SRA-E Lisbon 2017

Previous research showed that icon arrays and analogies improved understanding of medical risks among adults (Cosmides & Tooby 1996; Fagerlin et al. 2005; Galesic et al. 2009; Galesic & Garcia-Retamero 2013) and among children (Multmeier 2012; Ulph et al. 2009).
In our studies we measured the effect of icon arrays and analogies on the understanding of risk information among adolescents 11-15 years of age. Though there are a number of experimental studies among adults and some studies among younger children (6-11 years of age), there are almost no experimental studies among adolescents. We suggested that icon arrays and analogies would increase the accuracy of understanding risk as well as risk comparison among adolescents. We conducted two experimental studies.
In Experiment 1 among 213 participants we tested if icon arrays produced a higher accuracy in easy risk calculation tasks and decreased the ratio-bias effect. In addition, we measured if analogies were helpful for understanding some medical information. The icon arrays questions were adapted from (Galesic et al. 2009). There were two distinct icon arrays for treated and untreated people. There were four vignettes with two treatment risk reduction levels (20% and 60%) and two levels of the denominator size (100 and 1,000). The tasks with analogies were adapted from (Galesic & Garcia-Retamero 2013) and (Schwartz et al. 1997). There were two easy and two difficult medical problems. E.g. in easy medical problems respondents were expected to evaluate what individuals should first of all know if they get positive results from medical screenings.
In Experiment 2 among 157 participants we tested if icon arrays produced a higher accuracy in difficult tasks such as tradeoff and Bayesian problems. The tradeoff tasks were adapted from (Waters et al., 2006) and (Hawley et al. 2008). Respondents were expected to calculate the risk of getting two viruses after treatments and evaluate if the total risk of getting viruses after treatment was increased, decreased or had not changed compared to the total risk before treatment. The Bayesian task was adapted from (Brase 2009).
Overall, in both experiments icon arrays produced better understanding of risk information and more accurate risk comparison. The effects varied depending on the task difficulty and risk literacy of the participants. Icon arrays were more helpful for the low numeracy adolescents in complex tradeoff problems. Analogies were helpful for high literacy adolescent children.