (2002), "Developing and Validating Trust Measures for E-Commerce: An Integrative Typology," Information Systems Research, 13 (3), 334-359.
Milberg, Sandra J., Smith, H. J., and Burke, S. J. (2000), "Information Privacy: Corporate Management and National Regulation," Organization Science, 11 (1), 35-57.
Mothersbaugh, D. L., Beatty, S. E., Foxx II, W. K., and Wang S. (2012), "Disclosure Antecedents in an Online Service Context: The Role of Sensitivity of Information,"
Journal of Service Research, 15 (1), 76-98.
Neelamegham, R. and Jain, D. (1999), "Consumer Choice Process for Experience Goods: An Econometric Model and Analysis," Journal of Marketing Research, 36 (3), 373-386. Nolan, J. M., Schultz, P. W., Cialdini, R. B., Goldstein, N. J., and Griskevicius, V. (2008),
"Normative Social Influence is Underdetected," Personality & Social Psychology
Bulletin, 34 (7), 913-923.
Norberg, Patricia A. and Horne, D. R. (2007), "Privacy Attitudes and Privacy-Related Behavior," Psychology & Marketing, 24 (10), 829-847.
Pavlou, Paul A. and Gefen, D. (2004), "Building Effective Online Marketplaces with Institution-Based Trust," Information Systems Research, 15 (1), 37-59.
Pollay, R. W. and Mittal, B. (1993), "Here's the Beef: Factors, Determinants, and Segments in Consumer Criticism of Advertising," Journal of Marketing, 57 (3), 99-114.
Premazzi, Katia, Castaldo, S., Grosso, M., Raman, P., Brudvig, S., and Hofacker, C. F. (2010), "Customer Information Sharing with E-Vendors: The Roles of Incentives and Trust," International Journal of Electronic Commerce, 14 (3), 63-91.
Reichheld, F. F. and Schefter, P. (2000), "E-Loyalty: Your Secret Weapon on the Web,"
Harvard Business Review, 78 (4), 105-113.
Rousseau, D. M., Sitkin, S. B., Burt, R. S., and Camerer, C. (1998), "Not so Different After all: A Cross-Discipline View of Trust," Academy of Management Review, 23 (3), 393-404.
Schindler, Robert M. (1998), "Consequences of Perceiving Oneself as Responsible for Obtaining a Discount: Evidence for Smart-Shopper Feelings," Journal of Consumer
Psychology, 7 (4), 371-392.
Schumann, Jan H., Wangenheim, F., and Groene, N. (2014), "Targeted Online Advertising: Using Reciprocity Appeals to Increase Acceptance among Users of Free Web Services,"
Journal of Marketing, 78 (1), 59-75.
Smith, H. J., Milberg, S. J., and Burke, S. J. (1996), ""Information Privacy: Measuring Individuals' Concerns about Organizational Practices"," MIS Quarterly: Management
VIII
Statista (2016), "Leading E-Retailers Worldwide in 2014, Based on Retail Revenue (in Million U.S. Dollars)," [available at http://www.statista.com/statistics/287950/leading-e-retailers-worldwide-based-on-revenue/].
Thibaut, John W. and Harold H. Kelley (1959), The Social Psychology of Groups. New York: Wiley.
White, K. and Peloza, J. (2009), "Self-Benefit Versus Other-Benefit Marketing Appeals: Their Effectiveness in Generating Charitable Support," Journal of Marketing, 73 (4), 109-24.
White, K. and Simpson, B. (2013), "When do (and Don'T) Normative Appeals Influence Sustainable Consumer Behaviors," Journal of Marketing, 77 (2), 78-95.
White, Tiffany B. (2004), "Consumer Disclosure and Disclosure Avoidance: A Motivational Framework," Journal of Consumer Psychology, 14 (1), 41-51.
Wirtz, Jochen and Lwin, M. (2009), "Regulatory Focus Theory, Trust, and Privacy Concern,"
Journal of Service Research, 12 (2), 190-207.
Xie, En, Teo, H., and Wan, W. (2006), "Volunteering Personal Information on the Internet: Effects of Reputation, Privacy Notices, and Rewards on Online Consumer Behavior,"
Marketing Letters: A Journal of Research in Marketing, 17 (1), 61-74.
Zauberman, G., Kim, B. K., Malkoc, S. A., and Bettman, J. R. (2009), "Discounting Time and Time Discounting: Subjective Time Perception and Intertemporal Preferences," Journal
of Marketing Research, 46 (4), 543-556.
Zimmer, J. C., Arsal, R., Al-Marzouq, M., Moore, D., and Grover, V. (2010), "Knowing Your Customers: Using a Reciprocal Relationship to Enhance Voluntary Information
IX Appendix Appendix 1 Descriptive Statistics N Mean Std. Deviation Amazon.com 8 5,0909 ,85142 Apple 9 4,6566 1,47834 JD.com 9 3,1111 1,17724 Zalando 9 5,7778 ,57695 Suning Commerce 8 4,0682 ,66760 Otto Group 9 5,1515 ,95129 Lojas Americanas 10 3,9545 ,93425 Groupe Casino 9 4,3131 ,74843 Tesco 10 4,6545 1,27475 Walmart Stores 9 5,0505 ,86616 Total 90
Appendix 1: Descriptive Statistics Pretest
Appendix 2
X
Appendix 3
A seven-point Likert scale (1 = “strongly disagree” and 7 = “strongly agree”) was used for the following question.
Disclosure Through Tracking (Malhotra, Kim, and Agarwal 2004; Schumann, Wangenheim, and Groene 2014)
Given this hypothetical scenario …
1. I would probably allow the website to evaluate my surfing behavior. 2. It is likely that I would consent to an analysis of my surfing behavior. 3. I would be willing to agree to an evaluation of my surfing behavior.
Relevance Anticipation (Laczniak and Muehling 1993; Schumann, Wangenheim, and Groene 2014)
If I allow the website to evaluate my nonpersonally identifiable surfing information … 1. I will see online ads that are relevant to me.
2. I will receive useful information through online ads. 3. Online advertisements will be interesting to me.
4. Online advertisements will be worth paying attention to.
Distributive Justice (Schumann, Wangenheim, and Groene 2014; Wirtz and Lwin 2009)
1. It is fair to reward the website for providing low prices.
2. It is okay that the website asks for a favor in exchange for the low prices offered. 3. Providing the website a benefit in return for low prices is fair.
Trusting Beliefs (McKnight, Choudhury, and Kacmar 2002)
Benevolence
1. I believe that the website owner would act in my best interest.
2. I believe that the website owner shows me advertisements that fit me the most. 3. The website owner is interested in my well-being, not just its own.
Integrity
1. The website owner is truthful in its dealings with me. 2. I would characterize the website owner as honest. 3. The website owner would keep its commitments. 4. The website owner is honest and fair.
Competence
1. The website owner will be competent and effective in providing online advertisements.
2. The website owner will perform its role of showing advertisements very well. 3. Overall, the website owner will be a capable and proficient advertisement provider. 4. In general, the website owner will be very knowledgeable about providing
XI Internet Affinity (Neelamegham and Jain 1999; Schumann, Wangenheim, and Groene 2014)
1. I use the Internet more often than other people do. 2. I am interested in the Internet.
3. I am experienced in using the Internet. 4. In general, the Internet is important for me.
General Concern for Privacy (Dinev and Hart 2006)
1. I am concerned that information I submit on the Internet could be misused.
2. I am concerned that a person can find private information about me on the Internet. 3. I am concerned about submitting information on the Internet, because they could be
used in a way that I cannot foresee.
Attitude Towards Advertising (Pollay and Mittal 1993; Schumann, Wangenheim, and Groene 2014)
1. Overall, I consider advertising a good thing. 2. My general opinion of advertising is favorable. 3. Overall, I like advertising.
Appendix 4
Table 3: Descriptive statistics
Gender Frequency Percent Cumulative
Percent Age Frequency Percent Cumulative Percent
Male 120 57,7 57,7 15 - 24 130 62,5 62,5 Female 88 42,3 100 25 - 34 71 34,1 96,6 Total 208 100 35 - 44 4 1,9 98,6 45 - 54 1 ,5 99,0 55 - 64 2 1,0 100 Monthly Net
Income Frequency Percent Cumulative Percent
Total 208 100
< 1000€ 134 64,4 64,4 Highest Obtained
Degree Frequency Percent Cumulative Percent
1000€ - 1499€ 26 12,5 76,9
1500€ - 1999€ 17 8,2 85,1 Less than
high school degree 3 1,4 1,4
2000€ - 2499€ 14 6,7 91,8 High school degree 40 19,2 20,7
2500€ - 3000€ 8 3,8 95,7 Bachelor's degree
or equivalent 122 58,7 79,3
> 3000€ 9 4,3 100 Master's degree
or equivalent 43 20,7 100
XII
Appendix 5
Appendix 5a: Proof of Homoscedasticity (manipulation check benefit approaches)
XIII
Appendix 6
ANOVA
Sum of Squares df Mean Square F p-value
Between Groups 6,415 3 2,138 ,712 ,546
Within Groups 612,440 204 3,002
Total 618,855 207
Lavene Statistic ,176
Appendix 6a: ANOVA benefit approach
Model Summary R R Square Adjusted R Square Std. Error of the Estimate 0,527 ,278 ,256 1,49110 ANOVA
Sum of Squares df Mean Square F p-value
Regression 171,958 6 28,660 12,890 ,000
Residual 446,896 201 2,223
Total 618,855 207
Coefficients
Beta Std. Error t p-value VIF
(Constant) 2,300 ,882 2,607 ,010 Reciprocity ,202 ,293 ,691 ,490 1,501 Neutral ,308 ,296 1,042 ,299 1,532 Relevance -,035 ,294 -,118 ,906 1,518 Internet Affinity ,313 ,119 2,635 ,009 1,043 Privacy Concerns -,441 ,086 -5,120 ,000 1,062 Attitude towards Advertising ,337 ,068 4,938 ,000 1,082
Appendix 6b: Multiple regression benefit approach
XIV Coefficients
Beta Std. Error t p-value VIF
(Constant) ,855 ,841 1,017 ,310
Dummy Low Trust -,076 ,632 -,120 ,905 12,466
Control Variables
Internet Affinity ,170 ,105 1,628 ,105 1,075
Privacy Concerns -,361 ,075 -4,790 ,000 1,084
Attitude towards Advertising ,147 ,063 2,326 ,021 1,236
Relevance Anticipation, Distributive Justice
Relevance Anticipation ,466 ,104 4,488 ,000 2,594
Distributive Justice ,237 ,092 2,582 ,011 2,373
Interaction Effect: Low Trust,
Distributive Justice -,087 ,126 -,690 ,491 9,676
Interaction Effect: Low Trust,
Relevance Anticipation ,074 ,137 ,542 ,588 11,457
Appendix 6c: Manipulation check benefit approach
Coefficients
Beta Std. Error t p-value VIF
(Constant) 3,968 ,207 19,153 ,000
Reciprocity ,135 ,293 ,459 ,646 1,500
Neutral ,077 ,293 ,263 ,793 1,500
Relevance -,513 ,293 -1,750 ,082 1,500
Appendix 6d: Multiple regression mediation effect (Distributive Justice)
Coefficients
Beta Std. Error t p-value VIF
(Constant) 3,933 ,194 20,255 ,000
Reciprocity ,010 ,275 ,035 ,972 1,500
Neutral ,135 ,275 ,490 ,624 1,500
Relevance -,034 ,275 -,123 ,903 1,500
XV
Appendix 7
Appendix 7a: Proof of Homoscedasticity (manipulation check trust)
XVI
Appendix 8
ANOVA
Sum of Squares df Mean Square F p-value
Between Groups 8,530 7 1,219 ,399 ,902
Within Groups 610,325 200 3,052
Total 618,855 207
Levene Statistic ,445
Appendix 8a: ANOVA trust
Model Summary R R Square Adjusted R Square Std. Error of the Estimate 0,533 ,284 ,248 1,49945 ANOVA Sum of
Squares df Mean Square F p-value
Regression 175,928 10 17,593 7,825 ,000
Residual 442,927 197 2,248
Total 618,855 207
Coefficients
Beta Std. Error t p-value VIF
(Constant) 2,600 ,921 2,824 ,005
High Trust, Reciprocity -,118 ,418 -,283 ,777 1,771
High Trust, Neutral ,221 ,417 ,530 ,597 1,763
High Trust, Relevance -,260 ,420 -,619 ,536 1,783
Low Trust, Reciprocity
and Relevance -,440 ,417 -1,056 ,292 1,759
Low Trust, Reciprocity ,083 ,416 ,199 ,843 1,754
Low Trust, Neutral -,058 ,422 -,137 ,891 1,806
Low Trust, Relevance -,257 ,417 -,616 ,539 1,760
Internet Affinity ,319 ,120 2,653 ,009 1,052
Privacy Concerns -,458 ,088 -5,194 ,000 1,100
Attitude towards
Advertising ,333 ,069 4,846 ,000 1,086
Appendix 8b: Multiple regression trust
XVII ANOVA Sum of Squares df Mean Square F p-value Regression 234,093 10 23,409 11,986 ,000 Residual 384,761 197 1,953 Total 618,855 207 Coefficients
Beta Std. Error t p-value VIF
(Constant) 2,965 ,850 3,489 ,001 Reciprocity ,177 ,274 ,647 ,518 1,502 Neutral ,238 ,278 ,855 ,394 1,544 Relevance -,015 ,277 -,056 ,956 1,535 Control Variables Attitude towards Advertising ,256 ,067 3,841 ,000 1,174 Privacy Concerns -,368 ,082 -4,490 ,000 1,095 Internet Affinity ,186 ,116 1,604 ,110 1,126 Moderator Trust (trusting beliefs) Trust ,750 ,203 3,683 ,000 3,211 Interaction Effect: Reciprocity, Trust -,011 ,308 -,036 ,971 1,843 Interaction Effect: Relevance, Trust -,455 ,332 -1,368 ,173 1,673 Interaction Effect: Neutral, Trust ,028 ,316 ,090 ,928 1,814
