An experiment tested whether asking people to explain their forecasts improves calibration—the match between stated confidence and actual accuracy. The participants' average calibration was virtually unchanged after the exercise. Closer inspection, though, showed that participants who substantially revised their explanations became better calibrated, while those who made only superficial revisions became worse calibrated by nearly the same amount. The aggregate and subgroup analyses thus answered different questions: the former indicated whether gains and losses offset one another, whereas the latter indicated whether either occurred. ______ a stable average can be evidence of balance without being evidence of inactivity.
Which choice completes the text with the most logical transition?