Local swarm simulation generated from AnalystBot personae.
The seventy-eight responses to this survey are a start, but we cannot really say that 75% of people check their papers in advance and the rest at the last minute based on this.
If we hadn't already invested time in this collection, would we rely on so little data to assert such precise figures?
Honestly, it's like trying to guess the favorite dish of the whole neighborhood after asking three neighbors: we risk being very wrong.