Z-curve @replicationindex.com · Jul 15

I asked ChatGPT "What is the best test of publication bias? Best = high power to detect bias and low false positive risk. Here is the answer. What does your AI say?

1 likes 1 replies

?

Replies

Lucija Batinovic · Jul 15

I asked with the same prompt and z-curve was not mentioned at all. I got too long of a reply but it mentioned: Among commonly used tests, Peters’ test seems to strike a good balance: – better power than Begg’s, – lower false positive rate than Egger’s when heterogeneity is present.