Repository navigation
Idea: introduction of randomness in how some form logic is handled #1829
Description
Scenario:
A client form about gambling where the first question "have you gambled in the last 2 weeks" which is yes/no and show/hides a large number of follow up questions. The client realizes that if they select "no" that they're done the form and can move on to the next without answering any additional questions.
Solution:
A randomness check sometimes shows all the followup questions anyways regardless of the answer to the first question.
Rationalle:
In a study with large numbers of entirely online recruited and surveyed participants this method could improve data quality by discouraging "speedrun" skipping of forms to get to the end.
Activity
thomasbeaudry commentedon Oct 27, 2026
This is an interesting thought, but I don't think it would discourage people trying to "speedrun" they would still try it regardless, as they aren't familiar with the platform and are use to doing it for other ones. Even in a series instrument, I could see them trying the speedrun approach each time, with the rationale that not all forms would be built the same
gdevenyi commentedon Oct 27, 2026
I totally agree that they could still click-through speedrun, but this could discourage the "really easy" method at least a little.
With big studies its about improving things on the margin.
thomasbeaudry commentedon Oct 27, 2026
I agree with you there, but you are also runing the QOL of other users that are telling the truth. Your apporach could still be the best - like i said it's an interesting problem, that I haven't spent too much time thinking about