12 August 2017
→Controlling procedures: replaced "k" with "m" for consistency with other articles and with other section in this article.
10 August 2017
7 August 2017
6 August 2017
27 July 2017
17 June 2017
saying "higher" threshold seems misleading. the significance threshold is actually LOWERED. saying "stricter" avoids confusion.
16 June 2017
6 June 2017
→Large-scale multiple testing: rephrased to avoid false statements, e.g., that it is possible to identify false positives
13 May 2017
Undid revision 780044662 by Stateditor (talk) we don't need that arbitrary citation, given that we have a hyperlink to a full wiki article
12 May 2017
7 May 2017
→Large-scale multiple testing: fixed sentence fragment
→Large-scale multiple testing: a blog is not really a reputable statistical source
sentence was unnecessarily verbose and jargon-ish
→Large-scale multiple testing: the cited source makes no mention of multiple testing. in fact, that has been a criticism of the cited source.
6 May 2017
5 May 2017
17 March 2017
"look-elsewhere effect" appears to be a rather esoteric term compared to the more generally used "problem of multiple comparisons"
16 March 2017
14 March 2017
→Example: removed unsourced, unnecessary section. we already have examples. readers interested in coin-flipping analogy explaining error rates can see corresponding sections in articles about p-values or FWER
→Controlling procedures: no need to redefine the problem here
→Definition: misc minor corrections
→Definition: corrected overly-specific definition of multiple comparisons
→Controlling procedures: m
→Definition: comparisons aren't necessarily between 2 groups. for example, you could compare an observed value to a hypothetical null value.
10 March 2017
9 March 2017
4 March 2017
17 February 2017
→Controlling procedures: Journal cites: fix volume param, added 1 issue number using AWB (12142)m
16 December 2016
29 November 2016
24 November 2016
23 November 2016
22 November 2016
→Further reading: re-categorisation per CFDS, replaced: Category:Hypothesis testing → Category:Statistical hypothesis testing using AWBm
→Controlling procedures: removed sentence that was unclear and unsourced
10 November 2016
→History: weird to say the issue of multiple comparisons "came back" in the 1980s. Also, one of the citations in the sentence was from 1996--a far cry from the 1980s.
→History: even if we don't list the conferences, we should give the reader some image of how often it is, and making clear it is one conference and not a few
→History: minor fix of previous edit
→History: way too much detail. noting that the conference exists is relevant, but listing all those cities of conferences in specific years is not.
9 November 2016
8 November 2016
7 November 2016
→Controlling procedures: cleaned up somewhat misleading wording
→Controlling procedures: perfect negative dependence would inflate the FWER
→Post-hoc testing of ANOVAs: This entire section is misleading and unnecessary. Whether multiple comparisons are conducted after an omnibus ANOVA or in some other context is irrelevant to the application of most of the given procedures.
→Large-scale multiple testing: the sentence was self-contradicting