Psychology paper

Estimating the Reproducibility of Psychological Science

Open Science Collaboration (Brian Nosek et al.) · 2015

Read this as

New to this. Just the gist and why it matters.

How deep

The gist, about 2 minutes.

In one line

The short of it

A huge team re-ran 100 published psychology studies. Fewer than half reproduced, and the effects that did were about half the original size. This is the paper that made the replication crisis impossible to ignore.

Why it exists

The problem it tackles

Science is supposed to be reproducible: run the study again and you should get a similar result. Nobody had tested this at scale in psychology. The team asked a simple, uncomfortable question: how many published findings actually hold up when you repeat them carefully?

In plain words

The key idea

Take 100 studies from top journals and replicate each one with high statistical power and, where possible, the original materials and authors' input. Then count how many succeed and by how much the effects shrink.

Think of it like: Like a restaurant critic returning to 100 acclaimed restaurants to see whether the dish that earned the rave review is still as good the second time. A lot were not.

What they did

The method

270 researchers replicated 100 studies from three leading psychology journals, using larger samples and preregistered plans.

What they found

The results

About 36% of replications reached statistical significance, versus 97% of the originals, and effects were roughly half as large.

Does it hold up?

The honest part. What is solid, and where to be careful.

Judging what counts as a successful replication is itself contested; the paper uses several metrics rather than one.
A failed replication does not by itself prove the original wrong; context or hidden moderators can matter.
The 100 studies came from three journals in one year, so the exact percentage should not be over-generalized.

So what, for you

If you are a curious newcomer

Whenever you read a single exciting psychology finding, remember this: across 100 top studies, fewer than half held up when repeated. One study is a clue, not a fact.