One Brain, Two Settings

Adults learn with the same machinery as teenagers. Recalling something strengthens it, spacing practice out makes it last, and feedback on the task moves people forward, in a training room as much as in a classroom.

What changes at work is how much people already know, and that is where most corporate training goes wrong: it gives everyone the same course, which holds back the expert and loses the new hire.

How People Learn sets out each of these principles and its evidence in full. This essay is about what carries over to work, and what does not.

The core findings from school research hold for professionals. In a US trial with medical residents, material they were tested on was remembered far better six months later than material they had studied again, 39% against 26% (Larsen and colleagues, 2009).

In a trial with 537 urology residents in the US and Canada, short questions sent by email and spaced over weeks beat the same content studied in one block (Kerfoot and colleagues, 2007). In a later follow-up the advantage from spaced questions was still there after two years (Kerfoot, 2009).

A review of 23 studies with more than 3,000 health professionals found the same small-to-moderate benefit from spaced digital training (Martinengo and colleagues, 2024).

In one trial the learning reached the job. US clinicians who played a game of spaced questions about blood pressure got their own patients’ blood pressure under control slightly faster than clinicians who did not (Kerfoot and colleagues, 2014). It was measured in patients, not in a quiz.

Prior knowledge is the real difference between learners, and it cuts both ways. Support that helps a beginner, such as a fully worked example, can get in the way once someone knows the task, because they have to process guidance they no longer need. This is the expertise reversal effect (Kalyuga and colleagues, 2003).

The effect showed up in a study of apprentices at two Australian manufacturing firms: worked examples helped them while they were new, and once they had experience, solving problems on their own worked better (Kalyuga and colleagues, 2001).

A course pitched at the middle of a team is too much for the people who know the work and too little for the ones who don’t.

In my own classes, mixing levels in one group became a motivation problem too. Learners who had to wait for others to catch up, or were walked through things they already knew, switched off. They also trusted the trainer less. It felt like a waste of their time, sometimes an insult to their intelligence, and often they stopped coming.

Several ideas that corporate training runs on have little evidence behind them, and each one points design effort the wrong way.

Andragogy, the idea that adults need a different kind of teaching, has been studied for decades with results one review called inconclusive (Rachal, 2002). The trials above point the other way: methods first tested in schools also worked with doctors. What adults bring is more prior knowledge, and that is what a course should adjust to.

The 70-20-10 rule, that 70% of learning at work happens on the job, comes from executives looking back on their careers and has never been tested as a way to design training (Clardy, 2018). The rule says nothing about how to make learning on the job stick. Spaced recall and feedback can be built into the work itself, using the tasks people do every day.

Learning styles have no adequate support (Pashler and colleagues, 2008), yet a review across 18 countries found that 89% of educators believe in them (Newton and Salvi, 2020). Matching training to a style spends effort that should go into matching it to what someone already knows.

Microlearning is a format rather than a method. A short video watched once fades like any other single session. Short pieces work when they come back over weeks and make people recall what they learned.

Motivation is where the workplace does add something of its own. People keep learning when they have some choice, can see themselves improving and understand why it matters, and managers who support that autonomy have more self-motivated staff (Slemp and colleagues, 2018). That is a reason to give learners real choices and show them their own progress. It is not a reason to change how the learning itself is built.

So the design starts where each person starts. A short diagnostic finds what someone already knows. Beginners get worked examples and guidance that fades as they improve. Experienced staff go straight to realistic problems, with help only when they need it. Practice comes back in short, spaced pieces that ask people to recall and apply, not reread, and the feedback is built on a map of the mistakes people in that job usually make. It is the same engine a good school system uses, with different content.