[R-sig-eco] Standardising and transformation of explanatory/independent/predictor variables for multiple regression analysis

[Chris Howden]

The only time one might always consider using a transformation on a
response is when it's a ratio. There are 2 reasons for this. Firstly, many
methods will favour the part of the ratio that is above 1 since it is
larger. And secondly ratios aren’t symmetric and are dependent on how we
define it e.g. if we have 2 numbers 2 and 1 then it can be expressed as 2
ratios either 2/1 =2 or 1/2 = 0.5. So taken together this means we can get
very different models depending on how we define the ratio.

One way around this is to take the log of the ratio. The result is now
symmetric about 0 and we should get the same\similar model irrelevant to
how we define the ratio. E.g. log(1/2) = 0.3 and log(2) = -0.3. (although
the parameters might have reversed signs).

Chris Howden B.Sc. (Hons) GStat.
Founding Partner
Data Analysis, Modelling and Training
Evidence Based Strategy/Policy Development, IP Commercialisation and
Innovation
(mobile) +61 (0) 410 689 945
(skype) chris.howden
chris at trickysolutions.com.au




Disclaimer: The information in this email and any attachments to it are
confidential and may contain legally privileged information. If you are
not the named or intended recipient, please delete this communication and
contact us immediately. Please note you are not authorised to copy, use or
disclose this communication or any attachments without our consent.
Although this email has been checked by anti-virus software, there is a
risk that email messages may be corrupted or infected by viruses or other
interferences. No responsibility is accepted for such interference. Unless
expressly stated, the views of the writer are not those of the company.
Tricky Solutions always does our best to provide accurate forecasts and
analyses based on the data supplied, however it is possible that some
important predictors were not included in the data sent to us. Information
provided by us should not be solely relied upon when making decisions and
clients should use their own judgement.

-----Original Message-----
From: r-sig-ecology-bounces at r-project.org
[mailto:r-sig-ecology-bounces at r-project.org] On Behalf Of Scott Foster
Sent: Friday, 5 September 2014 9:04 AM
To: r-sig-ecology at r-project.org
Subject: Re: [R-sig-eco] Standardising and transformation of
explanatory/independent/predictor variables for multiple regression
analysis

Hi again Sam,

I think that you have it.  Extreme values will have more influence, due to
their placement in covariate space.  This is often countered with
transformation (of the covariates) but I tend to think that altering your
data for the sake of the model is the wrong way around.  Nevertheless, it
can be effective.  Especially when there is a reason (from the
application) to do so.  Using other methods may also help, without the
need to transform.

Note that this is not the same issue as transforming the outcomes
(responses).  There I would try my hardest not to transform at all --
transformation can do funny things to the statistical properties of the
outcomes (and it is those statistical properties that are of direct
interest).

Good luck,

Scott

On 05/09/14 01:12, SamiC wrote:
> Thanks Scott,
>
> That does help to clarify things.
>
> So if a covariate is highly skewed, extreme values will be more
influential.
> And this can be reduced through a transformation (which can be
> justified) or through other techniques (e.g. bootstrapping).
>
> Cheers
>
> Sam
>
>
>
> --
> View this message in context:
> http://r-sig-ecology.471788.n2.nabble.com/Standardising-and-transforma
> tion-of-explanatory-independent-predictor-variables-for-multiple-regre
> sss-tp7579048p7579052.html Sent from the r-sig-ecology mailing list
> archive at Nabble.com.
>
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
>

--
Scott Foster
CSIRO
E scott.foster at csiro.au T +61 3 6232 5178 Postal address: CSIRO Marine
Laboratories, GPO Box 1538, Hobart TAS 7001 Street Address: CSIRO, Castray
Esplanade, Hobart Tas 7001, Australia www.csiro.au

_______________________________________________
R-sig-ecology mailing list
R-sig-ecology at r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-ecology