Large Categories

In our survey, we capture migration. Sometimes, the individual is migrating from a village to another within the survey area. As they are already enrolled, there is no need to collect all their information again. In these instances, we simply have the interviewer select the name of the individual.

We have 1,00,000+ individuals enrolled, and the SS Limits do not permit a question with that many options. To get around this problem, we have a categorical question to select the village, and then 1 question per village with the list of individuals in that village as categories. This question is enabled if the corresponding village is selected.

We had to create about 100 questions (1 for each village). Currently we do not have a village with population larger than the SS limit of 15,000.

Is there a simpler way to achieve this?

When you send your interviewer to a particular respondent, preload the assignment with the particular village information obtained from the previous round.

Do not store any confidential data in the questionnaire itself!

This includes, for example:

  • respondents’ data (such as names, ages, number of children, etc);
  • interviewers’ data (such as name, interviewer ID, etc).

Best, Sergiy

The migration could happen from any village, so preloading data from just 1 village will not help.

If they are already registered, why not continue with the interview you already have for each registered informant?

It seems to me that the approach given to the problem is not the optimal one because the case in which the limits are exceeded can be faced, perhaps the informant should be taken as the basic unit and the villages used to monitor the migration of the informant.
For example, you may have a question to know how many times the informant has migrated in the last six months and with this answer, enable a list to know the village of origin (from where he emigrated) and the village of destination (to which he emigrated ).

Agree. Thats why the post.

This is a household level interview with individuals in a roster. We capture migration (In and Out). Out-migration is easy, just the name if the village is to be selected. For in-migration, we ask for the village from where they are migrating in. If the village is within our area, then we need to select the name of the individual from the list. Else all details are to be filled (new enrolment).

This may result in the information about the village collected by multiple interviewers, possibly resulting in duplication (and potentially different/contradictory information collected by each).

This is regularly reconciled. Out-migration and in-migrations are matched. Information about an individual is only collected at one point, not both places.