[R] scaling advice

[peter dalgaard]

On Jul 15, 2011, at 23:05 , Data Analytics Corp. wrote:

> Hi,
> 
> I have a consultants nightmare -- I was given a project that another consultant did and I was told to do the same calculations, but there's no documentation on what he did.  Basically, I have yes/no answers to survey questions about the effectiveness of product attributes by brands.  There are 44 attributes and 13 brands.  The other guy scaled the proportion of respondents who said Yes to be mean 0 and variance 1.0, apparently doing this by brand within each attribute.  He then created a matrix of 44 rows for the attributes and 13 columns for the brands.  No problem with this; I can always replicate this much.  But then he apparently rescaled this 44x13 matrix so that the rows all sum to zero and the columns all sum to zero.  None of the row and column standard deviations are 1.0.  This I can't see how to do.  How can I rescale the rows and columns so that they all sum to zero?  Any suggestions?
> 


If the _sum_ is zero, there must be both negative and positive elements, so it can't be a pure scaling. sweep()'ing out the row and column means would be the first thing to come to my mind.

-- 
Peter Dalgaard
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Email: pd.mes at cbs.dk  Priv: PDalgd at gmail.com