Re: st: why it worked so slowly to do multiple imputation on 204 ordinal variables

[Nils Braakmann <nilsbraakmann@googlemail.com>]

Dear Mandy,

i don't know if your computations are indeed as slow as you say. I'm
not sure what the -cycles-option does, but if that option increases
the number of regressions run by the factor 10, you calculate 5*10*204
regressions. In 48 hours that gives an average of about 6-7 minutes
per regression (if my back-of-the-envelope calculations are right).
While this seems a bit slower than I would expect, it doens't look
completely off.

Best,
NIls

On Mon, Mar 8, 2010 at 6:18 AM, Mandy fu <mandy.fu1@gmail.com> wrote:
> Dear all,
>
> I met some difficulty to conduct multiple imputation for 204 ordinal
> categorical variables. The computer worked unbelievably slowly. I
> would really appreciate if someone could give me some
> suggestion.
>
> The data set I have includes :
>
>  *450 observations totally
>  *204 ordinal variables (after imputation these 204 items are used to
> calculate 5 scales, which are the dependent variables in the
>          following regressions)  have missing values
>  *   4  independent variables (dummy variables)                 no
> missing values at all
>
> I used --ice-- and included all the 208 variables in the variable list
> to be imputed using ologit. I set up imputing 5 times and 10 cycles
> for regressions( m(5) cycle(10)).
>
> The problem is, the computer works unbelievably slowly when conducting
> the MI. It has taken two whole days until now and the computer is
> still working on it now. I cannot figure out why this assignment
> causes the computer worked so slowly. I don't think the computer has
> problem, since I have tried on three computers at different locations
> and all of them worked very slowly.
>
> Thanks for your time!
>
> Sincerely,
> Mandy Fu
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/