Effect Size

What effect size actually is

Effect size is a number that answers the question “how much?” If a study finds that a therapy reduces symptoms, the effect size tells you how big that reduction really is. A small effect might be technically real but barely noticeable. A large effect is the kind of change a person would actually feel.

This matters because “did it work” and “how well did it work” are different questions. A study can show that a treatment has an effect while that effect is too small to make any practical difference in someone’s life. Effect size is how researchers separate those two things.

How it works

There are several ways to express an effect size depending on what’s being measured. A common one for comparing two groups describes the difference between them in standardized units, so results from different studies can be compared on the same scale. Other measures describe how strongly two things are related, or how much a treatment changes the odds of an outcome.

The point of all of them is the same. They put the size of a finding into context. A treatment group and a comparison group might differ by an amount that’s tiny, moderate, or substantial, and the effect size is what tells you which.

What it isn’t

Effect size isn’t the same as a p-value. A p-value speaks to whether an effect is likely real rather than chance. Effect size speaks to how large that effect is. A result can be statistically detectable yet trivially small, which is why a low p-value alone doesn’t mean a finding is important.

It also isn’t proof that a treatment will work the same way for any given person. Effect sizes describe averages across groups, not guarantees for individuals.

Related terms you’ll see next

A p-value addresses whether a result is likely due to chance, which is a separate question from size. A meta-analysis pools effect sizes from many studies into one estimate. A randomized controlled trial is a common source of effect-size data. The placebo comparison is what a treatment’s effect size is usually measured against.

Why it matters when you read about mental health

Headlines love the word “significant,” but in research that word often just means “probably not chance.” It says nothing about whether the effect is big enough to care about. When you read that a treatment “significantly improved” something, the useful follow-up is “by how much.” Effect size is the answer, and it’s what separates a finding that changes lives from one that’s real but too small to notice.

<blockquote>
  <p><strong>Effect Size:</strong> Effect size is a number that describes how large an effect is, not just whether it exists. It tells you whether a result is big enough to matter in real life.</p>
  <p>&mdash; <a href="https://shrinktionary.com/terms/effect-size/">Shrinktionary</a>, a plain-English mental health dictionary, medically reviewed by a board certified psychiatrist.</p>
</blockquote>