TL;DR
This paper presents a longitudinal survey study investigating public attitudes toward automated COVID-19 contact tracing. The researchers found that while a majority of people are willing to adopt these technologies under "perfect privacy" conditions, trust evaporates quickly when data is shared with governments or managed by companies with poor privacy reputations.
What Problem Does This Paper Address?
To effectively suppress a pandemic like COVID-19, epidemiological models suggest that automated contact tracing apps require adoption rates of roughly 60–70%. However, these technologies sit at a tense intersection of public health necessity and individual privacy.
The authors address a critical gap: the "human factor" in technical design. While engineers were debating the merits of centralized vs. decentralized architectures (e.g., GPS vs. Bluetooth proximity tracking), there was little empirical data on what the public actually felt comfortable with. This paper seeks to understand the privacy values, concerns, and trust levels that drive a user's decision to install—or delete—a contact tracing app.
Methods
The researchers conducted a sequence of online surveys using the Prolific platform, starting in early April 2020.
- Structure: Each survey iteration involved approximately 100 participants.
- Protocol: To avoid "priming" participants to be overly privacy-conscious, the survey initially avoided the word "privacy," instead asking about "comfort" levels and "likelihood" of use across various hypothetical scenarios.
- Mixed Methods: The study combined quantitative 5-point Likert scales with qualitative free-response questions to capture the "why" behind participant sentiment.
- Scenarios: They tested variables including data sources (cell tower, GPS, proximity, credit card history), data-sharing entities (governments, tech giants, universities), and specific app features (notifications vs. public databases).
Datasets
The study relies on original survey data collected from participants primarily residing in the UK (26%), the US (16%), and various European countries (46.5%).
To contextualize the findings, the authors reference the global state of the pandemic during the initial survey window (April 1 to May 1, 2020), as shown in the table below:
Key Insights and Ideas
- The "Trust Gap" in Tech: Participants showed a clear hierarchy of trust. Mapping apps (Google Maps, Apple Maps) were viewed far more favorably for contact tracing than social media or messaging apps (Facebook, TikTok). Users perceived mapping apps as having a "legitimate" reason to already possess location data.
- Proximity Tracking isn't a Privacy Silver Bullet: Interestingly, the public did not necessarily view Bluetooth-based proximity tracking as more private than GPS location tracking. Some participants even viewed the "automatic exchange" of data between phones as a new security risk or a "subtle" invasion of privacy.
- Distrust in Government Permanence: A staggering 72% of participants doubted their governments would delete the data after the pandemic, and 65% doubted the data would be used only for COVID-19 mitigation.
- The "Police State" Threshold: While users were somewhat open to app-based tracking, they were overwhelmingly uncomfortable with "passive" surveillance methods like analyzing credit card histories (51-63% uncomfortable) or surveillance camera footage (69% uncomfortable).
Key Results
- Adoption Potential: Under "perfect privacy" conditions, 72% of participants said they would be likely to download a contact tracing app.
- Privacy Degradation: When the scenario changed to "app makers know location but don't share it," likely adoption dropped to 55%. If that data were shared with the government, adoption fell further to 49%.
- Feature Preferences: Users strongly desired features that notify them of exposure (a, b) but were split on more invasive features, such as automatically notifying authorities if a user breaks quarantine (g).
- Institutional Trust: University research groups were the most trusted entities to develop these apps (74% comfort), though the authors acknowledge a potential response bias given the survey was hosted by a university.
Limitations and Future Work
The authors emphasize that these results are preliminary and were released quickly to inform the rapid development of contact tracing protocols in mid-2020.
Key limitations include:
- Sample Size: Each weekly batch was limited to 100 participants, which may not be fully representative of the general population.
- Platform Bias: Prolific users may be more tech-savvy or have different privacy profiles than the average citizen.
- Rapidly Changing Sentiment: Public opinion in a pandemic is volatile; the authors noted that willingness to share data with the government dropped significantly between April and May.
Future work involves deeper statistical analysis of the longitudinal data to identify correlations between demographic factors (like political leaning or COVID-19 risk level) and privacy preferences.
Summary updated