Hello, 

Let's say we have an ordinal variable with categories 0 to k. We take each imputed value and subtract off the minimum ordinal value (0 in this case) and divide by the ordinal range (k) and then restrict these transformations to be between 0 and 1. With these probabilities, we draw a binomial random variable for each imputed cell with k trials and probability equal to the transformed imputation from the last sentence. Lower continuous imputed values lead to lower probabilities and, thus, lower draws from this binomial. Higher imputed values lead to higher draws from this binomial. 

Hope that helps!

Cheers,
matt.

~~~~~~~~~~~
Matthew Blackwell
Assistant Professor of Political Science
University of Rochester
url: http://www.mattblackwell.org


On Thu, Dec 13, 2012 at 12:35 PM, audigier <audigier@agrocampus-ouest.fr> wrote:
Dear all,

I don't understand how amelia handles imputation of ordinal variables. We can read in Amelia II: A Program for Missing Data:
"Imputations for variables set as ordinal are created by taking the continuously valued imputation and using
an appropriately scaled version of this as the probability of success in a binomial distribution. The draw from
this binomial distribution is then translated back into one of the ordinal categories."
This is not explicit enough for me. I don't understand what does "appropriately scaled" mean
and what's the link with binomial distribution. Can you give me more explanations ?

Thank you for your cooperation.

V.Audigier

--
Amelia mailing list served by HUIT
[Un]Subscribe/View Archive: http://lists.gking.harvard.edu/?info=amelia
More info about Amelia: http://gking.harvard.edu/amelia
Amelia mailing list
Amelia@lists.gking.harvard.edu

To unsubscribe from this list or get other information:

https://lists.gking.harvard.edu/mailman/listinfo/amelia