I doubt the exam itself is harder. However, the number of students taking the exam is far lower. Therefore, it is statistically more likely for the raw score/scaled score conversion to be considerably different.
In other words, since there's more people taking the exam on Saturday, you're more likely to get a consistant sample of pre-medical students. However, since there are fewer people taking the exam on Sunday and let's say, for example, they are considerably smarter (hence misrepresentative) than the general pre-med population, the average student will fare far worse on Sunday than on Saturday.
Am I making sense at all??
😉