Length of Exam Questions

Following on from David's interesting series of posts about exams and other forms of assessment, I was wondering about another issue - the ideal length for an exam question, particularly in the context of a supervised/in-school examination.  I've always kept the length of my questions fairly short (maximum two or three paragraphs) so students don't spend so much of their time reading the questions and deciding which questions to answer - I always give the students choice as to which questions to answer.  I realize that a 2-3 paragraph question is not possible in some courses where the facts, of necessity, have to be more detailed.  But I have also had colleagues who write very long fact situations with plenty of "red herrings" included to test the students on how well they cope with sorting through the detail in the questions.  Some of these colleagues have not expected any student to be able to "finish" the exam, but have notionally given points for how far each student can get.

Is there a better or worse approach here?  Does it matter what approach you take as long as you have prepared the students by explaining your expectations and showing them past exams? 

Comments

[Jonathan H. Adler]

I'm not a big fan of red herrings or of exams that feature gazillions of issues piled on top of one another, but I still tend to write rather long essay questions, and often have difficulty shortening them. For some of the classes I teach (e.g. Admin) there's a trade-off between writing richly detailed fact patterns that are more representative of the sorts of scenarios a young lawyer could encounter in practice, and hopelessly long fact patterns that take too long to read and digest in an exam setting. I try to find ways of writing shorter questions, but it can be difficult to do without hobbling the exam.

JHA

[Roger]

I always vastly preferred professors who wrote short exam questions. I had a three-question constitutional-law exam that fit on a single side of a single piece of paper.

I had a corporations exam that had the same one-paragraph question four times, with very slight changes in the facts. It actually worked pretty well.

[Orin Kerr]

I think balance is key: Long enough that there's a rich set of facts, but not so long that it takes a long time to read and digest. In effect, as short as the fact pattern can be while still giving students the facts they need to work with to apply the law.

I'm curious, though, what's the rationale behind giving students a choice as to what questions to answer?

[Jacqueline Lipton]

Orin: I wouldn't necessarily advocate a choice of questions in all subjects. In the last few years I've tended to teach "overview" type subjects like cyberlaw, international intellectual property law, and intellectual property survey. Because these courses tend to have a lot of breadth and not necessarily a lot of depth in any one area, I find it is perhaps more fair to let students choose the areas they want to focus on. When I teach more "nuts and bolts" subjects like contracts, secured transactions, sales etc, I tend not to give much or any choice of questions because I want the students to have a solid grounding of everything covered in the course.

[David S. Cohen]

I try to keep my questions straightforward and don't really try to put any unnecessary complexity, but I find that sometimes to get enough facts into the problem to give students something to analyze under many of the fact-specific doctrines of conlaw, the questions tend to be a page to a page and a half. Ideally, I'd shoot for under a page, but I think it's hard to do that and give students enough material to use.