TimelyScholarFree Tools › Prediction Interval Calculator
Turn a random-effects result into a 95% prediction interval: the range in which the true effect of a new, similar study is expected to lie.
FreeNo account neededYour data stays in your browser
| 95% prediction interval | 0.14 to 1.78Range for the true effect in a new, similar study |
|---|---|
| 95% confidence interval | 0.34 to 0.70Uncertainty about the average effect |
| Standard error of the pooled estimate | 0.1842Log scale, derived from the CI |
| t critical value | 2.201 (df = 11) |
The prediction interval crosses 1: although the average effect is statistically significant, the effect in some settings may be null or even in the opposite direction.
Estimate and 95% confidence interval.
Between-study variance and number of studies.
Alongside the confidence interval.
The confidence interval describes uncertainty about the average effect. The prediction interval also includes the between-study variation (τ²), so it shows how much the effect may vary from one setting to another.
When heterogeneity is present, the prediction interval can be much wider than the confidence interval and may cross the line of no effect even when the average effect is statistically significant. Presenting it routinely is recommended.
RevMan reports Tau² under the heterogeneity statistics; R (metafor, meta) and Stata report tau² or tau. For ratio measures it is on the log scale.
The method uses a t distribution with k − 2 degrees of freedom, which needs k ≥ 3. With few studies the interval is very wide.