That's right, but I think that using the mean could be the play or not depending on the amount of discrepancy and the discrepancy tolerance of each individual. In this case, assuming the same HFOV, the mean scalar_0 value would give a 13% increase in MDH for the 1x group (MDH 88%) and an 8% decrease for the second group (MDH 67%). A >13% difference in MDH might be tolerable.
For me, personally, a 25% change in MDH is too big of a difference. MDH 100% for the 1x group gives a scalar_0 value equal to MDH 75% for the second group. That's so noticeable for me that I just try to avoid using those iron sights (very annoying in the early game). Now, if I used the mean scalar_0 value, it'd be the same as using MDH ~125% for the first group and MDH ~86% for the second one. That's up to a 25% difference again.