On adult learning theory

Why I think it's garbage. This one is personal.

Photo by Dark Rider on Unsplash

I’ve been a bit quiet on here of late. This is because I have been in the quagmire of discovery learning. I’ve been knee-deep in my Masters thesis¹, which started as a school-based study to see if we could improve student indicators of wellbeing by improving their study skills. We thought some descriptive statistics about shifts in mean measures could be cool. Job done. Well, that was early days and it turns out there were correlations. And then it turns out those correlations were strong enough to warrant regression analysis. And then it turns out I would be learning structural equation modelling and how to use jamovi. This is where I have been for the past few months. If you’re in cognitive overload right now, good. It sets the scene for what I’m about to tell you.

The continual ‘upsizing’ of this project has meant that I didn’t have the luxury of laying down some foundational learning. My knowledge of statistics was fairly basic — I thought it was ok before I started down this path — classic Dunning Kruger. I have been engaged in full-blown discovery learning and I can tell you, at times, it’s the pits. There have been tears. Before I go on, I want to emphasise that my supervisor is the best. Together, we are pretty good at making sure I get just-in-time instruction, but we have both talked about the issues with my strange pathway: it’s almost the opposite of “p hacking”, where the results have offered up better and better methodologies, so the learning has been steep and sometimes circuitous.

At the same time as this, I’ve been doing some research into adult learning. I started here, with Linda Darling-Hammond’s review, which has clocked up almost 5,000 citations. I’m pretty sure that’s what NESA (the NSW teacher regulatory body) base their professional learning course accreditation requirements on, loosely at least. NESA has condensed these into requirements for courses that can be a minimum of 1.5 hours in duration. The key takeaways from Darling-Hammond’s review is that PL should:

1. Be content focused

2. Incorporate active learning utilizing adult learning theory

3. Support collaboration, typically in job-embedded contexts

4. Use models and modelling of effective practice

5. Provide coaching and expert support

6. Offer opportunities for feedback and reflection

7. Be of sustained duration

Firstly, it’s difficult to see how a course of 90 minutes could be ‘sustained,’ but this is more of a New South Wales problem. However, it does relate to my second point, namely that the studies cited all seem to feature inordinate investments of time (which is money), and it seems to me that schools could really only focus on that one whole-school project for the year. It makes sense that these case studies were hugely successful, with the amount of resources and energy allocated. But I wonder if this advice is helpful for schools that may not be able to make this kind of investment into their professional learning programs. It seems to amount to ‘pick something, throw everything at it, hope it’s the right thing.’

I hadn’t really heard of adult learning theory (Feature 2), being of the belief that learning is just learning. Adult learning theory posits that adults learn differently and that instructors need to act as facilitators rather than “supplying them with facts.” Adult learners must co-construct knowledge and need to see the relevance of something in the real world. Sound familiar? It’s basically constructivism for grown-ups. Some aspects glom well onto evidence-based practice, for example, seeing the relevance of a lesson is just good practice and has a basis in schema theory. However, these elements of sound practice do little good when learning something for the first time through facilitation. It’s true — adults don’t know everything. As far as I can tell so far, adult learning theory is largely vibes-based. Don’t let the high citation counts fool you.

I’ve also been reading Sam Sims’ research into effective professional learning. He points out that the causal evidence for each potential feature of effective PL is weak — in other words, we don’t know which factor has the impact. This spurred a random side-thought: what if the effect was interactive, for example what if the combination of the social factors of adult learning were the keys to inspiring and sustaining changes to practice? What if effective professional learning had more in common with the success of Alcoholics Anonymous, with its sense of mission and high accountability than with any of the individual factors that have been researched? Sims draws upon coaching as a sustainable and effective model, but professional learning is about more than just instruction; coaching doesn’t go far enough to lay the knowledge foundations that will produce sustained change and avoid lethal mutations².

Given my recent experience, I’m interested in answering a very basic question: how do adults learn things? My answer may not surprise you — it’s the same as everybody else. Now I’m going to get to what I thought was going to be the guts of this piece (I got sidetracked), which is where I review three modes of learning and explain how they did or didn’t meet my needs as an adult learner.

1. Textbook — Discovering Statistics Using IBM SPSS Statistics by Andy Field.

Ok it doesn’t help that I’m not using SPSS, but this is a fairly useful textbook if you have some background knowledge, which I didn’t. Pros: logically sequenced, self-referential i.e. refers to previous sections so students can fill gaps in knowledge. Cons: pictures of cats, cute stories about guitars. The language can be a bit digressive and I really hoped that the statistical information was going to be as easy to understand as the guitar story. The cute illustrations just made me hate Andy more when I hit cognitive overload. Verdict: a necessary evil. Start with a basic book on research methods first.

2. Video demo of regression using least squares method by StatisticsFun.

This sarcastically named video was the best example of dual coding I have ever seen. It helped me fundamentally understand what regression actually is. No Cons. Pros: the voice doesn’t conflict with the video, with no distracting side comments by the instructor. It’s SUPER slow, which is important for me — I didn’t reach overload before hitting pause to re-watch. For me, there was a bit of an Easter egg at the end — the last screen says ‘Party More, Study Less,’ and the comments section is full of glowing exclamations that the viewers learned more from this video than from a three-hour lecture. So yeah, I suppose you could party more if you were that way inclined.

3. ChatGPT-4

I have the paid version of ChatGPT and it does some ok analysis (not for SEM so much, but other things). Pros: It’s brilliant for explaining the key differences between related terms at the novice stage or refreshing my memory about definitions, or explaining terms I haven’t seen before in research papers. It’s good for sense-checking, helping me phrase my novice-understandings in more conventional ways, and software and syntax help with jamovi. Importantly, it’s open all the time and ready for my query. Cons: I think like Kirschner’s urban myths, it’s destroyed my long-term memory. Knowing it’s open and waiting for me means that I don’t make a conscious effort to commit understandings to long-term memory. My learning has been quite superficial as a result.

Adult learning is just as frustrating as kid-learning. I know the nature of postgrad study is ultimately independent learning, but I think the implications, based on my purely anecdotal experience (!), are reasonably clear. Here are my takeaways: First, don’t overestimate what adults know; second, tell them and show them, make them do it and give feedback. My third takeaway is the AA factor and it’s admittedly vibes-based: make the mission visible and take advantage of social reinforcement of good practice.

Because I haven’t written in a while, this post is a bit of a dump of my thinking on this whole process. I suppose the purpose of it is to tell you, dear reader, that if you’re also sitting at your screen in hot tears, staring at a spinning wheel on the jamovi interface, you’re not alone.

1. Just two terms, only 12K words, not the MRes

2. See the ADKAR model for more info on change management — I’m talking here about the A, D and K.