Timeline for What is the meaning of p values and t values in statistical tests?
Current License: CC BY-SA 3.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 11, 2017 at 22:10 | review | Low quality posts | |||
| Sep 11, 2017 at 22:46 | |||||
| Jan 3, 2015 at 23:37 | comment | added | Silverfish | A feature of this answer I like a lot is the clear explanation that p-values are calculated using a null model, even if we don't (subjectively) believe the null model is actually true. I think the fact test statistics are calculated under a model is a key point that many students struggle with. | |
| Jan 3, 2015 at 23:33 | comment | added | Silverfish | I also felt your conclusions are phrased too strongly. If we reject the null, we have significant evidence against it, but don't know that it's false. When we fail to reject the null, that certainly doesn't mean the null is true (though it may well be). As a more general comment I have the feeling the test you're describing, in quite abstract terms, is not likely to be clear to a learner who is just learning how to perform a test. The lack of a clearly defined test statistic doesn't sit well with the original question asking how to interpret t-statistic too. | |
| Jan 3, 2015 at 23:22 | comment | added | Silverfish | Since the original question is asking what a p-value is, I thought that getting that definition across clearly was important. Just saying "more extreme" isn't in itself very helpful without explaining what "more extreme" might mean - that's a weakness of most answers in this thread I think. Only whuber's answer and the "tea test" one seem to really explain why the "more extreme" matters too. | |
| Jan 3, 2015 at 19:45 | comment | added | cgreen | @Silverfish Could you elaborate? Probably it would be more correct to call the p-value the probability of making that observation OR a more extreme observation. But it sounds like you have a deeper criticism. | |
| Dec 30, 2014 at 13:46 | comment | added | Silverfish | I disagree with the p being defined as "We call the probability of observing this data in that alternate world the p-value" and also the strength of the conclusion being drawn (especially failure to reject the null). | |
| Nov 3, 2014 at 1:21 | history | answered | cgreen | CC BY-SA 3.0 |