We have observed that there is huge inconsistency among data related to time. In our multisectional research (16 in total) it was seen that the average time spent on each section is 2 seconds or less which is impossible! How reliable is this information? Could an official person please explain the possible reasons for this outcome? I also would like emphasize that our questionnaire/survey takes around minimum 30-45 minutes to be completed by the enumerator.
I look forward to your responses.
How was this calculated?
It was calculated from the paradata
Please show the data and code used to obtain these results so that they can be reproduced.
It is calculated through time differences between following Answer Sets. We thought that the time differences between two following Answer Sets would give the time differences which is the duration difference of following questions. Is this a correct interpretation? This ‘2 Seconds’ issue looks absurd from our side. We would like to understand how paradata works. We sincerely believe that there is a technical story behind this error.
According to the documentation on the paradata format, the timestamp included in each record in the paradata file is the timestamp of when the event occurred:
Date and time when the event occurred combined in a single timestamp (in UTC).
See also, this relevant discussion: