Stocks, flows, and risk response to pandemic data.

Abstract

During the coronavirus disease 2019 (COVID-19) pandemic, data regarding new infections were commonly presented and used to guide policy decisions (e.g., whether to close schools) and personal choices (e.g., whether to dine at a restaurant). In this manuscript, we highlight a critical aspect of pandemic data that can pose a challenge for people trying to reason about it. Data on infections-like much time series data-can be presented as either stocks (the total number of cases) or flows (the number of new cases over some interval). We show that seeing the same data presented in one format versus the other can shift judgments of risk and behavioral intentions. Specifically, when participants were shown data that depicted the number of new cases each day (flow) decreasing, they judged the current risk of COVID-19 to be lower than participants who were shown the same data as the total (cumulative) number of cases (stock), whichby its nature-continued to increase. Risk appraisal, in turn, predicted a wide array of behavioral intentions (e.g., likelihood of dining indoors at a restaurant). Thus, the choice of how to present pandemic data can lead people to different conclusions about risk and can have practical consequences for risky behavior. Public Significance StatementSeemingly minor visualization choices regarding how to present pandemic data can affect the public's judgments of risk and corresponding behavioral intentions. When the number of new coronavirus disease 2019 (COVID-19) cases is in a period of day-to-day decline, people judge the risk of COVID-19 to be greater when shown the data as cumulative totals (over time) compared to when the same data are presented as number of new cases (over time).

Linear Biases and Pandemic Communications

“…Although the total frame would exhibit acceleration, deceleration, and reacceleration while monotonically increasing, the rate frame would exhibit multiple reversals in the sign of the recent slope. Although the judgments and perceptions studied here likely influence intentions and behaviors (e.g., risk perceptions predict behavioral intentions 6,11 and affect adherence to COVID-19 protective instructions 24 ), research has identified attitude-intention and intentionbehavior gaps that may complicate public health responses, 25,26 so future research is needed to better understand how these translate into behavioral intentions and actions. In addition, the repeated measures in these studies occur in 1 session rather than over the course of several months or quarters and involve a ''well-behaved'' epidemic rather than a noisy one that exhibits jagged case counts (like the COVID-19 epidemic's unreliable weekend case counts); although in the present studies, these features provide the benefits of avoiding attrition and comparing estimates with counts that are uncontaminated by measurement errors, generalizing these results to judgments made with large time gaps between them and with noisier data may be useful.…”

“…These risk perceptions are especially important as they relate to downstream intentions to engage in behaviors with higher transmission risks, such as gathering indoors and visiting service providers. 11 However, stocks and flows does not fully characterize their forecasts of rates and totals in an epidemic, and even these forecasts have been linked to behavioral outcomes such as social distancing, mask wearing, and hand washing. 6 Exponential growth bias describes total forecasts well, with one caveat: estimated totals are higher based on a total (v. rate) frame even prepeak.…”

Risk perception, decision-making, and risk communication in the time of COVID-19.

“…For instance, Garfin et al (2021) found that perceiving greater risk of catching and dying from COVID-19 was associated with greater frequency of social distancing. Reinholtz et al (2021) found that greater risk perception was associated with lower likelihood of engaging in risky behaviors (e.g., dining indoors at a restaurant). In addition, although some studies found fairly high compliance with precautionary recommendations (Garfin et al, 2021), some found that intentions varied according to factors related to vulnerability to illness such as age, with older people indicating stronger intentions to get vaccinated (Reinhardt & Rossmann, 2021), and engage in social distancing and mask wearing (Garfin et al, 2021; Joslyn et al, 2021) than younger.…”

“…Similarly, different visualizations of the same data can influence both risk perception and behavioral intentions. Reinholtz et al (2021) demonstrated that when the number of new daily cases is declining, people shown a graph of This document is copyrighted by the American Psychological Association or one of its allied publishers.…”

“…Similarly, different visualizations of the same data can influence both risk perception and behavioral intentions. Reinholtz et al (2021) demonstrated that when the number of new daily cases is declining, people shown a graph of the cumulative case totals (reduced increase), perceive greater risk than those viewing the number of new cases showing a decrease. Moreover, those viewing the decrease in cases had stronger intentions to engage in risky behaviors.…”