@interviewaurvey
Really the best, design option would be to first ask for preferred candidates (up to 4) with a multiple_select and then continue with a filtered rank type question.
Rank01_RankType.xlsx (12.9 KB)
Update: A question type can be hidden by not setting a label or hint. So, the calculated select_multiple in the attachments can be changed to hidden by removing the label
Here are two other design options for your ranking issue.
Rank02_ChoiceFilter.xlsx (12.0 KB)
Rank03_PositionNo.xlsx (12.9 KB)
(Both tested with the Online validator)
You can find further ideas and examples with the search function of the forum, see
Search results for 'rank select' - KoboToolbox Community Forum.
For current problems with rank type, see e.g. First rank and then select from ranking - #4 by wroos.
There are further requirements to define:
- Must the sequence be sequential and start with 1? Probably a preference 1, 3, 4 or 2, 3, 4 are invalid.
- Probably the first answer should be mandatory, as you offer an option “Undefined”.
- At which position is “Other” allowed? Probably only after at the end of selected preferences.
- Is only one “Other” allowed?
- How do you need the export structure for further analysis?
You can find related checks in the attached files.
You should also test, what will happen if a position is selected twice and the user moves back and changes or deselects the first duplicate? Kobo might not always update everything, incl. the highlighted error. So, the user would need to clean up manually (e.g. by deselecting and reselecting or by “save”).
And I would like to recommend to test the export (addressing later analysis) before finally deciding on your design option.
