[EpiData-list] Please help: Analysis

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Tue Jul 18 07:29:27 CDT 2006


Elize,

You can do this as shown below (there is no COMBINE command in EpiData).

read yourdata
define aprov ____________
select prov1 <> .
aprov = prov1
savedata p1 aprov  // save only aprov; you can save other fields if you want
select
select prov2 <> .
aprov = prov2
savedata p2 aprov
* etc
close
read p1
append p2
append p3
append p4
freq aprov
* The denominator will be total number of mentions of province, not 
respondants

This is more or less what Epi6 did internally. You can simplify this 
without the select statements:

define aprov ________
aprov=prov1
savedata p1  //save all data
* etc

Jamie Hockin
Public Health Agency of Canada

> Hi there
> I am currently analysing my first survey using Epidata Analysis instead of
> DOS Epi Info (I hated Windows EpiInfo) and have been using Epidata quite a
> lot for data entry, so now I am giving Analysis a go!  In the DOS EpiInfo I
> have been able to use the "COMBINE" command to do a freq on a question where
> more than 1 option can be selected.
> Eg. Question: In which province do you primarily work? (You may choose more
> than 1 option)
> The provinces are then all listed and the respondents just tick
> In my Epidata REC file I have numbered all the options from 1-10. I repeated
> the question 4 times in the REC file, so that 4 options could be ticked (my
> data does not have more than 4 options ticked for this Question - I
> checked!) I have a dataset with 148 respondents. I know that my denominater
> will increase with the "combine" command, but I am interested in the
> percentages only.  
> What command in Analysis will assist me with this analysis problem?
>  
> Warm regards,
> Elize Webb
> School of Health Systems and Public Health
> University of Pretoria



More information about the EpiData-list mailing list