Satisficing versus optimizing in instructor selection

In my previous post, I argued both that instructor selection is relatively neglected in undergraduate courses, and that criteria that students routinely use for instructor selection may be suboptimal. I outlined some possible pointers regarding good instructor selection, based on my own understanding of how the process works. There were a large number of thoughtful responses to my post. The most interesting responses argued that people do, and/or should, satisfice with respect to instructor selection.

I raised the point that people tend to satisfice rather than maximize in one of my own comments on the post:

So, to clarify, although I think it “often” happens that people select instructors randomly, it doesn’t always happen. My impression is that there’s heavy availability bias: if students can easily access evaluations, or know friends who have taken classes with a particular instructor, they’ll use that data. If not, they’ll select randomly. In general, my impression is that students are risk-averse and satisfice on this count: as long as an instructor seems “good enough” they won’t generally try to look for a better one (this is particularly true for people in multi-course sequences with an instructor, we could call this a “status quo bias” or an “endowment effect” depending on your perspective). More proactive approaches, such as sitting in on classes that the instructors are teaching in the previous term, seem to be relatively rare.

My other point is that even when students are making proactive choices, the criteria they use may be suboptimal. I am not aware of high quality advice that would help students select instructors effectively. This is, I believe, in contrast with the vast (though possibly not very high quality) literature available on how to select a college or a major. If intra-institutional variation in instructor quality is comparable to inter-institutional variation, then there are probably unrealized gains in selecting instructors according to better criteria.

In response, Linkhyrule5 wrote:

One thing I would note is that good teachers are rare and not very “good,” but bad teachers are common and very very bad. This ties into Trevor’s post: once you’ve avoided the bad teachers, it’s probably not worth the much larger effort required to find an extra-bonus-good teacher that may or may not exist.

and elaborated in the same thread:

I should probably have qualified my earlier post, as the only data I have to draw on is my own anecdotal evidence.

Nevertheless, with the qualifications that I went to a private high school and am currently at an Ivy college:

Sample size: since freshman high school, 26 “teachers” and nine “professors.”
Definitions: a teacher/professor is good-I if I was particularly interested in his/her course (and not just the subject), and good-R if I retained particularly more from that course. A teacher/professor is bad-I if I was particularly uninterested in his/her course (and not just the subject), and bad-R if I retained particularly little from that course.

Of my teachers, 2 good-I teachers and 0 good-R teachers, with two “maybes”—not particularly good but above-average. 3 teachers who were both bad-I and bad-R—all of these are extreme cases, in which I learned almost nothing and loathed the class.

Of my professors, 0 good-I and no data on good-R (I’m a sophomore), but already 2 bad-I and I highly suspect bad-R.

Trevor had written:

A good instructor will often offer a good class that good students can take advantage of. A bad instructor will never offer a good class and no one can take advantage of it. Avoiding bad teachers is therefore always a move in a good direction, while selecting good teachers is only sometimes a move in a good direction.

I’ve been working in K-12 and college since 2005 as a sign language interpreter. The qualities of good instructors varies, but the qualities of bad instructors are shared. Late, unprepared, digressive without profit, argumentative, close minded, uses the class as political platform, uses the class as therapy session, uses the class as amateur comedy, unable / unwilling to earn respect (grudging or otherwise) of class—those are some of the things bad instructors share. Avoid those and you’re more likely to succeed in any class.

Dre wrote:

Going off of what others have said, I’ll add another reason people might satisfice with teachers.

In my experience, people agree much more about which teachers are bad than about which are good. Many of my favorite (in the sense that I learned a lot easily) teachers were disliked by other people, but almost all of those I thought were bad were widely thought of as bad. If you’re not as interested in serious learning this might be less important.

So avoiding bad teachers requires a relatively small amount of information, but finding a teacher that is not just good, but good for you requires a much larger amount. So people reasonably only do the first part.

Satisficing versus optimizing: the descriptive, the possible, and the prescriptive

This has got me thinking about the relative roles of satisficing versus optimizing in instructor selection. There are three separate but related questions:

Do people tend to satisfice, rather than optimize, with respect to selecting instructors? I think the general consensus seems to be that most people tend to satisfice. For instance, very few students sample classes by instructors that they plan to study under in a subsequent term. Very few students think deeply and carefully about what pitfalls to avoid in instructor selection. Most of them try to avoid instructors who are bad (in a superficial sense) -- instructors who speak unclearly, have terrible handwriting, or don’t grade “fairly” (fair grading is student lingo for easy grading).
Do there exist ways for people to do better than satisfice, without incurring huge costs? Dre’s comment, quoted above, suggests one obstacle: everybody agrees on the terrible teachers, but different people have different ideas of what constitutes good teaching. This is corroborated by the fairly polar responses on the student evaluations of all except very bad teachers (“he is a great teacher, very clear in class” versus “I can rarely understanding anything the teacher says” for the very same teacher). That being said, I think that for people who are aware of some of the basic ideas as I outlined in my preceding post, and take proactive steps, one can choose good instructors without incurring huge costs.
Should people choose to optimize rather than satisfice? This is the part that’s most open to debate. It depends on the degree of variation between good enough and great teachers in terms of absolute outcome differentials for human capital or signaling. Another related factor, that I didn’t mention in the post, is whether your peers are optimizing or satisficing. If the peers you really want to have are optimizing for instructor, it makes sense to optimize for instructor, so that you get those peers. If, however, they are merely satisficing, then optimizing for instructor will not necessarily optimize for peers.

Is optimizing for instructor inherently zero-sum?

There are two related points I want to make. In colleges where the number of seats in classes is fixed and most classes come close to filling their seat quotas, optimizing for instructor has the connotations of a zero-sum game. This is also true in the case of satisficing for instructor, though perhaps less so: if students unanimously protest against terrible instruction, departments can, at least in principle, remedy the problem (by either firing or retraining the terrible instructors). Converting mediocre instructors to great instructors, on the other hand, is relatively hard. If everybody competed for limited student slots with a great instructor, you’d just be displacing another individual.

There are three counterpoints to the perceived zero-sum connotation of optimizing.

Students who are actively looking for good instructors are likely to benefit more from good instructors than students who are more indifferent. In econ-math jargon, we’d say that student concern for finding good instruction is complementary to instructor quality (formally, the mixed partial with respect to student desire for a good instructor and instructor quality is positive).

Different students have different tastes for what constitutes a good instructor. Thus, one student’s “best” instructor may differ from another’s (drawing on Dre’s point again). To the extent this is true, it cuts both ways. On the one hand, it means that selecting good instructors would be more of a positive-sum game, and therefore, is socially useful. On the other hand, it’s harder to easily acquire information about who is the best instructor, and it may be hard to compile generic advice on that front

In the longer run, when universities and departments see that students are specifically interested in good instructors, they may work to either (a) improve the quality of instruction, or (b) reduce hard limits on enrollment allowing more students to take classes with the instructor of their choice. It’s unclear whether (a) is possible, and the merits of (b) are unclear.

The analogy with satisficing in charity, and why optimizers may remain a minority, but could still grow

Charity is one domain where many people tend to satisfice rather than optimize. People choose a charity to donate to, then check (using Charity Navigator or a similar service) whether the charity satisfies some minimal threshold (low overhead, no fraud, etc.). They then donate to the charity. Status quo bias, personal relationships, and many other factors play important roles. The effective altruism movement aims to change the norms surrounding charitable giving from satisficing to optimizing. For instance, Giving What We Can claims that some charities can be 1000 times as efficient as others. GiveWell puts emphasis on funding the right program and spends hundreds of hours doing the background research for its top charity recommendations. Proponents of effective altruism are nonetheless quite sanguine about the prospects of converting a majority of people to the mindset of optimizing in charitable giving. They do, however, think that the minority that does care about optimization can be better served, and can grow to include others who would care about optimization if they are made to consider the issue. For instance, when responding to the Money for Good study, GiveWell wrote:

Our goal isn’t to create a product that the majority of people like; it’s to create a product that some minority market loves. From what we’re seeing now, it’s still possible that the minority of donors interested in impact-focused research is quite large.

In the same way, I think there is a minority market that is interested, and a somewhat larger market that potentially could be, interested in optimization with regard to facets of educational experience such as instructor selection. The advice I gave in the preceding post is geared for that minority market.

Looking for thoughts

I’m most interested in people’s thoughts about the numbered questions 1, 2, and 3. However, I’d welcome thoughts on any of the other assertions I made as well. Thanks for reading!