Am New to Kobo Toolbox and I need some clarification.
In ODK, a new question cannot be added to a deployed form as it adds a new column to the data. However, you can change question types, restrictions etc and update the existing form without losing any data collected.
Now my question is, does this apply to in Kobo Toolbox where you are unable to update an existing form with a new question like in ODK?
I asked because with Kobo you can just add questions and redeploy but I want to know if it would affect data already collected.
Welcome to the community! Kindly please be informed that it’s possible to update a survey project (adding or removing a variable as required) during an ongoing survey. But there are certain things that should be kept in mind (which has been discussed in one of the previous post earlier):
should we collect some more details, like what will happen
a) on export and
b) on re-edit of the previous data on the server:
if a type is changed, like text to date?
if a new calculation is added?
if a new relevant is added?
if a constraint is added or changed?
if a choice item is added or removed?
More general question: Does a re-edit of the data record on the server always relate to the form version used when first entering the record. (As a re-edit on KoBoCollect device before sending or a viewing afterwards.)
Data mix on export (and re-edit) might even get chaotic
if a select is changed to another list?
a new group is defined for existing questions?
Of course, all those potential problems are good arguments for detailed testing BEFORE field work (incl. pre-test)… (Weak quality assurance always comes back to you.) Let’s try harder to avoid changing forms during field work?
I do hope you are keeping safe. Let me respond to each issue and @Kal_Lam can chime in.
On export your data will be defined based on the latest form definition. Please note, you can get previous data downloaded as you may already know however, the format of columns will be determined by the latest version of the form.
Changes on the data type affect the export such that the predominant definition takes defines the export definition e.g. when text data exists in a date column for excel, the data will be defined as text on export. As you rightly noted, changing the type definitely get so messy.
The calculation will only be effected on data collected using the form with the added calculation.
Data collected previously before the relevant will not take the new relevant definition.
Same to relevant, the constraint would not be applied to what was collected before the constraint was added.
You can do this, but be very keen not to mix the codes e.g. if your initial response options were
2 for Wearing a mask
3 for Using sanitizer
Be sure not to mix up when you remove something like wearing a mask and you end up with
1 for Handwashing
2 for using Sanitizer
In this case all data previously collected as 2 will now appear as using sanitizer.