TL;DR
This paper explores how people value their personal information in real-time, contextual situations using a smartphone app. It reveals that while people can assign monetary values to their data, these valuations are highly sensitive to context and prone to cognitive biases like anchoring and endowment effects, suggesting a complex landscape for any future "information bazaar."
What Problem Does This Paper Address?
The proliferation of smartphones and apps has created a massive ecosystem where personal data is constantly harvested and cataloged. Companies like Google and Facebook leverage this data for targeted advertising, but users often lack a clear understanding of the value of their own information. Existing research has largely focused on static or future-oriented valuations, failing to capture the dynamic, in-the-moment perception of data value. This paper aims to bridge this gap by investigating how individuals perceive and price their personal information in real-time, contextual scenarios, moving towards understanding the feasibility of a regulated marketplace for personal data.
Methods
The core of this research is a large-scale, contextual study conducted over ten days using a custom-built smartphone application. This app collected data on user activities, location, and companionship, and crucially, prompted users to assign monetary values to their personal information. To address potential biases in valuation, the study employed four different experimental versions of the app, each using distinct methods to elicit willingness to pay (WTP) and willingness to accept (WTA). These methods included open-ended price fields and sliding scales with varying price ranges (£5 and £10). The study also controlled for factors like income and prior valuations. Statistical analyses, including chi-square tests and ANOVA, were used to examine the influence of elicitation methods, demographic factors, and situational contexts on the assigned values.
Datasets
The study collected data from 101 participants who provided at least one valuation. The dataset consists of over 1,400 individual valuation entries, each associated with the user's current activity, location, and company. This data was collected in real-time via the smartphone application. No pre-existing datasets were used for training or evaluation; instead, the paper introduces a newly collected dataset of contextual personal information valuations.
Key Insights and Ideas
- Context is King (but not always): While situational context (activity, location, company) was expected to heavily influence data valuation, the study found surprisingly weak main effects. However, the interaction between situation and the elicitation method was significant, suggesting context matters more in how biases manifest.
- Biases are Pervasive: Participants exhibited clear signs of anchoring and endowment effects, meaning their valuations were influenced by the way questions were framed and by the perceived ownership of their data.
- WTP vs. WTA Matters: Participants were more likely to consider their information "free" when asked their willingness to pay to restrict access (WTP) compared to when asked their willingness to accept payment for access (WTA), highlighting the endowment effect.
- Demographics Play a Role: Older individuals were less likely to offer their information for free, and women were more likely to either provide information for free or not sell it at all.
- Social Situations Dampen Biases: Interestingly, in social situations (e.g., "being out and about" with friends), participants showed less susceptibility to anchoring and endowment effects, suggesting that shared experiences might lead to more stable valuations.
Key Results
- In 79.4% of cases, the demanded price for personal information did not exceed £10.
- 20.4% of valuations were zero, indicating a willingness to provide information for free.
- A chi-square test showed that different elicitation methods led to different valuation categories, though the influence was weak (uncertainty coefficient of .04).
- Women were more likely to provide information for free or not sell it compared to men.
- Older participants (>40) were less likely to provide information for free than younger participants.
- An ANOVA revealed a significant main effect of elicitation method and a significant interaction between situation and elicitation method on price demanded. Income was also a significant covariate.
Limitations and Future Work
The authors acknowledge that the study is a "first look" at the data and that further qualitative and experimental analyses are needed for deeper interpretation. The rough categorization of activities and company types might have masked more complex situational patterns. The mobile application was only compatible with Android devices, potentially biasing the sample. Future work could explore more granular situational contexts and investigate how these findings can inform the design of a regulated "information bazaar" where users can truly control and monetize their data. The study also emphasizes that while users can attach a price, this doesn't necessarily equate to a willingness to sell, underscoring the need for further research into the nuances of data exchange.
Summary updated