Promoting inquiry learning in the sciences is deeply ironic for this reason. The scientific method requires us to adapt our beliefs to the evidence.
For some, scare quotes are obligatory, but the evidence for explicit teaching is clear: decades of converging evidence, including randomised controlled trials, ecologically valid observational studies, meta-analyses, and large-scale assessment point in a particular direction and give teachers a clear directive: tell students first.
I see dead horses being flogged frequently in research into inquiry learning: small samples in articles with no limitations section to speak of; an unwillingness to compare inquiry to alternative instructional approaches; and poorly defined notions of “traditional teaching”.
A frequent explanation is that inquiry fails when it is poorly resourced – so where is this resourced utopian school system to be found?
When students report experiencing anxiety due to the chaotic learning environments engendered, they are blamed for being overly focused on grades.
Teachers rescuing students from the chaos are simply “bolstering” learning, not teaching. Students floundering are reframed as “help seeking”.
Rather than engaging seriously with learning outcomes, these are often sidestepped altogether.
Expertise Generalisation Syndrome, coined by Paul Kirschner, describes the phenomenon of individuals esteemed in a particular area trying to extend their expertise, Dunning-Kruger style, into other domains.
Quite commonly, this happens in education, where everyone who has been to school feels they have insight into what education should be and how students should be taught.
Personally, I feel the pain of their comments most acutely when the commenter in question comes from the tech world.
They tend to view children as mini-adults who should be exposed to “authentic” problem-solving tasks, like those encountered in the workplace; who should possess an innovator’s mindset; and who should approach their schoolwork with a cavalier attitude to failure.
Now, here’s the bit where you might ask, 'is Rebecca OK?'
I may have, at least partially and for now, changed my mind about whether explicit teaching should always come before problem-solving.
The science of learning requires me to consider new evidence. My recent investigations are a good case study in how the field addresses gaps, how the trail of breadcrumbs is followed in the research, and how we stand on the shoulders of giants, to mix a few metaphors.
For years, I have promoted the idea that gifted students need to be taught explicitly when learning new content.
Not being a gifted education expert, I have based this view largely on a literature review by CESE, which recommends explicit instruction of new content, even for gifted learners.
In relation to gifted learners, the expertise reversal effect is familiar, if a bit impractical, because it suggests that instruction should move towards problem-solving as students develop mastery.
Continuing with worked examples once schemas are well developed can introduce unnecessary cognitive load.
Given that this point of readiness will occur at different times in a class of 25 students, it is hard to know how teachers are meant to split themselves to give differentiated instruction.
It may be more feasible in subjects such as mathematics, where learning naturally progresses towards problem-solving, than in skills like writing, where the knowledge base is cumulative and less discretely staged.
In mathematics, teachers can sometimes move a subset of students on to problem-solving while continuing explicit teaching with the rest of the class.
In addition, relatively new research reinforces that prior knowledge has a substantial impact on cognitive load, something we already knew.
The science of learning requires us to update our understanding as new evidence emerges.
The complexity of a task is relative to a learner’s expertise, meaning element interactivity may be low for some students and high for others (that is, the number of ideas that must be held and coordinated in working memory at once).
Greg Ashman’s doctoral work examined this issue and concluded that for most learners, when learning new material, element interactivity will be high.
Under these conditions, beginning with explicit instruction rather than problem-solving is the most effective approach.
But taking all this into account, is it possible that if gifted learners are more likely to possess higher levels of prior knowledge, they might benefit from engaging in some form of problem-solving activity before explicit instruction?
Researchers have attempted to answer this question. The theoretical rationale behind a problem-solving-first approach is that by attempting to explore a problem initially, students may become more attuned to gaps in their understanding, which can then be addressed through subsequent explicit instruction (perhaps rather than experiencing overload by attending to all aspects of the instruction at once).
Some explanations proposed in the literature suggest that this approach may be more motivating for gifted learners, potentially stimulating curiosity.
Others propose that learners with high prior knowledge may be better able to recognise the deep structure of problems.
Lim, Jung, and Kalyuga (2023) found that a problem-first approach was more effective for gifted learners in their study, although not conclusively for any of the reasons proposed above.
Indeed, the precise mechanism remains unclear. Nevertheless, it may be that this sequencing is beneficial for some gifted students under certain conditions.
This still leaves us with several problems. Outside experimental conditions, teachers may struggle to know how long to let students “struggle” before it becomes truly unproductive.
Experiments are conducted under idealised conditions – time-boxed, constrained, well-defined. Small inefficiencies in instruction, when repeated daily, add up.
In contrast, explicit instruction like that observed by Rosenshine is notable for its efficiency. OECD analyses show that Australian students spend an above-average amount of instructional time in school, but this additional time doesn’t seem to result in better learning outcomes.
These ideas may also fail to translate well to the humanities, which are more concerned with long-term synthesis than with problem-solving.
These subjects are interpretive, cumulative, and in many respects unstable: the goal is to help students see the world differently.
Problem-solving first in these contexts tends to elicit guesses and vague opinions. Expecting students to make profound initial insights about the deep structure of the world is unrealistic, with the potential to fritter away valuable learning time that could otherwise be spent building knowledge foundations.
Thirdly, we face persistent challenges in identifying giftedness. Under-identification is common in schools. Many high-potential but underperforming learners may not yet possess the knowledge required for problem-solving approaches to be effective or motivating.
At the same time, over-identification remains an issue, and even streamed or selective classes are often highly variable in ability.
Teachers cannot reasonably be expected to conduct individual cognitive load diagnostics in real time.
So where does this leave us? The science of learning requires us to update our understanding as new evidence emerges.
Differentiation has long been a burden for teachers, yet we know gifted learners can experience negative academic and life outcomes when their needs are not met.
I don’t have all the answers, but streaming – only with genuine instructional differentiation for that group of students – may be one viable option, provided lower-performing students are not sacrificed at the altar of giftedness by being relegated to a single bottom class.
In mixed-ability settings, an approach that remains heavily weighted towards explicit instruction (Pareto’s 80/20 principle still holds up), with occasional, tightly constrained problem-solving-first episodes, may be workable, but only with strict limits on the duration of struggle and ideally supported through coaching and feedback.
I’m very cautious about advice that teachers change their approach to suit a subset of students, when we know explicit instruction first works for most learners.
Given the inefficiencies and potential for chaos that problem-solving can produce, I’m hesitant to flip all of my views based on this one study.
However, the challenge of gifted education is not going away, and it’s an area of national significance. Australia is not exactly the clever country, and it increasingly feels we are not the exporters of knowledge we used to be.
So, if the research presents me with a relatively low cost of meeting the needs of gifted learners, then I’m willing to listen.
This article was first published on the author's Substack. Read the original post here.
Further reading
Ashman, G., Kalyuga, S., & Sweller, J. (2020). Problem-solving or Explicit Instruction: Which Should Go First When Element Interactivity Is High? Educational Psychology Review, 32(1), 229–247. https://doi.org/10.1007/s10648-019-09500-5
Lim, S., Jung, J. Y., & Kalyuga, S. (2023). Effectiveness of invention tasks and explicit instruction in preparing intellectually gifted adolescents for learning. Instructional Science, 51(6), 921–952. https://doi.org/10.1007/s11251-023-09616-w
Zhang, L., & Sweller, J. (2024). Instructional sequences in science teaching: Considering element interactivity when sequencing inquiry-based investigation activities and explicit instruction. European Journal of Psychology of Education, 39(4), 3791–3801. https://doi.org/10.1007/s10212-024-00799-5ac
