How to increase the productivity of undergraduate research assistants?

I supervise a few undergraduates doing independent study and internships in Computer Science. Some of their tasks are to assist me with various back-burner research projects. I can offer them college credit in exchange for their work. However, their labor is difficult to utilize. Their commitment may only be for a few hours a week, they can take a long time to ramp up work, and they sometimes disappear mid-semester. What strategies are out there to make the most of working with this mercurial resource?

Answer

I've had a lot of undergrads in my lab (physics). Unlike Corvus, I'd say that certainly more than half have been net positives to the group's research, and several have been very productive. The two most important things, I would say, are:

1. 
   

   A lot of direct contact and guidance -- not by email, but actual conversation -- either (ideally) from you (the PI) or from graduate students committed to the project and to mentoring someone. Simply being clear about tasks to be done, etc., is insufficient; every new researcher comes in without a good grasp of the motivations of the field, the challenges of exploring something new, etc., and these are only surmounted by talking to people. Several times, I've had great conversations with undergraduates in which I point out that I'm thrilled that their experiment has failed, because now they can better understand what doing science is like, and they can pick themselves up, learn from what happened, and try again, and this is what makes them 'real' researchers.

   
   


2. 
   

   I require a commitment of more than a few hours a week. If people aren't putting this in, we chat about the impossibility of having a meaningful project, and, again, learning how to be a 'real' researcher. Sometimes people drop out (or I ask them to leave) if the motivation to do this isn't there. There's no point in mentoring someone who can't commit real effort, and it doesn't do them or you any favors.

   
   

Your question is a great one. Best of luck, and please don't think that mentoring undergrad research should just be a net-negative "service." It does take a lot of work, but it can pay off!