The effect of corporate transparency on personal data sharing:
An empirical investigation
Therese Anschütz
Student number: 10993096 January 28, 2016
MSc. Business Administration, Marketing track Supervisor: Joris Demmers
Statement of originality
This document is written by Therese Anschütz, who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Acknowledgements
Finalizing my thesis would have not been possible without the help of a number of people. Thus, I would like to thank them for their advice and support.
Firstly, I want to thank my supervisor, Joris Demmers, for his guidance and constructive feedback, which gave me confidence and valuable insights. Furthermore, I´d like to thank Quirijn Kolff and Marco Heemskerk from the Amsterdam Cheese Company for allowing me to conduct my experiment in their shop. Without their cooperation it would not have been possible to gain the interesting findings of this study. In this context, I´d also like to thank all my respondents for contributing to the results by filling out my survey. Last but not least, I´d like to thank my family and friends for being on my side. Especially, I´d like to thank Davide Perazzoli for his help and support in overcoming obstacles.
Abstract
Firms increasingly rely on personalizing their customers in order to provide them the best and most suitable services and offerings compared to competitors. However, persuading consum-ers to voluntarily share pconsum-ersonal information is a challenging task for firms. Corporate trans-parency is a possible option to foster the gathering of consumer information. Accordingly, disclosing information about the usage of collected data, defined as specific information, or general transparent messages, defined as generic information, should increase consumers´ propensity to respond in kind. A between-subjects field experiment based on a digital survey in a shop in the food sector investigates the effects of corporate transparency on consumers´ willingness to share personal data. The findings support the positive effect of disclosing ge-neric as well as specific information on consumers´ willingness to share, due to increased trust and decreased privacy concerns about the collection of data. The study further points out that also a third party disclosing generic information about a firm increases consumers´ shar-ing propensity, but the same effect is not found for specific information. The results show that both generic and specific information enhance willingness to share for both lowly and highly sensitive consumer´ information. The study provides valuable insights allowing a firm to gather information about its customers by being perceived as a trustworthy counterpart.
Key words: Corporate Transparency; Data sharing; Reciprocity; Consumer data; Trust;
Table of content
Table of content ... V
1 Introduction ... 1
2 Literature Review ... 3
2.1 Corporate transparency ... 3
2.2 The disclosure of generic firm-related information ... 5
2.3 The disclosure of specific data usage-related information ... 7
2.4 Exposure source of transparent information ... 10
2.5 Differences in requested information ... 12
2.6 Conceptual model ... 14
3 Method ... 15
3.1 Sample and design ... 15
3.2 Pretest ... 16
3.3 Procedure ... 18
3.4 Measures ... 21
4 Results ... 21
4.1 Analytical strategy ... 21
4.2 Hypothesis testing – Consumer data sharing ... 22
4.3 Moderating effect – Exposure source ... 25
4.4 Mediating effects – Privacy concerns and trust ... 26
5 Discussion ... 29
5.1 Evaluation of hypotheses ... 29
5.2 Scientific and managerial implications ... 34
5.3 Limitations and implications for further research ... 35
6 References ... 38
List of tables and figures
Figures
Figure 1 Conceptual model
Figure 2 Frequency – Information type on sharing propensity
Figure 3 Mediating effects – Disclosure of generic information by the firm
Figure 4 Mediating effects – Disclosure of generic information by the third party
Figure 5 Mediating effects – Disclosure of specific information by the firm
Figure 6 Mediating effects – Disclosure of specific information by the third party
Tables
Table 1 Experimental groups of the field study
Table 2 Frequency – Type of consumer information and sharing propensity
Table 3 Binary logistic regression – Type of consumer information on sharing propensity
Table 4 Frequency – Information type and source on sharing propensity
Table 5 Binary logistic regression – Information type and source on sharing propensity
1
Introduction
Over the last few decades, privacy has become a buzzword in the marketing context. The three biggest consumer concerns regarding their privacy and personal information are firms trading personal data without permission, consequences of insecure transactions and data theft (Acquisti and Grossklags 2003). In general, consumers are becoming more skeptical and dis-trustful about business practices (Dark and Ritchie 2007). At the same time, companies in-creasingly rely on customers´ personal information in order to optimize B2C relationships and to achieve higher profits and customer loyalty (Chou, Teng and Lo 2009; Ansari and Mela 2003). Acquiring consumer data is still a critical challenge for companies to achieve competi-tive advantages by profiling their consumers (Pentland 2009; Peppers and Rogers 2012). Es-pecially in online businesses, personalization systems are becoming an essential tool because they enable firms to customize products, contents and advertisements on their websites (Yang and Padmanabham 2005; Atahan and Sarkar 2007). However, also in offline businesses, as customization is becoming increasingly important, organizations are dependent on customer information as a critical input to their product and service development (Zimmer, Arsal, Al-Marzouq, Moore, and Grover 2009). The collection of a greater amount of data helps firms in providing better offerings to their customers. Thus, the benefits related to personalized offer-ings lead customers to accept to voluntarily disclose their personal data, even though some-times they find data collection as intrusive (Zimmer, Arsal, Al-Marzouq, Moore, and Grover 2009). Accordingly, companies are highly interested in decreasing distrust from their custom-ers (Berry 1995).
Previous work has demonstrated that corporate transparency can be a key driver in trust building and information exchange within the relationship between the firm and the cus-tomer (Urban, Amyx and Lorenzon 2009). Furthermore, academic research examined positive effects of transparency in the sense of fostering a reciprocal relationship: (1) when firms are
transparent about the purpose of their data collection (specific information), consumers are more likely to share their data (Zimmer, Arsal, Al-Marzouq, Moore, and Grover 2009; Dem-mers and de Vos 2015) and (2) when firms disclose information about their activities in gen-eral (generic information), consumers are more willing to provide some private information in return (Demmers and de Vos 2015).
Therefore, corporate transparency is an important area to study in the business context as it has the potential to predict behavior regarding reciprocity and the disclosure of infor-mation. This context is especially relevant for marketers, because achieving consumer data is a crucial challenge in times of increasing privacy concerns (Pentland 2009; Peppers and Rog-ers 2012). Thereby, transparency is seen as a key ingredient of relationship marketing, which implies that relationships are dependent on transparency and the subsequent reciprocity (Mur-phy, Laczniak and Wood 2007).
After these considerations, the goal of this thesis is to examine corporate transparency as a driver for information exchange in a reciprocity relationship between firms and consum-ers in a between-subjects field-experiment. More specifically, the experiment consists of a digital survey, which was conducted in a company in the food sector in Amsterdam. A total of 200 people from 36 different countries participated voluntarily to the study. The aim of the survey is to investigate the effects of being transparent about the usage of collected data (spe-cific information disclosure) and communicating general information about the origin of the products (generic information disclosure) on the consumers´ propensity to also provide in-formation about themselves. The findings of the study show support for a positive effect of corporate transparency on consumers´ willingness to share their data, both for generic and specific information disclosure. The effect is mediated by increased trust in the firm and de-creased privacy concerns on the collection of data. The study finds that also if a third party
found for specific information. Furthermore, the results show that the effects of corporate transparency enhance the sharing of lowly sensitive as well as highly sensitive consumer in-formation.
The thesis is structured in four chapters. The next section will provide a complete overview of the recent literature on corporate transparency and the related concepts that will be relevant in the study, leading to the conceptual model and hypotheses. In chapter 3 the methodology of the study will be described in detail. Afterwards, chapter 4 will illustrate the results of the field experiment, assessing if the hypotheses were supported. Finally, chapter 5 will discuss the results and the implications for managers and researchers, ending with direc-tions for further research.
2
Literature Review
This chapter will show the status quo of the current literature about the importance for a com-pany to be transparent when the collection of consumer data is crucial. Firstly, the concept of corporate transparency and its relevance will be discussed and evaluated. Secondly, the dif-ferent underlying mechanisms of disclosing generic and specific information will be ex-plained. Thirdly, the role of the exposure source regarding both information types will be out-lined. Finally, consumers´ willingness to share different types of requested information will be discussed. The chapter ends with a graphical illustration of the hypotheses that will be ana-lyzed and interpreted in the following chapters.
2.1 Corporate transparency
In the academic literature, there are several definitions of transparency from different fields such as marketing, finance, accounting, informatics, political science and communication. In marketing, transparency is defined as “openness and clarity of communications” (Murphy,
Laczniak and Wood 2007), “the access to information” (Prahalad and Ramaswamy 2004), “the availability of firm-specific information to those outside the firm” (Bushman, Piotroski and Smith 2003) and “the individual´s subjective perception of being informed about the rele-vant actions and behavior of the other part in an interaction” (Eggert and Helm 2003). In gen-eral, transparency consists of three specific dimensions: information disclosure, clarity and accuracy (Schnackenberg and Tomlinson 2014). This research highlights the component of information disclosure in order to influence consumers´ willingness to share personal data with the firm.
Transparency in a business context is generally described as the disclosure and sharing of information that is usually not shared in B2B and B2C relationships. Researchers suggest that transparency has beneficial effects for a company, but also a few drawbacks. In general, transparency has several positive effects according to the recent literature. Consumers value if a firm acts in a transparent way, which relates to better evaluations of the company, the prod-uct or brand (Granados, Gupta and Kauffman 2010). Other studies found that transparency leads to an increased trust in the firm (Eisend 2006), higher willingness to pay (Carter and Curry 2010) and a higher likelihood of purchase (Dapko 2012). A possible drawback of cor-porate transparency could be the reduction of information asymmetry, with customers gaining enough information to have more negotiation power than the firm (Hultman and Axelsson 2007). Especially cost transparency could be a threat for firms because it could make the mar-ket highly transparent and products could be easily compared with those of the competitors. Accordingly, it is difficult for firms to create close relationships to the customers because of low switching costs and commoditized products. With regards to consumer data collection specifically, previous literature claims that marketers could benefit from hiding their data col-lection practices in a way that customers share their data unconsciously (Acquisti and
Grossk-of corporate transparency: if the firm discloses information about their usage Grossk-of collected da-ta, which is in the following defined as specific information, consumers are more likely to also share personal data about themselves (Zimmer, Arsal, Al-Marzouq, Moore, and Grover 2009). Additionally, there is empirical evidence that not only specific information about the usage of the data increases the consumers´ willingness to provide information, but also infor-mation about general activities, which is defined as generic inforinfor-mation, is crucial to foster the sharing propensity of the customers (Demmers and de Vos 2015). Overall, the disclosure of specific and generic information enhances the sharing of personal information, but through different underlying mechanisms, as will be discussed and scrutinized in the following sec-tions.
2.2 The disclosure of generic firm-related information
Latest research examined how corporate transparency triggers a process of reciprocity, which leads to voluntary disclosure of personal data (Demmers and de Vos 2015). The existing liter-ature provides several explanations of reciprocity in the fields of psychology, communication and business. In general, reciprocity is a behavior studied by psychology theorists, which ex-plains how disclosure of one´s personal information increases the likelihood of disclosure by its counterpart (Collins and Miller 1994). In addition, reciprocity describes an interaction be-tween sender and receiver of a message. If the sender discloses some information, the receiver is also encouraged to disclose information (Collins and Miller 1994; Taylor and Hinds 1985). In the field of communication, reciprocity is defined as the extent to which a communication allows for mutual conversation or mutual action (Johnson, Bruner II and Kumar 2006). The concept of reciprocity applied to a business context suggests that when firms are transparent about themselves and their activities (generic information), consumers are more likely to share their data (Zimmer, Arsal, Al-Marzouq, Moore, and Grover 2009). In sum, by
enhanc-ing the information exchange in a mutual way, firms are able to predict the behavior of their customers. An information disclosure by the firm leads to a dyadic disclosure where also cus-tomers are more willing to exchange personal information (Miller and Kenney 1986). There-fore, I propose the following hypotheses:
H1a: The disclosure of generic information increases consumers´ willingness to share
per-sonal data.
Additionally, the disclosure of generic information is mediated by increasing trust in the firm (Demmers and de Vos 2015; Chou, Teng and Lo 2009). There is empirical evidence that consumers consider transparency when making purchase decisions, as recent findings illustrated that they have more trust in transparent companies (Eisend 2006). Thus, transpar-ency represents a critical driver to create and maintain a sustainable relationship between or-ganizations and their customers (Schnackenberg and Tomlinson 2014). A previous study ana-lyzed that if a company discloses a big amount information about its identity on its website, consumers will trust that company more. This leads to an increased willingness to disclose some personal information in return. In case the company only discloses a brief introduction and limited contact information, consumers are less willing to interact with the firm because they perceive higher risk and less trust. Therefore, trust serves as a mediator between the dis-closure of company information and consumers´ self-disdis-closure (Chou, Teng and Lo 2009).
Previous literature examines different types of trust: “experience-based trust” and “cue-based trust”, where the former one is built through repeated interaction (past experience, company reputation) and the latter one is developed from the information cues provided dur-ing a consumer´s first interaction with the company (Schoenbachler and Gordon 2002; Wang,
first time, they build trust basing on the information cues provided by the firm or another source, therefore identification information about the company should provide a cue for trust (Chou, Teng and Lo 2009). Behavioral research examines that trust further fosters dialogue (information exchange), while on the other hand distrust induces defensive behavior such as aversion to communicate (Schurr and Ozanne 1985). The relationship between generic infor-mation disclosure and trust is leading to the following hypotheses:
H1b: The effect of generic information disclosure on consumers´ willingness to share
person-al information is mediated by increased trust in the firm.
2.3 The disclosure of specific data usage-related information
Consumers experience a certain risk by sharing their data with a commercial counterpart, es-pecially in online surroundings (Dinev and Hart 2006; Malhorta, Kim and Agarwal 2004; Sheehan 2002; Taddei and Contena 2013). Accordingly, when companies are interested in collecting consumer data, they should act in order to decrease the risk awareness and privacy concerns of their customers. An appropriate way to decrease privacy concerns is to provide the consumer with information about why their private data is being collected and how it will be used, which implies the disclosure of specific information about the data usage. Indeed, previous studies supported the effectiveness of providing specific information to the custom-er, who is therefore more willing to share personal information with a commercial third party (Zimmer et al. 2010; Demmers and de Vos 2015). One limitation of these previous studies is that their findings were supported in a controlled setting. Therefore, it is unclear if the effect of providing specific information about the data usage also holds in uncontrolled settings such as a field study. Hence, I will investigate the effect in a real life setting:
H2a: The disclosure of specific information increases consumers´ willingness to share
per-sonal data.
In general, privacy is defined as the claim of individuals to decide for themselves when, how and to what extent information about them is communicated and shared to others (Olivero and Lunt 2002). The concept of privacy consists of four different dimensions: intru-sion (physically intruding in a person´s environment), disclosure (publicly disclosing private information), false light (falsifying facts) and appropriation (using personal identity without permission) (Prosser 1960). When consumers receive a request to share some personal data, they consider that there might be a misuse of their data. Consumers are interested in what firms know about them, how companies use collected data and that their information are ac-curate (Katz and Tassone 1990).
Furthermore, companies are confronted with a personalization-privacy paradox. Firms increasingly invest in personalization, while consumers are more concerned about their priva-cy and the possible misuse of their data. There is a critical dilemma because customers who desire greater information transparency regarding their privacy issues are also less willing to be profiled (Awad and Krishnan 2006). Furthermore, consumers are confronted with a tradeoff between their gains from being profiled (e.g. customized offerings from the compa-ny, discounts for sharing personal information with the firm) and their concerns for privacy (Chellappa and Sin 2005). They engage in a sort of cost-benefit analysis to determine their privacy-related behavior. Thereby, benefits from data disclosure such as access to websites, use of free Wi-Fi and discounts are often immediate and certain. In contrast, the costs of dis-closing personal information such as spam, intrusive ads and misuse of data, are more delayed than the benefits. Such a cost-benefit analysis regarding privacy situations is a construct from
Moreover, scholars underline that privacy is not seen as an absolute concept, consumers are willing to give up some privacy if they perceive some benefits in return (Chellappa and Sin 2005). Additionally, companies could also decrease the perceived costs by providing a com-plete privacy policy, which would claim to protect consumers´ privacy (Andrade, Kaltcheva and Weitz 2002).
Researchers found out that consumers would like to have more control, which could relate to decreasing privacy concerns (Phelps, Nowak and Ferrell 2000). Marketers should give consumers the possibility to gain at least some control, which could decrease privacy concerns and increase the willingness to share personal data. Interesting in this context is that only a few consumers would actively use the control over their shared information, but it is important for them to know that they have the opportunity to influence the information pos-sessed by marketers (Phelps, Nowak and Ferrell 2000). Consumers perceive a lack of control about their data, which results in an increasing attention towards privacy concerns. In general, consumers´ willingness to share personal data regarding privacy concerns depends on the fol-lowing factors: type of information requested, degree of control, potential benefits and conse-quences of sharing data and consumers´ individual characteristics (Phelps, Nowak and Ferrell 2000).
When firms use data collection practices, they arise privacy concerns in their custom-ers (Culnan and Armstrong 1999). This phenomenon links to the concept of procedural fair-ness, which is defined as a customer´s perception of his/her participation in a practice as being conducted under fair conditions (Thurston and McNall 2010). Procedural fairness is based on two key concepts: notice and consent (Culnan and Armstrong 1999). This means when con-sumers disclose personal information, they have the right to know why the information is be-ing collected, how it will be used and how the shared data will be protected (Culnan and Arm-strong 1999). A practical example of fairness is the chance given by firms to customers to
remove their names from mailing lists. Consumers who are aware of the possibility of remov-ing names have lower concerns for privacy (Greenberg 1987). In other words, companies are able to decrease privacy concerns by providing consumers with procedural fairness, which serves as an intermediary to increase trust (Milne and Gordon 1993). To sum up, the disclo-sure of specific information could be an instrument to reduce consumers´ privacy concerns and to gather information about them. Hence, I propose the following hypotheses:
H2b: The effect of specific information disclosure on consumers´ willingness to share
per-sonal information is mediated by decreased privacy concerns.
To mitigate the effects of the increased unwillingness to share personal data due to privacy concerns and trust, companies should be more transparent about themselves and their usage of collected data (Zimmer, Arsal, Al-Marzouq, Moore, and Grover 2009). There is less empirical evidence on how corporate transparency influences the behavior and attitudes of customers besides purchase decisions, such as the willingness to share private information with the firm. As explained earlier, the study from Demmers and de Vos (2015) found out that there are different underlying mechanisms on consumers´ willingness to share data when the company discloses specific or generic information in controlled settings. The present research will test whether these effects hold also in uncontrolled settings such as in a field experiment.
2.4 Exposure source of transparent information
When disclosing specific or generic information, not only the characteristics of the messages are relevant, but also who discloses the information is an important factor influencing the sharing propensity of the customer (Demmers and de Vos 2015). Thereby, the underlying
meaning of the exposure source varies between a disclosure of specific and generic infor-mation.
In case specific information about the usage and purpose of collected data is commu-nicated, the message itself is crucial to decrease privacy concerns (see chapter 2.3). Hereby, the source of disclosing the information has less influence on the sharing behavior. However, the one who is disclosing the information should be trustworthy and credible (Geer 2003).By disclosing a generic information, e.g. about the production process of a company, the one who is communicating takes an important influencing role on the consumers´ sharing behavior. The theory of reciprocity accounts for the direct information exchange between the sender and the receiver of a message (Miller and Kenney 1986) (see chapter 2.2). In case a third party would disclose transparent information about a company, it is possible that the consumer would also provide personal information, but sharing it with the third party instead of the firm. This assumption suggests that when disclosing generic information a reciprocity process favorable for the firm is initiated only when the firm itself is the source. Overall, the previous argumentation indicates that the exposure source is relevant when disclosing a generic infor-mation, but not in the case of sharing a specific information about the usage of the personal data. Similar assumptions were investigated in a controlled setting in a previous study by Demmers and de Vos (2015). Therefore, I want to test whether the theoretical constructs de-scribed above hold under real life circumstances:
H3a: The effect of generic information disclosure on consumers´ willingness to share
person-al information is dependent on the source of information exposure, with a higher willingness to share when the firm is the source.
H3b: The effect of specific information disclosure on consumers´ willingness to share
2.5 Differences in requested information
Firms are relying more than ever on personalizing their customers. Gathering a bigger amount of information leads firms to create suited offerings and increase consumer loyalty in offline as well as online businesses (Chellappa and Sin 2005; Chou, Teng and Lo 2009; Ansari and Mela 2003). Personalization is dependent on two factors: the firms´ ability to obtain data and the consumers´ willingness to provide some information to the company (Chellappa and Sin 2005). There is empirical evidence that consumers are not only concerned about their personal identification data, but also about their personal unidentifiable information that could be col-lected through cookies (Chellappa and Sin 2005).
Generally, there are different types of data that are relevant for online and offline per-sonalization purposes: user- (personal and demographic characteristics) and identification data (e-mail, addresses, phone numbers), usage data (users´ behavior), general habits (product preferences, interests, hobbies) and sensitive information (medical details, financial infor-mation) (Karat, Blom, and Karat 2004; Andrade, Kaltcheva and Weitz 2002; Phelps, Nowak and Ferrell 2000). There is empirical evidence that willingness to share personal data varies between the different types of information, because they evoke stronger or weaker privacy concerns. Specifically, they found that consumers are more willing to share demographic and information about general habits compared to sensitive, purchase-related and personal identi-fication data. This implies that consumers are less willing to disclose information such as an-nual income, credit cards they possess and social security numbers. On the other hand, infor-mation such as their hobbies, age, marital status, occupation and education are usually shared with less concerns about their privacy. In general, consumers want to protect personal infor-mation when they perceive that the inforinfor-mation will be used for marketing offers (Phelps, Nowak and Ferrell 2000). Especially for B2C context, consumers want to obtain as much
in-formation as they can get about a company, but they are generally less willing to disclose the same amount about themselves (Hoffmann, Novak and Peralta 1999).
Considering the argumentation of a varying propensity to share personal data depend-ing on the type of requested information, I want to investigate further the underlydepend-ing rationale. Disclosing specific or generic information before requesting various types of information could influence positively the sharing propensity. More specifically, I suggest that specific information about the purpose of collected data decreases concerns about a possible misuse of private data, which in turn relates to an increased likelihood to disclose highly sensitive in-formation such as medical or financial inin-formation, but also lowly sensitive inin-formation. On the other hand, the disclosure of generic information enhances the probability of sharing in-formation due to a reciprocity effect. However, with the disclosure of generic inin-formation, consumers do not know what will be the usage of their data and thus they will be less likely to share highly sensitive information. Therefore, I propose that the reciprocity effect will provide an increased willingness to share when the information asked is lowly sensitive, but not when highly sensitive. Basing on differences in perceived sensitivity of requested information in relation to corporate transparency, I propose the following hypotheses:
H4a: The disclosure of generic information enhances consumers´ willingness to share lowly
sensitive, but not highly sensitive information.
H4b: The disclosure of specific information enhances consumers´ willingness to share both
2.6 Conceptual model
FIGURE 1: Conceptual model
The aim of this thesis is to examine corporate transparency as a driver for information ex-change in a reciprocity relationship between firm and consumer. The effects of corporate transparency on trust and privacy concerns are also analyzed, to see if they mediate the rela-tionship described above. Furthermore, this thesis investigates if the effect of corporate trans-parency on consumers´ willingness to share personal data is depending on the disclosure source. Additionally, the effects of information disclosure on the propensity to share different requested information (lowly, moderately, highly sensitive) are examined.
H1a: The disclosure of generic information increases consumers´ willingness to share
per-sonal data.
H1b: The effect of generic information disclosure on consumers´ willingness to share
H2a: The disclosure of specific information increases consumers´ willingness to share
per-sonal data.
H2b: The effect of specific information disclosure on consumers´ willingness to share
per-sonal information is mediated by decreased privacy concerns.
H3a: The effect of generic information disclosure on consumers´ willingness to share
person-al information is dependent on the source of information exposure, with a higher willingness to share when the firm is the source.
H3b: The effect of specific information disclosure on consumers´ willingness to share
per-sonal information is not dependent on the source of information exposure.
H4a: The disclosure of generic information enhances consumers´ willingness to share lowly
sensitive, but not highly sensitive information.
H4b: The disclosure of specific information enhances consumers´ willingness to share both
highly and lowly sensitive information.
3
Method
This chapter illustrates the empirical part of the study. First, the characteristics of the collect-ed sample and the research design will be explaincollect-ed. Afterwards, a pre-test will be describcollect-ed, which was crucial for the development of the survey design. Subsequently, the exact proce-dure and measures of the conducted experiment will be outlined.
3.1 Sample and design
The sample of this study are customers or potential customers who were inside the shop of the Amsterdam Cheese Company. The data collection was performed by asking possible partici-pants in the shop to fill in a digital survey. The participartici-pants in this study were 200. The size of
the sample was chosen to ensure that it would be large enough to make the findings general-izable.
The quantitative research was carried out in a 2 x 2 + 1 between-subject field experi-ment, which indicates that each participant is only tested once. Two factors were tested in the experimental conditions: (1) information disclosure: generic versus specific and (2) exposure source: firm versus third party. Moreover, a second dependent variable was set to examine differences between requested information (either lowly or moderately sensitive). Consumers´ willingness to share personal data was measured as a dichotomous dependent variable (yes/no).
3.2 Pretest
Before creating the questionnaire, it was necessary to select transparent information which would be perceived as relevant, useful and neutral from customers and to assess which infor-mation would be more sensitive. Therefore, 11 messages (two specific and nine generic in-formation) were created in cooperation with the Amsterdam Cheese Company. It was im-portant to measure the neutrality of the transparent messages, considering that too positively or too negatively perceived information could bias the propensity of data sharing by influenc-ing the attitude towards the company. Furthermore, by choosinfluenc-ing not too positive nor too nega-tive messages, which are also perceived as relevant and useful, it was possible to make sure that the differences in willingness to share information would be due to the information dis-closure and not because of the characteristics of the messages. The relevance, usefulness and neutrality were measured on a bipolar five-point scale (not relevant/not useful/negative to relevant/useful/positive). It was also important to figure out which types of information con-sumers are in general more willing to share to avoid the bias due to a higher willingness to
of information people are in general more willing or less willing to share because they are perceived respectively as less sensitive or more sensitive regarding privacy concerns. In the pretest the information have been divided in three categories: (1) lowly, (2) moderately and (3) highly sensitive information. The categorization was made by asking participants to eval-uate their willingness to disclose each piece of information. The willingness was measured on a four-point Likert-scale (never willing to always willing). The tested specific and generic messages, chosen experimental stimuli and requested information for the field experiment are available in the appendix.
The pretest involved the evaluation of 24 participants with no missing values within the requested information. The results of the first part of the test showed that both the specific information (third party, firm itself) were suitable for the subsequent experiment. Further-more, it allowed to select the generic information, basing on the necessity to find a relevant and useful message, which is at the same time not too positive. The generic information tested were sometimes too positive or too less relevant and useful. The selected generic information was more relevant and useful than the average of the test, and at the same time in the range of moderate positivity. Both the specific and the generic information were perceived as relevant and useful (Mspecific, third party, relevance = 3.67, SD = .963; Mspecific, third party, usefulness =
3.54, SD = .884; Mspecific, firm, relevance = 3.71, SD = 1.042; Mspecific, firm, usefulness = 3.75,
SD = .989;Mgeneric, relevance = 3.71, SD = 1.042; Mgeneric, usefulness = 3.92, SD = .776).
Fur-thermore, the specific and generic information were perceived as moderately positive (M spe-cific, third party, neutrality = 3.46, SD = .884; Mspecific, firm, neutrality = 3.42, SD = .884; Mgeneric, neutrality = 3.92, SD = .776). A series of one-way ANOVA confirmed that there is no
signifi-cant difference between specific and generic information in the perceived relevance, useful-ness and neutrality (p > .05). Therefore, if there will be a relationship between corporate
transparency and willingness to share by the consumer, it is likely to be due to the information type (generic, specific) rather than differences in relevance, usefulness and neutrality.
The second part of the pretest showed that the participant´s willingness to share pri-vate data differs between the various requested information. Accordingly, it was possible to classify the information, which were measured on a four-point Likert scale, into three groups: (1) values between three and four were characterized as lowly sensitive information, (2) val-ues between two and three were defined as moderately sensitive information and (3) valval-ues between one and two represented highly sensitive information. In the appendix, the chosen information for the field experiment are shown, with their respective sensitiveness score. The information were sorted into categories, each with a category mean of sensitivity. Hereby it is important to mention that a single group of only highly sensitive requested information was not possible to set in the experiment because of a potential damage on the reputation of the Amsterdam Cheese Company if they asked very sensitive data to their customers. Both mod-erately sensitive groups include some highly sensitive information or information that are close to the high sensitivity range: monthly income (M = 1.92), favorite payment method (M = 2.38), money spent for groceries per week (M = 2.48) and the accommodation in Amster-dam (M = 2.04).
3.3 Procedure
To test whether generic or specific information disclosure influence the sharing propensity of the customer in an uncontrolled setting, a field study in a shop of the Amsterdam Cheese Company was conducted. The customers were asked to fill out a short survey about customers in the dairy industry. Before being asked, they were already assigned to one of the five exper-imental groups (see table 1). Thereby, only the experexper-imental groups (2), (3), (4) and (5)
expe-group (1) received no transparent information, which also indicates that the source of disclo-sure is not relevant.
TABLE 1: Experimental groups of the field study
Exposure source Firm
(Amsterdam Cheese Company)
Third party
(University of Amsterdam) Control group (1) No information disclosure, exposure source is not relevant Manipulated group
(disclosure of generic information)
(2) Generic information disclo-sure by the firm
(3) Generic information disclo-sure by the third party Manipulated group
(disclosure of specific information)
(4) Specific information dis-closure by the firm
(5) Specific information disclo-sure by the third party
If they agreed to take part, they were required to fill in a digital version of the survey. Firstly, the participants received generic or specific information by the firm or third party, except in the control condition where no information was disclosed. Afterwards they were asked to fill out several blocks of questions: (1) personal information (name, country, email address), (2) attitude towards the Amsterdam Cheese Company, (3) lowly sensitive infor-mation (gender, age, knowledge of company, impression of Dutch cheese, favorite cheese from the shop), (4) moderately sensitive information (food intolerance, favorite payment method, money spend in the shop and for groceries generally, monthly income), (5) moderate-ly sensitive information, specificalmoderate-ly for tourists (activities planned in Amsterdam, length and location of stay, other things bought in Amsterdam), (6) questions about extent of trust in the Amsterdam Cheese Company (modified measurement items from Chaudhuri and Holbrook 2001), (7) questions about extent of privacy concerns about information shared in the survey (modified measurement items from Smith, Milberg and Burke 1996) and (8) question about the recognition of the exposure source, either Amsterdam Cheese Company or third party
(manipulation check). After the question blocks (3), (4) and (5) the participants were asked how willing they would be to share the provided information with the Amsterdam Cheese Company. If they didn´t want to share the information with the company, they were told that their personal data won´t be saved in the customer database of the firm. The gathered personal information were not relevant for the purpose of the study. The obtained information were only shared with the Amsterdam Cheese Company if the participant agreed to share them with the company. As the majority of the customers came from outside the Netherlands, the com-pany was interested in the information to identify their customers and improve their products and services based on the information collected in the survey. An illustration of the conducted survey of the experimental group (2) can be found in the appendix.
Among the different conditions (see table 1), the participants experienced two types of manipulations: (1) information type and (2) exposure source. The information type was ma-nipulated through different messages, with the different groups receiving either generic or specific information. In the generic condition, they were told where and from which farmers the cheeses of the Amsterdam Cheese Company are produced and bought. In case of a specif-ic information disclosure, the partspecif-icipants were told what the exact purpose of the survey is and what kind of data they would be asked. Both of the information were created in coopera-tion with the Amsterdam Cheese Company. The experimental stimuli of the generic and spe-cific information, which were used in the field study, can be found in the appendix. Moreover, the information source was manipulated through using in the questionnaire a banner with a logo and an introductory sentence identifying the source being either the Amsterdam Cheese Company (firm) or the University of Amsterdam (third party).Furthermore, when I had to be identified as being from the Amsterdam Cheese Company as source of information exposure I wore an apron with the logo of the company as their employees do. The described
manipula-source (Groza, Pronschinske and Walker 2011; Liu, Austin and Jin 2011; Demmers and de Vos 2015) were similarly conducted in previous studies.
3.4 Measures
The survey consisted of different types of information requested to the customer. The catego-rization (lowly, moderately and highly sensitive) of these information was adopted from Phelps, Nowak and Ferrell (2000). The dependent variable (willingness to share data) was measured with a binary scale (willing – not willing) after the request of different types of in-formation. The first mediator, trust, was measured with a two-item seven-point Likert scale (Chaudhuri and Holbrook 2001). The second mediator, privacy concerns was measured using a two-item seven-point Likert scale (Smith, Milberg and Burke 1996). The modified meas-urement items of both mediators can be found in the appendix.
4
Results
In this section the results of the hypotheses testing will be illustrated. Firstly, there will be an overview of the sample and the manipulation check. Secondly, the general effect of the inde-pendent variable (corporate transparency) on the two deinde-pendent variables (willingness to share information and willingness to share different types of information) will be analyzed. Thirdly, the mediating effect of privacy concerns and trust on sharing propensity will be out-lined. Finally, the moderating effect of the exposure source on willingness to share infor-mation will be assessed.
4.1 Analytical strategy
The sample of 200 voluntary participants was composed by almost the same amount of men and women, with 52.5% of females and 47.5% of males. The age and nationality were quite
various. The age spaced between 17 and 74 years old, with an average age of 28.5. The partic-ipants were from 36 different countries, with British (18%), Germans (14%) and Spanish (8.5%) being the most represented in the sample.
Among the 200 participants, 25 did not identify correctly the information source from the experimental condition they were assigned to, therefore they were excluded from the analysis. In sum a total of 175 participants are included in the subsequent analysis. Moreover, there was one missing value found, which has been listwise deleted.
4.2 Hypothesis testing – Consumer data sharing
To test H1a and H2a about the effect of transparent information on consumer´s data sharing propensity, a binary logistic regression was performed. Therefore, the type of information, either specific or generic, was identified as a categorical predictor and the willingness to ei-ther share data or not was set as a dichotomous dependent variable. The control group, where no information was disclosed, was used to examine whether there is an effect of corporate transparency on the willingness to share private information without regard for differences in the requested information (lowly and moderately sensitive). The frequency analysis showed that both generic and specific information disclosure compared to the control group led to a higher sharing propensity (figure 2). The regression confirmed that this difference is positive-ly correlated with a disclosure of either generic (B = 1.105, SE = .218 and p < .001) or specif-ic information, (B = .979, SE = .203 and p < .001). Thus, these results show support for both H1a and H2a.
FIGURE 2: Frequency – Information type on sharing propensity
To test H4a and H4b a binary linear regression was performed. The low sensitive and the two moderately sensitive consumer information were grouped by the types of corporate infor-mation disclosure. Every group was then compared with the control condition. In general, both the disclosure of specific and generic information increased the sharing propensity for every type of requested information (table 2). The regression pointed out that generic infor-mation disclosure significantly increases the willingness to share by the customer, both for lowly (B = 1.313, SE = .402, p = .001) and moderately sensitive information requests (B = 1.020, SE = .354, p = .004; B = 1.056, SE = .394, p = .007). Also in the case of specific in-formation, the sharing of both lowly (B = .725, SE = .342, p = .034) and moderately sensitive information (B = 1.036, SE = .336, p = .002; B = 1.247, SE = .388, p = .001) was significantly increased (table 3). In the case of generic information, corporate transparency fostered more the sharing of lowly sensitive information by the customer (82%) compared to both moderate-ly sensitive information (67.2%, 73.1%). However, a two-tailed follow-up contrast test refut-ed the significance of the difference between the effect of lowly and moderately sensitive in-formation on willingness to disclose (p = .940). Therefore, the results provide support for a
48,70% 74,10% 71,60% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Control group Generic information Speci?ic information Sharing propensity (%)
significant increase of sharing for both lowly and highly sensitive information, refuting H4a. Instead, for specific information the effect of corporate transparency was more evident in the case of moderately sensitive information sharing, which has shown a higher difference in sharing propensity compared to the control condition (67.6% vs 42.5% and 76.7% vs 48.6%) than the lowly sensitive (71.6% vs 55%). A two-tailed follow-up contrast test refuted the sig-nificance of the difference in the willingness to share between lowly and moderately sensitive information (p = .091), providing support for H4b.
TABLE 2: Frequency – Type of consumer information and sharing propensity Type of consumer information Sharing propensity (%)
Control Generic Specific
Low sensitivity 55 82 71.6
Moderate sensitivity (I) 42.5 67.2 67.6
Moderate sensitivity (II) 48.6 73.1 76.7
TABLE 3: Binary logistic regression – Type ofconsumer information on sharing propensity
B SE p 95 % CI
Generic, low sensitivity 1.313 .402 .001 1.692 – 8.174
Generic, moderate sensitivity (I) 1.020 .354 .004 1.385 – 5.554
Generic, moderate sensitivity (II) 1.056 .394 .007 1.329 – 6.216
Specific, low sensitivity .725 .342 .034 1.056 – 4.037
Specific, moderate sensitivity (I) 1.036 .336 .002 1.459 – 5.444
4.3 Moderating effect – Exposure source
To test H3a and H3b a binary logistic regression was conducted. The control group was com-pared with four experimental conditions: generic information disclosed by the firm or third party and specific information disclosed by the firm or third party. Overall, as table 4 shows, the sharing propensity in all experimental groups is higher than in the control group. The re-gression showed that generic information, both disclosed by the firm (B = .984, SE = .297 and p = .001) and by the third party (B = 1.251, SE = .321 and p < .001) have a significant effect on sharing propensity of the consumer. However, when the third party discloses the generic information customers are more likely to provide their information (76.8 %) than when the firm itself discloses it (71.7 %). Instead, in the case of specific information, the disclosure by the firm encourages a higher sharing propensity (80.7 %) than the disclosure by a third party (60.6 %). The regression in this case shows that only the disclosure by the firm has a signifi-cant effect (B = 1.483, SE = .302 and p < .001).Subsequently, a series of one-way ANOVA was conducted to assess whether the difference in the effects on willingness to share between firm and third party as sources of information are significant. For generic information disclo-sure there is no significant difference between disclodisclo-sure sources (p > .05), thus H3a is not confirmed. For specific information instead the difference between firm and third party is significant (p = .001), therefore also H3b is not supported.
TABLE 4: Frequency – Information type and source on sharing propensity Information type and source Sharing propensity (%)
Control 48.7
Generic, firm 71.7
Generic, third party 76.8
Specific, firm 80.7
TABLE 5: Binary logistic regression – Information type and source on sharing propensity
B SE p 95 % CI
Control vs. generic, firm .984 .297 .001 1.493 – 4.790
Control vs. generic, third party 1.251 .321 < .001 1.861 – 6.559 Control vs. specific, firm 1.483 .302 < .001 2.438 – 7.961 Control vs. specific, third party .484 .282 .086 (NS) .934 – 2.819
4.4 Mediating effects – Privacy concerns and trust
To test the indirect effect about generic information predicted in H1b a simple mediation on PROCESS (Hayes, 2012) was performed with a bootstrapping test (n iterations = 5000). Be-fore undertaking the mediation test, a preliminary analysis was made to assess whether there were differences in trust and privacy concerns between exposure sources, as according to the moderation there were no significant differences in willingness. This analysis showed no sig-nificant difference in trust (Mfirm = 5.773, Mthird party = 5.554, p = .202) between sources, but
a significant difference in privacy concerns (Mfirm = 4.121, Mthird party = 3.071, p < .05).
Therefore, the subsequent analysis considered firm and third party disclosure separately. More specifically, generic information disclosure both from the firm (B = .8207, p < .001) and from a third party (B = .6357, p = .0011) positively affected trust, which in turn enhanced the propensity to share information (B = .4849, p < .001 and B = .5057, p = .001). In both cases, the indirect effects (B = .3980 and B = .3215) with a 95% confidence interval con-firmed the significance of the mediating effect, as there is no zero within the intervals (.1536, .7576 and .1079, .6468). Additionally, the analysis has tested the mediating effect of privacy concerns on willingness to share. A disclosure of generic information by the firm has no sig-nificant effect on privacy concerns (B = -.1587, p = .4934). Also, the indirect effect with a
was a reduction of privacy concerns (B = -1.2947, p < .001), which in turn led to a higher willingness to share (B = -.4015, p = .0031). The indirect effect also confirmed the mediation (B = .5198) with a 95% confidence interval excluding zero (.1468, 1.0968). The results pro-vide support for H1b, confirming the mediating effect of trust on the relationship between generic information disclosure and willingness to share information. However, the additional analysis conducted points out that also privacy concerns are mediating the relationship, but only when the third party is the source. A visual representation of the mediation models can be found below.
FIGURE 3: Mediating effects – Disclosure of generic information by the firm
Note: *p < .05, ** p < .01, *** p < .001
FIGURE 4: Mediating effects – Disclosure of generic information by the third party
For specific information, to test the effect assumed in H2b a further simple mediation on PROCESS (Hayes, 2012) was performed with a bootstrapping method (n iterations = 5000). As the moderating effect pointed out a difference in willingness to share between firm and third party, separate analysis have been conducted for the two sources. A preliminary analysis also pointed out a difference in trust between sources (Mfirm = 5.897, Mthird party =
5.286, p < .05), but no significant difference in privacy concerns (Mfirm = 3.321, Mthird party =
3.200, p = .531). The mediation showed that privacy concerns are decreased both when the firm (B = -1.0189, p < .001) and the third party (B = -1.1148, p < .001) are disclosing infor-mation, which in turn both led to an increased willingness to share (B = -.2606, p = .0411 and B = -.3527, p = .0029). The indirect effects confirmed in both cases (B = .2655 and B = .3932) the significance of the mediation, with 95% confidence intervals excluding zero (.0070, .6466 and .1172, .7973). An additional analysis was run to test if also trust has a me-diating effect on specific information. Trust was found to be positively affected by specific information either when the firm (B = .9729, p < .001) or the third party (B = .3848, p = .0420) disclosed, which in turn led to an increased sharing propensity (B = .5567, p < .001 and B = .6175, p < .001). The indirect effects supported these findings (B = .5416 and B = .2376) with 95% confidence intervals without zero (.2903, .9021 and .0197, .9182). The re-sults thus supported the mediation effect predicted in H2b, as privacy concerns influenced the relationship between corporate transparency and consumers´ sharing propensity. However, the additional analysis pointed out that also trust is mediating this relationship. A visualization of the mediation models can be found below.
FIGURE 5: Mediating effects – Disclosure of specific information by the firm
Note: *p < .05, ** p < .01, *** p < .001
FIGURE 6: Mediating effects – Disclosure of specific information by the third party
Note: *p < .05, ** p < .01, *** p < .001
5
Discussion
In this final chapter, the results will be discussed and evaluated considering their scientific and managerial implications. Furthermore, the limitations of the research will be outlined. Finally, directions for further research in the field of corporate transparency and data sharing will be suggested.
5.1 Evaluation of hypotheses
willing-the hypowilling-theses proposed. Considering willing-the main effect predicted in H1a and H2a, willing-the empirical findings provided support for the positive link between disclosure of both generic and specific information and consumers´ willingness to share. The findings indicate that consumers value when a firm is providing them with transparent messages, which increases the likelihood of providing the firm with personal information. Moreover, it is interesting that the sharing pro-pensity was slightly higher when generic information about the place of production (74,10 %) than specific information about the usage of requested data (71,60 %) were disclosed. The concept of reciprocity could account for the enhanced willingness when generic information were disclosed (Collins and Miller 1994).Hypotheses H4a and H4b investigated how differ-ent types of requested information influence the willingness to provide personal data. In gen-eral, both generic and specific information increased the likelihood to share for every type of requested information (lowly and both moderately sensitive information). However, there was no significant difference in the sharing between lowly and moderately sensitive information for both generic and specific information disclosure. This means corporate transparency leads to an increased willingness to share lowly as well as highly sensitive information regardless of the type of transparent messages. The results for H3a and H3b provided no support for the moderating effects predicted. In the case of generic information, the empirical evidence did not support a higher disclosure of consumer information when the firm itself is transparent. Oppositely, the disclosure from the third party fostered more the willingness to share. How-ever, a significant difference between exposure sources was not found. For specific infor-mation the results were also refuting the hypothesis: the findings show a significant difference between firm and third party in fostering consumers´ information sharing, with a higher effect of the firm as a source. Interestingly, the results of these two hypotheses show exactly the opposite of what the two hypotheses proposed. The disclosure of generic information seems
cific information is effective only when disclosed by the firm. These findings suggest that specific information on usage data is believable only when the one who is collecting personal data is the same that provides the information. The empirical findings of H1b and H2b about the mediating effects of both types of information supported the mediating effect of increas-ing trust for generic information and of decreasincreas-ing privacy concerns for specific information for both sources of information. However, the results show that privacy concerns are also re-duced with generic information disclosure, but only when the firm is the source. Furthermore, trust is also positively mediating the relationship between specific information disclosure and sharing propensity, regardless of the exposure source. Taking a deeper insight on these find-ings, it can be seen that in the case of specific information, trust and privacy concerns posi-tively influence the sharing propensity of consumers, but the overall effect is not significant if the source is the third party. This implies that a disclosure of specific information has positive effects, limiting privacy concerns and fostering trust. However, observing the total effect, only if the firm is disclosing the information this will end up in an increased sharing propensi-ty. An overview of the empirical findings can be found in table 6.
TABLE 6: Overview of hypotheses tested
Hypotheses Empirical
findings
General effects
H1a: The disclosure of generic information increases
con-sumers´ willingness to share personal data.
Supported
H2a: The disclosure of specific information increases
con-sumers´ willingness to share personal data.
Supported
H4a: The disclosure of generic information enhances
consumers´ willingness to share lowly sensitive, but not
highly sensitive information. Not supported
H4b: The disclosure of specific information enhances
sensitive information. Supported
Mediating effects
H1b: The effect of generic information disclosure on
con-sumers´ willingness to share personal information is
mediat-ed by increasmediat-ed trust in the firm. Supported
H2b: The effect of specific information disclosure on
con-sumers´ willingness to share personal information is
mediat-ed by decreasmediat-ed privacy concerns. Supported
Moderating effects
H3a: The effect of generic information disclosure on
con-sumers´ willingness to share personal information is depend-ent on the source of information exposure, with a higher
willingness to share when the firm is the source.
Not supported
H3b: The effect of specific information disclosure on
con-sumers´ willingness to share personal information is not
de-pendent on the source of information exposure. Not supported
Comparing this study with a previous research by Demmers and de Vos (2015), there are sim-ilar as well as opposing findings. The general positive effect of disclosing generic or specific information on willingness to share personal data are confirmed in both studies, which lead to a strong empirical significance of the findings in controlled and uncontrolled settings. Con-sidering the moderating effect, the current study is contradicting previous findings. Demmers and de Vos (2015) found empirical evidence that the disclosure of generic information in-creases consumers´ willingness to provide information when the firm is the source, while the current study found no significant difference between sources. Furthermore, they supported that the disclosure of specific information on consumers´ sharing propensity is not depending on the exposure source. Oppositely, this research found a positive effect of information dis-closure by the firm, but no significant positive effect on sharing propensity when a third party discloses. Furthermore, the studies differ in their findings about the mediation effects of trust and privacy concerns. The previous study found that increased trust has an effect only on
ge-that also generic information reduces privacy concerns, but only when the firm is disclosing, and specific information is increasing trust regardless of the source.
Furthermore, the present study contributes to the research by adding empirical evi-dence to the existing literature, sometimes contradicting previous findings. Phelps et al. (2000) found out that the willingness to share information varies between different types of information. Their research showed that consumers are less willing to share highly sensitive information because of high privacy concerns about their data. Oppositely, lowly sensitive information evokes less privacy concern, which leads to a higher motivation to share their data. This study adds that corporate transparency, either through generic or specific mation disclosure, enhances the probability to share lowly as well as highly sensitive infor-mation. Hoffmann et al. (1999) supported in their paper that consumers are interested in ob-taining as much information as they can about a company, but they are less willing to disclose the same amount of information about themselves. Conversely, this study investigated an in-formation exchange under the concept of reciprocity, which means that when a firm is dis-closing transparent messages, this increases consumers´ behavior to respond in kind signifi-cantly. Awad and Krishnan (2006) highlighted the phenomenon of a personalization - privacy paradox, which means that firms increasingly invest in personalization systems, while con-sumers are more concerned than ever about their privacy and a possible misuse of their data. Furthermore, they described a critical dilemma of the use of personalization because custom-ers who desire greater information transparency about privacy issues are also less willing to be profiled. In this context, Andrade et al. (2002) added that companies could decrease priva-cy concerns through providing a complete privapriva-cy polipriva-cy. Additionally, Culnan and Arm-strong (1999) investigated the principle of procedural fairness, which indicates that consumers want to know why the information is being collected and how they will be used and protected. Extending the crucial aspect of privacy concerns examined by previous researchers, this study
found out that it is advantageous for firms to actively seek to reduce privacy concerns of their customers. Providing consumers with information about the purpose of the data collection before they even ask or worry could increase the information exchange between firm and cus-tomer and foster procedural fairness. The present study found empirical evidence of specific information about the usage of collected data as an option to decrease privacy concerns re-gardless of the exposure source, which indicates that even a third party has a significant influ-ence on consumers´ privacy concerns. More interestingly, there was also support for reducing privacy concerns when generic message with no information about the usage of data were disclosed. This implies that actively providing transparent messages to the customer helps to decrease worries about providing data and to overcome the unwillingness to be profiled.
5.2 Scientific and managerial implications
Previous studies started to consider corporate transparency within different facets: transparen-cy leads to better evaluations of the company, the product or brand (Granados, Gupta and Kauffman 2010); it increases trust in the firm (Eisend 2006); it fosters a higher willingness to pay (Carter and Curry 2010) and a higher likelihood of purchase (Dapko 2012). Demmers and de Vos (2015) were the first in the scientific community to focus on the effect of corporate transparency on consumers´ willingness to share personal data in a controlled setting. There-fore, the study by Demmers and de Vos (2015) was crucial for the current research, which aimed to foster the strength of their empirical findings by conducting a field experiment. The current study provides an additional academic contribution to earlier findings by confirming a significant increased willingness to share data after receiving transparent messages.
Furthermore, the study provides several beneficial implications for practitioners. As mentioned before, collecting personal consumer data is a crucial challenge for companies
exchange by disclosing information first, they enhance the possibility of data sharing by the consumers voluntarily. The ability to profile their customers may provide marketers with ad-ditional advantages such as fostering loyalty towards the firm through more suitable offerings. Furthermore, an interesting finding of the present study was that a disclosure of either generic or specific messages enhances the likelihood of sharing lowly as well as highly sensitive in-formation by the customer. This implies that firms could gather a great amount of beneficial information about their consumer in an honest way. The study further found out that a disclo-sure of generic information by a third party is also enhancing consumers´ willingness to share personal data. Therefore, a firm could benefit from a partnership with another company, by disclosing information about each other and favoring the gathering of consumer information for both firms. Nevertheless, there should be a “fit” between the companies, either in being in the same product category or by having a similar brand concept (Lanseng and Olsen 2012). Furthermore, using endorsers disclosing firm information as a third party could be far more valuable for marketers than simply attracting customers. However, there should be a “fit”, “congruence” or “belongingness” between the endorser and the firm to ensure the effect of disclosing information on consumers´ willingness to share (Till 1998).
5.3 Limitations and implications for further research
Although the field study presents some positive aspects, such as the size of the sample which contributes to make the findings generalizable, there are also a few limitations. Firstly, the chosen experimental stimuli of the specific and the generic information were slightly per-ceived as positive within the pretest, which could lead to a biased sharing behavior of the par-ticipants in the final study. Nevertheless, messages that were perceived more neutral were also rated as less relevant and useful, e.g. contact information of the company. Thus, the transparent information were selected considering the best possible combination of neutrality,
