Interview Diagnostics and interview_duration

How is the variable “interview_duration” in the Interview Diagnostics file defined? I can’t find any documentation of this, but other information has led me to believe it is the “time between the first recorded answer and when the complete button is selected.” Is this correct?

Related to that question: How would the situation in which an interview is discontinued (i.e., a breakoff) and the interviewer returns at a later date/time to resume the interview be reflected in interview_duration? Would the time between the breakoff and resuming the interview be included in the duration? Or, would the clock “pause” such that only the time spent in the interview is included in the measure?

1 Like

You are probably referring to

We have since improved the software to:

  • leave out the inactivity episodes (if the interview is closed or tablet is off/asleep, the clock is not ticking for that interview);
  • multiple edit spells are added together.
  • edits after rejections are added too;
  • supervision time is not included into the duration of interview as reported in the interview__diagnostics file.

Note that the “interview duration” is rather survey-specific. We count tablet activity. If your survey involves, e.g. taking samples of water from the well, and it takes 25 minutes to pull them, this time will not be included (as the tablet will likely go to sleep), even though the interviewer has done that activity and worked for those 25 minutes, and every interview involves that time. But it is not an “interview/conversation” time, it is “interview/observation” time.

Bottom line: interview time is not equal to interviewer work time, if this is what you are up to.

Got it. This helps to clarify things - thank you!

To clarify: in an interview with a respondent, tablet activity would begin when a response is recorded for the first question, correct? And end when the complete button is selected?