Audio Audit and Consent

Survey Management
1

Dear colleagues,

We are in the process of implementing a survey with audio audit for the first time. We are introducing a question where we ask for respondent’s consent to being (potentially) recorded. However, I found myself with no ways to end the interview in the event they answered “no” and the interview happens to be one we picked for audio audit.

Any ideas or suggestion? does such a feature exists?

Thanks

D

2

Tangentially, is there a way to automate this in a batch upload? or I need to manually assign the interviews to audio audit? if the second, which type of user has the rights to do so? HQ? Supervisors?

Thanks again,

D

3

If “ending a survey” means “asking no more questions” than all subsequent questions should be tied to the consent question with an enabling condition. This is easiest to do if you put the conditions on the subsequent sections, that would propagate the condition to all the questions and subsections inside them, of course. See for example a public questionnaire “QUESTIONNAIRE_KASAI_DRC”.

I don’t see how this is related to the audio audit.

4

Thanks Sergiy,

We are going to do a subsample of interviews with audit. However we are going to ask for consent to every one since we don’t want to signal we are doing any specific interview as audit.

Including the consent for everyone and subsequently ending the survey (i.e. “asking no more questions”) if they answer no (which we know how to do), might induce a higher no-response rate even for people that wouldn’t be recorded in the first place.

So, my question is, for those not being recorded who answer that don’t want to be recorded, is there anything we can use to signal the questionnaire to continue?

5

In short, the interview questions should not depend on whether the interview is being observed or not, that means there is no logic that should be tied on the audio audit status inside the interview, and hence that status is not obtainable in the logical expressions that you include in the Designer.

The answer to your question is ‘yes’: have another variable that you preload that will determine whether the consent to recording is critical for interview continuation or not, and then you branch accordingly, but:

  1. this will require you to make a decision (record or not record) upfront, when the assignment is created and not change it afterwards (which is a possibility in Survey Solutions).
  2. this will result in interviewers being able to probe whether recording is activated or not, by first entering “no consent given” answer (regardless of what the respondent says) and checking whether the interview is permitted to continue. If no - then the recording is demanded, and they will be extra super proper. If yes, then the recording will not take place, and they will go back to their kitchens to fill the rest.

For the reason of 2 I’d say this is not recommended.

What is the exact wording of your consent question?

and this is the price for the audio audit. And even if you get the consent, you may still be collecting different responses with audio audit then without it. It’s just two different surveys, even though they are having the same questions. And not just audio audit, but mere presence of other people during interviews may change the respondent’s perception, willingness to reveal the truth, or other quality of answers they communicate. (It seems to me it is more studied in this context, rather than audit, see e.g. https://journals.sagepub.com/doi/full/10.1177/0049124120914938 and related studies).

Best, Sergiy

6

Many thanks Sergiy, this is exactly the type of reasoning I was going through, but failed miserably at putting as elegantly in writing as you did.

I will translate this to our partners and act accordingly.

Have a great day.

D