Keywords

1 Introduction

Web forms have been a known component on websites since the 1990s. They are often used by web service providers to collect personal data of their users, which is either directly required for the functionality of the service, or serves other purposes such as enabling data analytics (e.g. for personalized advertisements or service improvements). Usually, the users are asked for a lot of information, often of personal nature, while some items are labeled as optional and others as mandatory. When filling in the web form, users have to decide, which data they want to share. The only way for the users not to share data requested in mandatory fields is either not to use the service at all, or to provide information that is fake but has a semantic that the service provider accepts (e.g. a wrong birthday but still an existing date). Users have more power in deciding whether to share information or not when the fields are optional.

A number of studies have been dedicated to the research on users’ behaviour and perception regarding the web form fields. In particular, several studies focused on researching a link between the users’ willingness to provide data by filling in the web form fields that are not mandatory and the users personality traits [5, 11]. The paper by Preibusch et al. [12] provides a list of reasons which might explain, why users provide data for optional fields on the website form. This list, however, is only partially supported by existing studies. Furthermore, these studies have been performed in specific contexts, such as creating an account on social lending networks. Hence it is unclear whether their results can be generalised to other types of web services.

In order to check the relevance of their list, Preibusch et al. conducted a quantitative study in [12]. One of the goals of the study has been to find out the reasons why the participants filled in the optional fields. Their results confirmed the relevance of some of the reasons from their initial list. Furthermore, the study found anecdotal evidence for additional reasons such as user extroversion or feeling compelled to complete all the fields in the form. The authors, however, did not elaborate on the additional reasons. Furthermore, being only one of several research goals of the study, the reasons of filling in the forms were the focus of only one open question, which did not allow clarifying follow-up questions. The study also focused on a specific context of the mTurk platform. As such, the participants expected that their input data will be used for statistical research (which provided additional motivation for some of them to input more), that the more data they input, the more rewards they would gain from the mTurk platform, and the authors themselves recognize that the active users of mTurk might be more inclined to fill in web forms out of interest than the general population.

In this work we conduct a user study to verify, whether the reasons for filling in the optional fields of the web forms are consistent with the initial list by Preibusch et al. in the context of interacting with a commercial website. Concretely, the scenario for our study was that the participants had to register a user account on a mock website of the Deutsche Bahn (German railways) companyFootnote 1. They were told, that the goal of the study is to evaluate the usability of the new design proposed by the company. After filling in the registration form, semi-structured interviews were conducted with the participants, where they were asked to explain why they filled or not filled in the optional form fields. The interviews were qualitatively analysed using the reasons from the initial list of Preibusch et al. as pre-existing categories. The results show which reasons have been mentioned by our participants and whether additional reasons have been mentioned that cannot be assigned to the initial list. We found correspondences for three out of ten items in the list in our interviews. We also found evidence for two additional categories of reasons that the participants gave when asked to explain why they filled in an optional field.

We furthermore looked at the reasons why the participants were reluctant to share additional data, and at the countermeasures they used, such as providing fake data, in order to avoid sharing more than they would want.

2 Methodology

In this section we describe the user study that we have conducted, and the methods we used to analyse the resulting interviews.

2.1 User Study

We first by describing the study design and the demographics of our participants.

Mock Registration Website. For our user study we set up a mock registration website that contained a cloned and modified registration form of the “Deutsche Bahn” company on a local virtual machine. The DNS entries in the host file on the operation system were manipulated, so that the participants were not able to tell that the website is not online. For the same purpose, the Internet connection status bar was hidden.

The form on our mock website resembled the design of the original Deutsche Bahn registration website, but contained a different set of form fields, namely, eight mandatory fields and six optional fields. The fields that were included in the form were chosen as common fields on the websites in Alexa top 50Footnote 2. Namely, we included such optional fields as title, date of birth and phone number. In additional to these fields that were commonly encountered on the websites, we also chose to include two optional fields that we rarely used in web-forms, namely, the marital status and country of origin. We further included an optional checkbox that asked whether the participants consent to using cookies.

Participants. The study consisted of 16 participants, with eight women and eight men. The youngest participant was 23 years old, and the oldest 58 years old, with 36.5 as the mean age. In order to prevent priming the participants towards thinking about their privacy, the participants of the study were told that they are going to participate in a study done in collaboration with the Deutsche Bahn, and that the goal of the study was usability evaluation of a new registration form for the Deutsche Bahn website. The participants were offered either 10 Euros or one credit point reimbursement for their participation in the study. Most of the participants rated their IT knowledge highly: When asked to agree or disagree with the statement that their IT knowledge is good, 11 participants answered that they either “strongly agree” or “agree”, three neither agreed nor disagreed, and three disagreed with this statement. All but one participants answered either “strongly argee” or “agree” to the statement that their privacy is important to them (the remaining participant did not answer that question), and all but three answered “strongly agree” or “agree” to the statement that they take active measures to protect their privacy (out of the remaining participants, two neither agreed nor disagreed, and one did not answer the question).

Study Design. After welcoming the participants, the study consisted of two parts.

Registration. In the first part, the participants were told to fill in the registration form on the mock website that we have set up. At the beginning of the study every participant had to read the same study description, describing the goal of the study, namely, usability evaluation of the registration form. The participants were then told to register themselves using the modified registration form. It was furthermore stated, that no questions during the registration process are allowed in order not to interfere the process. Still, the participants were encouraged to think out loud during the registration. The registration process was completed when the participant clicked on the send button.

Follow-up Questions. The follow-up questions were asked in form of a semi-structured interview. After the registration, the participants were told that we have an exclusive access to their registered dataset to further discuss their perceived usability of the registration form. To obfuscate the real intention of our interview and to be in compliance with our communicated research goals, the introductory questions started with usability topics. Afterwards, based on the displayed data type fields, the participants were asked a set of questions about why they have or have not filled in the optional fields.

The study concluded by debriefing and gathering the demographic data.

2.2 Analysis Methodology

Our main research goal was to find out, which reasons for filling in the optional fields from the Preibusch et al. initial list [12] were mentioned by our participants, and whether there have been any reasons not on this list. For this purpose, the interviews were transcribed and analysed using qualitative semi-open coding approach. We took the list of Preibusch et al. as the pre-defined categories and classified the participants responses in the interviews according to these categories. In case we encountered responses that could not fit into the pre-defined categories, we assigned them to new categories. Each transcript has been analysed by two independent authors, and the findings were then discussed and agreed upon among the authors. The categories were supplemented with the quotes from the interviews, translated from German to English.

As additional research goals, we decided to consider the reasons that the participants gave for not providing their personal data to websites, and the countermeasures they used when a website requested some kind of personal data they did not want to disclose. For these goals, the interviews were analysed by two authors using open-coding approach, and the resulting categories were further discussed among the authors and agreed upon. As with the main research goal, we provide a quote supporting each one of the categories, translated from German to English.

3 Results

In this section we describe the evaluation results of our study.

3.1 Reasons for Filling in Optional Data Fields

We first describe the findings relevant to our main research goal. We provide the list of the pre-existing categories and specify whether we found any correspondences to them in our dataset. We further describe the new categories that were derived from our analysis.

Pre-existing Categories. We first describe the correspondences we found in our interviews to the list of Preibush et al. in [12].

Over-disclosure by Accident. Commonly, the users do not distinguish between optional and mandatory forms, either due to the website’s design or due to not paying attention to the clues that point that a field is optional. As such, a significant number of participants in our study reported not seeing the red star that appears only near to mandatory fieldsFootnote 3, and then mentioned that they would not have filled the data if they have seen that it is optional.

“It was not intended, I would not have filled it in if I did not think that I had to input it.”

Over-disclosure by Proxy. This item relates to the cases, where the autocomplete function of one’s browser ends up filling in more data than the user intended to. As the participants in our study used a lab computer to fill in the form, over-disclosure by proxy was not relevant for them.

Limit Disclosure is Costly. It has been suggested, that some users fill in all the fields in the form, since distinguishing between optional and mandatory fields requires too much time or effort, for example, if the website requires sending the filled form first before telling whether there is data missing in some of the mandatory fields. However, none of the participants named this reason for filling in optional forms explicitly.

Building Social Capital. The studies on websites that maintain a public or semi-public (i.e. open only to friends on social networks, or to recruiters on job hunting websites) have shown [7], that some users provide more data in their accounts in order to create a better image of themselves. In our study, however, the participants did not have to create a public profile of themselves, hence, they could not build social capital based on the data they provided. Therefore, as expected, none of them has mentioned this reason.

Expecting Monetary Return. The data provided by the users is often used by the companies to provide additional offers to the users such as personalised advertisements. Hence, it has been suggested that the users might input their data in order to be able potentially to benefit from such offers. A number of our participants mentioned, that they disclose such data as their date of birth, expecting special offers sent to them on their birthday, or expecting information on discounts tailored to their interests.

“Okay, it can also present a benefit, if I, for example, register myself somewhere or fill in some form, and in this way the personalized offers can be tailored to me. This can be an advantage.”

Note that although the participants interacted with the mock website, they did not attempt to surf the website in order to find the information about the exact benefits they might get from disclosing additional data. They also did not mention that they tend to research the potential benefits of data disclosure on other websites they use before they actually input their data on these websites. Hence, their expectations relied more on their reasoning and previous experience than on the information provided by the service prior to the data disclosure.

Expecting Non-monetary Return. Similar to monetary benefits, the companies might provide additional features to the user based on their input data, such as personalised recommendations of products or services or additional functionality. Some of our participants mentioned expecting such a non-monetary return in form of an additional functionality in exchange for providing additional data, such as getting phone notifications when the transport is late if the phone number is provided.

“...while booking a bus trip in Germany on the Internet, one has to input the phone number in order to be notified about the delays. And I see a benefit in this, that I leave my phone number, although generally I am reluctant. This would be an example where I see that it makes sense for me to leave my phone number.”

Similar to the expectations of monetary return, our participants neither attempted to find out whether the Deutsche Bahn provides additional functionality in exchange of disclosing optional data prior to the registration, nor did they mention researching potential benefits of data disclosure before providing their data on other websites.

Expecting Infrastructure Improvements. Preibusch et al. suggest, that the companies can use the information gathered from the users to better adjust their services to the demands of their customers. Hence, expecting such adjustments, the users might choose to provide additional data. However, none of our participants mentioned such motivation for disclosing data on web forms.

Acting Reciprocally/Altruistically. Studies have shown [11] that people who generally tend to act reciprocally also provided more data by filling in the fields in the study questionnaires. Since, however, our study focused on filling in the registration forms on commercial websites, it is not surprising that our participants did not mention the motivation to act reciprocally or altruistically as their reason for providing additional data.

Personality. Preibusch et al. suggest that for some users their personality might influence their decision to input more data, for example, if the user enjoys filling in the questionnaires. Indeed, the study in [12] included a significant number of participants who mentioned that they enjoyed participating in the surveys or find the activity of filling in the forms fun and interesting. However, none of our participants mentioned their personal preferences as a motivation for providing more data. It is worth noting, however, that a number of participants mentioned their personality traits as the reason not to provide their data on the websites, which we describe in Sect. 3.2.

New Findings. We further describe additional reasons mentioned by our participants but not included in the initial list in [12].

“It makes sense for them to request this information” Several users mentioned filling in the fields, that they expected to be mandatory, even though the fields were marked as optional. The expectations of the participants were either due to their previous experience with similar services, or due to their assumption that the particular data is required for the service functionality.

“Maybe for some... maybe at the Espirit online shop, there I would think, why are they interested in my date of birth, they are only interested in what I order. [...] They do not need to know my date of birth. And here I thought, that it might be relevant for ordering the train ticket. I would relate the date of birth to the registration.”

“Country of origin... I saw that it is not mandatory... I deliberately filled it in, because I think that this is an important category for the classification. This was just my interpretation.”

“No, I think, when I fill something in, do they really need this, or not? And all that I filled in is important... so, in my opinion.”

As with the case of expecting monetary or non-monetary return from providing additional data (see Sect. 3.1), the participants in this category relied on their own reasoning in deciding whether the requested data is indeed required by the service instead of attempting to get this information from the service provider itself.

“I trust that they have their reasons for requesting this information” Similar to the previous category, some participants claimed to disclose optional data if they believed that there was a good reason for the service provider to request the information. However, while the participants in the previous category based their beliefs on their own reasoning, others relied more on their trust in the service provider to use their data responsibly.

“Now, for example, I have an airline in mind, they need some data in any case. I do not have any problems with it, since I trust that the data stays confidential with them.”

Similar to the previous category, the participants neither attempted did not attempt to find out the reasons why the service collects the requested data.

Filling out Fields as Default Behaviour. Some of the participants claimed, they generally tend to fill in all the forms on the website, unless they have a particular reason not to. While these claims can be considered close to the pre-existing categories “over-disclosure by accident” and “limit disclosure is costly”, we still decided to categorize them separately, since the participants neither claimed to overlook the indicator and disclose more than they intended to, neither mentioned making a conscious decision to save time or effort by filling in all available fields.

“I just did not see any disadvantage, so I thought, I fill this in.”

In particular, some stressed that they would disclose the information if the website is trusted.

“And the fields I do not fill in, these are, for example, address stuff, but I have no concerns with the Deutsche Bahn.”

3.2 Other Findings

We describe the findings for our additional research goals, namely, by providing an overview of the reasons that our participants mentioned for not disclosing their personal data, and the countermeasures they mentioned using when confronted with the request to share more data than they wanted.

Reasons for Not Filling in Optional Data Fields. The responses of the participants who were reluctant to share their data can be grouped into two categories.

Personal Feelings. A number of participants mentioned that they did not share their data due to their personality, or because they “had a bad feeling” sharing more than they considered absolutely needed. As such, this group focused on their subjective feelings and personal preferences:

“I do not like disclosing it, but this really a very personal and subjective thing!”

Concrete Threats. Another group mentioned specific threats that they wanted to protect themselves against, such as spam mails or phone calls, or identity theft:

“I do not like it when people just call. I have experienced this a couple of times, that someone just calls me, and I do not like it.”

“Some [companies] really try [to protect the data], but then it’s like, yeah, we have been hacked, or... and this is just great. Then they have all the data, all the credit cards... this did not yet happen to me, but... this is why I do not have a lot to do with the Internet services.”

Countermeasures. We asked our participants what would they do if there is a registration form on some website with mandatory fields that require data the participants do not want to disclose. The responses can be grouped into following categories:

Boycott the Website. The most obvious solution mentioned by several participants was that they would refuse to use a website, if it required data considered too private by the participants. In particular, looking for alternatives that provide a similar service but either require less data or are more trusted not to misuse the collected data has been mentioned:

“I already had this, that I wanted to register, for example, in the online shop, and then I did not want to fill in the data. And then I did not register, and bought it at Amazon for a couple of euros more.”

Input Fake Data. A solution also mentioned by our participants was to input fake data, if the real data is considered too private to disclose. The types of data that is faked, as well as the settings in which fake data is given, varies. As such, a number of participants mentioned that they input fake data often, aside from the situations when it could hinder the functionality offered by the service:

“So I am always the one who under circumstances also inputs fake data, when it does not suit me. This is possible.”

Some have mentioned that they are reluctant to input fake data into the websites owned by governmental institutions:

“Actually, always, except for, I would say, official institutions, where it has to be correct.”

Another approach that has been mentioned in the interviews was to input fake data, which, however, is not misleading. One particular example is the date of birth: as the website’s intention is to find out, whether the user is older than 18, the specific age is assumed to be irrelevant, hence, fake data can be given.

“There is the Rotkäppchen sparkling wine, and when one goes to this website, then one has to input the date of birth. Maybe minors under 18 years old are not allowed to visit the website. So I could imagine. And when I look at something on the website, then I just click on some number. I mean, I am not under 18, but I just click something, since it does not matter whether I am 30, or 40, or 50 years old, for me to go there. ”

It is worth noting, however, that a number of participants claimed that they never input fake data due to their personality traits.

“No, I am very honest.”

Avoid Registration, but Still Use the Service. One possible solution to avoid filling in unwanted web form fields was to look for the ways to use the website functionality without registration.

“I actually never register, and continue without login. [...] They do not know who I am, what my name is, where I live and so on, and I do not have to remember any login and can always do that in another way, so to say.”

Use Throw-away Contact Information. The reluctance to fill in contact information has been often mentioned by our participants, either due to privacy reasons, or in order to avoid unwanted advertisements. Hence, in order to be able to register on the websites, that demanded the user’s e-mail address, some of the participants mentioned registering a separate address just for the registration purposes, that is not checked as often as their regular address.

“The e-mail address is in any case a second email address, so it is not an important one. When too many junk gets there, then it will not be read.”

“Then one can have a spam e-mail. Then they can spam me as they want, that does not bother me.”

While the practice of using throw-away phone numbers appears to be much less frequent than using throw-away e-mail adresses, it has been mentioned as well. In particular, one of the participants reported registering a phone number from an Internet phone company, so that the calls to this number went to the participant’s email instead of going to their regular phone.

“When one has to input the phone number as a mandatory field, then I often input a Sipgate phone number, that lands in a normal mail box. [...] This is a Voice-over-IP phone number, there I get at most an e-mail, when someone calls it. But my mobile phone does not ring.”

4 Related Work

For describing the related work we focus on research that studied the factors that influence the data disclosure of the users and the tools that aim to prevent the users from disclosing too much data. We furthermore describe the works in other domains that study the mental models of the users and the motive for their behaviour concerning various security mechanisms.

Reasons and Factors that Influence Data Exposure. A number of studies focused on the topic of web forms and optional fields. As such, Preibusch et al. conducted a quantitative user study in order to study the users [12] and construct our original list of reasons to expose their data. They also conducted a user study trying to gauge additional responses, but the context was also limited (the users thought that the purpose of the study was to gather and analyse their data). Their further findings include quantitative analysis whether users are likely to fill in optional fields, whether the presence of mandatory fields increases their likelihood to enter data and how long does it take to enter data.

The personality traits of the users that influence their data disclosure have been the topic of several studies. As such, Egelman [5] studied the personality traits that help predict the decision making and risk-taking attitudes of the users. The focus of other studies was more specific. As such, Adams et al. [2] studied the trade-offs that the users consider acceptable for disclosing their personal data, and Ackerman et al. studied the users attitudes towards providing data in e-commerce [1]. The study in [11] focused on the dependencies between the personality traits such as fairness or desire to act reciprocally and filling in the forms. All those studies strengthen the assumption that attitudes and personality traits should be more focused when trying to understand differences in privacy behavior.

Other researchers studied disclosure of personal data on social lending sites [4]. They argue that this exposure is related to the theory of descriptive social norms. It means that either the similarity of context, social proximity, and mimicry of success factors leads to people exposing their data because of social norms and less because of rational decisions. Kramer conducted a similar study where they look at the specific privacy in Facebook [8]. Furthermore Korff et al. studied the effect of differences in the choice amount by changing the number of chechboxes and choice structure by varying the sensitivity of personal data items presented on privacy behavior [7]. They expect the amount and the structure to have a similiar effect on the privacy behavior compared to all day decisions like shopping. Acquisti et al. studied the extent to which the users are ready to sacrifice their privacy in exchange of a monetary return.

Tools that Prevent Data Exposure. A number of researchers focused on the development of different tools to support more privacy-aware behaviour of the users in the process of filling out web forms. Knijnenburg et al. conducted a study where they compared new and more detailed forms of auto-completion tools with a traditional one [6]. The main purpose was to revive the privacy calculus for filling out web forms. They proclaim that users may skip this privacy calculus out of convenience and therefore use the traditional auto-completion tools. Krol et al. developed a tool for alerting users when they are about to fill in an optional form [9], thus making people more aware of unnecessary data exposure.

Mental Models of Privacy-Preserving Behaviour in Other Domains. Aside from web forms, a number of papers studied the reasons why the users do not engage in privacy-preserving behaviour in various domains. As such, a qualitative study have been conducted by Renaud et al. [13] in order to derive the mental models of users regarding e-mail encryption. The study in [3] researched the reasons mentioned by the participants for not using password managers, and the study in [16] considered the reasons that prevent smartphoned users in engaging in various secure behaviour such as setting a screen lock or installing an anti-virus software. A general overview of mental models in security is provided by Volkamer et al. in [15], stressing that understanding the mental models and comprehension of security mechanisms of the users is cruical in supporting the users in their privacy-related decisions.

5 Conclusion

In this chapter we summarize our findings, as well as discuss their implications and possible directions of future work.

5.1 Summary

As the web-based services attract more users, the websites also tend to gather more personal data. The users are seldom provided an explanation on what the purpose of the data collection is, and often the website design makes it hard for the user to notice, which data is mandatory to provide for using the service. Hence, users result in filling in the optional fields on the website forms, providing more personal data than needed for their intentions.

We have conducted a study to find out the reasons, why the users fill in optional fields on the websites. We based our assumptions on what these reasons are on existing literature, namely, on the list provided in [12]. The reasons on this list, however, were either not confirmed in an empirical study at all, or the study was done in a specific context (such as the study of user’s behaviour on social networks or providing data for a research survey) which is not directly transferable to other types of websites and services. Our study focused on finding out whether the aforemendtioned reasons would be relevant for the scenario where the users have to fill in the registration form on a website of a company that provides commercial services, which is one of the most common contexts encountered on the web. In our study we asked the participants to register an account on our mock registration website, which, as they were told, belonged to the Deutsche Bahn company (German Railways) that assigned our research group to conduct a usability study of their new registration form. After the participants registered an account, they were asked to explain what data they decided to share and why.

We found, that only three out of ten reasons from the initial list in [12] were mentioned by our participants when asked to explain why they filled in the optional fields in the forms. Namely, the reasons that were mentioned by our participants were over-exposure by accident (i.e. not being able to notice an indicator that shows whether a field is optional or mandatory), expecting monetary return (e.g. special birthday offers, if the date of birth is provided) and expecting non-monetary return (e.g. a phone notification for a delayed transport, if the phone number is provided).

We have further identified three categories that were not explicitly present in the initial list by Preibusch et al., but mentioned by our participants. In the first category, the participants decided to fill in optional fields because they believed that the service required the particular data in order to provide the necessary functionality. Despite the fields being marked as optional by the service, the participants in this category strongly relied on their own reasoning to decide, whether it makes sense for the service to request a particular piece of data, hence, whether they should provide this data. The second category, on the other hand included the statements from the participants that generally relied on trust in the service. Even if the participants noticed that some fields were optional and they could not themselves think of a good reason for the service to require some particular data, they still decided to fill in these fields since they trusted that the service would not request the data unless it had a good reason to do so. The third category consisted of the statements that concerned the default behaviour of the users. Especially if the service itself was found trustworthy, the participants decided to fill in all the fields, since they saw no disadvantage in doing otherwise.

Further findings indicate, that many of our participants were reluctant to disclose their data, due to either concrete concerns of data misuse, or a general feeling of uneasiness. Moreover, we provided a list of countermeasures that the participants would use if the website requests some data they are not comfortable sharing, such as providing fake data, registering a separate e-mail which the user rarely checks for providing it on the website or boycotting the website entirely.

5.2 Discussion and Future Work

Our findings indicate following factors that determine whether the users are likely to input their optional data. The first factor is the users’ trust that the service would not collect data without good reason, is unlikely to misuse it and is capable of ensuring its security against external attacks. The second important factor is transparency meaning that it is be important for some users to understand what their data is used for before they decide to disclose it. Note, that the factor of transparency has also been found relevant in privacy-related decisions in other domains, such as in deciding to install a smartphone app if the permissions that the app requests make sense to the user [10]. The final factor is awareness, meaning that many provide more personal data then they would want to, only because they did not notice an option to do otherwise.

Note that all these factors are reflected in the EU General Data Protection Guideline [14] (GDPG). As such, Art. 5 states, that “Personal data should be [...] collected for specified, explicit and legitimate purposes and not further processed in a manner that is incompatible with those purposes; [...] adequate, relevant and limited to what is necessary in relation to the purposes for which they are processed; [...] processed in a manner that ensures appropriate security of the personal data, including protection against unauthorised or unlawful processing and against accidental loss, destruction or damage, using appropriate technical or organisational measures (‘integrity and confidentiality’)”, which corresponds to the factor of trust as expressed by our participants. The guideline further demands that the users are provided with “the purposes of the processing for which the personal data are intended as well as the legal basis for the processing” (Art. 13), which corresponds to the factor of transparency. As our results show, our participants relied on their expectations of which benefit they would get from disclosing additional data, or what the purpose of collecting specific information was, instead of attempting to find out this information from the service itself. Still, they were more likely to disclose the data if they could think of a purpose behind its collection. The factor of awareness is addressed with the guidelines requiring the consent of the users for data processing (Art. 6). The guideline defines consent in Art. 4 as “any freely given, specific, informed and unambiguous indication of the data subject’s wishes by which he or she, by a statement or by a clear affirmative action, signifies agreement to the processing of personal data relating to him or her”. The participants in our study, on the other hand, overlooked the information on the website, thus providing more data than they would otherwise do.

Given our findings related to the GDPG, an important direction of future work is the improvement of communication between the service providers and the users. As such, as trust in the service provider has been shown to be an important factor for the decision making of the users, tools for trust assessment (e.g. in form of an evaluation, to which extent the service provider complies to the GDPG) and communication, possibly from independent institutions, would be helpful in supporting the users. Furthermore, our study has shown that the users rely on their considerations on what the potential benefits of their data disclosure would be, or how the service could use their data, while deciding which data to disclose. Hence, input from the service provider with this information can help the users make a more informed decision. Finally, as a number of users tend to overlook the indicators for optional fields, providing more data than they would want to, more visible indicators on the website would make sure that accidental disclosure without the users explicit consent is minimized.

Our study has further shown, that there is a discrepancy in the participants attitudes towards data disclosure. As such, while some of the participants filled in the data without having any concerns, others claimed being reluctant to disclose their data. Given that all the participants had to interact with the same website, it would be interesting to investigate the further differences between those two groups that influence their decision making and attitudes towards data disclosure. Furthermore, an interesting direction of future work would be investigating other contexts in which the users have to decide whether to disclose data. As such, it would be interesting to compare, whether the user behaviour and reasons for either disclosing or not disclosing data differ while interacting with a trustworthy website such as a well-known Deutsche Bahn company, as opposed to interacting with a small and unknown online shop or other service that might be deemed less trustworthy by the participants.

The prevalence of various countermeasures, such as using fake data, that the participants use in order to avoid filling in the mandatory fields shows that the reluctance of sharing personal data, even at the expense of the user’s convenience, is a significant factor in decision making for many users. These findings suggest that collecting too much data without providing a sufficient explanation can be detrimental for the web services as well. On the other hand, since the countermeasures mentioned by our participants are not an optimal solution for every user, better tools for supporting the users who do not want to disclose their personal data are needed.