CODEWITHSUNDEEP
intervalsbayesian
Günal
Günal

Reputation: 751

Bayesian interval significance test

As far as I know, if a confidence interval of a parameter covers 0, then this parameter is said to be statistically insignificant at this level. I am wondering whether this is the case for a Bayesian interval. In other words, if a Bayesian interval of a parameter covers 0, then can this parameter said to be statistically insignificant?

Upvotes: 0

Views: 98

Answers (1)

Chris Jackson
Chris Jackson

Reputation: 871

In a Bayesian framework, you talk about posterior probabilities rather than significance. So if an equal-tailed interval for a parameter is centred around positive values but still covers 0, then you can say there's over a 2.5% probability that the parameter isn't actually positive.

Upvotes: 1

Related Questions