Abstract
Procedures for conducting power analyses at the level of combined probabilities are described. Specific procedures allow the determination of the relative power gain as the number of studies in the meta-analysis increases, the calculation of the number of studies needed to reject a false null hypothesis at a given level of power, and the estimation of the number of additional studies required to reverse a conclusion of "not significant" at a given level of power.
| Original language | English |
|---|---|
| Pages (from-to) | 595-599 |
| Number of pages | 5 |
| Journal | Psychological Bulletin |
| Volume | 98 |
| Issue number | 3 |
| DOIs | |
| State | Published - Nov 1985 |