Research DeskGlossary
What Is Effect Size in Research? A Plain Guide
Effect size tells you how big a finding is, not just whether it's unlikely to be chance. Here's how to read it.
The short answer
Effect size is a number that quantifies the magnitude of a finding, such as the difference between two groups or the strength of a relationship. Unlike a p-value, which only indicates whether a result is unlikely to be chance, effect size tells you how large or practically important that result is.
The evidence
- 1 observational study
- 5 reviews
Key takeaways
- Effect size measures the magnitude of an effect, while statistical significance only tells you whether a result is unlikely to be chance.
- A result can be statistically significant yet so small that it has little practical or clinical importance.
- Effect sizes come in standardised forms, such as Cohen's d, and unstandardised forms, such as the raw difference between group means.
- Reporting effect sizes alongside p-values supports better interpretation, comparability and meta-analysis.
- In practice, effect sizes are still under-reported, even in journals that request them.
In this piece, 7 sections
What is effect size?
Effect size is a measure of the magnitude of an observed effect. In a study comparing two groups, it can be the raw difference between their average outcomes, such as a 15-question improvement on an exam. It can also be a standardised measure, which transforms the effect onto a common scale so results from different studies can be compared1.
The distinction matters because statistical significance alone does not tell you whether a finding is large enough to care about. A p-value indicates whether an observed finding is unlikely to have occurred by chance, but it does not quantify the magnitude or practical importance of that finding2. You can have a statistically significant result that is so small it has no real-world value.
Why isn't the p-value enough?
The p-value is the most commonly used measure of statistical significance, but it fails to convey information about the clinical or practical relevance of a finding3. In biology, null hypothesis significance testing has been criticised for the same reason: it does not provide the magnitude of an effect or the precision of the estimate4.
A practical example helps. If an educational intervention improves exam scores by an average of 15 out of 50 questions compared with another intervention, the absolute effect size is 15 questions, or 30% better1. That number tells you something a p-value cannot: how much better one approach was.
Absolute versus standardised effect sizes
Effect size can refer to two different things1:
- Absolute effect size is the raw difference between group means. It is useful when the variable has intrinsic meaning, such as number of hours of sleep or questions answered correctly.
- Standardised effect size is calculated to put the effect on an easily understood scale. This is useful when measurements have no intrinsic meaning, such as scores on a Likert scale, or when studies have used different measures.
Common standardised measures include Cohen's d for the standardised mean difference and r for correlation coefficients. These two classes can be calculated from almost all study designs and are essential for meta-analysis4.
How effect sizes are actually used
Effect sizes serve several practical purposes. They help researchers determine the sample size needed for follow-up studies, and they allow effects to be examined across studies5. They also support the interpretation of biomarkers, diagnostic tests and clinical interventions by quantifying differences, associations and agreement2.
In veterinary medicine, reporting effect sizes alongside confidence intervals and the number needed to treat has been proposed as a way to help practitioners make informed, patient-centred decisions3. The same logic applies across fields: an effect size and its confidence interval enable a more effective assessment of relationships within data than p-values alone4.
How often are effect sizes reported?
Not as often as they should be. A survey of articles published in 2009 and 2010 in the Journal of Experimental Psychology: General found that effect sizes were reported for fewer than half of the analyses, and no article reported a confidence interval for an effect size6. For t-tests, two-thirds of articles did not report an associated effect size estimate6.
This is a known gap. Several reviews argue that routine reporting of effect sizes alongside p-values should become standard practice, to improve interpretation, comparability and clinical relevance24.
How to read an effect size
When you encounter an effect size in a study, ask three questions:
- Is it absolute or standardised? An absolute effect size is in the original units, such as points or hours. A standardised effect size is on a common scale, such as Cohen's d.
- How large is it? For standardised measures, conventional benchmarks exist, but context matters. A small effect in one field may be meaningful in another.
- How precise is it? Look for a confidence interval around the effect size. The interval tells you the range within which the true effect is likely to lie4.
If a study reports only a p-value, you are missing the magnitude of the effect. That is a reason to read with caution, not a reason to dismiss the study outright.
Related reading
Questions, answered
What is the difference between effect size and p-value?
What is a good effect size?
There is no universal answer. For standardised measures such as Cohen's d, conventional benchmarks exist, but what counts as meaningful depends on the field and the question. An absolute effect size is easier to judge because it is in the original units, such as points or hours1.
What is Cohen's d?
Cohen's d is a standardised measure of the difference between two group means. It expresses the difference in units of standard deviation, which allows effects from different studies or different measures to be compared on a common scale4.
Why are effect sizes important for meta-analysis?
Meta-analysis combines results across studies, which requires a common metric. Standardised effect sizes such as d and r can be calculated from almost all study designs, making them essential for pooling findings and comparing effects across research4.
Do all studies report effect sizes?
No. A survey of articles in one psychology journal found effect sizes were reported for fewer than half of the analyses, and no article reported a confidence interval for an effect size6. Under-reporting remains a recognised problem across fields.
References
The Research Library- 01
Gail M. Sullivan, Richard Feinn (2012). Using Effect Size—or Why the P Value Is Not Enough. Journal of Graduate Medical Education.
- 02
Nigussie N, Ordak M (2026). Effect size measures in laboratory medicine. Laboratory medicine.
- 03
Pugliese BR, Hallowell KL, Vigneshwar A et al. (2026). Veterinarians' brief guide to confidence intervals, standardized effect size, and number needed to treat: understanding the impact of treatments and disease. Frontiers in veterinary science.
- 04
Nakagawa S, Cuthill IC (2007). Effect size, confidence interval and statistical significance: a practical guide for biologists. Biological reviews of the Cambridge Philosophical Society.
- 05
Daniël Lakens (2013). Calculating and reporting effect sizes to facilitate cumulative science: a practical primer for t-tests and ANOVAs. Frontiers in Psychology.
- 06
Fritz CO, Morris PE, Richler JJ (2012). Effect size estimates: current use, calculations, and interpretation. Journal of experimental psychology. General.
Unleashed Newsroom
The newsletter
New pieces from the newsroom, in your inbox, with every claim still linked to its source.
Educational content, not medical advice. It can't account for your circumstances; talk to a qualified professional about your own health. Researched with AI assistance and reviewed by a human editor before publication. Read our standards. Spotted an error? Write to hello@unleashed.vision.