Does Explicit Instruction Work For Students with Learning Difficulties?

Unguided or discovery learning imposes unfair cognitive burdens on students with ADHD and dyslexia but in different ways

Twenty years ago, there were two views on learning that seemed to be fairly uncontested. Firstly, students learn best when they discover knowledge for themselves and secondly, explicit instruction is somehow inauthentic, a relic of authoritarian teaching that stifles creativity and critical thinking.

The final part of that calculus was that if students with learning difficulties are struggling, they simply need more opportunities to explore and construct meaning at their own pace. Often this manifested itself as teaching that was low in expectation and high in sentiment.

However, clear guidance and explicit instruction is not just helpful for students with learning difficulties, but arguably more helpful. A new computational study from the University of Haifa provides compelling evidence that students with ADHD and dyslexia face specific cognitive challenges that make explicit guidance not just helpful, but essential. The findings reveal something profound about the cognitive processes underlying different learning profiles, with significant implications for how we design instruction.

The Complexity Gap

The researchers presented participants with a sophisticated learning task that mirrors the complexity of real-world environments. Rather than the simplified laboratory conditions that dominate educational research, participants had to discover which features of multidimensional stimuli predicted rewards. Much like a student trying to discern the underlying patterns in a complex mathematical concept or the structural elements that make a persuasive argument effective.

Crucially, the study compared performance under two conditions: explicit-assist tasks where participants were told which dimension to focus on (analogous to direct instruction), and no-explicit-assist tasks where they had to discover the relevant patterns themselves (analogous to discovery-based learning).

The results were striking. While both students with ADHD and dyslexia performed above chance levels in both conditions, their learning profiles diverged in telling ways from their neurotypical peers. More importantly, computational modelling revealed that similar surface-level performance masked fundamentally different underlying processes.

The Discovery Tax

The researchers designed an elegant experiment that captures something crucial about real learning environments like classrooms: complexity. Rather than the sanitised laboratory conditions that dominate educational psychology research (Robert Slavin called these kinds of experiments “doomed to succeed”), participants faced multidimensional learning tasks where they had to discern which features predicted success, much like students trying to understand what makes an effective argument or identify patterns in mathematical relationships.

The critical comparison was between explicit-assist conditions (where participants received clear guidance about what to focus on) and no-explicit-assist conditions (where they had to discover the relevant patterns themselves through exploration).

The results were clear. Students with ADHD showed significantly worse performance when forced to discover patterns without guidance. They could learn when given explicit direction about what mattered, but struggled markedly when that scaffolding was removed. In computational terms, they were paying what we might call a "discovery tax": an additional cognitive burden that their neurotypical peers didn't face.

Students with dyslexia revealed a different but equally compelling pattern. They struggled in both conditions compared to controls, but the unguided discovery condition exposed their core vulnerability: rapid forgetting of recently learned associations. When cognitive resources were diverted to pattern discovery, their ability to consolidate learning deteriorated further

Intriguingly, this pattern emerged most clearly in the complex, no-explicit-assist condition, suggesting that cognitive load may exacerbate the decay effect. When explicit instruction reduced the processing demands, the difference was less pronounced.

This finding resonates with broader research on dyslexia showing difficulties with consolidating statistical patterns and maintaining learned associations over time. It's not that students with dyslexia can't learn; rather, their learning may require more frequent reinforcement and review to maintain acquired knowledge.

The Strategy Selection Problem

The computational modelling reveals something profound about why unguided discovery fails these students. Individuals with ADHD showed significantly reduced use of optimal Bayesian inference strategies, in other words, the sophisticated cognitive processes that allow learners to systematically integrate prior knowledge with new evidence.

Instead, they relied more heavily on simpler reinforcement learning mechanisms. This isn't a deficit in intelligence; it's a difference in cognitive architecture. But it means that when learning environments demand complex strategy selection (as discovery-based approaches inevitably do) these students are operating with fundamentally inadequate tools.

Think of it this way: students who are taught using explicit instruction are given a clear picture of the thing they are learning, and end up using a catalogued system to locate and connect new knowledge. ADHD learners, without clear guidance and using self-discovery, are sent on a scavenger hunt with no map, expending cognitive effort just figuring out where to look, rather than actually learning.

The Forgetting Accelerator

For students with dyslexia, there are some marked differences. For them, unguided discovery creates what the research reveals as heightened "decay rates", essentially, an acceleration of forgetting, and not the good kind of forgetting. New learning interferes more rapidly with previous learning, creating a cognitive environment where knowledge struggles to consolidate.

This finding illuminates why discovery approaches often fail these students so dramatically. Not only are they being asked to identify relevant patterns without guidance, but the cognitive effort required to do so actively undermines their ability to retain what they've learned. It's a double penalty: harder to learn initially, and harder to remember afterwards.

The explicit instruction condition didn't eliminate this challenge, but it significantly reduced the cognitive load, allowing these students to direct their limited processing resources toward consolidation rather than pattern detection.

The Equity Imperative

These findings have profound implications for educational equity. When we insist on discovery-based approaches for students with learning difficulties, we're not providing "rich authentic learning experiences", rather we're imposing additional cognitive burdens that these students' neurological profiles make unnecessarily difficult.

This isn't about lowering expectations or "dumbing down" instruction. Students with ADHD and dyslexia in this study learned successfully when provided with appropriate scaffolding. The issue is that unguided discovery asks them to simultaneously master content while managing cognitive processes that their peers handle more automatically.

This to me is why explicit instruction, scaffolding and adaptive/responsive teaching is such a powerful set of levers to help ALL students learn. When we provide clear guidance about what to attend to, systematic modeling of thinking processes, and structured practice with feedback, we level the cognitive playing field. We're not doing the thinking for students. We're removing the barriers that prevent them from thinking effectively.

As the researchers note, this has profound equity implications. Students from disadvantaged backgrounds, who often lack the cultural capital to navigate discovery-based approaches, may benefit most from explicit instruction. The same appears true for students with neurodevelopmental differences, though for different computational reasons.

Limitations

One limitation of the study is that the participants were adults in their twenties, not school-aged children. However for me, this doesn't diminish its value. If anything, it underscores how enduring these cognitive challenges can be. The fact that differences in learning strategies and memory consolidation persist into adulthood highlights the importance of designing effective instructional support early, before these difficulties compound over time.

Additionally, the study’s relatively small sample sizes (around 20 participants per clinical group) limit its statistical power. This raises the possibility of Type II errors and means that some non-significant findings (e.g., task-by-group interactions) may reflect underpowered tests rather than true null effects. Larger samples would be needed to more confidently detect subtle group differences or interactions.

Questions for Practice

However this research raises important questions for educators: How might we design instruction that provides explicit guidance while still fostering independence? How can we build in the kind of systematic review and reinforcement that benefits students with rapid decay rates? How do we balance cognitive load considerations with the need for appropriately challenging tasks?

As we continue to refine our understanding of learning differences, one thing becomes clear: effective instruction isn't about choosing between explicit and implicit approaches. It's about understanding the cognitive demands of different learning environments and matching instructional support to student needs.