62% of American Adults Say They 'Barely Understand Statistics' (Survey of 1,000 People)

Explanation

When 1,000 American adults were asked, "How well do you understand statistics and p-values?", 62% answered either "not at all" or "limited knowledge." The breakdown was 25% for "not at all" and 37% for "limited." The survey, conducted by Assistant Professor Mark Ramos (health policy and administration) at Penn State University and postdoctoral researcher Samuel Anyaso-Samuel at the National Cancer Institute, was published in PLOS One.

The survey targeted 1,000 people selected to reflect the demographic composition of the U.S. population. There were only two questions: "How well do you understand statistics and p-values?" and "If you understood better, how often would you base decisions on reported statistics?" These two questions were developed based on input solicited from attendees at the 2025 Joint Statistical Meetings held in Nashville.

Among the figures, the most striking is not the 62% itself. While only 11% said they use statistics on a daily basis, 90% said they would base decisions on statistics "if they understood better." In other words, this is not a case of many people avoiding statistics, but rather many people wanting to use them but unable to read them.

A p-value is a number expressing "if there were truly no difference at all, how likely would a difference as large as the one observed in the data occur by chance?" It is often interpreted as "the smaller the p-value, the larger the effect," but this is incorrect. A p-value answers not the size of an effect but whether it can be explained by chance, and with a large enough sample size, even a small difference can produce a small p-value.

What problems arise from being unable to read statistics? The researchers point out that it affects the interpretation of information that appears routinely in daily life, such as precipitation probabilities in weather forecasts and health-related risk figures. What does "30% chance of rain" mean? When a risk is said to be "twice as high," twice what original number?—the figures used for decision-making usually come in the form of statistics.

This survey itself also needs to be read carefully. What is measured here is not actual statistical ability, but how respondents self-reported their own understanding. Because it is self-reported, people who actually understand more may have rated themselves lower, and vice versa.

Paper: PLOS One, DOI: 10.1371/journal.pone.0350282
Source: Phys.org (August 21, 2026) https://phys.org/news/2026-08-adults-lack-basic-statistical.html