Connecting different questionnaires

I am working on a questionnaire which is divided into two sections. The main questionnaire is separate (may be administered by a different surveyor) from the supplementary questionnaire. Can I link the two? Can I use enabling or validation conditions that work across the two questionnaires?

How do you imagine this working on two disconnected tablets?

Through the common database? It would be good to know if there is a way around this.

By way of an answer, let me sketch a scenario and propose a few ways of handling it.

Scenario. Suppose you are doing a facility survey. One instrument captures information about the facility and enumerates the people who work there. Another instrument captures information about each person who works there.

Implementation ideas.

  1. One questionnaire, one interviewer. A single questionnaire would capture both facility-level and employee-level data. This would be done by a single interviewer. Validations could be done easily since the same questionnaire captures both facility and employee data.
  2. Multiple questionnaires, multiple interviewers, without linking. One questionnaire would capture facility-level info; another would capture employee-level info. Different interviewers could administer facility- and employee-level instruments, respectively. But because the facility and employee interviews for a given facility know nothing about one another (e.g., employee interview does not have access to data in the facility interview), the two instruments are not linked. This solution does not require internet, or involvement of HQ to make assignments based on data in the facility instrument. (Here, I assume that the employee instrument is done in “census mode”.)
  3. Multiple questionnaires, multiple interviewers, with linking. Same set-up as 2 above. But linking is done through making assignments. First, interviewer 1 completes the facility questionnaire and enumerates all employees. Then, interviewer 1 syncs completed interview to the server. Next, either someone or some program creates employee assignments based on data contained in the facility interview. Then, the supervisor syncs to get these assignments, and distributes them to interviewers 1 through N in their team. If facility information is preloaded into the employee assignments, there is some scope for validating that employee info is consistent with facility info. This solution would require internet, and involvement of HQ to create new employee assignments, either manually or via some script.

Hi Arthur,

We are going to have a survey like your 3rd scenario: We will have a classical household interview part first in the beginning of an interview. Then, we want to interview some individuals of the household that meet some criteria. For example, we first identify two household members who migrated in the last 12 months in the household interview part using one tablet. Then, we want to do individual interview with these two household members separately using two tablets (i.e. two enumerators do the individual interviews separately simultaneously). The individual interviews are linked to the household interview in this process. I think our case is similar to your 3rd scenario. Do you think this is possible? It would be great that you can provide an example in the Public Questionnaires.

Thanks very much for your time!

Percy, it is definitely possible!

Provided the following is satisfied:

  1. you have internet connectivity in the household.
  2. you have written the “some program” that was hypothetically mentioned in the 3rd scenario.
  3. the link is directed one way only FROM household TO person (and never back).

Best, Sergiy

Thanks much Sergiy for the prompt reply. This is such a great news! World Bank has started doing such household-individual combined interviews in the LSMS+ surveys. I think this is a trend in the future. Could we please have more guide or even an example for this issue?