Closed i-gao closed 5 years ago
It feels like a json specification of the full batch would be a more versatile option than a UI based solution to this.
i.e., we need to be able to insert two survey screens between some rounds and add a pre-survey.
See #349 Templatizing Surveys
I did it, but it's not very stable. You can check it in newarch-flexible branch.
I mean, now we can have/not have pre-survey/mid-survey in 1 round.
It was hard because of all timers, timeouts, batch live time, etc.
You must resave all your templates before using it.
Important: order of tasks is ANOTHER now. First task - for first exp round (pre/mid surveys too), second task - for second exp round , other tasks - for other random rounds (non experiment)
In general, it would be great if we could build a batch template like we build Google Forms, where we can schedule what the sequence of screens to display would look like.
For example, to get our current setup, we would specify this sequence:
Task1 Mid-Survey Task2 (experimental round) Mid-Survey Task3 Mid-Survey Task4 Mid-Survey Task5 (experimental round) Mid-Survey Post-Survey
We might then be able to specify new permutations, like Pre-Survey Task1 (unmasked) Mid-Survey 1 Mid-Survey 2 Task2 (masked) Task3 (masked) Mid-Survey 1 Post-Survey Demographics-Survey
These new survey types would be templatizable survey items.
This flexibility would mean a lot to the next sequence of experiments we hope to conduct, which do not follow the task/mid/task/mid/final structure.
Thank you!