ABX vs. ABXY 2017-06-30 14:51:42 I've noticed several double-blind testing tools have (or give the option of) ABXY testing. For those who haven't seen this, it involves randomized duplication of the A and B samples as X and Y. Therefore, the options to the test subject are:A is X, B is Y or A is Y, B is XUnder conventional ABX, the A and B are compared with X, with X being a duplicate of either A or B chosen at random. The options for ABX are therefore:A is X or B is XI'm struggling to see the logic of using ABXY over ABX. Statistically speaking, I can't see how it has any effect on alpha or power.Is there a particular methodological/statistical reason why it is used, for example, by the ABX Comparator tool?