Multiple comparison problem, illustrated. #metascience
Bearded men more likely to be sexist. It's science!
Bearded men more likely to be sexist. It's science!
How did they find that out?
By doing analysis until p < .05.
However, not everyone is this honest about the reporting. Most researchers don't tell you all the steps that led up to the magical number .05.
By doing analysis until p < .05.
However, not everyone is this honest about the reporting. Most researchers don't tell you all the steps that led up to the magical number .05.
What about replications?
They found nothing, of course.
They found nothing, of course.
And which study do researchers remember and cite?
The cool study that find something interesting of course, not the boring replication. After all, science is about finding new and interesting things, not about finding things that are true! (End of sarcasm.)
The cool study that find something interesting of course, not the boring replication. After all, science is about finding new and interesting things, not about finding things that are true! (End of sarcasm.)
• • •
Missing some Tweet in this thread? You can try to
force a refresh
