There is something fascinating about science. One gets such wholesale returns of conjecture out of such a trifling investment of fact.

Mark Twain


Please note, we do not take anonymous comments, or comments that espouse illegal behavior.

Memory Leaks

All of us have had the same problem; we can’t recall a memory that we know we have. Sometimes we have an image of a thing, but can’t put a word to it. Sometimes the memory feels like an echo in our mind but the connection is lost. In the programming of computer software, that’s called a memory leak. Typically, in order to store information, a programmer creates a glob of memory to insert the requisite data and then creates a pointer to allow retrieval. If the pointer is lost, the memory doesn’t go away, just the ability to retrieve it. Memory fills up, while usefulness decays away. If that sounds like old age, join the group. The real question in education is how to enhance retrieval in order to promote mastery. And a different look at subject in a paper by Karpicke and Roediger1 offers a clue.

The experiment is simple: study and test word pairs from two languages: Swahili and English. Now break the experiment into four parts.

  1. Repeat study and testing of all word pairs (ST)
  2. Repeat study of all word pairs but drop testing of successfully recalled words (STN)
  3. Repeat testing all word pairs but drop study of successfully recalled words (SNT)
  4. Drop both study and testing of successfully recalled words (SNTN)

If this list is examined closely, the standard learning pattern espoused by many educators is the last: once information is learned, drop further study and testing. Mastery has been achieved.

The different methods of learning took the same number of trials until the students (college students) had learned all the word pairs; four trials. The learning curve looked approximately the same for each method. And obviously, the last method which dropped both further study and testing once recall had happened was the fastest. But is it the best?

The students were then tested one week later. The results are significant. Those students who used repeated testing (retrieval) recalled about 80% of the word pairs on average. Those students who dropped testing after successful retrieval but continued study recalled about 36% on average. Those students who dropped both testing and study after successful retrieval recalled about 33% on average.

As the authors said, “Repeated retrieval practice enhanced long-term retention, whereas repeated studying produced essentially no benefit.” In order for students to achieve long term mastery, they must constantly practice retrieval through testing during the learning process. And the testing does not need to be teacher initiated. Students can be trained to self-test. (Remember flash cards.)

So, as we head to personalized learning, it is important to remember that study alone does not achieve mastery. Teachers that use only a lecture style of teaching are fooling themselves and harming their students. And further, assigning study without testing will not achieve the long term goal. Repeated retrieval is necessary to reinforce the ability to master a subject. The testing effect is real.

Of course, some of us learned that images are powerful methods of retrieval. We wrote the subject down so that hand-eye coordination and subsequent images acted as part of the pointer to that blob of memory. There are many ways to drive retrieval. Because, unlike some programming languages today, we don’t have garbage collection.

Just for fun, if you want to follow up, try Roediger’s page at the University of Washington at St. Louis.


Leave a Reply

You can use these HTML tags

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>