Wikipedia:Featured article candidates/Margin of error/archive1

Margin of error

--Fadethree I am self-nominating this because I have seen some amazing misinterpretations of this concept by both members of the press and the voting public. I think the topic is particularly germane given the election seasons in the US and elsewhere, which is why I am rushing to self-nominate. That said, I am very, very open to feedback, especially relating to how to make the page more accessible to the interested reader. I hope that the Do-It-Yourself Excel heading is appropriate; I think it would be useful. (<--edit: moved to Wikibooks.) I look forward to hearing from you all. 23:59 03 Oct, 2004 (UTC)

• Certainly it is timely during a presidential election campaign. The article still needs some work in some of the "displayed" TeX. Michael Hardy 00:17, 4 Oct 2004 (UTC)
  • Thanks, Michael. I have replied to explain the TeX, and the TeX you mention has been moved to Wikibooks. Best, Fadethree 19:12, 7 Oct 2004 (UTC)
• Object. An overall good article that has a big problem. It uses an overly conversational tone (eg "Let us use an running example..."), much of the article must be rephrased. Keep in mind that wikipedia is not a textbook but an encyclopedia. Also the do-it-yourself section needs to be removed, put that on Wikibooks if desired. Support. —siroχo 06:34, Oct 4, 2004 (UTC)
  • Thank you for the excellent recommendations, Siroxo. I prefer the conversational tone to make the topic less intimidating, but I understand that it is informal, so I have changed it. I have moved the do-it-yourself section to Wikibooks and provided a link in this article. Thanks again for your comments; let me know if you have other suggestions. Fadethree 08:19 04 Oct, 2004 (UTC)
• Support - much better! Object because it is written from the point of view that everyone is wrong about margin of error. Perhaps if the misconceptions section was moved downwards then that would sort out the POV problem. Also, I don't know what margin of error until about halfway down the article. Again, moving the first section to after the second section might solve this problem. - Ta bu shi da yu 09:12, 4 Oct 2004 (UTC)
  • Thank you, Mgm, for your support. I agree with ta bu shi da yu (big fish??), however, that there is sufficient POV to try some changes. I have shifted the misinterpretation section down and kept the example in place. I was going for a pedagogical approach by addressing readers at the level of their misconceptions, but this may indeed come off as POV. (continued...)
  • Your point about not knowing what the margin of error is until halfway down... this is a bigger challenge. Many difficult concepts cannot be truly understood until at least halfway down. The problem with "margin of error" is that it sounds like it should be much more than it is. Really, all it is, like I noted in the opening, is a transformation of the sample size that tries to reflect sampling variance. It picks a random percentage (50%) and reports the 99% confidence interval radius at that point regardless of whether that percentage is actually reported by the poll. Many users then take the margin of error and apply it to other reported percentages, though this is inappropriate. There are thus many opportunities for confusion that I am trying to address. I have made changes to the opening paragraph that I hope make this clearer. Let me know what you think. Fadethree 10:12, 4 Oct 2004 (UTC).
    • I'm feeling a little dense when reading this article, because I still don't know what margin of error is and why it is important. I'm no statistical genius! What would be helpful for me would be if you could try to sum up what margin of error is, and why it is important (without referring to misconceptions, as I have none as I don't have the first idea what it's all about). - Ta bu shi da yu 04:57, 5 Oct 2004 (UTC)
      • Thanks again for the comment; it is helpful. I added these lines to the definition: "The margin of error is an expression of the extent to which a poll's reported percentages may vary if the poll were to be taken again. The larger the margin of error, the less confidence one has that the polls reported percentages are close to the "true" percentages in the population.." I should have thought about saying it like this before. Let me know if you like how it reads now. Fadethree 06:13, 5 Oct 2004 (UTC)
        • Sorry for taking so long to strike my objections. This is very much better, and with the new graphic I most definitely support this article! Good work on a difficult subject (well, for some people like myself). - Ta bu shi da yu 05:39, 6 Oct 2004 (UTC)
• Support. It made a confusing mathematics topic easier to understand. I don't think the article suggests "everyone" is wrong about the margin of error. It only says it's often misinterpretted. And I think that could be considered a fact instead of POV, especially when Fadethree can prove it by linking a news article. That said, the article's language might benefit from some less loaded word choices. [[User:MacGyverMagic|Mgm|^(talk)]] 09:44, Oct 4, 2004 (UTC)
• I did not peruse the article entirely, so I will not vote at present. But I did notice that "margin of error" is in bold at every instance: I would imagine that using bold in the first case alone would suffice. -- Emsworth 01:46, 5 Oct 2004 (UTC)
  • Thanks, Emsworth. Neutrality took care of this and made some other very helpful edits. I kind of like having the phrase in bold throughout the article; I think it allows a user to skim for information easier. Featured articles should follow stricter conventions, however. Best, Fadethree 05:50, 5 Oct 2004 (UTC)
• Picture added. I hope you like it. I also hope that this article becomes featured while polls are still a hot topic in the US. I would imagine it would get a lot more hits if it coincided, say, with the aftermath of the third debate. Still, I understand that this is a lengthy process with many other articles in line. Best, Fadethree 08:45, 5 Oct 2004 (UTC)
• Comment: this may be a restatement of something said above, but I have a problem with the equation given in the lead, which is only valid for a 99% confidence interval. Surely the lead must be the generalized form of the equation. Jgm 18:35, 5 Oct 2004 (UTC)
  • Thanks, Jgm. 99% is the standard for the vast majority of the polls reported by the media. The last line in the definition follows, "The numerator of the above equation can change for 95 percent certainty (0.98) and for 90 percent certainty (0.82) depending on the standards of the polling organization." I also have a note in the Arguments Against... section as follows: "There is no agreed-upon confidence level. Most pollsters use 99 percent, but many use 95 percent or 90 percent; this makes their polls look more accurate." I would rather keep the equation simple in the beginning, because defining "confidence" takes a little work. Your thoughts? Fadethree 21:28, 5 Oct 2004 (UTC)
  • Note, there was brief discussion on this in the article's talk page, and I made the changes to the equations a while back. Fadethree 17:35, 11 Oct 2004 (UTC)
• We should use a made up and neutral example, so we need not use disclaimers like "It should be clear that the choice of poll and who is leading is irrelevant to the presentation of the concepts." These disclaimers are ugly in the flow of a well-written article. ✏ Sverdrup 10:14, 6 Oct 2004 (UTC)
  • This is a great point. I've made a reply at the margin of error talk page, and I'd like to solicit some other opinions as well. I frame it as an issue of accessibility and audience there, but perhaps this is an issue of standards for style. Fadethree 14:51, 6 Oct 2004 (UTC)
• Support. It would be timely to see this on the main page before the election. --Redquark 05:34, 8 Oct 2004 (UTC)
• Support Wolfman 07:31, 12 Oct 2004 (UTC)