Risk preferences and risk perception affect the acceptance of digital contact tracing

The COVID-19 pandemic highlighted the crucial role of non-pharmaceutical interventions, including digital contact tracing applications (DCTAs), in controlling disease spread. DCTAs leverage Bluetooth technology to anonymously record proximity between devices, alerting users of potential exposure to infected individuals. While DCTAs offer a powerful tool to mitigate pandemics and avoid harsh lockdowns, their effectiveness hinges on widespread adoption. However, in Western countries, adoption rates remain low, largely due to privacy concerns. Understanding these low adoption rates is vital for policymakers seeking to promote DCTA use as a less restrictive alternative to widespread lockdowns. This study aimed to identify the psychological factors driving or hindering the acceptance and compliance with DCTAs. The research focuses on the interplay between individual risk perception and preferences, and broader social preferences and values in shaping DCTA usage. The study uses data from a representative sample of the German-speaking Swiss population to investigate the predictive power of psychological factors on both DCTA acceptance (willingness to use and recommend) and compliance (following DCTA recommendations).

Existing research demonstrates a link between risk mitigation behaviors during pandemics and individuals' risk perception and preferences (Van der Pligt, 1996; Weber and Milliman, 1997). Studies have shown that knowledge about the virus and perceived risks are crucial determinants of engagement in preventive behaviors (Kwok, 2020; Zhong, 2020; Abdelrahman, 2020; Betsch et al., 2020). Individual differences in risk aversion also play a role, with risk-averse individuals being less likely to engage in risky behaviors. The use of DCTAs involves potential data security risks, which are a significant concern for many (Ienca and Vayena, 2020). Beyond individual factors, the societal perspective highlights the common-pool resource aspect of public health systems, where DCTA usage represents cooperation to prevent system overload. Social preferences, particularly prosociality, have been shown to influence preventive health behaviors during the pandemic (Campos-Mercade, 2021; Dryhurst, 2020; Zettler, 2021). Trust in government and identification with communities also influence cooperation in common-pool resource dilemmas. The potential divergence between individual and societal perspectives is also noted; a decision that appears rational from an individual risk perspective may have negative consequences for public health.

This preregistered study employed a nationally representative online survey of the German-speaking Swiss population (N=757, after excluding 91 participants due to low data quality). Data were collected in June 2020, shortly after the launch of the Swiss federal contact tracing application, 'SwissCovid'. Participants received a small remuneration for completing the survey. The study measured DCTA acceptance and compliance using four-item scales assessing willingness to use and recommend, perceived effectiveness, and data security (for acceptance), and willingness to self-isolate, report a positive diagnosis, contact the hotline, and get tested (for compliance). Social preferences were assessed using standardized measures including social value orientation, honesty-humility (HEXACO), and identification with local versus global communities (IWAH). Risk preferences were assessed using items from the German SOEP, supplemented with an item for data security risks. COVID-19 risk perception was measured using three items concerning infection and severe illness rates in Switzerland. Covariates included demographics, DCTA comprehension, technology affinity, and support for pandemic political measures. Bayesian linear regressions were used to test the preregistered hypotheses, employing Bayesian projective predictive model selection to identify the most parsimonious predictive model. Missing data on income and wealth were imputed using the sample medians, a step that was not part of the preregistered analysis plan.

The study found high compliance (mean score of 4.34 on a 1-5 scale) with DCTA recommendations, but comparatively low acceptance (mean score of 3.75 on a 1-5 scale). Bayesian regression analyses revealed that risk perceptions and preferences, rather than social preferences, were the most significant predictors of DCTA acceptance. Specifically, higher health-risk perception and lower data-security-risk perception were positively associated with DCTA acceptance. Conversely, perceiving COVID-19 as a high economic risk was negatively associated with acceptance. Regarding risk preferences, higher aversion to health risks and greater tolerance for data-security risks increased DCTA acceptance. Social preferences (honesty-humility, social value orientation, and identification with local/global communities) had minimal impact on DCTA acceptance, as indicated by a Bayes factor favoring a model excluding these variables (BF = 20,271,851). The model including comprehension of DCTAs and support for political measures showed that these factors had higher effects on DCTA acceptance than risk perceptions and preferences. Around half of the respondents correctly understood absolute numbers of COVID-19 related deaths and infections. A majority (92%) viewed COVID-19 as a global problem, while fewer (42%) saw it as a problem for themselves or their immediate surroundings. Age and gender differences were observed in mental well-being, support for political measures, and DCTA acceptance and comprehension.

The findings highlight the dominance of individual risk considerations over social factors in shaping DCTA acceptance. The strong positive association between health-risk perception and DCTA acceptance aligns with previous research, emphasizing the need to effectively communicate the personal health risks of COVID-19 to encourage uptake. The negative association between data-security concerns and acceptance underscores the importance of transparently addressing data privacy issues. The unexpected negative correlation between perceived economic risk and DCTA acceptance might stem from individuals prioritizing economic stability over pandemic mitigation or a general societal divide between those focusing on health versus economic impacts. The substantial impact of DCTA comprehension and support for political measures suggests the necessity of clear communication campaigns and building public trust in governmental pandemic management. The high compliance levels, despite low acceptance, indicate a willingness to follow public health guidelines when alerted, but this compliance is undermined by low application usage. Therefore, interventions targeting both risk perception and DCTA comprehension are essential.

This study demonstrates the critical role of individual risk perceptions and preferences in influencing DCTA acceptance, highlighting the need to tailor communication strategies emphasizing personal health risks and data security. The study's findings provide valuable insights for policymakers seeking to increase DCTA adoption and thus enhance pandemic mitigation efforts. Future research should explore the effectiveness of various interventions—such as choice architecture, graphical risk communication, social comparison, and risk contextualization—in improving DCTA uptake.

The cross-sectional design limits causal inferences, and the reliance on self-reported data may be subject to biases. The study was conducted in a specific cultural context (German-speaking Switzerland) and may not fully generalize to other populations. The imputation of missing income and wealth data might have introduced some bias into the analysis. While the covariate selection process was robust, it's important to note that certain interactions might have been overlooked.