sdm-help-list > RE: Number of datasets needed
Jan 8, 2014  02:01 AM | Nobody
RE: Number of datasets needed
The minimum number of datasets to conduct an ES-SDM analysis may be very difficult (if not impossible) to estimate, as it would require a power analysis based on the randomization test, accounting for different sample sizes, potential heterogeneity...

... However, as a rule of thumb you may think about 10 studies for a standard coordinate-based meta-analysis. Less studies may be required when effects are strong, and much less (e.g. 4 or 5) when all t-test maps are available. Conversely, more studies may be required for detecting weak effects.

Threaded View

TitleAuthorDate
Nobody Dec 31, 2013
RE: Number of datasets needed
Nobody Jan 8, 2014