When training fails, our first instinct is often to add something.
Another explanation. Another example. Another module. Maybe a quiz with three answers nobody could possibly get wrong.
But what if content isn’t the problem?
I recently reviewed 50 studies from psychology, behavioral science, and learning research. I chose them for their influence, historical value, and usefulness at work.
Several of them highlight that training isn’t always the answer.
Here are five that changed how I think about performance.
Get the complete 50-study field guide
Want the other 45? I compiled all 50 studies, along with their limitations, workplace applications, and practical ideas, into a free field guide.
A layout problem can become a learning problem
Picture a diagram on one side of a screen and its explanation on the other. Your eyes bounce back and forth. Meanwhile, your brain tries to hold one piece in memory while searching for the next. That mental juggling has a cost. Paul Chandler and John Sweller called it the split-attention effect. When people must combine related information from different places, the layout itself can interfere with learning.
Split attention—and related attention problems—can appear when we:
- Put a legend far from the visual it explains
- Separate instructions from the task
- Make employees switch between systems to assemble one answer
- Show visuals that compete with the narration
- Add decoration that steals attention
The lesson is not “make everything shorter.” It is this:
Make important relationships easier to see.
Try this: Find one screen that separates information people must understand together. Put those pieces side by side—but don’t duplicate information that already makes sense on its own.
Remembering beats reviewing
Rereading feels productive.
The words look familiar. The page makes sense. We nod along and mistake recognition for mastery.
Then we need the information three days later—and it has vanished.
Henry Roediger and Jeffrey Karpicke compared repeated study with retrieval practice: asking people to recall what they had learned.
Repeated study performed better after a very short delay. Retrieval produced stronger retention later.
That difference matters.
If people need knowledge on the job, they should practice retrieving it from memory—not merely looking it up again.
A quiz should not be the inspection station at the end of a course. Used well, it can help create the learning.
Try this: Before reviewing something you have already taught, ask learners to recall it without looking. Then provide the correct answer and useful feedback.
Effort helps when success is possible and feedback follows. Blind frustration does not.
Impact becomes real when it has a face
Organizational psychologist Adam Grant and his colleagues studied university fundraising callers.
Some callers met a scholarship student whose education had been funded by their work.
The script stayed the same. So did the pay.
But the callers could now picture someone who benefited when they picked up the phone. After meeting the student, they worked longer and raised more money.
Now compare that with the language we often use to introduce workplace training:
- Maintain compliance
- Reduce organizational risk
- Support strategic priorities
- Complete this requirement by Friday
Technically accurate. Emotionally dead.
Grant studied direct contact with a beneficiary, not whether a good story can rescue a dull compliance course. But the design question travels well:
Can employees see the person their actions affect?
Who stays safe? Whose work becomes easier?
Try this: Replace “This training is important because…” with a real example of someone affected by the behavior.
Don’t just explain the requirement. Give the consequence a face.
Information cannot fix a friction problem
Imagine you know exactly what to do.
Then you meet the form.
The instructions are vague. The button you need is buried three menus deep. One approval takes four days, and nobody knows who owns it.
Would another course help?
Eric Bettinger and his colleagues studied families applying for college financial aid. One group received personalized information. Another received the same information, plus hands-on help completing and submitting the application.
The help changed behavior. Information alone did much less.
The workplace parallel is hard to miss. We prescribe training when the real obstacle is:
- A confusing form
- A buried button
- Too many approvals
- An unclear process
- A lack of time
- A system designed like an escape room
If people already know what to do, more content is unlikely to save the day.
Try this: Before building training, watch someone perform the task. Count the clicks, decisions, delays, handoffs, and chances to give up.
You may not have a knowledge problem. You may have a friction problem.
Sometimes the right intervention is not better instruction. It is fixing the system.
A badly framed norm can make behavior worse
In a well-known field experiment, households learned how their energy use compared with their neighbors’.
The heavy users cut back.
But some households that had used less energy began using more. Once they saw the average, they drifted toward it even though that meant doing worse.
A small sign of approval, a smiley face, stopped that boomerang effect.
Now consider this workplace message:
Thirty percent of employees still click suspicious links.
The intended message is clear:
This behavior needs to change.
But employees may hear something else:
Plenty of people do it. I wouldn’t be unusual.
Never make the behavior you want to stop sound normal.
When you use a social norm, make it truthful and pair it with clear approval of the desired behavior.
If your data support it, you might say:
Most employees report suspicious messages when something doesn’t look right. Thank you for helping protect your coworkers.
What if the desired behavior is not common yet?
Don’t invent a statistic. State the standard:
Reporting a suspicious message is exactly the response we want. Every report helps protect the people around you.
Try this: Review one message about poor behavior. Does it make the problem sound normal—or the solution?
The Bottom Line
These studies examine different problems, but they point in the same direction:
Performance depends on more than content.
People must notice what matters. They must recall it when needed, see a reason to act, have a workable path, and believe the behavior belongs here.
Training can help.
But it cannot repair a cluttered design, bring an empty mandate to life, remove a broken process, or overpower a culture that rewards the wrong behavior.
Sometimes the smartest move is to improve the training.
Sometimes it is to fix what stands in the way.
What workplace learning problem should I investigate next?