Your looking at your end of course survey and find that you have been bashed by a couple of students for grading "too hard" and being "unfair in grading". With near perfect prose, which they didn't show in the classroom, they go on a long winded discussion on how poor of an instructor you are. I almost want to give them an A for the quality of their writing. With some luck you also might find an equal amount of positive comments but they are usually shorter such as ,"Great instructor!" and "Love the Class". The good news is that student evaluations have no reflection on your skill as an instructor. One could make an additional argument that high ratings doesn't necessarily mean your a good instructor.
A study discussed in Inside Higher Ed helps show that there is almost no correlation between instructor quality and student evaluations. Previous studies showed only a mild correlation based on flawed research and limited examples. A recent study shows shows there is almost no correlation and it is unwise to use such information in professor evaluations for promotion and pay.
Students come to each classroom with different skill sets and abilities. While the study didn't address this issue but those who give the highest grades are likely to have higher ratings while those that challenge students are likely to give poor ratings. The grades are seen as a scorecard and those professors that push their students to learn stand in the way of a high GPA.
The study should make us reconsider student evaluations for specific instructors. Is the purpose of higher education to make students feel happy with the class or is it to have them learn more? If we are seeking to raise student learning for a higher performing graduate that is capable of meeting the demands of work and society then it may not be wise to determine merit based on how much the students like or dislike an instructor as described in a few short questions.