Services & ResourcesWolfram ForumsMathGroup Archive

[Date Index] [Thread Index] [Author Index]

Re: Levene and BrownForsythe Tests

• To: mathgroup at smc.vnet.net
• Subject: [mg122157] Re: Levene and BrownForsythe Tests
• From: Andy Ross <andyr at wolfram.com>
• Date: Mon, 17 Oct 2011 08:09:07 -0400 (EDT)
• Delivered-to: l-mathgroup@mail-archive0.wolfram.com
• References: <201110161104.HAA14700@smc.vnet.net>

The definition present in 8.0 and 8.0.1 use the squares of differences 
as you say. That definition is a bit obscure compared to the more common 
interpretation you mention. The code was updated some time ago to use 
the more common definition and should be present in the next minor 
release of Mathematica.

Sorry about the confusion.

Andy Ross
Wolfram Research Inc.

On 10/16/2011 6:04 AM, Tony Harker wrote:
> Can a statistician out there please tell me whether the Mathematica
> implementation of the Levene and Brown-Forsythe tests for equality of
> distribution variances is correct? The Mathematica scheme defines an
> auxiliary data set formed from the squares of the differences between the
> data and the data means or medians, whereas other sources use the modulus of
> that difference.
>
>
>
>    Tony Harker
>
>
>
> Department of Physics and Astronomy
>
> University College London
> Gower Street
> London
> WC1E 6BT
>
> Tel: (within UK)    020 7679 3404
>       (overseas ) +44 20 7679 3404
> E:    a.harker at ucl.ac.uk

• References:
  • Levene and BrownForsythe Tests
    • From: "Tony Harker" <a.harker@ucl.ac.uk>