Monday, January 23, 2017

Is the absence of grade inflation a problem?

Grade inflation is commonly seen as a problem. An average grade between B+ and A in an undergraduate class is typically viewed as an indicator that instructional rigor is missing and that grades do not properly reflect student learning (or, rather, the lack of it). To avoid this problem, instructors are frequently admonished by academic departments to aim for an average class grade somewhere in the C-range. I’ve been thinking a bit about this and have a question that requires a bit of backup.

Let me start by stating a couple of rules that I think most people would agree with:

  1. Graded assignments should measure students’ knowledge and skills in the subject matter of a class. A more sophisticated–or at least more Benjamin-Bloomsy–version of this rule would state that graded assignments measure whether, and to what extent, students attain the learning objectives that the instructor has set for the class.
  2. Learning goals, assignments, and grading criteria should be transparent, so that students know what to do in order to succeed. In other words, instructors should tell students what they want them to learn and how that learning will be evaluated.
  3. Readings, class activities, and the like should provide students with opportunities to learn the things that graded assignments evaluate. This implies that assignments should actually test what students can learn in the class.
  4. Teaching should generally be effective in that students who participate in the class can learn the skills and knowledge that the class is meant to teach them. This means a number of things. Of course, it implies rule 3: the class should teach what it tests. But it also means that students should receive the support, inside and outside the class, that enables them to be successful. It also means that instruction should be able to address the learning needs of a diverse set of students, giving all the same opportunity to learn.
  5. Instructors should continuously improve their teaching.
  6. Students should be reasonably motivated to learn. This is connected to rule 4 in that effective instruction should include effective efforts to motivate students to learn.
If all of these rules are followed, or if all these conditions (if we want to think of them as such) are met, I don’t see how we cannot see grades go up over time. In addition, if a letter grade of B indicates that a student has achieved a learning objective, and an A indicates that the student excels in that objective, I don’t see how a course average can be below a B in the long run if all six rules are met.

Assume that a class has an average of a B or a B+ and the instructor is asked to lower the grade to a C. What can the instructor do?

The most common response is that the instructor can grade “more rigorously”. Essentially, the instructor would redefine the meaning of different letter grades. Class performance, for example a term paper, that was previously deemed good enough for a B would receive a C. Assuming rules 2 (the change in grading criteria is transparent), 4 (instruction is effective), and 6 (students are motivated), it is not clear whether such a strategy would in fact lower the average grade, since motivated students would adapt to the new circumstances and produce better work. (A good thing, of course, but not a cure to higher grades.) Maybe the strategy would work for a semester or so as students are not yet aware of the new standards (in violation of rule 2), but over time grades should creep up again. Of course, it is possible that most students would be fine with a C average in the class and refuse to put in the extra work to keep up with the higher standards, but this would violate rule 6.

Another possibility that an instructor can pursue is to increase the emphasis on areas of performance in which students tend to have weaknesses. For example, the instructor could put a higher weight on grammar, mechanics, and style in writing assignments. While this might be effective in lowering the average course grade, it would be in violation of rules 3 and 4 if the instructor would not also provide additional instruction in grammar, mechanics, and style.

A different approach to combatting grade inflation is to “curve” the grade. This can be done formally, by calculating the distribution of student scores and using it to scale scores up or down to a desirable new distribution, or informally, as instructors (to take an example) compare student papers or other responses to each other to determine what is excellent in this particularly group of students, good, satisfactory, and so on. The problem with curving, however, is that it violates rules 1 and 2, since the grades not only measure what students know and/or can do (knowledge and skills), but also what other students who happen to be in the same class know and/or can do. An excellent performance may receive a low score in a class in which most students perform excellently as well (or slightly more excellently), while a mediocre performance may receive an excellent score in a class of poor performers. In a class of all equally excellent students, all students would receive a C.

I suspect that absence of grade inflation is most commonly due to the following reasons: First, assignments do not really measure what is being taught (see rule 1). For example, students receive low scores in a political science class for papers that have good or excellent political science content but that are badly written; the instruction is on political science, and only little time is spent on effective teaching of good writing. Second, instruction is not as effective as it should be and does not improve over time (rules 4 and 5). Third, students are not sufficiently motivated to keep up with rising or high standards (rule 6).

From whichever angle I examine this question, I find that the lack of grade inflation and the presence of low average grades presents a problem, or at least points to underlying problems. Motivated students who receive effective instruction and are evaluated on the basis of criteria that are connected to transparent learning goals should demonstrate good performance. If they don’t, something is wrong.

Grade inflation may point to a problem if instructors give good grades that do not reflect student performance (see rule 1). But the absence of grade inflation points to problems as well. If we focus on grade inflation and not on learning goals, quality of learning activities and assignments, quality of instruction, and student motivation, we’re doing it wrong.

What am I missing? (Honest question!)

[OK, I've been bad: no lit. review, references, hyperlinks. I promise (but don't promise) another post on lit by people who have actually studied this and didn't just write a lazy reflective essay.]

Update, January 25: Conversations among (some) colleagues suggest an additional way to lower course averages – you make it really difficult for students to document common illnesses for which you don't commonly go to the doctor, then penalize undocumented absences. As a result, students come to class sick, infect other students, and voilà: lower grade average. Added bonus: the instructor may catch the occasional bronchitis or pneumonia and may be able to cancel a few classes. Win-win!

No comments: