[R] Simple question on binning data

[Frank E Harrell Jr]

On 05/14/2010 07:35 PM, kMan wrote:
> Wow! This definitely contributed to my evening.
>
> If you could indulge, I would like some clarification on this matter of
> binning and distortion, particularly wrt time series (perhaps related to
> long-memory processes?). I had thought binning was standard practice in
> spectral analysis and ANPOW.
>
> ...<snip>
> Bert is correct.  Binning is seldom needed and usually distorts.  It is the
> statistical equivalent of a former governor from Alaska.
>
> Frank
> ...<snip>
>
> Sincerely,
> KeithC.
>
>

My experience in time series is limited so I'll defer to others.
Frank

-- 
Frank E Harrell Jr   Professor and Chairman        School of Medicine
                      Department of Biostatistics   Vanderbilt University