Talk:Holm–Bonferroni method

The description is correct, it is statistics that can feel backwards! A researcher typically looks for some difference, e.g. "this new drug is better than the usual drug". The null hypothesis says: "Nothing special here, the two drugs give the same results". Therefore, the REJECTED null hypotheses (those with the low p-values) are the interesting results. Lj phys (talk) 14:28, 25 January 2018 (UTC)

I'm with the previous comment. It looks like in

• Reject the null hypotheses $H_{(1)}\ldots H_{(k-1)}$ and do not reject $H_{(k)}\ldots H_{(m)}$

we're rejecting the hypotheses with the lowest associated p values. (But not quite sure enough to edit.) Eac2222 (talk) —Preceding undated comment added 19:34, 19 June 2017 (UTC)

Is the direction of the inequality in the equation in the third bullet correct? It seems opposite to the Bonferroni correction. If it is right, I think a bit more explanation would be useful.130.215.14.21 (talk) 18:25, 3 November 2015 (UTC)

I'm a bit confused about the bit on adjusted p-values. what does capital 'N' denote - this variable appears to not be declared. Is this what previously was referred to as 'm'? Cheers 193.62.66.241 (talk) 10:35, 15 October 2013 (UTC)

I found this concise explanation to be extremely helpful. Thank you to whoever authored this page!-- 129.2.60.183 (talk) 19:51, 16 November 2007 (UTC)

I find this explanation of the Holm-Bonferroni method a little confusing - I take it from the logic that it is the null hypothesis that is being rejected at each step. The way it reads sounds like the experimental hypothesis that is being rejected. Interlope (talk) 01:59, 7 September 2008 (UTC)

This is an excellent description of the method. However I am not sure which test is meant in the line "Each intersection is tested using the simple Bonferroni test". I can't find any references to a "Bonferroni test" that is suitable for a single pairwise comparison. Could someone more knowledgable than me provide more detail? BruceMcAdam (talk) 10:24, 11 November 2008 (UTC)

I have now finished introducing major additions to the article (see the following diff: https://en.wiki.x.io/w/index.php?title=Holm%E2%80%93Bonferroni_method&diff=630373644&oldid=627604519). These are based on the work done by students in the Tel-Aviv University course "multiple comparisons". You may see a relevant revision history here: https://en.wiki.x.io/w/index.php?title=User:Talgalili/sandbox/Holm%E2%80%93Bonferroni_method&action=history). The work was done by: https://en.wiki.x.io/wiki/User:Shay66

Tal Galili (talk) 13:50, 20 October 2014 (UTC)

weighted Holm-Bonferroni

w(i) does not appear to be defined. 99.96.68.14 (talk) 17:49, 23 February 2023 (UTC)

Adjusted p-values

There seems to be an error in the example, revisited in this section. Applying the given formula, ${\widetilde {p}}_{(i)}=\max _{j\leq i}\left\{(m-j+1)p_{(j)}\right\},$ to the example in which $p_{(1)}=0.005,p_{(2)}=0.01,p_{(3)}=0.03,{\text{ and }}p_{(4)}=0.04$, I get ${\widetilde {p}}_{(1)}=4\times 0.005=0.02,{\widetilde {p}}_{(2)}=\max\{0.02,3\times 0.01\}=0.03,$ ${\widetilde {p}}_{(3)}=\max\{0.03,2\times 0.03\}=0.06,{\text{ and }}{\widetilde {p}}_{(4)}=\max\{0.06,1\times 0.04\}=0.06$.Izmirlig (talk) 19:57, 14 February 2024 (UTC)