Here's another recent bit of research, published in Science, demonstrating the value of doing a test ("retrieval practice"), immediately after learning something. Those who did the tests, instead of other study techniques, had much better retention 1 week after learning something.
This finding is consistent with my believe that a lot of study time is inefficient, because it encourages the studier to push forward to "get more reading done", before the reading which was just done has been consolidated well or reviewed. Even though "pushing forward" may seemingly cover more pages of text, it accomplishes less long-term learning. A much more efficient use of time is to pause, review, and do a test on the material at just the right interval. If the interval is too short, the tests will be too easy, and the review will also be an inefficient use of time. If the interval is too long, too much will have been forgotten already, the tests will be too hard, and it will make you have to go over the same material again, another inefficiency.
I don't believe findings such as this one necessarily contradict ideas about flexible or "constructivist" approaches to education. Nor do I believe it has anything to do with the controversial area of standardized tests for children or professional schools, etc. Standardized tests are not a mechanism for education, they are assessment tools (how good they are as assessment tools would be a subject for a good debate).
There are some instances in which pausing frequently to review could disrupt a larger thematic appreciation of a subject or experience--it would be like pausing a movie every ten minutes to answer questions about the plot or characters--so, of course, sometimes this technique would have to be set aside.
I think that flexible, personalized educational approaches are extremely important--but this evidence about the merits of retrieval practice testing can be applied to any such style. Its immediate value is in helping people use their time more efficiently for many study tasks.
A previous post also deals with the subject of study efficiency, and actually cites a more densely technical analysis showing more or less the same result, though it adds information about the frequency with which one should optimally pause to test oneself: http://garthkroeker.blogspot.com/2010/04/optimal-learning-training-schedules.html