So let's look at his experiment more closely to see what I mean. Tolman's lab set up mazes for three different groups of rats, with each rat released into the same maze every day for 22 days. Each rat in the first group was released into a maze where food was placed at the end for all 22 days. Each rat in the second group was released into a maze every day for 22 days, but no food was ever left at the end. The third group was given the same 22-day test in a maze, except food was placed at the end of the maze starting on the 11th day. It was the performance of this third group that got the researchers' attention. The rats in all three groups were measured for speed (time from beginning to end of the maze) and number of mistakes (wrong turns). On day 12--the first day after the third group found food at the end of the maze--the third group's performance improved dramatically. While the first group's performance had been the best up until that point, the third group's improved immediately and surpassed both of the other groups.
The scientists interpreted the results of these experiments to mean a couple things, at least: one, that those rats in the third group were learning even without obvious incentives (this has been called "latent" learning, and we will discuss that in a later blog post); and two, that as all the rats were learning the maze, they were creating cognitive "maps." Those rats that were rewarded with food from the beginning learned the route but not as fully as the rats who first explored the whole maze without any reward, since the first group always stopped exploring once they found the reward. Once the reward was introduced to the third group of rats, they had the advantage of 11 days of wandering around and learning the entire maze, and thus were better able to retrace the route to the food.
The concept of "desirable difficulty" was not part of the discussion about this set of experiments at the time--and it wasn't part of the questions that engendered the research to begin with--but it seems that we can read back on the experiment what scientists have learned about effort and learning: those rats that learned the maze best had to work harder for it, and did so by learning about the larger context of the maze through trial and error. And, Tolman and his colleagues have offered us the useful analogy of "mapping" for the way information is stored and processed in our brains.
In our next post we will take up the question of how these results have been interpreted, and what the debate means for learning theory.