Excess significance and publication bias in general are so common as to be the default; p=0.13 is pretty bad looking (with a single-tailed test, that’d be below the 0.10 threshold Ioannides suggest for publication bias tests due to their low power to detect bias).
Excess significance and publication bias in general are so common as to be the default; p=0.13 is pretty bad looking (with a single-tailed test, that’d be below the 0.10 threshold Ioannides suggest for publication bias tests due to their low power to detect bias).