In quantitative research, statistical significance tells us whether an observed effect is unlikely to be due to chance. However, statistical significance does not tell us how large or meaningful that effect is. Effect size addresses this limitation. It provides a measure of the magnitude of a relationship or difference, helping researchers interpret practical and theoretical importance. This article explains effect size conceptually, introduces common measures, and clarifies its role in social science and management research.
What Is Effect Size?
Effect size refers to a quantitative measure of the strength or magnitude of a relationship, difference, or association between variables.
While hypothesis testing answers:
Is there evidence of an effect?
Effect size answers:
How large is the effect?
These are fundamentally different questions.
Why Effect Size Matters
Statistical significance is influenced by sample size. In large samples, even very small differences may become statistically significant. Conversely, meaningful effects may fail to reach significance in small samples.
Effect size helps researchers:
- Assess practical importance
- Compare results across studies
- Evaluate theoretical strength
- Interpret findings beyond significance levels
Without effect size, statistical conclusions remain incomplete.
Example: Training Program Study
Suppose a researcher tests whether a new training program improves productivity.
Two scenarios are possible:
- The improvement is statistically significant but very small.
- The improvement is statistically significant and substantial.
Both scenarios may produce the same significance level, but the practical implications differ. Effect size distinguishes between these cases.
Common Measures of Effect Size
Different research designs use different effect size measures. Below are some commonly used ones.
1. Cohen’s d (Difference Between Means)
Cohen’s d measures the standardized difference between two group means.
Cohen’s d expresses the difference in terms of standard deviation units.
General interpretation guidelines (often cited):
- 0.2 → small effect
- 0.5 → medium effect
- 0.8 → large effect
These are rough conventions, not universal rules.
2. Correlation Coefficient (r)
The correlation coefficient reflects the strength of association between two variables.
Values range from -1 to +1:
- Values near 0 → weak relationship
- Values near ±1 → strong relationship
The magnitude of r itself represents an effect size.
3. R-squared (Explained Variance)
In regression analysis, R² represents the proportion of variance in one variable explained by another.
For example:
- R² = 0.30 means 30% of the variation is explained by the model.
This is also a measure of effect size.
Effect Size and Statistical Significance
Effect size and statistical significance serve different purposes.
- Statistical significance evaluates whether an effect likely exists.
- Effect size evaluates how large that effect is.
It is possible to observe:
- Significant but trivial effects
- Non-significant but potentially meaningful effects (especially in small samples)
This distinction reinforces the importance of reporting both.
Effect Size and Statistical Power
Effect size plays a central role in statistical power.
Power increases when:
- The effect size is larger
- The sample size is larger
- Variability is lower
In fact, sample size calculations often require specifying an expected effect size in advance.
Thus, effect size connects:
- Hypothesis testing
- Power analysis
- Sample size determination
Common Misunderstandings About Effect Size
A common misunderstanding is that effect size determines importance automatically. In reality, interpretation depends on context, theory, and practical consequences.
Another misconception is that effect size replaces statistical testing. Effect size complements, rather than replaces, hypothesis testing.
Effect Size in Research Reporting
Best practice in research reporting involves presenting:
- Statistical significance
- Effect size
- Confidence intervals
Together, these provide a fuller understanding of findings.
Conclusion
Effect size measures the magnitude of a relationship or difference in research. While statistical significance addresses whether an effect likely exists, effect size clarifies how meaningful that effect is. Understanding and reporting effect size strengthens interpretation, transparency, and theoretical insight in social science and management research.
This discussion builds on earlier explanations of hypothesis testing and statistical power, which clarify when effects are detected and how study design influences results. It also connects to sample size determination, where expected effect size influences required sample size.
Leave a Reply