Hello,
I have a rather technical question. We are currently evaluating how to create our design for a MXD Objective Case study, which we will conduct using Lighthouse Studio. We are aware that Sawtooth can generate a design for us that works very well. Looking forward to publishing our study, however, we think it's easier for us to approach potential reviewer comments if we construct the design by ourselves.
To the question:
We have 18 attributes in total, a number that we cannot change as it is bound to external factors. Following standard procedure, we tried to come up with a BIB-Design, which exists but requires at least 51 blocks, a number way too large for any respondent to handle.
Is there a way to assign subsets of blocks to respondents (randomly or with some structure) that works well in practice, and that allows stable estimation on the individual level? We will have approx. 1200 respondents, 18 attributes, (up to) 6 attributes per task, and (up to) 12 tasks per respondent.
Any comments or links to literature on the subject would be much appreciated!
Best regards,
Felix