Just looking for some quick verification on a Sparse MaxDiff set up.
Have a list of 36 items (attributes) and setting the number of items per set to 6 with the number of sets to 12. 1) Will this setup ensure that each respondent will see each benefit twice? 2) I've set the number of versions to 72 (6x12); is this over-engineering for a Sparse MaxDiff, or should I leave it at 300? 3) Additionally, my expected sample is n=1000; is this a) enough sample and b) with this sample would LC or HB estimation be preferred? My hunch, based on reading is that LC would work better in this instance.