Seasonal influenza causes illness and death in the United States and globally and imposes heavy burdens on health systems. Continuing efforts have been devoted to making more accurate influenza forecasts. However, little research has been devoted to how well users can understand and use these forecasts. This study extends Tversky’s theory of conceptual-spatial congruence in visualization (2011, https://doi.org/10.1111/j.1756-8765.2010.01113.x) to examine how well people can use four displays to make four influenza-related decisions. Two of the decisions involve categorical choices and two continuous ones. Two of the displays involve categorical visualizations and two continuous. Following Tversky, we hypothesized that participants’ performance would be higher when there was natural congruence between the choice and the display, as categorical or continuous. We measured performance in terms of comprehension, decision confidence, and perceived helpfulness. We also hypothesized that performance would be better for individual participants who reported greater familiarity with each display, subjective numeracy, and subjective visual literacy. This study was conducted on MTurk in early October 2020 in U.S. with 301 valid responses, after attention checks. We found that participants had better comprehension with the bar chart, for all four decisions, but did not perceive it as more helpful or have more confidence in their responses – only partially consistent with our extension of Tversky’s theory. Contrary to our expectations, participants who reported greater familiarity with a display performed less well; however, they had greater confidence in their decisions and rated the display as more helpful. Our study extends a prominent theoretical framework for visualization to an important empirical domain. Our approach offers a general method for testing the usability of visualizations. Our results indicate the need for testing displays and for helping users align their performance and confidence.