Question

How should a reader interpret a single failed replication?

DOdortiz· 6 days ago

One failed replication doesn't falsify an effect — there's sampling error, hidden moderators, and protocol differences. But a systematic pattern across many studies says something about a field. How do you all calibrate between 'this one study is noisy' and 'this literature is fragile'?

Psychology

1 Reply

Accepted answer

WEweizh6 days ago

I weight the aggregate far more than any single attempt. The value of this paper isn't 'study X is wrong' — it's the distribution: effect sizes shrinking by ~half and many CIs missing the original is a field-level signal, not a verdict on one lab.