While listening to an audiobook called “Range” by David Epstein, I’ve come across the concept related to learning called desirable difficulties. These are features of the learning process that, as the name implies, create difficulties for the student but improve the long-term learning outcome. There is compelling evidence that increasing the difficulty of tasks is beneficial in the long-term, even though it slows down the initial progress.
This is counter-intuitive, and it creates a conundrum, both for the student and for the teacher: one needs to trust the process to continue viewing the difficulties as desirable even in the face of decreased performance, e.g. relatively poor test results. There is some consensus, though, on how desirable difficulties can be created by the teacher or by the students themselves (the key assumption here is that everyone agrees that the difficulties are, in fact, desirable).
One tactic is retrieval practice, which is, basically, testing. Again, it’s been shown that spending some learning time on testing, including self-testing, is beneficial. It sounds like a truism, but exerting effort in retrieving the information that needs to be learned helps with the learning. Flashcards is a typical example of a retrieval practice tool, and progressively increasing the size of the stack of flashcards is a desirable difficulty.
Not surprisingly, feedback is important for learning, i.e. the student needs to receive correct information about their performance. Surprisingly, though, delaying the feedback, or the test itself, is a desirable difficulty. This idea clashes somewhat with the huge body of research that shows that immediate feedback is beneficial for building skills, being a key characteristic of so-called “kind learning domains,” e.g., classical music, golf, chess, etc. I think there is no logical problem here, though. Kind learning domains facilitate reliable immediate progress, while “wicked domains”, where feedback is delayed, are conducive to better long-term learning. I should note, that delayed feedback can be intentionally used in a kind domain, and that wicked domains are sometimes characterized by misleading feedback, which is definitely not conducive to learning and what makes these fields “wicked” in the first place.
Another neat technique for introducing a desirable difficulty is interleaving. Here is an example from the “Range” book. Suppose that you are studying painting styles of van Gogh, Picasso, Monet and Kandinsky with the goal of being able to identify the author of a painting by their style during an upcoming visit to a museum. If you are using flashcards with reproductions of various paintings on one side and the painter’s name on the back to self-test your knowledge, it would be more difficult, but beneficial in the long-term, to make a deck of flashcards containing the works of all these painters, rather than studying them one-by-one.
When I was a graduate student, one of the professors in our department used to joke that one qualifies for a post-graduate degree not so much on the basis of acquired knowledge or on the level of contribution to the field, but on the certain amount of suffering one accumulates during the studies, e.g., suffer for two years – get a Master’s degree, suffer for four more years – get a PhD. It appears that there is some truth in this joke – the amount of struggling along the way correlates with slower initial progress, but also with deeper knowledge down the road.
