Jump to content

Talk:Pseudo-R-squared

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Feedback from New Page Review process

[edit]

I left the following feedback for the creator/future reviewers while reviewing this article: Great work!.

HᴇʀᴘᴇᴛᴏGᴇɴᴇꜱɪꜱ (talk) 14:56, 26 July 2022 (UTC)[reply]

Logistic regression is not always heteroscedastic

[edit]

If it were, what's your source? 68.134.243.51 (talk) 03:55, 8 December 2022 (UTC)[reply]

Discussing the same value under two names

[edit]

I think the authors made a mistake in not realizing that what is called R2L (Cohen) here is mathematically the same as McFadden. In R, deviance(glm()) = 2*logLik(glm) and then the two formulae offered are just paraphrases of each other: (deviance(m0)-deviance(m1))/deviance(m0) is of course the same as 1-(deviance(m1)/deviance(m0)), and if deviance(m) is replaced with 2*logLik(m), the two formulae are identical. Thus, at this point the article is needlessly confusing. — Preceding unsigned comment added by 98.185.254.231 (talk) 20:05, 7 January 2023 (UTC)[reply]

Wiki Education assignment: PSY8052-Advanced Statistics II

[edit]

This article was the subject of a Wiki Education Foundation-supported course assignment, between 16 January 2023 and 5 May 2023. Further details are available on the course page. Student editor(s): Bmpohl (article contribs).

— Assignment last updated by Bmpohl (talk) 21:09, 4 May 2023 (UTC)[reply]