Footprint of publication selection bias on meta‐analyses in medicine, environmental sciences, psychology, and economics