It is a fairly common idea in the ideology of many of those who run our education system that if you give students the ability to chose their professors or teachers, they will chose the best professors or teachers. The idea is to make education a marketable commodity with professors and teachers as service providers and students as consumers. There is a deep and fundamental flaw in this view. Markets are indeed extraordinarily good at satisfying consumer demand. The problem is that too many students are not demanding a quality education, but rather the highest possible grade, possibly with the least amount of effort. In other words too many students value the credential rather than the education it is meant to represent. Thus, the market system for education works against the Public Interest, putting an upward pressure on grades and a downwards pressure on standards.
What are some solutions to this quandry? The problem is often that grades for particular courses consist only of a percentage. In most schools and universities those percentages in a particular course do not differentiate between different professors or teachers. Thus a grade given by a challenging professor and one given by an easy professor are difficult to distinguish. The proposal in TFA might help the situation, but I think there is another way. What if each professor got a score not based on the evaluation by students but rather by how his students scored in other courses, especially those that follow his own course. This score for a professor would be like an adjustment factor for his grades. Let's say most students in one professor's Calculus II class who get 75% usually go on to get an 85% score in Calculus III. Thus, this professor's grades would be deemed better than another professor's grades whose 75% students usually go on to score 65% in Calculus III.
This system would reduce pressure on professors to raise grades, especially if students understood this rating system. All that would matter would be that the professor be consistent year after year. It might seem complicated to implement but in our world of computers and databases, I don't think it would be impossible to create. It wouldn't be necessary to follow all of a professor's students, only a few in order to gain a correlation. Indeed, all it would initially require would be for each professor or teacher to be given a unique code which would be attached to each grade given to each student. The rest would be data mining by whatever authority has access to the data.