Timeline for Why would parametric statistics ever be preferred over nonparametric?
Current License: CC BY-SA 3.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 30, 2015 at 14:47 | comment | added | conjugateprior | +1 for noting that we have absolutely no chance of estimating anything of value in a regression without some assumptions about the regression function. | |
| Jul 30, 2015 at 14:46 | comment | added | conjugateprior | @cagdas-ozgenc It tells you that p-values are conditional on the model. But it's not clear how things could be otherwise... | |
| Jul 30, 2015 at 14:41 | comment | added | en1 | @dsaxton what you say is true about testing different hypotheses but people still interpret them the same way. Then there's also regression, where the insights you get between nonparametric and parametric analyses are pretty much the same. | |
| Jul 30, 2015 at 14:22 | comment | added | Cagdas Ozgenc | Yes, and that adds a model bias. What does that tell about the p-values researchers reporting? | |
| Jul 30, 2015 at 14:16 | history | answered | dsaxton | CC BY-SA 3.0 |