[R-sig-ME] Does lmer with default family require normal distribution of IV

[Reinhold Kliegl]

Just as an addition: The distribution has an impact on power. If I
recall correctly, as far as interactions between covariates
(predictors) is concerned power is highest for uniform distributions
and pretty bad for normal distributions; for details see McClelland &
Judd (1993, Statistical difficulties of detecting interactions and
moderator effects, Psychological Bulletin). That's one of the
advantages of experimental design over field studies and one reason
why you need lots of data if you want to get significant interactions
between continuously measured covariates (which usually have some kind
of normal-looking distribution.)
Reinhold Kliegl

On Mon, Mar 7, 2011 at 10:29 AM, ONKELINX, Thierry
<Thierry.ONKELINX at inbo.be> wrote:
> Jerry,
>
> Neither the response nor the covariates must be normally distributed. The assumption of normality is on the residuals of the model. Although normality is not needed for the response, if it helps to get normally distributed residuals.
>
> Best regards,
>
> Thierry
>
> ----------------------------------------------------------------------------
> ir. Thierry Onkelinx
> Instituut voor natuur- en bosonderzoek
> team Biometrie & Kwaliteitszorg
> Gaverstraat 4
> 9500 Geraardsbergen
> Belgium
>
> Research Institute for Nature and Forest
> team Biometrics & Quality Assurance
> Gaverstraat 4
> 9500 Geraardsbergen
> Belgium
>
> tel. + 32 54/436 185
> Thierry.Onkelinx at inbo.be
> www.inbo.be
>
> To call in the statistician after the experiment is done may be no more than asking him to perform a post-mortem examination: he may be able to say what the experiment died of.
> ~ Sir Ronald Aylmer Fisher
>
> The plural of anecdote is not data.
> ~ Roger Brinner
>
> The combination of some data and an aching desire for an answer does not ensure that a reasonable answer can be extracted from a given body of data.
> ~ John Tukey
>
>
>> -----Oorspronkelijk bericht-----
>> Van: r-sig-mixed-models-bounces at r-project.org
>> [mailto:r-sig-mixed-models-bounces at r-project.org] Namens Jerry Zhu
>> Verzonden: maandag 7 maart 2011 5:46
>> Aan: r-sig-mixed-models at r-project.org
>> Onderwerp: [R-sig-ME] Does lmer with default family require
>> normal distribution of IV
>>
>> Hi Folks,
>> I have repeated measurements for each subject. The
>> DV(response variable) is "Value", a square root
>> transformation of which makes it normally distributed. The
>> IV(predictor variable) is "Delay", which is not normally
>> distributed. Then I ran a model:
>> lmer(sqrt(Value)~Delay+(Delay|Subject))
>>
>> Does this model require a normally-distributed Delay?
>>
>> I feel that lmer is a linear model, so it requires normal
>> distribution of both IV and DV. Am I right?
>>
>> Thanks for your reply!
>>
>>       [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> R-sig-mixed-models at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>