« The Possible Relationship Between Course Design and the Curve | Main | I think I agree with this »

Wednesday, September 14, 2016

Some Resources for Grading

Given the article in the New York Times the other day and Prof. Preis’s Post, I want to share some observations about and resources for grading  from both a statistical and educational perspective.  First, it’s important to get on the table that there is a persistent and deep-seated misunderstanding that there is something “scientific” about assigning grades based on a “bell curve”  or “normal curve.” Neither concept is relevant to the actual scores law students receive on exams.  An equally problematic misconception is that just because you can put things in rank order means you know the interval between them.   Imagine an A,B,C, and D. grading scale.  If the top five scores on an exam are 99 and there are only five A’s, then the first student getting a B (student number 6) could have scored 98.  

  How much less does student number 6 know the material than students number 1-5?   Hard to tell without a lot more information.  But what if the top five A’s were 98, 90, 85, 72, 70 and the first B a 65.  Would you feel comfortable that the two “B” students probably had the same level of competence?   Remember how little we like rank order when it comes to ranking colleges.

The whole concept of a comparative grading scheme rather than measuring against external measures of success is problematic. Imagine the catastrophic results if the Navy did this with pilots who land on aircraft carriers—advanced students through training by comparing their skills to the group rather than to objective criteria of generally agreed upon success.   (they don’t). Sometimes students (and faculty) point to curving grades as protecting students from variations between classes in that it guarantees that some students will get A’s and B’s and that not everyone will get an F. But comparative grading also can mean that until the bar exam, it can be hard for the student (and for us) to  assess how much "torts" or "contracts" he or she really knows.   

Also, looking back to yesterday,  by artificially limiting high grades, we run the risk of discouraging students who really are achieving at levels very similar to their peers with higher grades but who may drop back from intensive study once they realize that they will always be at the middle of the pack.

What to do? Well, first of all, it’s helpful to remember that there is an entire  literature devoted to grading.  Second, it's likely that the focus on learning outcomes is going to be very helpful here because a core concept is that before you can effectively assign grades, you have to know what you want to measure. 

And however you grade, you don't do have to do it by hand.   Look here, here, and for some ideas about using Blackboard and Excel here and here and a video here.


Posted by Jennifer Bard on September 14, 2016 at 04:25 PM in Teaching Law | Permalink