Reply to Nuijten et al.: Reanalyses actually confirm that US studies overestimate effects in softer research