Talk:Data transformation (statistics)

From Wikipedia, the free encyclopedia

Contents 1 Reads like a cookbook 2 Used for more than just regression analysis 3 This page needs references. 4 This isn't explained very well

[edit] Reads like a cookbook

I think this article reads too much like a cookbook. There's a decided air of "how-to-do-it" about the whole thing, which is inappropriate in an encyclopedia. Perhaps it should be rephrased, to make it clear that we are considering alternative assumptions about the prior distribution (is it normal? log-normal? the square root of a normal variable? etc.) DavidCBryant 14:02, 12 August 2007 (UTC)

[edit] Used for more than just regression analysis

Transformations are used to meet the assumpations of most parametric tests, surely not just regression analysis (pearsons correlation, ANOVA, t-tests etc)? The first paragraph suggests this is the most common procedure. Maybe change to read something like'...meet the assumptions of parametric testing...'Celticbattlepants 13:40, 26 September 2007 (UTC)

[edit] This page needs references.

Not all statisticians agree on the skewness and kurtosis guidelines. 137.150.105.56 00:12, 26 October 2007 (UTC) rouenpucelle

[edit] This isn't explained very well

this does not even show how to use the transformation after the regression analysis or which transformation to use! 124.181.73.210 (talk) 15:07, 5 May 2008 (UTC)