Communicators often use their intuitions to decide what information to present and don’t always use numbers. For example, federally required Vaccine Information Statements list common side effects, but provide no numbers about percentages of people expected to experience the side effects. Research in our lab (and others), however, has demonstrated that providing this information increases willingness to vaccinate and take medications while reducing overestimation of side-effect likelihoods.
However, not everyone uses numbers well* or pays attention to them. In a now-classic study, we demonstrated that the more numerate (numeracy is like literacy, but with numbers) better understand the meaning of numbers in context (e.g., how good or bad is a 3% risk) and are less influenced by how numbers are presented. For example, those who scored better on a math test perceived a 10% chance of a negative event and 10 chances out of 100 of the same negative event as equally risky, whereas those who scored worse perceived 10% as much less risky than 10 of 100.
As a result, the format used to present numbers matters, and particularly to people lower in numeracy, who are also more likely to have lower income, education, and health literacy. Thus, well-designed communications involving numbers may reduce inequalities.
Over the past 20 years, we have done research on best methods for communicating with numbers (here’s a recent review). For data visualizations, in particular, Box 1 includes some recommendations, with more detailed descriptions below.
1. Provide Numbers and Data Visuals
Similar to the studies above concerning vaccines and prescription drugs, across three experiments, when data were shown with graphics about relatively familiar topics, people scored 75-80% correct on comprehension tests (green bars below), compared to 30-50% correct when no data were shown (gray-black bars below).
2. Provide cues that help people evaluate information
People often can’t make sense of numbers shown in isolation. For example, knowing that 93% of pneumonia patients at hospital A survived treatment is hard to evaluate; stating that that survival rate is only fair (vs. good or excellent) can help. In four studies, consumers given labels to interpret numeric health information (e.g., percent of patients receiving recommended treatments) made better decisions.
Our study on vaccine hesitancy further suggests that vaccine-hesitant people benefit from interpretive labels. Compared to participants who saw only numbers (e.g., “21% chance of muscle pain”), those who received numbers plus an interpretive label (e.g., “very common”) were more willing to vaccinate, particularly if they were vaccine-hesitant (pictured below) and/or less numerate (not pictured).
3. Provide only important information
Communicators sometimes want to provide all the information to decision makers, but including unnecessary information can distract and reduce decision quality. For example, when choosing a hospital, out-of-pocket costs are more important than the quality of hospital food. However, including this unimportant information resulted in only 48% of participants choosing the highest quality option vs. 62% when it was eliminated.
Similarly, in a study on data visualizations, we provided participants with a graph of wildfire acres burned in the US from 1980 to 2020. Some graphs included a trendline (depicted below).
Participants did not need the trendline to tell that wildfire rates were increasing. However, adding it impeded their ability to identify when the most acres burned. Only 79% of participants correctly identified the years on the graph depicted above vs. 86% when the blue trendline was removed.
4. Provide information in a format that matches the communication goal
A critical first step in communication is to choose the goal of the communication. It’s important because providing information in a format that matches the goal increases comprehension. For example, all participants viewed graphs showing the number of authorized immigrants (see below). Some participants also saw the total foreign-born population graphed, as below; others saw a line tracking the subcategory, unauthorized immigrants, instead.
When graphed with total foreign-born people, participants understood the total better but understood less about the unauthorized immigrants subcategory (see green bars below). The opposite occurred when we graphed both subcategories (gray-black bars below). Thus, to maximize comprehension of total immigrants, you should choose one way to present the data (i.e., graphing total immigrants), but to maximize understanding of subcategories, you should choose the other way.
In sum, communicators should begin by identifying a communication goal, and then identify the critical numbers and data visuals that support the goal: Provide interpretive aids to help people understand and evaluate those numbers and visuals, and format that information to eliminate distracting or unimportant information and emphasise the important information. This approach will aid policymakers in effectively communicating with those who would benefit particularly from their interventions.
For more on how to present numbers so that they matter, check out:
Innumeracy in the Wild:Misunderstanding and Misusing Numbers
*Test your numeracy on our website
This work was supported by grants from the National Science Foundation (SES- 2017651) and USAFacts. The content of this post is solely the responsibility of the authors and does not necessarily represent the official views of the National Science Foundation or USAFacts.