Talk:Vuong's closeness test

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

Gaps in Definition[edit]

There are a number of components of the equations which are never defined (e.g. capital L) — Preceding unsigned comment added by 129.170.195.76 (talk) 21:30, 11 June 2019 (UTC)[reply]

Contradictory?[edit]

The article states, "They can be nested, non-nested or overlapping." But then it also states, "Vuong's test for non-nested models has sometimes been used to determine whether zero-inflation is present in data. As a given model and its zero-inflated counterpart are not non-nested, this is an erroneous use of the test." This is confusing and seems to be contradictory. The introduction states that the test can be used for nested models, but the conclusion states that it is a test for non-nested models. Which is true? — Preceding unsigned comment added by 204.74.0.112 (talk) 15:09, 11 December 2018 (UTC)[reply]

I've had an initial look at the paper, [1], and it claims "They can be nested, non-nested or overlapping.", so I'll remove the statement in the last paragraph from "As a given model and its zero-inflated counterpart are not non-nested, this is an erroneous use of the test." and remove the Expert-subject|statistics|date=November 2008 note from the page.
I'll note this on Wikiproject Stats for any further comment.
Newystats (talk) 10:27, 21 March 2019 (UTC)[reply]

Note[edit]

A preprint that seems to relate to this article is presently freely downloadable at http://ideas.repec.org/p/clt/sswopa/605.html Melcombe (talk) 09:06, 6 August 2008 (UTC)[reply]


References[edit]

  1. ^ Vuong, Quang H. (1989). "Likelihood Ratio Tests for Model Selection and non-nested Hypotheses". Econometrica. 57 (2): 307–333. JSTOR 1912557.