Talk:Design effect

Fix definition

I'm pretty sure design effect is used in the context of systematic random sampling as well as cluster sampling and possibly has a broader definition; Design effect = V1/V2 where V1 = The variance of the sampling system in question V2 = The variance of simple random sampling. Greenmachine2000 (talk) 05:55, 1 April 2012 (UTC)

I totally agree. I've been working on extending this article to be much more extensive. Tal Galili (talk) 12:28, 3 June 2021 (UTC)

This was addressed in the updated article. Tal Galili (talk) 09:12, 7 May 2024 (UTC)

GA Review

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

This review is transcluded from Talk:Design effect/GA1. The edit link for this section can be used to add comments to the review.

Nominator: Talgalili (talk · contribs) 14:14, 9 May 2024 (UTC)

Reviewer: Phlsph7 (talk · contribs) 08:04, 11 May 2024 (UTC)

Hello Talgalili and thanks for all your improvements to this article. However, despite the improvements, the article fails criterion 2b since there are too many unreferenced paragraphs. Examples are the paragraphs starting with "In other words, $Deff$ measures the extent", "Even so, in various cases a researcher might approximate", and "An example of a design which can lead". According to criterion 2b, these passages require inline citations "no later than the end of the paragraph". This was already pointed out at the featured article review a few days ago at Wikipedia:Featured_article_candidates/Design_effect/archive1. I suggest that you add all the relevant references before a renomination. It would probably best to have a peer review before that.

A few other observations

• WP:EARWIG detects potential copyright violations with http://www150.statcan.gc.ca/n1/en/pub/12-001-x/2004002/article/7751-eng.pdf. This should be checked in more detail before a renomination as well.
• The article lacks a short description
• many headlines in the table of contents have problems, such as "Haphazard weights with estimated ratio-mean (?'"`UNIQ--postMath-000000C2-QINU`"'?) - Kish's design effect" and "Spencer's Deff for estimated total (?'"`UNIQ--postMath-000000EB-QINU`"'?)"
• To have a more impersonal encyclopedic style, replace first-person pronouns like "we" and "our" with alternative formulations
• this can lead to an increase both the variance and bias of the weighted estimator. add "of" before "increase"
• In their works, Kish and others highlights several known replace "highlights" with "highlight"
• Reasons for non-response are varied and depends on the context replace "depends" with "depend"
• https://stats.stackexchange.com/questions/525741/how-to-estimate-the-approximate-variance-of-the-weighted-mean/525770#525770 and https://github.com/pewresearch/pewmethods are probably unreliable sources.

Phlsph7 (talk) 08:04, 11 May 2024 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Followup to QA review

Hey Phlsph7, thanks for the review!

First, I'd like to mention that this article has been in the works for 2 years, and has just been accepted in the Wiki Journal of Science (link), so it has gone through a peer review process (which has greaty helped in improving it!).

Hi Talgalili and thanks for putting so much time and effort into this article. Phlsph7 (talk) 07:23, 12 May 2024 (UTC)

Hey Phlsph7, thank you for all the responses! It will take me time to get to it, but I'll take care of them in the future. Tal Galili (talk) 14:26, 12 May 2024 (UTC)

Some questions and followups: I suggest that you add all the relevant references before a renomination.

Could I please ask you to add "citation needed" in whichever sections you think are missing citations? This article already has 52 citations (most of them are referenced numerous times). So more specific direction would help. For example, the sentences you mentioned are based on references given in previous paragraphs (so I can continue to introduce the same references over and over, but knowing which sentence appears to most require it - would be very helpful).

I added citation needed tags at the end of the relevant paragraphs. Feel free to remove any or all if you disagree. Basically, every claim in the body of the article that is not a trivial statement like "the sky is blue" needs to be supported by a reference, including claims in lists and tables. How many references the article in total has is not so important. The lead paragraphs do not need references if they only summarize sourced material in the body of the article. Phlsph7 (talk) 07:23, 12 May 2024 (UTC)

It would probably best to have a peer review before that.

Would you still recommend it even after it went through the Wiki Journal process? (editor + two academic peer reviews + a "math expert" editor). I'm fine doing it, just checking.

I recommended a peer review because the article failed 2 reviews in a row. On a superficial look, the article seems to me quite technical, possibly too technical for a general encyclopedia. An example would be all the technical vocabulary that is being introduced at every step. The section structure is unconventional, with countless sub, subsub, and subsubsubsections while some section titles are variable names. This is not explicitly prohibited by the GA criteria, but it could raise flags for other reviewers.

I haven't given the article an indepth review so the points I raised are not exhaustive but just issues that stood out to me. I don't know whether the article would be ready for GA once they are addressed or how the Wiki Journal process differs from the GA process. Phlsph7 (talk) 07:23, 12 May 2024 (UTC)

copyright violations...

Could you share a link to where this is mentioned? The review process of the wiki journal indicated no copyright issues.

You can find Earwig's Copyvio Detector at https://copyvios.toolforge.org/. It's not foolproof and only shows possible copyvios so you have to use your own judgment and consider its results in the light of WP:CLOSEPARAPHRASE. Phlsph7 (talk) 07:23, 12 May 2024 (UTC)

The article lacks a short description

What's the issue with the first three paragraphs? (they are the abstract)

I meant WP:SHORTDESC. This is not about the lead paragraphs but about adding the Template:Short description to the article. Phlsph7 (talk) 07:23, 12 May 2024 (UTC)

many headlines in the table

I have no idea how to fix it. Any suggestion on who to ask?

I assume it is because of the math formulas used in the section titles. You could try reformulating the section titles so you do not need to use math formulas in them, which is not common anyways.

For this and other general questions, you can try the Wikipedia:Teahouse. Phlsph7 (talk) 07:23, 12 May 2024 (UTC)

are probably unreliable sources.

The two links are good resources. The github onen is the location of the Python package (it's the source of truth for that comment)

In that case, it should be fine. Phlsph7 (talk) 07:23, 12 May 2024 (UTC)

The stackexchange link is a supportive link. The original reference is Sarndal, ... (1992). Model Assisted Survey Sampling. I'll add that reference too.

See the entry "Stack Exchange" at Wikipedia:Reliable_sources/Perennial_sources. If it is not required then it would be better to remove it. Phlsph7 (talk) 07:23, 12 May 2024 (UTC)

I'll do fixes based on your other comments. Thanks upfront for your help! Tal Galili (talk) 18:26, 11 May 2024 (UTC)