Statistics shape decisions long before most people realize they are making them. A weather forecast influences whether someone carries an umbrella; a baseball player’s batting average can affect how a team values performance; and medical risk estimates can alter decisions about screening, treatment, or lifestyle. Yet a new national survey analysis led by Penn State researchers suggests that most U.S. adults feel poorly equipped to interpret the numbers increasingly used to explain science, health, politics, and everyday life.
The study, published in PLOS One, found that 62% of respondents reported having no or only limited knowledge of statistics. Only 11% said they regularly use statistics in daily life. At the same time, the survey revealed a striking contradiction: roughly nine out of 10 participants said they would be more likely to base decisions on reported statistics if they understood them better. The findings suggest that the problem may not be widespread rejection of data, but a gap between people’s desire to use evidence and their confidence in interpreting it.
“While statistics are not hard to understand, they are even easier to misunderstand,” Mark Ramos, an assistant research professor of health policy and administration at Penn State, said. Ramos conducted the analysis with Samuel Anyaso-Samuel, a postdoctoral fellow at the National Institutes of Health’s National Cancer Institute. Their work examines how Americans perceive their own statistical literacy, particularly their understanding of concepts that frequently appear in scientific reporting but rarely receive detailed explanation.
The data came from a nationwide survey of 1,000 U.S. adults. During the 2025 Joint Statistical Meetings in Nashville, Tennessee, a market-research and survey company invited conference attendees to submit questions for inclusion in a survey of a representative sample of the U.S. population. Ramos and Anyaso-Samuel contributed two questions: how much respondents understood about statistics and p-values, and how often they would use reported statistics to make decisions if they understood them better. The researchers then analyzed the responses to assess Americans’ self-reported familiarity with statistical reasoning.
One-quarter of the participants said they had no understanding of statistics at all, while 37% described their familiarity as limited. Slightly more than one-quarter said they had learned some statistics in school. These responses do not necessarily measure what participants can objectively calculate or explain, because the study assessed perceived knowledge rather than performance on a statistics examination. However, self-confidence can strongly influence whether people engage with evidence, question a claim, or avoid numerical information altogether.
The survey focused in part on p-values, one of the most commonly cited—and most frequently misunderstood—quantities in scientific research. A p-value is calculated under a statistical model that assumes a particular null hypothesis, often the idea that there is no difference, association, or effect. It estimates how unusual the observed data, or data more extreme, would be if that assumption were true. A small p-value can provide evidence against the null hypothesis, but it does not prove that a finding is important, that a hypothesis is true, or that a result will be replicated.
That distinction matters because statistical significance is not the same as practical significance. A very large study can produce a small p-value for an effect too minor to matter in real life, while a potentially meaningful effect may fail to reach a conventional significance threshold in a small study. A p-value also does not tell readers the probability that a research hypothesis is true, nor does it measure the size of an effect. To understand a study responsibly, people may also need to consider confidence intervals, sample size, study design, sources of bias, and whether the result is consistent with other research.
Ramos emphasized that statistics are not a system for eliminating uncertainty. They are a framework for reasoning under uncertainty, using data to estimate patterns and compare possible explanations. Weather probabilities, medical risk estimates, polling margins of error, and research findings all contain uncertainty because observations are limited and the world is variable. Statistical literacy therefore involves more than recognizing numerical terms. It includes asking what was measured, who was studied, how large the uncertainty is, what alternative explanations exist, and whether the evidence justifies the conclusion being presented.
The researchers argue that statistical literacy should be accessible to anyone who wants to evaluate claims affecting personal and public life. News reports, social media posts, advertisements, and policy debates routinely invoke percentages, risk ratios, averages, and research findings, often without explaining their limitations. A number can appear authoritative even when it is presented without a denominator, a comparison group, or information about uncertainty. Improving public understanding could help people distinguish robust evidence from exaggerated claims and become more discerning consumers of science. Penn State offers free online materials for STAT 200, an introductory statistics course, as one resource for readers seeking to build that foundation.
Subject of Research: People
Article Title: Self-reported perception of statistical literacy: Evidence from a National Survey of U.S. Adults
Web References: https://doi.org/10.1371/journal.pone.0350282; https://online.stat.psu.edu/stat200/
References: PLOS One, DOI: 10.1371/journal.pone.0350282
Keywords: Statistical literacy, statistics, p-values, statistical reasoning, scientific evidence, public understanding of science, U.S. adults, survey research, data interpretation, health communication
