Abstract
We studied the effects of presentation formats (frequency, 1-in-N, and visual) and numeracy level on students' understanding of prenatal screening results, as well as their risk assessment for having a fetus with Down syndrome. Frequency format (vs. 1-in-N and visual formats) improved participants' ability to accurately assess the chances of having a fetus with Down syndrome, and was associated with lower risk estimates. High numeracy levels were associated with a better ability to judge risk likelihood. For individuals of low numeracy levels, however, the frequency format significantly facilitated accurate understanding of probability information. This suggests that presenting information in frequency format is particularly beneficial for certain populations.
We thank Hilary Louise Bekker, Wolfgang Gaissmaier, Mirta Galesic, Shimon Saphire-Bernstein, Anita Todd, and two anonymous reviewers for comments on earlier drafts, and Amy Krilla for excellent research assistance. The first and second authors contributed equally to the preparation of the article. The study was supported by the Center for Health and Wellbeing, Princeton University.
Notes
1Responses of “1/181,” “1:181,” “0.55%,” and “one out of one hundred and eight-one” all would be considered correct. In addition, participants who responded “less than 1%” (or some variation of that statement) were considered correct, but note that this does not apply to participants who gave a specific numeric response outside the 0.5%–0.6% range that happened to be less than 1%.
2As with the comprehension variable, participant responses were translated into a numerical percentage, so responses of “1/181,” “1:181,” “0.55%,” and “one out of one hundred and eight-one” all would be considered correct. In addition, participants who responded “less than 1%” (or some variation of that statement) were considered correct, but note that this does not apply to participants who gave a specific numeric response outside the 0.5%–0.6% range that happened to be less than 1%.
Note. The F value is the result of a univariate analysis of variance test for differences in the mean level of the dependent variable in the corresponding row, across the three conditions. The ^ and + symbols identify significant pairwise comparisons. If two groups are not significantly different for a given variable, they receive the same sign. Group means flagged with a different symbol are significantly different from the other groups.
a The means for this variable also can be read as a percentage, as answers were scored 1 if correct and 0 if incorrect. Thus, for example, 36% of participants in the probabilistic condition gave the correct answer.
b Answers were scored 1 for “low,” 2 for “medium,” and 3 for “high.” Responses of “I don't know” were not included in the analysis.
c Answers were scored 0 for “no chance,”1 for “less than 1%,” 2 for “2–3%,” and 3 for “4% or greater.” As no participants select the “no chance” option, a lower mean response for a group indicates that the group as a whole was more correct, relatively speaking.
∗p < .05. ∗∗p < .01. ∗∗∗p < .001.
Note. Participants were given four choices: No chance, less than 1%, 2–3%, 4% or above. The correct response was less than 1%. As reported in the text, there were main effects for presentation format, F(2,235) = 4.04, p < .05, and for the numeracy median split, F(1,235) = 11.52, p < .001. The interaction was not significant, F(2,235) = 2.57, p = .08.