Americans’ perceptions of privacy and surveillance in the COVID-19 Pandemic

OBJECTIVE
To study the U.S. public's attitudes toward surveillance measures aimed at curbing the spread of COVID-19, particularly smartphone applications (apps) that supplement traditional contact tracing.


METHOD
We deployed a survey of approximately 2,000 American adults to measure support for nine COVID-19 surveillance measures. We assessed attitudes toward contact tracing apps by manipulating six different attributes of a hypothetical app through a conjoint analysis experiment.


RESULTS
A smaller percentage of respondents support the government encouraging everyone to download and use contact tracing apps (42%) compared with other surveillance measures such as enforcing temperature checks (62%), expanding traditional contact tracing (57%), carrying out centralized quarantine (49%), deploying electronic device monitoring (44%), or implementing immunity passes (44%). Despite partisan differences on a range of surveillance measures, support for the government encouraging digital contact tracing is indistinguishable between Democrats (47%) and Republicans (46%), although more Republicans oppose the policy (39%) compared to Democrats (27%). Of the app features we tested in our conjoint analysis experiment, only one had statistically significant effects on the self-reported likelihood of downloading the app: decentralized data architecture increased the likelihood by 5.4 percentage points.


CONCLUSION
Support for public health surveillance policies to curb the spread of COVID-19 is relatively low in the U.S. Contact tracing apps that use decentralized data storage, compared with those that use centralized data storage, are more accepted by the public. While respondents' support for expanding traditional contact tracing is greater than their support for the government encouraging the public to download and use contact tracing apps, there are smaller partisan differences in support for the latter policy.

: Distribution of survey weights. Target proportions for age, gender, region, race, and income come from the 2018 American Community Survey. Weights were calculated using iterative proportional fitting a nd t rimmed a t t he 5 th a nd 9 5th p ercentiles. M issing d emographic i nformation was i mputed f or the purpose of generating sample weights using multivariate imputation by chained equations (MICE).  Fig. 3: Demographic predictors of support for surveillance policies, measured on a scale from 0 to 100. Each respondent was asked to rate their support for four policies: traditional contact tracing and three randomly selected among the remaining options. Plot shows coefficients from a single OLS regression of support for any surveillance policy on all categorical demographic characteristics listed, with 95% confidence intervals. Standard errors are clustered at the level of the respondent (N = 2, 511). Baseline categories are as follows: male (gender), white (race), Independent/other (party), 18-29 (age), less than $50k per year (income), did not graduate high school (education), Midwest (region).    Outcome is binary, coded as "likely" (1) if the respondent answered 60 or above on a continuous scale when asked how likely they would be to download the app described, and zero (0) if the respondent entered a number below 60. Question was asked of all respondents who reported owning a smartphone (N = 1, 883).

No information
After vaccine found  Fig. 9: Marginal mean probability a respondent is likely to download the app for each conjoint attribute value. Outcome is binary, coded as "likely" (1) if the respondent answered 60 or above on a continuous scale when asked how likely they would be to download the app described, and zero (0) if the respondent entered a number below 60. Question was asked of all respondents who reported owning a smartphone (N = 1, 883)        Respondents were asked each of these questions after reading a description of a hypothetical contact tracing app. For the first three questions, respondents were asked to answer on a scale from 0 to 100. Respondents were then asked about their confidence t heir p ersonal d ata would b e p rotected w hen u sing t he a pp ( 0 i ndicates " not a t a ll confident" and 3 indicates "very confident). T he l ast f our q uestions a sked r espondents t o i ndicate t heir a greement with a provided statement (-2 indicates "strongly disagree" and 2 indicates "strongly agree"). 25 Table 11: Association between trust in institutions and support for surveillance policies. We report results from two regression models where we predict support for the surveillance policies using responses to questions related to trust in institutions or actors. The "mean trust in institutions" variable is the average level of trust respondents have in the three institutions they were randomly assigned to evaluate. Both linear regression models control for the type of policy respondents were asked to evaluate. Model 2 also includes demographic variables including gender, race, party identification, a ge g roup, i ncome l evel, l evel o f education, and geographic region.
(1)   Pearson correlation Fig. 16: Pearson correlation between trust in institution or actors and support for surveillance policies. For the institutions or actors that are not marked with an * , respondents were presented with three randomly selected actors and asked much confidence they have in each to act in the best interest of the public using a 4-point scale (0 = no confidence at all; 3 = a great deal of confidence). For the institution or actors marked with an * , respondents were asked how much they trust advice related to COVID-19 coming from that institution or actors using a 5-point scale (-2 = distrust a lot, 2 = trust a lot); respondents had to evaluate all three institution or actors. 0.032 N 2,057 * p < 0.05, ** p < 0.01, *** p < 0.001   Percentage who responded 'yes' Fig. 17: Respondents' predicted outcomes of the surveillance policies. For each surveillance policy the respondent read about, they were presented with nine outcomes (the order was randomized) and asked if they think each outcome would happen if the policy were adopted. The answer choices included "yes," "no," and "don't know." The figure p resents t he p ercentage o f r espondents w ho a nswered " yes" f or e ach p olicy and outcome. Respondents were presented with the following five s tatements s uch t hat t he o rder o f t he s tatements were randomized. They were asked to select all the statements that are true of the hypothetical contact tracing app that they read about. Respondents could refer to the description of the app (available as collapsible content) while answering the manipulation check question.

Statement
Proportion answered correctly The app will track your location data. 0.65 The app will send you the names of infected people you have been in close contact with.

0.69
All user data will be stored on a central server.

0.45
Apple and Google are building this app. 0.77 Public health experts say that at least 60% of smartphone users needs to use this app for it to be effective at limiting the spread of COVID-19. 0.67