barely significantp = 0.05
Fifth, as a check for potential publication bias, we used Rosenthal’s ( 1979 ) fail-safe N (FSN), which indicates the number of studies with null results that would be required to lower a significant relationship to a barely significant level ( p = 0.05).