borderline significancep = .05
This method estimates the number of hypothetical unpublished studies with an average effect size of 0 which, if they were included in meta-analysis, would increase the p value for the meta-analysis to borderline significance ( p = .05).