Talk:Wilks's lambda distribution
|WikiProject Statistics||(Rated Stub-class, Low-importance)|
Can anybody make this article simple and easy to understand format, along with day-to-day language examples. Thank you.
I would like to suggest including a one-sentence description of what it's used for. Unfortunately, I don't know that. —Preceding unsigned comment added by 184.108.40.206 (talk) 20:28, 24 March 2009 (UTC)
In general I think Wikipedia statistics articles authors need to 1) decide on a format for different tests and 2) stick to that format. Here is my proposal, which makes sense mathematically.
- 1) State name of test/theorem/result, and give some history.
- 2) State a precise definition. Attempt to make the definition self-contained when possible.
- This means:
- a) Expand and substitute other definitions in place except for truly universally understood things (like mean, standard deviation, etc). For example, just write the equation in place versus referring the reader to the XXYY result.
- b) Where other definitions are necessary, make sure that appropriate links point to these definitions.
- c) Avoid buzzwords and terminology. This partly goes back to a) and b); if terminology is absolutely required, then link to it, but never leave the reader guessing with phrases like, a generalization of Foo's bar.
- d) "Is a generalization of," and "Is a special case of" are not appropriate definitions! The word "essentially" should never appear in a definition.
- e) Do not explain a definition by example or special case. Go straight for the general case. a) and b) should make this possible.
- f) Make sure every single assumption is listed.
- 3) Give some intuition, provide examples.
- 4) Give other related work, and most importantly do a pro/con analysis. —Preceding unsigned comment added by 220.127.116.11 (talk) 19:19, 2 April 2011 (UTC)
This article seems more-or-less OK as far as it goes, but there are crucial omissions if one wants to know how to use this in the hypothesis tests for which it is intended. I think I'll look this over for a couple of days and maybe add something on that. Michael Hardy (talk) 18:36, 9 April 2014 (UTC)
Why is this called "Wilks's lambda distribution" and not just "Wilks' lambda distribution". The latter is what Kanti Mardia calls it. I have just read Wikipedia claiming that "Chicago Manual of Style recommends this (the former adding -'s) more modern style", but is looks odd to me. — fnielsen (talk) 13:42, 17 February 2016 (UTC)