A research study published last week by Science Magazine is drawing a bit of attention. The Flypaper fellows are talking about it, the NY Times wrote about it (via Pittsburgh Post-Gazette), and the Northwest Evaluation Association tweeted about it:
All three give the impression that taking a test helps with learning. It's fair to say the assumption is that we're talking about standardized variety, or some kind of hybrid multiple choice and extended response format.
Not so. Not even close.
The "test" we're talking about here is simply writing down everything you remembered after reading a for a given amount of time (I believe it was 45 minutes in the study). Free-recall. No multiple choice. No grading of the material. That was a more effective way of studying than putting together a concept map. That's what the study really looked at.