Gender bias found in how scholars review scientific studies