@ALRAINBOW
The reason to upsample to 24/96 is simple. At least some versions (don't know which you have) of the foo_ABX plugin, produce a noticeable pause when you switch from one sample rate to another. So you could keep track of whether the pause is short (same sample rate) or long (different sample rate). If you use a good sample rate converter, there is no harm in upsampling. To keep future skeptical questions to a minimum, just upsample.
You may want to think of the possible outcomes from continuing with the ABX, and be prepared. I can think of 3, each with pros and cons for you. For each outcome, you can report it here or not.
If you decide not to proceed, most people will wonder why, and you'll be the subject of ridicule (being chicken, etc.). But most of those who would ridicule you already are, so it may be a push. That's one outcome.
If you fail to pass (generally considered > 5% chance of guessing), you can repeat the test with another track or other conditions, or just give it another try. But you may also be convinced that in the past you really couldn't "hear" a difference, but rather had "perceived" one due to biases. Then some here will say "I told you so", but many will welcome you into the fold of the converted. Many people here (and on other sites) have had such an experience and may tell you about it. That's the second outcome.
The third outcome is one you really need to consider. What if you pass the test? Most (or all?!) of those pushing you to do the test will
not accept this outcome! They will be certain that you
either cheated or screwed something up. You have to know that saying "I didn't cheat" won't work. Both people who cheat and people who don't, say that. There may be some effort to figure out how you cheated or how you screwed up, but some simply won't accept a pass. You may want to ask beforehand, something that any honest science-oriented person can answer: Under what circumstances/conditions will you accept a passing score (e.g. < 5% chance of guessing)? Everyone who says "it won't happen", "it can't happen", "it's not possible, so I don't have to answer that", will never be satisfied. Know that ahead of time.