Top read of 2008: Modern myths of learning: You only remember 10% of what you read

Main points: 

CRM calamityBest General Feature 2008

You only remember 10% of what you read...don't you? Hot on the heels of Martin Shovel's expose of the Mehrabian misunderstanding, Donald Taylor sweeps away another myth.

There's an old maxim in training:

You remember 10% of what you read
You remember 20% of what you hear
You remember 30% of what you see
You remember 90% of what you do

It's easily remembered. It's widely repeated. It's completely wrong.

I was delighted when Martin Shovel recently exposed how Albert Mehrabian's work on communication is often misrepresented. (See chapter 11 of Max Atkinson's 'Lend Me Your Ears' for more on this). There are too many myths and half-truths in the world of learning. They deserve to be uncovered and we need to ask ourselves hard questions about why the learning profession seems to love myth-making so much.


Photo of Donald H Taylor"A glib, uncritical parroting of the unproven is not only useless in itself, it also risks painting learning and development as a mumbo-jumbo profession."

The 'you remember 10%' myth was brilliantly pulled apart by Dr Will Thalheimer in May 2006, when he exhaustively detailed how the myth was created. (A presentation (PPT file) by Tony Betrus and Al Januszewski of SUNY Potsdam did much the same in November 2002, but Thalheimer's posting seems to have had the greater impact.)

The beginning of Thalheimer's article is unequivocal:

"People do NOT remember 10% of what they read, 20% of what they see, 30% of what they hear, etc."

He goes on to say that that information, and similar pronouncements are pseudo-scientific 'you remember' tosh.

Thalheimer is clearly angry.

He is right to be angry because a glib, uncritical parroting of the unproven is not only useless in itself, it also risks painting learning and development as a mumbo-jumbo profession, supported by nothing more than pseudo-scientific bunkum.

Dr Thalheimer's quest for the truth began when he saw this graph:


Problem solving graph


Look familiar? This graph or versions of the data are all over our field like a rash (see here, here and here for examples).

Despite the citation at the bottom, the graph does not appear in cognitive science. The cited author, Dr Michelene Chi, disowned it. Dr Thalheimer decided to look further.

He traced the first publication of the figures to D. G. Treichler, an employee of Mobil Oil Company, writing in 1967. However the NTL Institute for Applied Behavioral Science has laid claim to the figures, saying they are based on research 'in the early sixties' and bizarrely adding that 'we no any [sic] longer have - nor can we find - the original research that supports the numbers'.

Thalheimer has pulled NTL's argument on this to pieces pretty convincingly, not least by pointing out that all the percentages are perfectly round: what research into human activity ever resulted in six different round numbers?

It's bad enough that a national body lays claim to these figures without any proof of their validity. It's worse that the erroneous 'research' is then misattributed to genuine academics such as Dr Michelene Chi to give it a gloss of respectability. Far worse though, is the unholy melding of these numbers to Dale's Cone.

Edgar Dale developed his cone in 1946. It ranges through 11 different audio-visual media of increasing concreteness: from visual symbols to direct purposeful experience. Here is the cone as reproduced by Thalheimer (and by Betrus and Januszewski) from Dale's book 'Audiovisual Methods in Teaching':

Dale's cone of experience

No numbers. No reference to what people see or hear. Dale's Cone is just a useful way of considering the impact of more or less abstract experiences on learning.

But then somebody, we don't know who, coupled Dale's Cone with the spurious non-data of Treichler. The result: a series of Technicolor pyramids, bar graphs and semi-circles. They look great. They are irresistible to those who prefer copy-and-paste to reflect-and-consider. And – because the original research doesn't exist, and Dale never used it anyway – they are all irredeemably, utterly wrong.

Here's a typical one, revealed as spurious by Betrus and Januszewski in 2002, and still on the Computer Strategies LLC website today:

Dale's cone of experience chart

Remember: the words inside the pyramid aren't Dale's. The numbers on the left aren't based on real research. The citation is spurious, and yet this pyramid, or versions of it, are commonly trotted out as 'Dale's cone of learning'.

The very worst of it? Some of these diagrams are produced by people who really should know better. Academic bodies such as North Caroline State University, services for educators such as, and one individual - working for the good of others - who put a lot of work into producing two different pyramids, with the specific aim of making the diagrams available for free, general use, under creative commons.

What does this tell us about our profession?

First, arguably, it tells us that we do not have a profession. A real profession would have more concern about what was acceptable data rather than adopting things uncritically because they look pretty. Also, sadly, it tells us that many people working and writing in learning and development don't seem to want to take the time to stop and think.


"We do not have a profession. A real profession would have more concern about what was acceptable data rather than adopting things uncritically because they look pretty."

Just look at those numbers: 10%, 20%, 30%. Do they seem at all likely? A moment's reflection says they don't. A questioning mind would ask what exactly is meant by, for example, '10% of what you read'. How was the experiment carried out? What period of time passed before retention was tested? How familiar were the individuals with the subject matter covered? What was their reading age? Was a control group used? But nobody has asked those questions, and so both the bastard off-spring pyramids and the graph at the beginning of this article have been repeated endlessly, gaining credibility with each repetition.

Betrus and Januszewski's presentation was in 2002. Thalheimer's post was in May 2006, and has been referenced many times online. And yet many in our field continue to trot out the same learning myth, because they can't be bothered to check their facts – in September 2007, for example, Suite101 gave us the full range of suspiciously round figures.

We have to ask ourselves a question. Do we want to work in a profession based on solid data, where evidence is assessed critically? If not, what do we expect others to think of the field of learning and development? If we want to work in a real profession, we need to look beyond the surface and to question and to check. How serious are we about L&D, and how seriously do we want to be taken? It's up to us.

Donald H Taylor is chairman of the Learning and Skills Group and the Learning Technologies conference. He blogs at

Click here to read Martin Shovel's feature Mehrabian Nights: An informative tale about (mis)communication

Please note that this article first appeared on earlier this year.




Merry stuff to all

Nice one Donald, good graphics etc

I think what you are doing is very valuable. There are tons of good methods and facts in training and human resource management. They can be got through well written books in general.

Unfortunately the web is a different story. Some developments seem to latch onto as many myths as they can so as to hook as many customers as possible. The result: more myths spread.

Here is a great source:

I reckon NLP must be just about the most prolific spreader of myth out there today.

Of course NLP includes the same sort of myths you are talking about here. The article linked gives some interesting insights into why the myths spread and why people fall for them.

NLP is a great anti-exemplar to learn from.


DonaldHTaylor123's picture

Sue Beatt says “until someone comes up with some qualitative research that proves or disproves the theory, I'll keep designing training that is varied and stimulates as many senses as possible” – which is exactly as it should be: an L&D professional using a pragmatic approach.

The whole point of this article is this: we do not have to justify what we do in training and learning with spurious theories and invented data. Yes, we need good models and research, but they must be soundly based, and where there are gaps in our understanding, a pragmatic approach (such as Sue’s) beats invented nonsense every time.

Many of the comments below echo this point. A few, though, suggest that they believe the figures above to be right but unproven. That looks to me like an Emperor’s new clothes theory: one that people have invested too much in to let go.

Let’s use real research and data where they exist, and where they don’t, let’s not kid ourselves that if we repeat someone else's nonsense long enough time it will become true.

Andy B.'s picture

Reading Sue Beatt's excellent (IMO) reply to this article I was prompted to read the article itself again, and what did I find?

The article itself is as "pseudo-scientific" as anything it complains about.

Take this comment for example:

"Just look at those numbers: 10%, 20%, 30%. Do they seem at all likely? A moment's reflection says they don't."

A moment's reflection by whom?
Since when has "a moment's reflection" qualified as a "scientific" evaluation process?

Even more to the point, why should we regard Dr Thalheimer's findings as more reliable than anyone else's?

I'm not arguing here over the right and wrong of the figures. I simply distrust any presentation which says I have to believe or disbelieve such and such without giving me adequate evidence. Which, in my opinion, is exactly what this article does.

So, has it suddenly become "scientific" to use pseudo-science to disprove pseudo-science?
Pull the other one, it plays a subliminal version of "Silent Night"!

suebeatt's picture

Having read both Donald's posting and the presentation he mentions, my concern is that both the posting and the presentation pull the theory apart with nothing more than questions and statements. i.e. there doesn't seem to be any research behind the debunking except to trace where the figures actually came from.

I don't know whether the figures are right or wrong and I'm not venturing an opinion either way. But I do know that it takes more to to disprove any theory than someone saying, no matter how forcibly, that it's wrong.

The tone of Donald's posting reminds me of when my parents used to say "because I said so".

Yes, people used to believe the world was flat, and it wasn't until someone proved it wasn't by sailing 'over the edge' that people started to understand and believe the truth.

So, until someone comes up with some qualitative research that proves or disproves the theory, I'll keep designing training that is varied and stimulates as many senses as possible - just in case!!

You talk about the off-spring of the pyramids being endlessly repeated and gaining credence in the process.

Perhaps 90% of people are right not to believe what they read?

Andy B.'s picture

The trouble with skeptics is two-fold:

1. They tend to be skeptical about everything and everyone EXCEPT "scientists" and themselves (i.e. skeptics)
2. They frequently throw the baby out with the bathwater

Mehrabian's research has been misrepresented, it's true, but the original research was carried out in an appropriate manner, the results were replicated a few years later by a team at Oxford (IIRC) University under Michael Argyle, and similar research, under Rosenthal ("thin slices") and the "Myron Fox" experiment indicate that there is something genuine going on, not just the freak outcome of a single piece of research.

Unfortunately, in rejecting the overstated version of Mehrabian's conclusions, self-styled skeptics frequently reject the whole deal - and consequently end up as misinformed as the people they are criticising.

Just a thought

fledermouse's picture

Thank you for the great article, Donald!

I love myth-busting! (Especially since I am a part-time university student who actually does research!)

When statistics such as the misquoted ones by Mehrabian are offered, I always try to interpret these with a critical mind. Since I know there are all types of learners: audio, visual and kinesthetic, how can there be a one-size-fits all approach such as this 10%, 20%, etc. to all learners?

People learn in different ways, so there cannot be any statistics on how EVERYONE learns - unless it is a very, very loose guideline.

When statistics are offered, and we are tempted to use them for our own purposes, we need to first engage our three best tools: common sense, critical thinking and real research.

Gill hits on an interesting point about caring and putting down. However I think this needs to be balanced with the general acceptance of some of these theories, there is widespread almost global buy in to Mehrabian and other theories trotted out here. They are constantly used as source documents and course material by esteemed accreditation bodies. Ask anyone who has tried to challenge them and they will be quietly told 'not to cause a fuss'. So I suggest it is widespread belief that is being debunked and that is where the glee comes from, its not necessarily aimed at individuals per se. In order to undo widespread belief is is sometimes necessary to pull them apart in a structured way and to ram a point home one can also use ridicule and highlight inconsistencies and ambiguities - ask any barrister. This isnt a bad thing, its the cut and thrust of argument, if a theory is to stand up at all it should be able to withstand ridicule and glee from the protagonist.

> And I'm also wondering what great harm has been caused by our learners
> discovering that learning by doing has great merits, or that body, voice
> and words all play a part in the messages we send?

Millions of pounds and Euros and rupees at every call centre that existed, in training and lost time and sitting meaningless exams having to regurgitate these facts. Millions of disillusioned call centre agents and sales people who dared to query these figures which didn't stack up with their world view and were politely put down by management and trainers alike. Millions of customers who've had to endure poor scripts and overly empathetic sales agents and complaints agents, saying how "they understand their complaint and are very sorry but the computer says 'no'" and thousands of trainers who've been told not to query content and get on and train.

Once upon a time people believed the world was flat, compassion can only go so far in dispelling that myth. Don't get me started on "The Creation"....;-)

I'd be interested to hear what others think because you can bet your bottom dollar it'll be decades before this debunk becomes established as 'fact''s picture

It’s taken me some time to compose a response to this article having felt in turn shocked and then downright angry at the tone. By this in particular I mean the ‘delight’ at ‘exposing’ one myth or another ‘brilliantly pulled apart’

I’ve got no qualms about putting right a misconception – this particular discussion has been around for some time, as has the misrepresentation of Mehrabian and the over simplification of the left-right functions of the brain to name but a few.

And I’m also wondering what great harm has been caused by our learners discovering that learning by doing has great merits, or that body, voice and words all play a part in the messages we send?

I work with many trainers of varying ages and levels of experience, but many who are young, keen and with an enormous amount of passion to make a difference in their organisations.

Only last week I suggested to a couple of ambitious new trainers keen to take on some professional development to look at some of the past discussions on this site on the subject. I feel saddened to think that they might have come across this particular discussion and found the almost gleeful putting down of anyone who has in the past made reference to some unfounded research.

There is a lot of talk in the original article and the responses about the type of profession L&D is and wants to be. Yes, I’d like us to have great access to solid data where helpful. But I’d go for understanding, a genuine desire to help others to learn, compassion in helping those who might make ‘mistakes’, and an energy and enthusiasm to make a difference to people’s lives over academic ability any day.

chamberlainandco's picture

An excellent article that highlights the fact that received wisdom (RW) is usually received tosh; one only has to watch QI to see RW debunked every week.

Training is an art not a science, just as Economics is not a science. The excitement of training lies in the fact that delegates arrive with unique genes, formative experiences and whatever is going on for them that day. As a trainer my job is to work with what is in the room .I don't expect my GP to assume a one size fits all approach to diagnosis and cure.

Back to top Back to top