Talk:Type III error

Latest comment: 6 months ago by 2601:404:D400:4AF0:5496:6ACA:2DD3:3E29 in topic Wording

Not Raiffa, but Kimball

edit

It seems that the attribution to Howard Raiffa, in this sense of the meaning of the term "Type III error" is wrong. see Four types of error currently (as at 22 June 2006) under construction. It seems that the "naming rights" for the term, in that specific meaning, must go to:

  • Kimball, A.W., "Errors of the Third Kind in Statistical Consulting", Journal of the American Statistical Association, Vol.52, No.278, (June 1957), pp.133-142.

Once I have finished my work on creating Four types of error, I will follow this upLindsay658 02:01, 22 June 2006 (UTC)Reply

David

edit

Didn't David first coin the term "type III" error in technical writing? Shouldn't this section indicate that her statement is the canonical origin of the term, even if it has subsequently been reinterpreted and had its meaning extended/generalized by others? Hobsonlane (talk) 23:49, 23 September 2015 (UTC)Reply

Merge

edit

I have just merged the section into Four types of error, and I will do my best to also merge the others two (type I and type II)Lindsay658 03:31, 26 June 2006 (UTC)Reply

Type III error

edit

Shouldn't this redirect here? ~ BlueNovember 11:07, 27 April 2007 (UTC)Reply

No, that's a different thing. - dcljr (talk) 19:51, 7 May 2020 (UTC)Reply

Wording

edit

This wording in the first paragraph seems incorrect/inaccurate: "i.e. when the correct hypothesis is rejected but for the wrong reason." It seems like the wording in the Mosteller section below would be more accurate (or less confusing?): "correctly rejecting the null hypothesis for the wrong reason". This seems more consistent/analogous with/to a Type I error, in which the hypothesis is rejected. (But importantly: not identical to a Type I error, as the former quote implies--since in the case of a Type III, the rejection is ultimately correct.) Consensus to change?

2601:404:D400:4AF0:5496:6ACA:2DD3:3E29 (talk) 00:29, 1 May 2024 (UTC)Reply