You are assessing a one-month training program for meat inspectors. At the end of every day, students inspect 100 cattle and the number of mistakes is counted. As the month progresses, the students tend to get better and the trainers gradually recognize the usual abilities of each student. One of the trainers always criticizes his students when they make more mistakes than expected, and never praises them when they make fewer mistakes than he would have expected. When asked about this practice, the trainer related a previous observational study he had conducted. Whenever a student made a surprisingly good score (given the student's current state of training, experience and previous performance), the instructor praised him. He then noted that, more often than not, the next day's test score was worse than it was the day before when the student was praised. In contrast, whenever a student made an unexpectedly poor score (given current expectations), the trainer sharply criticized the student. He then recorded that, more often than not, the next day's score was better than the previous day's score for which the student was criticized. Wanting all scores to increase, the trainer concluded that praising good scores was a hurtful and criticizing bad scores was helpful. Was this trainer reasonable? Explain your answer with regard to the possible impact of regression to the mean. (Answer in 100 words or less.)