Clicked Gallery

What is effect size?

Highlighted from a real textbook passage. Explained by Clicked.

Used in a sentence

University Course Reader · STEM

The drug showed a statistically significant but small effect size, raising questions about its practical benefit.

The reader highlighted one term — in an article or a paper PDF. Clicked made the statistics term “effect size” easy to understand:

Explained in three depths

Same facts, different vibe — Slang mode 😎

The Clicked way

●○○

Overview

Effect size is how big a difference or relationship actually is: how much the drug lowered blood pressure, how far the new teaching method moved scores. It answers a different question from statistical significance, which only says the difference is probably not luck. A result can be real and tiny at the same time, and effect size is the number that tells you which you have.
●○○

Overview

Effect size is the "okay, but how much" number. Science news loves the word significant, and in statistics that word means one thing only: probably not a coincidence. Not big. Not important. Not "buy this supplement." A finding can be significant and microscopic at once, and most of the time nobody puts the microscopic part in the headline. Effect size is where they keep it. 😎

A quick take — often all you need.

●●○

Detail

Effect size measures how big a finding is. Statistical significance asks whether a difference is likely real rather than chance; effect size asks how much there is, in units you can weigh: minutes of sleep, points on a test, dollars saved. The two come apart often, and sample size is the reason. With enough participants a study can detect almost any real difference, however small: a trial of a million people can find that a sleep app adds forty seconds of sleep and call the result solidly significant. Nothing there is wrong. It is just forty seconds. Researchers therefore report effect size alongside significance, often in standardized form so studies can be compared, with rough labels of small, medium, or large. Context outranks the labels in both directions. A tiny effect can matter enormously at scale: a drug trimming a fraction of a percent off heart attack risk saves thousands of lives when millions take it. A large effect on something trivial matters not at all. The reading habit worth building is one question asked twice: is the difference real, and is it big enough to care about? Significance answers the first. Effect size answers the second, and news stories routinely report only the first.
●●○

Detail

Effect size is how much a thing actually does, and once you ask for it, health headlines never read the same. The trap is built into the language itself. A huge trial can prove a microscopic improvement is real, which means "significant in a study of two million" can translate, with a straight face, to "we are extremely sure it barely works." Watch the percentage trick too. "Doubles your risk" sounds like a siren, but doubling one-in-a-million gets you all the way to two-in-a-million, and the scary version of that line somehow never mentions the starting point. Flip side: cheap, tiny improvements applied to whole countries genuinely add up, which is how public health earns its keep. The fix is one reflex. Whenever a finding gets announced, request the magnitude in things you can actually count, and if the response keeps hiding behind the word significant, then the response, translated, is probably embarrassing. 😎

Want more? One click digs deeper.

●●●

Analogy

Effect size is the brightness of a star; significance is whether your telescope can pick it out. Build a big enough telescope and you can detect the faintest speck in the sky. Detection is a fact about your equipment as much as the star, and it says nothing about how bright the star is. That is the trap in "the study found an effect." A study with a million participants is an enormous telescope, and finding a speck through it proves the speck exists, nothing more. Before acting, you want the other number: not "did the telescope see it," but "how bright is it, really." One is significance. The other is effect size.
Deepest level reached.
●●●

Analogy

Effect size is your raise in actual money. Imagine HR announces, with total confidence, that you are getting a raise, verified by the finance department, checked three times, absolutely beyond dispute. It is twelve cents a month. Every part of the announcement was true, and the department's certainty was real. Certainty was just never the interesting question. The interesting question fits on a coin. That is a significant-but-tiny finding: a difference proven beyond doubt that would not cover a stick of gum. So when the announcement sounds triumphant, do what you would do with HR. Nod politely, then ask to see the number. 😎
Deepest level reached.

Unfamiliar concept? A real-world example makes it click — fresh analogies on tap.

AI explanations may contain errors · Not professional advice

Formal definition — The same term, explained the usual way

Effect size is a quantitative measure of the magnitude of a phenomenon, such as the difference between group means or the strength of an association, reported independently of statistical significance. Common standardized measures include Cohen's d, expressed in standard deviation units, the correlation coefficient r, and odds or risk ratios for binary outcomes. Because p-values conflate magnitude with sample size, effect sizes are reported alongside significance tests and are required by many journals; conventional thresholds, such as d values of 0.2, 0.5, and 0.8 for small, medium, and large effects, are heuristics whose practical meaning depends on the domain.

Want Clicked to explain terms like “effect size” directly in your browser — including on PDFs?

Add to Chrome — Free

50 free Explanations · No credit card required