Ends are not means

Embed from Getty Images

At the recent researchED Melbourne events, both John Sweller and I talked about biologically primary versus biologically secondary information. This distinction, credited to David Geary, has been taken up by Sweller as a basis for cognitive load theory.

Briefly, we have evolved to acquire biologically primary knowledge and abilities and so we pick them up seemingly effortlessly. They include the ability to speak our mother tongue and certain types of folk knowledge about human relationships and the physical world. Biologically secondary information has been created by human cultures in relatively recent times and so evolution has not had time to produce mental modules to acquire this information. Instead, biologically secondary information co-opts systems designed for primary information. And the distinction between the two is often fuzzy. Mathematics, although largely secondary, incorporates primary number sense.

Sweller stresses that generic problem solving is primary and reduces down to one main strategy: means-ends analysis. This is the process by which we compare our current state to a desired state in order to work out what steps to take next. It is pointless to try to teach this strategy to students because they already possess it.

Schools should therefore be in the business of teaching domain-specific problem solving strategies, such as how to solve a particular class of algebra problem. When schools think they are teaching generic problem solving skills they are, in reality, teaching domain-specific heuristics. For instance, the prompts to, ‘Write down everything you know,’ and, ‘Draw a diagram,’ have moderate value in solving physics problems but are only likely to hamper attempts to rebuild a relationship after an affair.

I also think that means-ends analysis is something we engage in across the education system. And often this does not lead to improvements. We have developed domain-specific solutions to problems because means-ends analysis is inefficient and not always effective, particularly when we try to jump too quickly to the ends.

I was reminded of this when chatting to Tom Bennett at researchED. He made the observation that if you enter a classroom with immaculate behaviour you will probably notice the good quality relationships between the teacher and students. However, this does not necessarily imply that teachers should focus on constructing good quality relationships. The pathway towards immaculate behaviour may be more about routines, boundaries and quality teaching, with the good relationships emerging as a result.

I remember a time in the U.K. when policymakers observed that schools with good behaviour had fewer exclusions and decided it would therefore improve behaviour if schools were required to reduce exclusions.

This is where research and experts should help. Instead of simply trying to emulate ends we should, through systematic study, find the non-obvious pathways towards those ends. Yet one of the ironies of education is that it is often researchers and experts who shove the coconuts on our heads and tell us the planes will come and it is the folk knowledge of practising teachers that leads them to be sceptical about such claims.

Fortunately, there are some experts who recognise the problem. Paul Kirschner has worked hard to point out it is a fallacy to assume that asking children to behave like scientists is necessarily the best way to teach them science. Yet inquiry learning, rooted as it is in the idea of students copying the behaviour of experts, seems ever more popular.

Similarly, the teaching of writing often seems to focus on complex products rather than the building blocks that the products are constructed from. ‘Writers write!’ Is a facile observation based on the copy-the-experts principle. Fortunately, alternative approaches are starting to become popular.

Means-ends analysis is what you do when you don’t know what to do. It’s time for education to start applying some domain specific knowledge.


4 thoughts on “Ends are not means

  1. Pingback: De positie van het Platform O2032 over kennis | onderwijs_2032 science check

  2. Pingback: Spaced practice, retrieval practice and dual coding: some practical applications in the teaching of history | History Lover

  3. Pingback: Evolved minds and education: evolved minds | logicalincrementalism

  4. Reblogged this on No Signaling Allowed and commented:
    Very thought-provoking blog post by Greg Ashman. Here he mentions John Sweller’s contention that generic problem solving methods boil down to means-ends analysis which is highly inefficient compared to domain-specific problem solving methods. Greg’s insight here is that good results fall out not necessarily out of trying to reach the good results i.e trying to directly reduce the difference between the current state and the goal state, but rather by applying multiple domain-specific methods which may not bear any resemblance to state difference reduction. It may be that good results are a quality emerging from achieving subgoals that may not bear any direct obvious relationship to the end goal state.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.