Decision Trees

While watching this video (Big Data in Education MOOC, Week 1 Video 3 at around 8:25), I tried to think of an example of when to use a Decision Tree.   I've never actually tried using a decision tree before (outside of class assignments and stuff). 

In my work, I use weighted sums to compute P(learned).  (aka latent knowledge estimation! (related entry))

The decision trees appear to be useful if you already have the final P(learned) but maybe you want to predict one of the components (such as the skill level of the student, or the time spent on the exercise, or other dimensions that could be used to calculate P(learned)).

'Just a thought.