[EpiData-list] Re: Defining confidence intervals in analysis

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Thu Apr 19 15:39:52 CDT 2007


>
> >I would like to know how one could change the confidence Interval (CI)
> to reflect 90% instead of 95% which I assume is the default. I am
> working in Epidata Analysis V2.0 (build 108). 
>
>   
The command "set table CI HEADER="(90% CI)" only changes the text in the 
tables, not the calculation. Until now the statistics have NOT been 
adapted to other confidence intervals than 95%. For a given table it 
should be quite easy though given you have the raw numbers. Enter these 
into e.g. www.openepi.com and get the desired 90% CI.

For analysis the possibility of users defining CI (e.g. 90 95 99) could 
be implemented within the next few years.

Just now for the v2 of EpiData Analysis we are working on some "subtle" 
internal issues, which is why seemingly very little progress was made 
recently. The implementation of the Upper Case fields <AAAAA> revealed 
need for quite many changes, since all internal parts were made only for 
uppercase handling.  The problems mentioned in the Mantis database and 
on the www.epidata.dk/testing.php page that the latest test build does 
not save data properly has been solved.

Regards Jens Lauritsen
EpiData Association


More information about the EpiData-list mailing list