ranking of teams without school names or conferences. All other information would be available to them - W-L records, SOS/SOR, head-to-head results, wins & losses against ranked opponents, team stats, and any other computer model numbers. The only thing they couldn't do is watch games (in addition to not knowing the logo on the helmet or the conference affiliation). You could design the blind test to include scenarios that would vet out/expose what criteria they favor, and thereby back into a model/system of evaluation they are essentially using.

You could then go back to the previous seasons in the CFP era and apply that same criteria to the teams in those years and see where the rankings land relative to how they were actually ranked. It would be comical to see their brand and conference biases utterly exposed in that effort.