Bryan Caplan  

Forgetting: The Basic Facts

The Economics of the Cadillac ... Forgetting How to Drive in the...
You don't know most of what you've learned.  Why not?  Because human beings forget.  How and why do they forget?  The main story is that people forget what they don't use.  A bit of googling turned up a nice meta-analysis, "Factors That Influence Skill Decay and Retention" (Arthur et al., Human Performance 1998).  Defining the problem:
Skill decay refers to the loss or decay of trained or acquired skills (or knowledge) after periods of nonuse. Skill decay is particularly salient and problematic in situations where individuals receive initial training on knowledge and skills that they may not be required to use or exercise for extended periods of time.
So what's known about skill decay?  Common sense checks out.

1. Time matters.  "There is an increase in the amount of skill decay as the length of the nonpractice interval increases."

2. Physical skills decay slower than cognitive skills.  "[P]hysical tasks display less skill decay than cognitive tasks, and the difference in decay is close to half a standardized unit... across all retention intervals."

3. Speed tasks decay more slowly than accuracy tasks.  "Across all retention intervals, the amount of skill decay for accuracy tasks was over three times higher than that of speed tasks (i.e., [effect size] = -1.00 and -0.32, respectively)."  Learning to do something rapidly stays with you longer than learning to do something correctly.

4. Ability to transfer knowledge decays faster than mere retention.  "[S]kill decay [is] negatively associated with the level of similarity between the original learning and retention contexts."

Big takeaway:
[T]he amount of skill loss ranges from a d [effect size] of -0.1 immediately after training (less than one day) to a d of -1.4 after more than 365 days of nonuse. That is, after more than 365 days of nonuse or nonpractice, the average participant was performing at less than 92% of their performance level before the nonpractice interval.
The results of this study suggest that the similarity of the training (acquisition) and work (retention) environments plays a major role in the retention of skills and knowledge over periods of nonuse or nonpractice, providing additional support for a basic tenet in training-program design - that is, to enhance retention, trainers should try to ensure the functional similarity of both the training device (acquisition) and actual job equipment (retention) and the environment in which both are performed.
The authors don't connect their findings to pedagogical reform, but I'm happy to pick up the slack.  If you really want kids to acquire a cognitive skill, don't just teach it to them and move on.  You have to maintain the skill not just with practice, but with distributed practice.  The iconoclastic flip side: Cognitive skills that aren't worth endlessly practicing probably aren't worth learning in the first place.

P.S. Don't forget these facts.  They're important!

COMMENTS (10 to date)
Doug writes:

I don't practice my hobbies nearly enough ... I guess this article is the motivation I needed!

Matt M writes:

Khan Academy knows this, and they incorporate it into their instruction. The Calculus lessons do a great job of re-testing you on skills learned in earlier blocks during later ones. The site tracks skills you've gained and knows to periodically test you on old skills so they don't expire. At each level, your "end of block test" for a new skill also includes several questions from prior skills. If you fail at these, you're required to do more review of that prior skill to move on. You must do "maintenance" on acquired skills to keep them checked off or you lose that check and have to re-earn it. Many other sites and teaching methodologies pay lip-service to this concept of review or do it half-heartedly, but Khan Academy does it in a very rigorous and deliberate way.

[html corrected. Please take a look at our FAQ at for creating links.--Econlib Ed.]

Joao Eira writes:

I am fascinated with memory research and its impressive success in finding general rules that govern how everyone's memory works, not just a tiny subset of the population.

Anyone interested in getting up to speed with the basic findings (and they may be basic but they will change your intellectual life if you know them) might start by googling "spacing effect" and start from there. You can download Anki, read some of the Wozniak articles to know how to use Anki (even if they were meant for Supermemo, the principles are the same), and you'll quickly see how useful it all is.

Felipe writes:

My college calculus teacher used to say you learn each course's material while taking the next one. Which is why nobody remembers anything about multi variable calculus but still remember a lot of single variable calculus (as you may guess, mv calc was the las course in the series).

ThomasH writes:

Of course there is a tradeoff between practicing to retain skill X, Y, and Z and practicing the ability to continuously learn new skills. Presumably the real value of a PhD in economics is to demonstrate that one can learn how new models work and how to make new models, not that one can operate the models one was trained on.

Hazel Meade writes:

I suspect that skills that have been put to use at one point are easier to relearn than skills that were merely taught but never practiced.
In other words, if you take a class on AutoCAD, and then use it for a while, you're more likely to remember how to use AutoCAD than if you take the class and then never use it.

So, you don't necessarily have to continue ot practice all your skills continuously, as long as you put some time in using the skill in practice after having learned it.

Njnnja writes:

This is why I went into math. It doesn't matter how much I forget since I can just re-derive everything from first principles!

JK Brown writes:
trainers should try to ensure the functional similarity of both the training device (acquisition) and actual job equipment (retention) and the environment in which both are performed.

The more succinct version used in police, military, and self defense training is: Train as you'll fight.

A Country Farmer writes:

The last sentence made me cackle :)

bill writes:

Nice post. I had the good fortune of learning to speak German in Germany in a class where the students came from all over the world. What that meant was that the teacher could only speak German to us. Whereas if I had studied in the US, the teacher would have "explained" German to us in English. So every single lesson was repeated practice of prior lessons and the teacher had to innovatively show us the new concept using previously learned concepts. It's the best way to learn.

Comments for this entry have been closed
Return to top