Bryan Caplan  

Low Transfer of Learning: The Glass Is Half Full

The Subtle Value-Added of Fred... Living in Bubbles...
Teachers like to think that no matter how useless their lessons appear, they are "teaching their students how to think."  Under the heading of "Transfer of Learning," educational psychologists have spent over a century looking for evidence that this sort of learning actually occurs. 

The results are decidedly negative.  Learning is highly specific.  One decent summary:

[T]ransfer is especially important to learning theory and educational practice because very often the kinds of transfer hoped for do not occur. The classic investigation of this was conducted by the renowned educational psychologist E. L. Thorndike in the first decades of the 20th century. Thorndike examined the proposition that studies of Latin disciplined the mind, preparing people for better performance in other subject matters. Comparing the performance in other academic subjects of students who had taken Latin with those who had not, Thorndike (1923) found no advantage of Latin studies whatsoever. In other experiments, Thorndike and Woodworth (1901) sought, and generally failed to find, positive impact of one sort of learning on another...

Thorndike's early and troubling findings have reemerged again and again in other investigations...

Most educational psychologists are dismayed by what they've discovered about Transfer of Learning.  (Here's one eloquent example).  After reviewing the National Assessment of Adult Literacy, though, I realized that researchers' moping is premature.  Despite everything, the glass is half full. 

How so?  Well, the NAAL finds that American adults' literacy and numeracy is shockingly low.  Many high school graduates have trouble correctly interpreting a t.v. guide or calculating the total savings from $.05/gallon savings on 140 gallons of oil.  If knowledge reliably transferred, we would have to conclude that most American adults are simply unemployable.  After all, if people readily generalize from specific knowledge to general knowledge, their lack of general knowledge strongly suggests their lack of specific knowledge.

The reality is happily different.  Despite their low literacy and numeracy, Americans perform a wide variety of jobs competently.  How? 

Step 1: Specialize. 

Step 2: Practice, practice, practice. 

If they work for the post office, they learn how to correctly sort mail.  If they work at McDonald's, they learn how to make fries and make change.  If they install automatic garage door openers, they learn lots of details about garage doors.  The result: We get mail, food, and convenient access to our garages.

Wouldn't it be better if this task-specific competence naturally blossomed into across-the-board excellence?  Sure.  But that's probably not on the menu.  Instead of bemoaning American workers' mediocre literacy and numeracy, we should be grateful that millions of Americans who learn little in school still manage to learn useful trades on the job.  Seriously.

Comments and Sharing

COMMENTS (14 to date)
blink writes:

So what do you recommend? Should we revamp the curriculum and stop teaching such subjects? If we did, could that lead to an even worse situation than our present signaling equilibrium? After all, even studying Latin is better than repeatedly copying the dictionary or memorizing long strings of decimal approximations of transcendental numbers.

Fonzy Shazam writes:

Two questions:

Is there more recent, high quality evidence? Going back over a century loses something at least aesthetically.

Assuming learning skills and knowledge that will be applicable to people throughout life or at least job/career enhancing is generally determined late in the education cycle, is there something to the idea that these various studies that typically do not yield much including associative learning are nevertheless important? I'm thinking from the standpoint of casting a wide net. Think of the college freshman who takes a lot of electives trying to find his calling. Do we lose something as we lose the more esoteric subjects; a la, Latin? I realize this isn't in direct opposition to your main point that the evidence of "learning how to think" seems more a dream than reality.

Tracy W writes:

I think the educational psychologists still have a case for moping. After all, by their evidence, governments around the world are wasting billions on education. And you presumably become an education psychologist in the first place in the hope that you'll improve educational outcomes, and this evidence here is depressing.

Devon writes:

It isn't the exact same transfer of learning argument, but does the literature cited above also test the notion that learning one language makes it easier to learn another? That strikes me as a lower bar for Latin-supporters, for example, to get over.

Michael Rulle writes:

Well stated. In the examples I would have preferred more diversity. For example, econmics professor :).

Kidding aside, the concept of specialization in learning is more universal and not related to measured IQ as I inferred from your comment.

People often believe, either about themselves or about others, if one is proficient at one thing (e.g., learning in school), that would make them proficient at other things if they had applied themselves (e.g., head of marketing at a company). That is clearly not correct.

I definitely agree with your point on specialization, practice and learning. I just think it applies more universally than a linear walk down the IQ curve to people working at McDonalds. Speaking of which, try to buy a franchise from the latter. Acceptance rate is approximately the same as Harvard.

Thomas DeMeo writes:

It seems odd to assume that Latin is a particularly magical topic that has a greater impact than others.

Let's say we took a set of identical twins. One went to school for 16 years. One did chores and worked at the family farm. Would the one that went to school perform better at writing software, or designing a widget? Would the other twin be able to deal with working in an office environment?

To take the last comment first, software design and widget design and the office are all "school tasks." The problem may be in aggregating all of "school" under one rubric. We were required to take shop classes in junior high (Cleveland Public, 1950s-60s). Those of us who were told that we were college-bound were not encouraged to do well and were allowed to blow it off because it did not count against you for the academic honor society rankings. Thirty years later, I was working with robots. Who knew? So, whether learning transfers or not, a broad range of it seems helpful, at least.

As for Latin, learning one language does make learning others easier. I grew up with Hungarian, took German, (later Italian and Japanese) and then 40 years later as a campus safety officer, taught myself enough Farsi to greet and clear a room of Persian Poets. I also maintain that I learned first Fortran and then Java because they are languages. But does learning Latin help you learn accounting ... or the steps required to "drop and hook" a semi-truck and trailer?

liberty writes:

The problem is the way that we define and think about learning and education.

Every human can learn - this is part of what it means to be human. Even the highly "learning disabled" it turns out can learn, although they may be better at different things and require different kinds of education. Videos abound on the web about the miracles that have occurred when parents stuck it out. (e.g., this girl)

We need to realize that everyone is different, and a one-size-fits-all education system or way of thinking about education won't work. Of course outcomes are bad. Of course you'll find that learning one thing won't help you learn another if you teach kids the first thing in a boring, repetitive, memorize-this, kind of way.

Here is a Ted talk on education and creativity.

JKB writes:

This "low transfer of learning" is a term that confuses the real problem. Perhaps the inability to properly name the problem is a part of the problem. The problem is a lack of assimilation of the learning. The inability to use knowledge learned in one context in an unrelated context of your life. Our education system, perhaps inadvertently, works against assimilation with the discrete topic periods which condition students to put away one type of learning and focus on the another type.

Add to that the spoon-feeding that is endemic especially in the first decade or so of "education.' This teacher prompted "learning" creates good little test takers but at the price of thinking.

To assimilate, one must think about (consider purposely, judge, associate and organize) and struggle with the knowledge. In essence, you keep what you kill. Most of the rest goes in one end and out the other.

Floccina writes:

If school is mostly signalling, why can’t we signal while learning useful stuff? I understand that what is learned must be easily testable to serve as signaling and this eliminates some areas of useful knowledge but it seems it could be much better than it is.

To me it seems that the big waste is that signaling has squeezed out education.

Philo writes:

@ Floccina:
“[W]hy can’t we signal while learning useful stuff?” Because of specialization—the Division of Labor--there isn’t much stuff that would be useful to all, or even most, of us. These qualify: mastery of English, especially reading, and of arithmetic. But what else can you specify?

Is it learning that doesn't transfer or is it teaching that doesn't transfer? I'm sure learning transfers when students learn on their own.

Ken B writes:

As Devon notes, some transfer DOES occur. Learning calculus does help in learning abstract algebra later. Some of the mental skills overlap. Chess players pick up Shogi pretty quickly. Bridge players learn backgammon quickly it seems from my experience.

You need to be very careful in your definitions.

BG writes:

The paper that Caplan quotes doesn't seem to have the same trajectory as his article. It concludes:

"In summary, a superficial look at how research on transfer casts its vote is discouraging. The preponderance of studies suggest that transfer comes hard. However, a closer examination of the conditions under which transfer does and does not occur and the mechanisms at work presents a more positive picture. Education can achieve abundant transfer if it is designed to do so."

Comments for this entry have been closed
Return to top