Timeline for Bayesian Analysis in the Absence of Prior Information?
Current License: CC BY-SA 4.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 10, 2022 at 10:21 | answer | added | utobi | timeline score: 2 | |
| Nov 10, 2022 at 6:00 | history | tweeted | twitter.com/StackStats/status/1590585043877261313 | ||
| Nov 9, 2022 at 17:04 | answer | added | Christian Hennig | timeline score: 2 | |
| Nov 9, 2022 at 15:26 | history | edited | Richard Hardy |
edited tags
|
|
| Nov 9, 2022 at 14:29 | history | became hot network question | |||
| Nov 9, 2022 at 9:31 | comment | added | Sextus Empiricus | Whether you paint on a blank, beige, or vanilla canvas, the painting will look very similar. Professional painters might argue about different qualities of canvas and how pigments can have different adhesion. But that doesn't destroy the analogy and even allows to expand it, we could say that the quality of the analysis is more important than the bias/colour of the naked prior information that is used. | |
| Nov 9, 2022 at 8:26 | comment | added | dipetkov | Is the question about "confidence in prior information" only relevant when using Bayesian approaches? For example, how confident are researchers in the assumptions necessary to do power analysis & sample size calculations for study design? | |
| Nov 9, 2022 at 7:52 | comment | added | Xi'an | The prior distribution is not a "truth" we should aim for. It constitutes a reference measure to which the posterior distribution is compared to assess the information contained in the data. This is why several priors can coexist, why credible intervals are not confidence interval, and the probabilistic meaning of the posterior statements remain mostly epistemic. And definitely not seeking which prior fit the data best. | |
| Nov 9, 2022 at 6:53 | answer | added | Tim | timeline score: 8 | |
| Nov 9, 2022 at 6:28 | history | asked | stats_noob | CC BY-SA 4.0 |