Course contentsShow
← Data Science
Lesson 387 of 2,14510. Hypothesis Tests for MeansFree lesson

Confidence Intervals for Effect Sizes

Constructing and interpreting confidence intervals around effect size estimates to quantify uncertainty.

Confidence Intervals for Effect Sizes

You'll learn how to quantify the uncertainty around your effect size estimates using confidence intervals.

Why Effect Sizes Need Confidence Intervals

You've learned to calculate effect sizes like Cohen's d to measure the magnitude of a difference. But just like any sample statistic, your effect size estimate has sampling variability—if you collected a different sample, you'd get a slightly different value. A confidence interval around your effect size tells you the range of plausible values for the true population effect size.

Understanding Effect Size Confidence Intervals

Think of it this way: if Cohen's d = 0.65 from your sample, you're saying "the effect size is about two-thirds of a standard deviation." But how confident are you? A 95% confidence interval might be [0.42, 0.88], meaning you're 95% confident the true population effect size falls within that range.

The interpretation is parallel to other confidence intervals: if you repeated your study many times, about 95% of the intervals you constructed would contain the true population effect size.

What They Tell You

Confidence intervals for effect sizes reveal:

  • Precision: A narrow interval (e.g., [0.60, 0.70]) means you estimated the effect size precisely. A wide interval (e.g., [0.20, 1.10]) signals high uncertainty.
  • Practical significance: Even if your p-value is small, if the confidence interval includes very small effect sizes, the practical importance might be questionable.
  • Consistency with benchmarks: Does your entire interval fall in the "large effect" range, or does it span from "small" to "large"?

Key Takeaway: Confidence intervals for effect sizes quantify the uncertainty in your magnitude estimates, helping you understand not just whether an effect exists, but how precisely you know how large it is.