Separating the roles of "teacher" and "assessor"

Question

Teachers at the university level (at least what I've seen in the US) are responsible for both

1. teaching students the material for a course, and

2. assessing the students' understanding of the course material.

This seems to encourage the creation of exams that assess whether or not students can solve very specific kinds of problems rather than assess a broad understanding of a course's content^[1]. Because of this it doesn't seem too unreasonable to have different people in charge or assessing students than are in charge of teaching them.

Are there any institutions that have a broad policy like this, where the grading and assessment is more centralized and out of the hands of individual instructors? By broad, I mean more than just having common finals for large classes with multiple instructors. Maybe even broad to the extreme of having a separate department for designing and grading assessments?

Have there been any studies into the whether or not this is a good idea? Have education researchers thought about this idea or written about the pros and cons of this before?

---

[1]: To provide a specific example of what inspired this question, I just proctored and graded an intro to calculus midterm (as a teaching assistant). This was one of the problems:

Use the squeeze theorem to evaluate the limit $$\lim\limits_{x\to \infty} \frac{\cos(x-2)+3}{2x}\;.$$

Many of the solutions to this problem were identical to what the primary instructor had written up in their grading key. So these students could remember the procedure the professor would go through when answering an example problem like this in lecture, and would themselves write out these steps like a recipe. But this didn't indicate to me that they really knew what the squeeze theorem meant.

The idea then is that having separate people teaching the material and grading the students would avoid situations like this. The teacher would have to teach calculus much more broadly to prepare students for a wider range of questions/problems that they don't have direct control over.

Answer 1

A question that occurs with a project like this (broader than one department, as you put it) would be: Who is qualified to make those assessments? Probably not any other department at a particular college, certainly -- the one department is, by definition, where all the experts in that subject work.

To some degree this actually is done in places, in the context of increasing corporatization of the university. For example: A college or state might outsource required tests to a company like Pearson or the College Board. Or a college might pay for a one-time development of an online course, say (either by resident faculty or an outside party), and then from that point on students learn from those packaged materials and present faculty only give grading and feedback on assignments (or vice-versa). This is particularly done at for-profit schools.

One problem with this is that it takes away faculty governance, control, and academic freedom. Outside agencies might be even less interested in upholding strong standards in the discipline than professional faculty; indeed, in some contexts the primary goal is strictly higher graduation numbers. The result of that might end up like, e.g., NYC public schools (which currently boast about a 90% graduation rate, and the majority of those students need grammar-school level remediation when they reach college).

Gary Rhoades, "We are all contingent", AFT on Campus, Summer 2014:

Reduced faculty control is particularly clear in traditional distance education, massive open online courses, and decisions to outsource segments of the general education curriculum to Pearson or other third-parties, often for-profit providers. In each of these cases, academic managers have developed structures and created mechanisms to bypass effective faculty decision-making related to the curriculum. Exceptions can be found, and collective bargaining agreements can build in protections against this. But the pattern is a strong one, and the growth of these high-tech infusions is linked to the furthering of contingency in academe. The model of high-tech employment is a just-in-time, just-in-the-classroom, just-at-the-manager’s-will type of employment.

Furthermore, the work of the instructor becomes less rewarding and more stressful in that type of environment. UNH Department of English website, "Stress, Control, and the Deprofessionalizing of Teaching", October 2009:

It is a Faustian bargain. When teachers lose control of decisionmaking—when they prepare students for tests they have no role in designing (and often no belief in), when they must abandon units they love because there is no longer time, when they must follow the plans designed by others, when they are locked in systems of instruction and evaluation they don’t create or even choose—they will not be relieved of stress. Their jobs are not made easier, they are made harder and more stressful. While some find a way to resist, others acquiesce, though they feel, as one teacher put it, that “the joy is being drained out of teaching.”

There's actually a particular phrasing in the literature for this decontextualized-division-of-labor, but at the moment it's escaping me. Perhaps someone else can chime in with that wording.

Answer 2

1 AP

An easy example is AP Calculus (or chemistry or etc. if I can be allowed the sneaky broadening to general academia). This is an example where the teacher is trying to prepare students for the test but doesn't know what will be on it, nor design it themselves. Of course instructors will have to do their own tests for the school grades.

My experience at a (pretty good) public school in the 80s was that calc (and chem and etc.) teachers tried to teach for general excellent insight and understanding into the topic. Even to the extent of a little coverage outside the requirements, as class was capable of it. (well...in chem/bio etc. not in calc.)

But all the subjects, even calc, I had the impression teachers were trying to teach a basic course content, regardless of the student sitting for the exam. Only in week or two before the test was there a little practice on old tests to see the format and get used to style of them (which differed from normal tests during school year.) However, the textbooks (Thomas Finney Calc, M, S, and S Chem were designed for AP usage (said so in preface), maybe with a few clearly asterisked extra topic chapters.)

Lately, I get the impression, from "expert teacher" blogs in calc and chem, (with growth of AP and such) that there is a bit more gaming the system, teaching very strictly to the curriculum and eschewing topics that "aren't on the test". [Even in some cases to the detriment of the student versus a college course. Has been some reduction of AP coverage versus standard state school coverage...less diffyQs and rotation volumes in calc, less algebraic chemistry problems...even though college courses still cover this. IOW, AP moved before rather than following college. I have tried urging teachers to teach what they think a normal college class should be (not saying Cal Tech, but Perdue, maybe). But the gaming aspect has gotten strong with teachers and even students.

2 Multi-section classes

My experience at military academy (still true today) is that they don't really do lecture hall classes. Even for subjects like chem, calc, physics, navigation, EE etc. that all 1000 kids in a year group take (that many colleges would choose to do in large lecture halls). So you get a bunch of profs teaching many 30 kid classes (like high school). But the overall content is prescribed AND the tests are common to the overall group. (done by some committee or master teacher or such.) The point of all this exposition is that teaching is very much divorced from assessment. And you can feel it and it comes across (teachers even saying they haven't seen the test when pushed for hints).

---

P.s. However, all that said, I think there is an assumption in the question that divorcing assessment from testing will lead to more thoughtful questions. I disagree with this as it is possible to combine the roles and have tests that are thoughtful. Or to divorce them and have more mechanical questions. This isn't just a theoretical point...I really think these are very independent variables, not much correlated. I think it is a course design issue, not a who designs tests issue.

P.s.s. I don't completely agree either with the implicit praise for concept questions versus learning how to do typical manipulative problems. This can be argued rather than assumed. Just because you are super sharp on definitions and answering tricky concept questions does not mean you will succeed at several step integrations (calc) or stoichiometry problems (chem). These are important skills for future use (thus worthy of learning) and one needs to practice solving these problems and be tested on them to develop the skill. I'm sure there is a balance (some concepts, some tricks and skills). But I would just be wary of assuming the balance needs to shift to more concept. It sounds fancier especially from a teacher point of view but may not be optimal objective, particularly for math service courses.