R help - Feb 2009 - MDS with ranking data (and transformation)

Dear Sirs and madams :-)

I am trying to teach myself multidimensional scaling. To that effect I have
collected a survey asking people to rank 10 philosophers and politicians
according to their preference. I have collected 61 answers. The data is
organized in ten columns and 61 rows. the columns are "choice_1",
"choice_2", "choice_3" etc. The cells is the name of the
philosopher

I guess I need to put the data in some other format, e.g. with colloumns:
"philospher_1", "philospher_2", "philospher_3"
etc. and then have the cell
hold the particular ranking (score) for that philospher (i.e. a number
between 1:10)

I guess such a transformation would also allow me to do clusteranalysis? -
But how to do it???

Anyways: what I have done so far is to compute a 10*10 matrix in a
spreadsheet application. I do this by
countif(range_choice1=philospher1)*10 for each philospher.
In ?range_choice2? I multiply by 9, and in ?range_choice3? I multiply with 8
etc.

The corresponding matrix I import to r and do 
dist(t(matrix)
and then I use cmdscale$points to draw a plot. It looks nice but I am almost
sure I'm doing it wrong. And I would certainly like not having rely on a
spreadsheet.

So my question is: how to transform the data, and is it true that my current
'spreadsheet' method is wrong? Also: should consider discarting some
data,
e.g. only using the top 3 choices?

Cheers in advance.

Andreas
-- 
View this message in context:
http://www.nabble.com/MDS-with-ranking-data-%28and-transformation%29-tp22025457p22025457.html
Sent from the R help mailing list archive at Nabble.com.

On Sun, 15 Feb 2009, achristoffersen wrote:
>
> Dear Sirs and madams :-)
>
> I am trying to teach myself multidimensional scaling. To that effect I have
> collected a survey asking people to rank 10 philosophers and politicians
> according to their preference. I have collected 61 answers. The data is
> organized in ten columns and 61 rows. the columns are "choice_1",
> "choice_2", "choice_3" etc. The cells is the name of
the philosopher
>
> I guess I need to put the data in some other format, e.g. with colloumns:
> "philospher_1", "philospher_2",
"philospher_3" etc. and then have the cell
> hold the particular ranking (score) for that philospher (i.e. a number
> between 1:10)
>
> I guess such a transformation would also allow me to do clusteranalysis? -
> But how to do it???
Something like

dat <- t( apply( your.data, 1 , order ) )
colnames(dat) <- paste( 'philosopher', 1:10, sep='_' )

HTH,

Chuck

>
> Anyways: what I have done so far is to compute a 10*10 matrix in a
> spreadsheet application. I do this by
> countif(range_choice1=philospher1)*10 for each philospher.
> In ?range_choice2? I multiply by 9, and in ?range_choice3? I multiply with
8
> etc.
>
> The corresponding matrix I import to r and do
> dist(t(matrix)
> and then I use cmdscale$points to draw a plot. It looks nice but I am
almost
> sure I'm doing it wrong. And I would certainly like not having rely on
a
> spreadsheet.
>
> So my question is: how to transform the data, and is it true that my
current
> 'spreadsheet' method is wrong? Also: should consider discarting
some data,
> e.g. only using the top 3 choices?
>
> Cheers in advance.
>
> Andreas
> -- 
> View this message in context:
http://www.nabble.com/MDS-with-ranking-data-%28and-transformation%29-tp22025457p22025457.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
Charles C. Berry                            (858) 534-2098
                                             Dept of Family/Preventive Medicine
E mailto:cberry at tajo.ucsd.edu	            UC San Diego
http://famprevmed.ucsd.edu/faculty/cberry/  La Jolla, San Diego 92093-0901