Dear ODK users,
our research group has been conducting a panel household survey in Malawi
using ODK. The second round is proceeding at the moment. I always wanted to
share my experiences here, but somehow never managed.
I thought it might interest people though how we solved the problem of
having the basic data (household location, roster etc.) preloaded, so that
the interviewers don't have to enter this data again. This is a feature
that doesn't seem supported by ODK so far (not sure how it's going to be
with survey). Basically what we did was, create one instance and fill in
placeholders for all variables that need to be prefilled. Then I wrote a
script that multiplies and renames this instance as well as replaces these
placeholders with the actual values from our database. As our household
roster was a loop I did this for every household size separately.
We then modified the instances.db and added in all the newly created
instances with the appropriate filepath and copied all the instances on the
tablets. Downside to this that all the interviewers have all the instances
on their tablets and need to scroll quite a bit when going to edit saved
I hope a feature like this will be integrated into ODK at some point as it
would make our life much easier. Another feature I would actually like to
is that changing the instance name also changes the name of the file and
the folder of this instance on the tablet. We have been backing up our data
by copying them to computers and it is quite difficult to keep a good
overview with the current naming scheme.