Learning Trends Sound Convincing. That Is Exactly the Problem.

New learning trends, technologies and frameworks appear constantly. Some are useful. Some are old ideas in new packaging. Some are supported by evidence – under specific conditions. The real challenge for Learning and Development is not keeping up with every trend. It is knowing how to judge which claims deserve our trust, our budgets and our learners’ time.

In a meta-analysis of training evaluation research, this was the average correlation between participants’ affective reactions – essentially, how much they liked or were satisfied with a training programme – and subsequent workplace transfer.

The estimate was based on only six correlations, so it should not be treated as a universal constant. But the underlying lesson remains important.

People can enjoy a learning experience without necessarily learning. They can learn something without applying it afterwards. And they can participate enthusiastically without changing anything about their performance or behaviour.

Satisfaction is useful information.
It is not evidence that learning has occurred. A passed final test is not evidence of transfer.

The distinction sounds straightforward. In Learning and Development, we cross these boundaries more often than we realise.

When plausibility becomes a learning strategy

Consider some familiar claims:
Microlearning improves retention.
Self-directed learning mainly requires motivation and autonomy.
Personalised learning is more effective because it matches individual needs.
AI makes learning faster and more effective.

Each sounds plausible. That is precisely why claims like these spread. Learning myths rarely survive because they are obviously absurd. They survive because they simplify. A complex problem becomes a memorable model. A correlation becomes an explanation. A finding observed under specific conditions becomes a general principle. A technology becomes a learning method.

Eventually, plausibility becomes strategy. Budgets are allocated. Platforms are purchased. Learning programmes are designed. Success metrics are chosen. Learners invest their time. At that point, an attractive claim can become an expensive decision.

A claim is not evidence. And evidence is not yet a decision.

Suppose somebody tells you: ‘Microlearning improves retention.’
Before turning that statement into a learning strategy, several questions need to be answered. What exactly counts as microlearning? Retention of what? Compared with which alternative? For which learners? For which learning objective? Over what period? And what was actually measured? Even if research showed a positive effect under particular conditions, one question would still remain:

Does this evidence justify the specific L&D decision we are about to make?

That final step is often overlooked. Evidence that something can work does not automatically mean that
it is what we should do under the specific conditions in our organisation?
Between a research finding and a design recommendation lie interpretation and analysis. And this is where some seemingly evidence-based arguments start to become fragile. Evidence-based practice demands more than finding a study that supports an idea. It requires us to examine what was actually studied, what was found, under which conditions and whether those findings are relevant to the decision in front of us.

That distinction is also at the heart of my OEB masterclass, ‘
Learning Myths in Corporate Learning: What Really Works and What Just Sounds Good’.
The aim is not to replace one list of fashionable methods with another. It is to develop a way of thinking that can be applied to the next learning trend, framework, technology or vendor claim.

Are they engaged, or are they learning?

The same problem arises when we evaluate learning that already exists. The room is lively. People are discussing. Polls are being answered. The digital board is filling with Post-it notes. Participants give excellent ratings on the feedback form. Something is clearly happening. But what?

Visible activity is not the same as learning.

The ICAP framework, developed by Michelene Chi and Ruth Wylie, offers a useful distinction. It focuses not simply on whether learners are active, but on the type of cognitive engagement an activity elicits. Are learners clicking and responding? Or are they explaining ideas, making connections, developing arguments and solving problems? Engagement can support learning. It can foster interest, participation and persistence.

But the relevant question is not: Are they busy?
It is: Are they cognitively engaged? For example, are they analysing a problem and working towards a solution?

Good learning is not the same as easy learning

Good instructional design should remove unnecessary difficulty. Confusing instructions, irrelevant information and poorly structured materials consume cognitive resources without improving learning. But reducing unnecessary cognitive load is not the same as removing all cognitive effort.

Retrieving knowledge from memory is more demanding than rereading. Explaining an idea requires more thought than recognising it. Solving a problem requires learners to do something that simply watching somebody else solve it does not.

Research on retrieval practice and desirable difficulties shows that some learning conditions that feel harder during practice can improve later retention. The important word is desirable. Difficulty does not, in itself, make learning effective. A challenge must serve the learning objective and remain manageable for the learner.

The aim is not to make learning harder. The aim is to make learners think. AI makes this distinction even more important

Today, generative AI makes information easier to access than ever before.

The problem is no longer simply access to information, but how to curate and evaluate it.
Ask a question, and a fluent, convincing answer can appear within seconds.

But there is a considerable gap between ‘I got an answer’ and ‘I learnt something’.

This does not mean that AI is bad for learning. Recent meta-analytic evidence suggests that GenAI-supported interventions can produce positive learning outcomes on average. But averages hide an important part of the story. Effects vary substantially across learning contexts, subjects, learners and instructional designs.

Current research on AI and higher-order thinking points in the same direction: AI can support reasoning, reflection and critical thinking when it is embedded in appropriately designed and scaffolded learning activities. Unreflected use can create very different conditions.

So the question is not: Does AI support learning?

It is: What are learners doing with AI?

Are they generating explanations? Comparing alternatives? Testing their reasoning? Evaluating the quality of an answer? Using feedback to revise their thinking? Or simply outsourcing the task?

The technology alone cannot answer that question. This brings us back to a simple principle:
The pedagogy should determine the tool, not the other way round.

Five questions before the next learning investment

When I examine a learning claim, I start with five questions.

1. What exactly is being claimed?
What is supposed to change?

2. What outcome is actually being measured?
Participation? Satisfaction? Learning? Transfer? Performance?

3. What evidence supports the claim?
A testimonial? A case study? A correlation? An experiment? A systematic body of research?

4. Under which conditions does the effect occur?
For whom, doing what and in which context?

5. Does the evidence actually justify the proposed decision?

The questions are simple.Applying them consistently takes effort.They also expose one of the most common misunderstandings about evidence-based practice.Evidence is not something we add afterwards to justify a decision we have already made. It should help us make the right decision.

From learning myths to better L&D decisions

Participants will leave with a practical way to test the next learning trend, technology or vendor claim before it turns into an investment decision. This means developing the ability to distinguish a convincing claim from a supported effect, and a supported effect from a justified decision.

We will put established assumptions, current trends and popular learning claims under scrutiny. Not to dismiss them. To examine them properly. We will look at topics such as self-directed learning, microlearning, AI-supported learning and persistent models that continue to shape L&D decisions.

Participants will work with decision criteria they can use to assess the next framework, technology, learning method or vendor claim that lands on their desk. This matters particularly if you design learning, commission programmes, select learning technologies or make decisions about L&D investments.

When resources are limited, expectations are high and new technologies promise simple solutions to complex learning problems, the ability to evaluate those promises becomes a strategic capability.

Facts instead of FOMO

Examining the gap between what sounds convincing about learning and what the evidence actually allows us to conclude has become the central theme of Yvonne’s work as a learning scientist, L&D strategist and Learning Myth Detective. Her first book, Lernmythen aufgedeckt – Wie wissenschaftliche Evidenz effektives Lernen und Praxistransfer im Unternehmen fördert, examines persistent misconceptions about learning and workplace transfer. It received the getAbstract International Book Award 2025 in the Business Impact category. Its follow-up, Neue Lernmythen aufgedeckt – Wirksames Lernen braucht Fakten statt Buzzwords, takes the investigation further, examining current claims about engagement, self-directed learning, microlearning, knowledge, AI and other ideas shaping corporate learning today.

Her keynotes challenge the assumptions behind such claims. Her masterclasses turn evidence-checking into a practical skill. Through consulting and sparring, she works with L&D teams to apply the same scrutiny to their own learning strategies, programmes, technologies and investment decisions.

The purpose is not to make L&D more sceptical. It is to make L&D decisions more deliberate, more defensible and more likely to lead to meaningful learning and transfer.

We do not need to follow every trend. We do not need to reject every trend.
We need to know how to examine one.
Trends do not belong in the dock. They belong under scrutinity.




Written for OEB 2026 by Dr Yvonne Konstanze Behnke.

Join Yvonne Konstanze for her Masterclass “Learning Myths in Corporate Learning: What Really Works and What Just Sounds Good” at OEB26.