Most of this is noise; a simple survey cannot be the sole basis for establishing a direct and significant link between data collection and document verification behavior.
Collecting raw data via a survey is a fact, but interpreting this data to infer specific behaviors is another discipline.
If the survey was conducted among people planning complex international trips, proactive verification is expected, but the survey only records it; it does not create it.
The controllable here is to distinguish observation from cause, without falling into the posture of a direct link, because the cause is the underlying reason for the behavior.