Dr Mircea Zloteanu πΊππ#statstab #234 Not all alphas can be justified
Thoughts: When using discrete distributions, with few outcomes, your alpha cannot always take the values you want. Same with the distribution of p-values under the null.
#statstab #189 Post Hoc Power: Not Empowering, Just Misleading
Thoughts: "the observed power is a 1:1 function of the P value" If you need a ref for why "post hoc" power is nonsense (paywalled).
#power #posthoc #QRPs #errorcontrol #poweranalysis
https://www.sciencedirect.com/science/article/pii/S0022480420305023
#statstab #186 Are Multiple Contrast Tests Superior to the ANOVA?
Thoughts: Focuses more on the statistical aspects of the comparison than the theoretical ones. But insightful for newbies.
#nhst #anova #ttest #errorcontrol
#typeI #pvalues
https://www.degruyter.com/document/doi/10.1515/ijb-2012-0020/html?lang=en
Are Multiple Contrast Tests Superior to the ANOVA?
Multiple contrast tests can be used to test arbitrary linear hypotheses by providing local and global test decisions as well as simultaneous confidence intervals. The ANOVA- F -test on the contrary can be used to test the global null hypothesis of no treatment effect. Thus, multiple contrast tests provide more information than the analysis of variance (ANOVA) by offering which levels cause the significance. We compare the exact powers of the ANOVA- F -test and multiple contrast tests to reject the global null hypothesis. Hereby, we compute their least favorable configurations (LFCs). It turns out that both procedures have the same LFCs under certain conditions. Exact power investigations show that their powers are equal to detect their LFCs.
#statstab #84 Using the data twice
Thoughts: A very good overview of conditional and unconditional inferences based on the modelling decisions we make. I can't say I've seen many discuss this.
#errorcontrol #inference #modelling #stats
https://www.alexpghayes.com/post/2020-05-04_using-the-data-twice/
#statstab #74 When to adjust alpha during multiple testing: disjunction, conjunction, and individual testing by @MarkRubin
Thoughts: Error control = NHST. But doing it right is a nightmare. This paper made sense to me.