We’re running a DCE and have now enough respondents to generate priors. Is there a way to edit the design and input the priors while data collection is ongoing?
This is fine and indeed best practice in Health Preference research. It’s also very easy to do
The standard pattern is to collect experiment data with a design that has no priors or at a minimum, assumed directionality e.g. price or risk.
Then build a model and use the model estimates (the beta coefficient) as new priors, regenerate the design and republish the study.
1 major caveat - the experiment structure and content must not change, only the design. Doing so may substantially change the meaning of the experiment and render application of priors meaningless or worse, you’ll have two datasets that wont merge cleanly as they differ
In surveyengine all you do is go into the experiment and either regenerate using Ngene with the new priors (or upload). Then republish.
Perhaps make a note of the versions for pre and post publish. The version numbers will be in your data and may be used to partition the data to see if the updated design is performing as well or better.
Good Luck!