The arbitrary decimal that can kill a year of scientific research

Science · Published 12 October 2026 Swipe more like this →

In short: Imagine spending a year on grueling fieldwork, only to have your results rejected because a single number was 0.002 too high. This is the reality of the p-value, a statistical gatekeeper that often prioritizes a rigid threshold over actual scientific meaning.

The p-value has long functioned as the gold standard for determining 'significance' in quantitative research. However, its widespread misuse has created what some scholars call a 'tyranny.' The core issue lies in how researchers interpret these numbers: a p-value is often treated as a binary switch for truth rather than a nuanced measure of uncertainty. When a result falls just slightly above a common threshold, it is often dismissed entirely, regardless of the underlying data's potential importance. This rigid reliance on p-values can distort how we interpret data and publish new findings. One well-known instance of this data issue happened when a researcher detected notable activity within the brain matter of a deceased fish.

The human cost of this rigid system is tangible. Researchers in diverse fields—including urology, critical care, and educational research—have reported instances where potentially groundbreaking work was discarded because it failed to meet an arbitrary decimal requirement. For example, a study might show a promising trend that is practically meaningful, but because the p-value was marginally over the limit, the work is deemed 'insignificant' by peer reviewers and journals. Specific experts in urology, critical care, and education have all experienced this frustration. These researchers saw their manuscripts rejected because of a tiny margin of error. This highlights how a single decimal point can lead to the loss of resources and time.

The problem is further complicated by how models are specified and results are interpreted, which can lead to deceptive conclusions even when 'significant' results are achieved. Rather than abandoning the p-value entirely, experts argue for a fundamental shift in statistical education and publication norms. The goal is to reframe the p-value as just one small component of a broader conversation regarding evidence, uncertainty, and the actual meaning of the data. The current goal is to move toward a more holistic view of research. Scholars want to change how statistics are taught to students. They believe we must look at the broader context of the evidence provided.

General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.

Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?

Related

• Why the scientific gold standard of statistical significance is under fire
• A small p is not the chance your hypothesis is true
• Why the p-value cutoff might be misleading your scientific conclusions
• Statistical significance is not enough: why effect size matters more than p-values
• How researchers accidentally or intentionally manufacture scientific breakthroughs through p-hacking
• The paleo diet's central premise about starch may be historically inaccurate
• The pea-sized gland that manages your social life and water balance
• The mathematical complexity hidden within a stack of pancakes
• A rapid breakdown of the 2023 Oxford Mathematics undergraduate admissions test solutions
• Understanding the dual role of the pancreas in digestion and glucose regulation
• How Blizzard designed a global hit in just six weeks
• The Paradox of Panic: Living in Fear of Your Own Body

More in Science · Latest · A–Z · All topics