Direct Instruction, Cognitive Load, Rosenshine, Desirable Difficulties and Responsive Teaching (and thanks to @Tom_Needham_)

I loved this blog by @Tom_Needham_ and I was really taken by his linking of Direct Instruction, Cognitive Load Theory and Rosenshine’s principles of instructions (link to @teacherhead). They fitted together so well but I wondered where desirable difficulties, in particular, fitted in here. Having recently finished @HFletcherWood’s excellent responsive teaching book I wondered where this might fit in too. Tom gave me permission to use his original document and I came up with this:

Cognitive Load Theory and Desirable Difficulties do seem to be in conflict with each other. CLT is about ensuring that working memory is not overloaded when learning (keeping things as simple and “clean” as possible) and desirable difficulties is a deliberate introduction of strategies that make learning harder. The kicker here is that introducing desirable difficulties (such as spacing, interleaving and retrieval practice) should increase the long term retention and transfer of learning.

For more on CLT, try this by @Olicav or this by @mrbartonmaths.

For more on desirable difficulties, try this by Robert Bjork or this by @greg_ashman.

I think desirable difficulties and cognitive load theory can be reconciled. Desirable difficulties should only be introduced at a late(r) stage into an instructional sequence (stage 4 above). At that late stage, implementing desirable difficulties should ensure that this knowledge will be available, with good storage and retrieval strength, in Long Term Memory which helps minimise cognitive load during future topics that link with the current topic. This also helps pupils build organised schema. See this by @DavidDidau for more on schema.

To give a concrete example:

In September, I taught my Year 11 pupils about flame tests and precipitate reactions for metal ions and non metal ions. After reading Responsive Teaching, particularly the chapter on unit planing, I became a lot more aware of how this links with a future topic – ionic and covalent bonding (to be covered in Feb/March). Because of this increased awareness , in September I explicitly taught the electron configuration of the metal ions as well as some non metal ions. At the end of that unit I would expect pupils to:

• be able to draw the electron structure of the atoms in groups 1, 2, 6 and 7 (from the first few periods)
• know the link between number of electrons in outer shell and group number
• know the difference between an atom and an ion
• know how many electrons are lost/gained when atoms in group 1/2/6/7 form ions
• “work out” the chemical formulae for some ionic compounds using the common ions table (and how to use bar modelling to check the plausibility of their answers –more here by @chipps_sci)

From this they would have a basic understanding of ionic bonding which involves transfer of electrons and formation of positive and negative ions.

The key here is understanding why I have covered more than I have needed to for this particular topic. Using the idea of Horizon Knowledge (see this by @HFletcherWood) I wanted to commit the above knowledge to Long Term Memory to minimise intrinsic load and aid schema building when I then teach Ionic and Covalent bonding later in the year.

Once I have finished the topic in September I used my optimum spacing graph to work out when to revisit the topic (more here):

The spaced retrieval (followed by further spaced appearances via the use of @adamboxer1’s retrieval roulette) are the desirable difficulties that will help commit the knowledge to LTM. These desirable difficulties should mean that by the time Feb/March comes, this knowledge should be “fingertip” knowledge, having both a high storage and high retrieval strength.

Feb/March- teaching ionic and covalent bonding. To start I will have a few multiple choice questions ensuring that pupils do have the background knowledge at their fingertips. These are, in effect, hinge questions because if some pupils are not secure then I can’t go any further until they are. However, I would be confident that pupils would be clear how (for example) sodium and chloride ions are formed. Then, and this really follows the 1, 2 and 3 model in @Tom_Needham_’s document, I would model how ionic bonding works using dot and cross diagrams. Then all the pupils would have a go at a common problem. This would be followed by a boundary example (an example which would be as challenging as it could get- idea from @mrbartonmaths). This would look like this:

This is the I and the We element. Before handing over to pupils to practise ionic dots and crosses diagrams, I would ask a hinge question to ensure they are ready for independent practice.

Then on to the You element:

Nothing earth shattering in this. However, one of the key concepts in this topic is to understand why ionic compounds have such high melting points and why they don’t conduct electricity when they are solid. This blog by @ChemDrK reminded me that pupils won’t necessarily “get this” because of the way we teach dot and cross diagrams. However, by this stage I would expect pupils to have a well formed schema of ionic bonding and this can keep intrinsic cognitive load down when explaining the formation of the ionic lattice. Having used desirable difficulties to build up the retrieval (and storage) strength for the background knowledge that they need, I will be confident that intrinsic cognitive load can be managed. Then they can think about the 3D structure of the ionic lattice rather than “what is an ion?”, “where have the electrons gone?”, “why is that one negatively charged and that one is positively charged?” etc. Instead they can think about what I need them to think about.

Covalent bonding would be covered in a similar I, We, You pattern.

Before moving on to stage 4, I would give pupils an exit ticket just to check they can do ionic and covalent bonding dot and cross diagrams.

This will make me more confident that the pupils are ready for the interleaved examples. If any pupils show errors in these tickets then next lesson I can work with those while the rest of the class attempts the interleaved examples. So next lesson pupils would have to draw dot and cross diagrams for a mixture of ionic and covalent structures. These are similar to @mrbartonmaths’s Same Surface Different Deep as the pupils must first decide if it is ionic or covalent before carrying out the correct procedure and drawing the correct dot and cross diagram.

This is cognitively more demanding. A desirable difficulty. But I am banking on the pupils having the correct knowledge easily retrievable from their long term memories. This will help minimise intrinsic cognitive load. Some pupils may need to still rely on their notes but as this blog by @adamboxer1 shows, that is another useful way of managing cognitive load.

After finishing this topic there is then a measured spacing gap before we revisit ionic and covalent bonding for a spaced retrieval activity. The desirable difficulties here should mean that pupils are ready to show their knowledge in the GCSE exam. And beyond.

I am starting to get more comfortable with the link between desirable difficuties and cognitive load, particularly using desirable difficulties in an earlier topic to reduce intrinsic load in a future topic. It is like when you are watching a 6 part television series (say Line of Duty) and something seemingly inconsequential is shared in episode 2. However, as the plot develops in episode 6, that earlier scene is revealed to have been integral to the plot and has enhanced the audience’s understanding of it.

Feedback, as always, is welcome.

Advertisement