Talk:Equivalence test

Section "Comparison between t-test and equivalence test​"

This section looks suspiciously as though it was included by the author(s) of the cited reference, with mediocre translation and mention of the "second Figure" like something copied from an academic document.

13:17, 30 May 2019‎ Xor2k . . (9,256 bytes) (+3,360)‎

—DIV (1.129.107.53 (talk) 12:53, 18 October 2020 (UTC))

For some reasons, my english has been poor recently, I apologize for that. What would be a better way to refer to that figure? "The second Figure" refered to the second figure in the Wikipedia article, although, by coincidence, it is the second figure in the academic article as well. In articles, Figure is also written with a capital F as figures have names (Figure 1, Figure 2, ...) for reference, which I now changed into a lower case f, since it is unnamed in the Wikipedia article. I changed it to "the figure below", which may actually be below or right, depending on which device one is using to see the website (either desktop or mobile). This solution is also not scalable when more figures are added to the article. I've also seen "the second figure" in some other Wikipedia article, which is the reason I've written it originally, but it would be really cool if there had been a more general, standard solution.

Indeed, I'm one of the authors of the article. Should that pose a problem? Although the math behind that article really is not rocket science at all, I think an easy way to comprehensively and quantitatively compare t-test and equivalence test is of great general interest and really belongs here. Especially the acceptance diagram from the second (Wikipedia article) figure adds a macroscopic level to the current microscopic-only understanding of the difference between t-test and equivalence test of the first (Wikipedia article) figure. Also, it should be compliant to the Wikipedia rules, since it is not original research. The original research has been done in the peer-reviewed academic article.

Our research showed that at least in the field of transportation, the t-test is often used incorrectly. Also, I have never heard about the equivalence test before doing research on the subject, neither at school nor in the statistics lectures I had during my math studies. According to Statistical hypothesis testing#History, "Modern hypothesis testing is an inconsistent hybrid of the Fisher vs Neyman/Pearson formulation, methods and terminology developed in the early 20th century.", which matches my experience and which is the reason why I highly disliked hypothesis testing altogether and specialized in optimization during studies instead.

I want to contribute to make hypothesis testing more approachable both with the academic article as well as with my Wikipedia contribution, which, by the way, took much more effort than it looks like. I have e.g. reworked the entire code of the figure, although it only looks like $\sigma$ instead of $\nu$. I can also provide more material, e.g. code/formulas for the second figure, which I probably would rewrite once again, this time in Python, since the current R version is, even after rework, (language-intrinsically) not really nice to look at, friendly speaking.

Best from Berlin, --Xor2k (talk) 00:32, 19 January 2021 (UTC)

I took some time and redid the second figure. Source code is available in the image description or here: https://gist.github.com/xor2k/2256abe4106cc8ac04a24f42bd36f2c6 I have also added a simulation in addition to the analytic results. It optionally runs on a GPU, so rendering the image just takes a few seconds. --178.6.226.247 (talk) 12:17, 30 October 2021 (UTC)

Broken References

The edit by Vjayvr on 13:25, 3 July 2022‎ (and other edits probably as well) have probably broken a lot of references: new refs were added without adjusting reference numbers. For this reason, it is not a good idea to have reference numbers anyway. Better use named references. I'll try to figure out what might make sense and fix it, but would be good if somebody else had a look as well. --Xor2k (talk) 19:25, 10 April 2023 (UTC)

Okay, I figured out what was going on: basically the edit by Vjayvr on 13:25, 3 July 2022‎ broke all references. The comment of Vjayvr was "I've corrected a few grammatical errors.", so it looks like the user was not aware of what was done. Reverting the edits on top of the latest version was a lot of work. Please be more careful when editing Wikipedia articles in the future - or leave it alone. Seriously, nobody needs grammar edits if it technically breaks major parts of the article. I don't want to discourage newcomers, but to my understanding this really was an act of carelessness. Xor2k (talk) 21:31, 10 April 2023 (UTC)

I noticed that the first reference referred to the wrong PMID, etc. and had the wrong name for the journal. Suspect that there's more to correct than just what was broken previously, as this mistake goes right back to the creation of the page. 37.10.45.42 (talk) 14:02, 25 August 2023 (UTC)

A section on nonparametric equivalence tests would be interesting

Maybe this paper could be a start https://pubmed.ncbi.nlm.nih.gov/19565562/ Biggerj1 (talk) 21:34, 2 July 2022 (UTC)