A recent Penning-trap measurement of the masses of 46 V and 46 Ti leads to a Q EC value that disagrees significantly with the previously accepted value, and destroys overall consistency among the nine most precisely characterized T = 1 superallowed β emitters. This raises the possibility of a systematic discrepancy between Penning-trap measurements and the reaction-based measurements upon which the Q EC values depended in the past. We carefully re-analyze (n,γ) and (p,γ) reaction measurements in the 24 ≤ A ≤ 28 mass region, and compare the results to very precise Penning-trap measurements of the stable nuclei 24 Mg, 26 Mg and 28 Si. We thus determine upper limits to possible systematic effects in the reaction results, and go on to establish limits for the mass of radioactive 26 Al, to which future on-line Penning-trap measurements can be compared. We stress the urgency of identifying or ruling out possible systematic effects.