Imagine if a friend believed cake was more nutritious for children than vegetables, and any time you tried to explain otherwise they closed down the discussion saying the two were a false dichotomy: children should just eat both.

This is what happens when the overwhelming evidence in favour of teacher-directed, rather than student-oriented learning achieves air-time in New Zealand.

For most of the last century, generations of New Zealand students benefited from classrooms where teachers mostly led. Nowadays, teachers are encouraged towards being facilitators; "guides on the side" not "sages on the stage".

Trivially, of course, children do benefit from both approaches. It is not as though Kiwi children have learned nothing since we embraced child-centred learning more than a decade ago. There are calories (and comfort) in cake.

However, what matters is whether children would learn more if their staple classroom diet was led by teachers, not themselves. There are finite hours in the school day. If we use them sub-optimally, children incur the opportunity cost.

And that cost is significant, because despite the superficial appeal and widespread adoption of child-centred learning, the overwhelming weight of evidence finds that for school-age students it generates less achievement than learning led by a teacher.

Intuitively, this also makes sense: Children do not know what they do not know, and we employ teachers to teach them.

The OECD's education director Andreas Schleicher drew attention to the superiority of teacher-led instruction in the lead-up to their recent PISA data-release. In his words "teacher-directed instructional practices tend to better predict student achievement than student-oriented learning". He then went on to systematically slay all the common objections.

There is undoubtedly a place for child-centred learning. After all, almost all doctoral-level study is student-led. However, relative independence works for postgraduates because experts learn differently from novices. Doctoral students are relative experts: Most school students are not. Instead, school students benefit from much more scaffolding, modelling and practice, all provided by a teacher.

Similarly, at the other end of the continuum in early childhood, babies and toddlers endlessly amaze us with what they learn naturally – everything from how to read facial expressions, to how to walk, talk and solve spatial problems. However, many of the things we want children to learn in school are not natural. While walking and communicating have been features of human evolution for millions of years, alphabetic writing and the number system have not. There is little wonder they require teaching explicitly.

On average, school students are also less motivated than doctoral students. Child-centred learning is often proffered as the solution to problems of motivation because it permits students to pursue their own interests. In certain circumstances, it may work. For example, when a student has already gained enough knowledge to begin learning independently, perhaps through an end of term project. However, to suggest or assume that child-centred learning is necessarily more motivating than teacher-led learning is madness.

There can be few things more motivating than learning from an expert who, through their carefully planned explicit teaching, enables a whole class to feel successful. There can also be few things more demotivating than to be a novice working unaided on a task.

And yet, nowadays, whole class, explicit teaching is widely derided in favour of learning centred-on or led-by the student.

Another example of where "just do both" frequently has the effect of closing-down discussion is in the debate over whether schools should teach knowledge or competencies.

The New Zealand curriculum communicates that knowledge, and the disciplines that organise it, are largely incidental: schools should focus on competencies.

However, competencies such as thinking, deciphering fact from fiction and problem solving are not generic. Though crucially important, these competencies cannot be taught, acquired or applied in the abstract. You need knowledge to think critically.

Even in an area as seemingly skills-based as sport, the key to expertise is knowledge. Take a cricket batsman as an example. A batsman's skill derives from the hours and hours of deliberate, broken-down practice which have rendered knowledge - of footwork, timing, shoulder position and arm swing - automatic in long term memory.

Thus, knowledge and skills are not a dichotomy. They are also not a false dichotomy that can be swept away with a "just do both". Instead, knowledge stored in long-term memory is the route to skilled performance. And the distinction between a dichotomy and a route matters, because hours spent practising skills might well be better spent learning knowledge.

So next time someone tries to close-down a discussion with a deft "just do both", check that they understand opportunity costs and the relative merits of veges and cake. Do not let enticing language about compromise shield preferences from evidence and debate.