[EpiData-list] outbreak table and calculations

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Thu Aug 12 14:51:35 CDT 2010


  Sandy asks why particular defaults were chosen for the tables command 
and outbreak analysis.

The reason is simple - we had to make a decision and usually I tend to 
work with 1 being indicator of a given status, e.g. case or exposure and 
0 being the "non-case" or "non-exposure". As simple as that, in other 
words it is a matter of preference.

I could have argued that we went over several epidemiological text books 
and found the most common choice of values. But the reality is that 
there is no uniform standard we could use.

The import decision is therefore to be very specific about how the 
values are interpreted and used. - and to add options - e.g. the /SA 
option - for further control by users. It is good practice to study the 
command reference for the commands you are using. Many options exist and 
are thoroughly explained in the command reference, which you find in the 
help menu of analysis.

The key design choice in all commands is to document exactly the codes 
used. You should also notice that in all commands the number of 
informative observations and the number of total observations is always 
readily seen in the output. Should there be a command, where this is not 
the case let us know on the list or report to the flyspray database.

Regards
Jens Lauritsen
EpiData Association
On 12-08-2010 18:39, epidata-list at lists.umanitoba.ca wrote:
> Hi Jamie,
> Thanks for the suggestions. The command: /sa worked. Will this always be
> the case? Will I always have to reverse the values? Why is the program
> taking the higher values of the variables to mean case and exposed as
> the default setting? I do have the latest version of analysis.
> Sandy
>


More information about the EpiData-list mailing list