Are interactive and tailored data visualizations effective in promoting flu vaccination among the elderly? Evidence from a randomized experiment.

ABSTRACT

OBJECTIVE: Although interactive data visualizations are increasingly popular for health communication, it remains to be seen what design features improve psychological and behavioral targets. This study experimentally tested how interactivity and descriptive titles may influence perceived susceptibility to the flu, intention to vaccinate, and information recall, particularly among older adults. MATERIALS AND METHODS: We created data visualization dashboards on flu vaccinations, tested in a 2 (explanatory text vs none) × 3 (interactive + tailored, static + tailored, static + nontailored) + questionnaire-only control randomized between-participant online experiment (N = 1378). RESULTS: The flu dashboards significantly increased perceived susceptibility to the flu compared to the control static+nontailored dashboard, b = 0.14, P = .049; static-tailored, b = 0.16, P = .028; and interactive+tailored, b = 0.15, P = .039. Interactive dashboards potentially decreased recall particularly among the elderly (moderation by age b = -0.03, P = .073). The benefits of descriptive text on recall were larger among the elderly (interaction effects b = 0.03, P = .025). DISCUSSION: Interactive dashboards with complex statistics and limited textual information are widely used in health and public health but may be suboptimal for older individuals. We experimentally showed that adding explanatory text on visualizations can increase information recall particularly for older populations. CONCLUSION: We did not find evidence to support the effectiveness of interactivity in data visualizations on flu vaccination intentions or on information recall. Future research should examine what types of explanatory text can best support improved health outcomes and intentions in other contexts. Practitioners should consider whether interactivity is optimal in data visualization dashboards for their populations.