Hello!

I’m trying to do somewhat complicated counterbalancing of my questions.

I have 4 question and answer combinations to display followed by 2 questions for them. The types of combinations are such that we have some changes to the answer as we are looking at effects of these changes on trust in the answers:

• no change to the answer
• adding an indication of certainty (‘probably’)
• adding a first person indication of uncertainty (“I’m not sure, but”)
• adding a third person indication of uncertainty (“It’s not clear, but”)

Each of these sets of 4 question and answer combinations are themselves within 4 different categories (food, sports, medicine, and acronyms), with 2 items in each of the categories (for example, food has Iron Soy Jelly and Hasperat). (yes, these are all made up items)

We want to have each participant have two each of the question and answer combinations (no change, certainty, first person, third person) for all 8 questions (2 for food, 2 for sports, 2 for medicine, and 2 for acronyms), but we do not want people to see question and answer combinations from each item (such as Iron Soy Jelly) one after the other, to try to reduce how closely they see the same question with a slightly different answer. Ideally, we’d also avoid having questions from the same category next to each other, but that’s less of a problem.

I imagine this must be possible, but I have no idea how to make it work. I saw that there was the ability to randomize within blocks and to randomize selected sets of blocks, but could not find a way to make sure that the 2 selected items from within a set of blocks were separated by the 2 selected items from within another set of randomized blocks.

I saw a similar question from four years ago, but the only answer mentioned javascript and boolean embeded data and I had no idea what that meant specifically.