Eliciting Risk Perceptions: Does Conditional Question Wording Have a Downside?

Jeremy D. Strueder,Jane E. Miller,Xianshen Yu,Paul D. Windschitl
DOI: https://doi.org/10.1177/0272989x231223491
2024-01-20
Medical Decision Making
Abstract:Medical Decision Making, Ahead of Print. BackgroundTo assess the impact of risk perceptions on prevention efforts or behavior change, best practices involve conditional risk measures, which ask people to estimate their risk contingent on a course of action (e.g., "if not vaccinated").PurposeTo determine whether the use of conditional wording—and its drawing of attention to one specific contingency—has an important downside that could lead researchers to overestimate the true relationship between perceptions of risk and intended prevention behavior.MethodsIn an online experiment, US participants from Amazon's MTurk (N = 750) were presented with information about an unfamiliar fungal disease and then randomly assigned among 3 conditions. In all conditions, participants were asked to estimate their risk for the disease (i.e., subjective likelihood) and to decide whether they would get vaccinated. In 2 conditional-wording conditions (1 of which involved a delayed decision), participants were asked about their risk if they did not get vaccinated. For an unconditional/benchmark condition, this conditional was not explicitly stated but was still formally applicable because participants had not yet been informed that a vaccine was even available for this disease.ResultsWhen people gave risk estimates to a conditionally worded risk question after making a decision, the observed relationship between perceived risk and prevention decisions was inflated (relative to in the unconditional/benchmark condition).ConclusionsThe use of conditionals in risk questions can lead to overestimates of the impact of perceived risk on prevention decisions but not necessarily to a degree that should call for their omission.HighlightsConditional wording, which is commonly recommended for eliciting risk perceptions, has a potential downside.It can produce overestimates of the true relationship between perceived risk and prevention behavior, as established in the current work.Though concerning, the biasing effect of conditional wording was small—relative to the measurement benefits that conditioning usually provides—and should not deter researchers from conditioning risk perceptions.More research is needed to determine when the biasing impact of conditional wording is strongest.
health care sciences & services,medical informatics,health policy & services
What problem does this paper attempt to address?