Jump to content

Talk:Repeated measures design

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

This article seems to be about crossover designs, which is a very small subclass of repeated measures designs (although very important)

[edit]

WARNING: The introduction needs significant editing. This article seems to think that repeated measures designs are identical with crossover studies. An experienced editor should flag this article with a WARNING, imho. Kiefer.Wolfowitz (talk) 18:25, 20 June 2009 (UTC)[reply]

I removed the worst confusions (crossover=repeated measurements) and the toxic references. There are two on-line references given (to lecture notes), which did not horrify me, so I left them, but I am sure that better on-line references can be found. Kiefer.Wolfowitz (talk) 19:09, 20 June 2009 (UTC)[reply]
I think what is missing in this article is an explanation what Repeated measures design IS. —Preceding unsigned comment added by 93.133.174.126 (talk) 19:17, 26 December 2009 (UTC)[reply]

This article needs much more editing on the advantages and disadvantages if anyone would like to do that. counterbalancing, order effects (of which practice effects is a part of), latin squares, and time interval between treatments are all not mentioned and are big parts of repeated measures design. also i think that the crossover studies should just be under "see also"

in addition, longitudinal studies are not a subset of Repeated measures design. they can be used in independent groups design as well. — Preceding unsigned comment added by 128.189.223.183 (talk) 21:17, 2 November 2011 (UTC)[reply]

Minke, 2006

[edit]

Can someone please put a more complete reference in for "Minke, 2006"? --Slashme (talk) 07:22, 29 January 2010 (UTC)[reply]

Oh, I've just figured it out: The date is a typo. --Slashme (talk) 07:44, 29 January 2010 (UTC)[reply]


Assumptions of repeated measures ANOVA

[edit]

I believe that the assumption of normality is incorrect. The sample mean of each condition should be normally distributed. Normality of the dependent variable is sufficient to guarantee normality of the sample mean, but it is not necessary (see the Central Limit Theorem). Bradknox (talk) 04:15, 30 April 2013 (UTC)[reply]

Partitioning of error

[edit]

In the section Partitioning of error, one might be more explicit about what n and k refer to. Presumably, n refers to the number of subjects/participants, and k to the number of measurements. A second comment relates to the definition of dftotal for repeated measures ANOVA, i.e., dftotal = (k-1) + (n-1) + ((n-k)-(n-1)). I doubt whether the third term at the right-hand side of the equation (pertaining to dferror) is correct. Instead of ((n-k)-(n-1)), this is normally written as (k-1)(n-1). But I'm not sure about this. — Preceding unsigned comment added by LKK346 (talkcontribs) 19:31, 8 July 2015 (UTC)[reply]

I agree, degrees of freedom for error term must be (k-1)(n-1), as in 'Foundations of Clinical research, Applications to Practice PEARSON ISBN-10 1-292-02036-9, p.479 — Preceding unsigned comment added by 91.180.76.172 (talk) 13:23, 10 January 2018 (UTC)[reply]

Wiki Education assignment: PSY8052-Advanced Statistics II

[edit]

This article was the subject of a Wiki Education Foundation-supported course assignment, between 12 January 2022 and 7 May 2022. Further details are available on the course page. Student editor(s): Stargazer.em (article contribs).