Friday, 23 March 2012

Whatever happened to Learning Theory?


I’ve already blogged about B.F.Skinner, and –coincidentally – he has just celebrated his 108th birthday. But it led me to think about how learning theory in general seems to have drifted slowly out of our undergraduate psychology curricula, out of our animal and experimental psychology labs, and out of the list of high impact journals. I don’t mean just ‘behaviourism’, I mean learning theory and all that embraces - from schedules of reinforcement and behaviour analysis, to associative learning and cognitive inferential models of conditioning – in both animals and humans.

In 2010, the BPS Curriculum for the Graduate Basis for Chartered Membership of the Society listed ‘learning’ as a topic under Cognitive Psychology (that would have jarred with Prof. Skinner!), and not under Biological Psychology. Interestingly, 10 years ago it was listed under both cognitive and biological psychology. In my own institution I know that learning theory has become a relatively minor aspect of Level 1 and Level 2 teaching. Until 2 years ago, I offered a final year elective called ‘Applications of Learning Theory’, but despite its applied, impact-related title the course usually recruited less than 10 students. I usually had to begin the first two lectures by covering the basics of associative learning. If these students had been taught anything about learning theory in Years 1 and 2 they had retained none of it. This state of affairs is quite depressing in an institution that twenty five years ago had one of the leading animal learning labs in the world, inhabited by researchers such as Nick Mackintosh, Tony Dickinson, John Pearce, and Bob Boakes, to name but a few.

I haven’t done anything like a systematic survey of what different Psychology Departments teach in their undergraduate courses, but I suspect that learning theory no longer commands anything more than a couple of basic lectures at Level 1 or Level 2 in many departments. To be fair, most contemporary Introduction to Psychology texts usually contain a chapter devoted to learning (e.g. 1,2), but this is usually descriptive and confined to the difference between instrumental and classical conditioning, coverage of schedules of reinforcement (if you’re lucky), and a sizable focus on why learning theory has applied importance.

So why the apparent decline in the pedagogic importance of learning theory? I suspect the reasons are multiple. Most obviously, learning theory got overtaken by cognitive psychology in the 1980s and 1990s. There is an irony to this in the sense that during the 1980s, the study of associative learning had begun to develop some of the most innovative inferential methods to study what were effectively ‘cognitive’ aspects of animal learning (3, 4) and had also given rise to influential computational models of associative learning such as the Rescorla-Wagner and Pearce-Hall models (5,6). These techniques gave us access to what was actually being learnt by animals in simple (and sometimes complex) learning tasks, and began to provide a map of the cognitive mechanisms that underlay associative learning. This should have provided a solid basis from which animal learning theory could have developed into more universal models of animal consciousness and experience – but unfortunately this doesn’t appear to have happened on the scale that we might have expected. I’m still not sure why this didn’t happen, because at the time this was my vision for the future of animal learning, and one I imparted enthusiastically to my students. I think that the study of associative learning got rather bogged down in struggles over the minutiae of learning mechanisms, and as a result lost a lot of its charisma and appeal for the unattached cognitive researcher and the inquisitive undergraduate student. It certainly lost much of its significance for applied psychologists, which was one of the attractions of the radical behaviourist approach to animal learning.

A second factor in the decline of learning theory was almost certainly the decline in the number of animal labs in psychology departments – brought about in the 1980s and 1990s primarily by a vocal and active animal lib movement. This was certainly one factor that persuaded me to move from doing animal learning studies to human learning studies. I remember getting back into work one Monday morning to find leaflets pushed through the front door of the Psychology building by animal lib activists. These leaflets highlighted the cruel research carried out by Dr. Davey in Psychology who tortured rats by putting paper clips on their tails (7). At the time this was a standard technique used to generate stress in rats to investigate the effects of stress on feeding and drinking, but it did lead me to think hard about whether this research was important and whether there were other forms of research I should be moving towards. It was campaigns like this that led many Universities to either centralize their animal experiment facilities or to abandon them altogether. Either way, it made animal research more difficult to conduct and certainly more difficult for the interested undergraduate and postgraduate student to access.

In my own case, allied to the growing practical difficulties associated with doing animal learning research was the growing intellectual solitude of sharing a research topic with an ever decreasing number of researchers. In the 1980s I was researching performance models of Pavlovian conditioning – basically trying to define the mechanisms by which Pavlovian associations get translated into behaviour – particularly in unrestrained animals. Eventually it became clear to me that only me and maybe two or three other people worldwide shared this passion. Neither was it going to set the world on fire (a bit like my doctoral research on the determinants of the fixed-interval post-reinforcement pause in rats!). To cut a long story short, I decided to abandon animal research and invest my knowledge of learning theory into more applied areas that held a genuine interest for the lay person. Perhaps surprisingly it was Hans Eysenck  who encouraged me to apply my knowledge of learning theory to psychopathology. During the 1980s, conditioning theory was getting a particularly bad press in the clinical psychology literature, and after chairing an invited keynote by Hans at a BPS London Conference he insisted I use my knowledge of conditioning to demonstrate that experimental approaches to psychopathology still had some legs (but only after he’d told me how brilliant his latest book was). This did lead to a couple of papers in which I applied my knowledge of inferential animal learning techniques to conditioning models of anxiety disorders (8,9). But for me, these were the first steps away from learning theory and into a whole new world of research which extended beyond one other researcher in Indiana, and some futile attempts to attach paper clips to the tails of hamsters (have you ever tried doing that? If not – don’t!)(7).

I was recently pleasantly surprised to discover that both the Journal of the Experimental Analysis of Behavior and the Journal of Applied Behavior Analysis are still going strong as bastions of behaviour analysis research. Sadly, Animal Learning & Behavior has now become Learning & Behavior, and Quarterly Journal of Experimental Psychology B (the comparative half traditionally devoted largely to animal learning) has been subsumed into a single cognitive psychology QJEP. But I was very pleasantly surprised to find that when I put ‘Experimental Analysis of Behaviour Group’ into Google that the group was still alive and kicking (http://eabg.bangor.ac.uk). This group was the conference hub of UK learning theory during the 1970s and 1980s, affectionately known as ‘E-BAG’ and provided a venue for regular table football games between graduate students from Bangor, Oxford, Cambridge, Sussex and Manchester amongst others.

I’ve known for many years that I still have a book in me called ‘Applications of Learning Theory’ – but it will never get written, because there is no longer a market for it. That’s a shame, because learning theory still has a lot to offer. It offers a good grounding in analytical thinking for undergraduate students, it provides a range of imaginative inferential techniques for studying animal cognition, it provides a basic theoretical model for response learning across many areas of psychology, it provides a philosophy of explanation for understanding behaviour, and it provides a technology of behaviour change – not many topics in psychology can claim that range of benefits.

(1)      Davey G C L (2008) Complete Psychology. Hodder HE.
(2)      Hewstone M, Fincham F D & Foster J (2005) Psychology. BPS Blackwell.
(3)      Rescorla R A (1980) Pavlovian second-order conditioning. Hillsdale, NJ: Erlbaum.
(4)      Dickinson A (1980) Contemporary animal learning theory. Cambridge: Cambridge University Press.
(5)      Rescorla R A & Wagner A R (1972) A theory of Pavlovian conditioning: Variations in the effectiveness of reinforcement and nonreinforcement. In A H Black & W F Prokasy (Eds) Classical conditioning II: Current research and theory. New York: Appleton-Century-Crofts.
(6)      Pearce J J & Hall G (980) A model for Pavlovian learning: Variations in the effectiveness of conditioned but not of unconditioned stimuli. Psychological Review, 87, 532-552.
(7)      Meadows P, Phillips J H & Davey G C L (1988) Tail-pinch elicited eating in rats (Rattus Norvegicus) and hamsters (Mesocricetus auratus). Physiology & Behavior, 43, 429-433.
(8)      Davey G C L (1992) Classical conditioning and the acquisition of human fears and phobias: Review and synthesis of the literature. Advances in Behaviour Research & Therapy, 14, 29-66).
(9)      Davey G C L (1989) UCS revaluation and conditioning models of acquired fears. Behaviour Research & Therapy, 27, 521-528.

11 comments:

  1. You are right to bemoan the absence of modules on learning from the core, BPS undergraduate curriculum. Where offered, such modules are usually delivered by an isolated behaviourist, ploughing his/her furrow, and kicking against the mainstream. I should know!

    I think it is important to distinguish between "learning theory", "associative learning" and "radical behaviourism" or "behaviour analysis". The latter are synonymous with Skinner. Many of the individuals listed may not see themselves as out and out Skinnerians!

    Skinner would be proud to know that there has, however, recently been a massive increase in postgraduate courses in applied behaviour analysis. They are mainly clustered around behavioural centres such as Bangor, Ulster, and Cardiff, as well locations in Ireland like Dublin, Galway, and Maynooth. Several countries in Europe now offer similar courses where, 10 years ago, none existed. Professionally approved by the Behavior Analyst Certification Board (BACB: www.bacb.com), these courses have largely risen in response to the growing demand for evidence-based services and intervention for autism and other developmental disorders.

    "E-bag" continues to be a great conference. I gave my first ever talk there way back in 1994. There is now also a European Association for Behaviour Analysis (http://www.europeanaba.org/) and a European Journal of Behavior Analysis, while the "parent organisation", the Association for Behavior Analysis International (http://abainternational.org/) is simply huge - 5000+ behaviorists attend its annual convention!

    For the future, the challenge remains of ensuring that a critical mass of behaviour analysts exists to teach, supervise students and conduct research. But, in just my short time as an academic (I obtained by PhD in 1996), a lot has changed, and for the better.

    ReplyDelete
  2. For other current applications of radical behaviorism and its developments, have a look at i.e.

    www.obmnetwork.com and the Journal of Organizational Behavior Management
    www.contextualpsychology.org and the upcoming Journal of Contextual Behavioral Science

    There's also some writing on Skinner and society that could be of interest, such as
    Dinsmoor, J. A. (1992). Setting the record straight: The social views of BF Skinner. American Psychologist, 47(11), 1454.

    http://dc417.4shared.com/download/kiBXK2rO/Dinsmoor__1992__Setting_the_re.pdf?tsid=20120302-092522-a3e0a9c

    ReplyDelete
  3. Psychology Press and Oxford University Press both seem to think that there is a market for books on applications of learning theory:

    http://www.amazon.co.uk/Clinical-Applications-Learning-Theory-Haselgrove/dp/1848720084/ref=sr_1_cc_1?s=aps&ie=UTF8&qid=1332506611&sr=1-1-catcorr

    http://www.amazon.co.uk/Associative-Learning-Conditioning-Theory-Applications/dp/0199735964/ref=tmm_hrd_title_0?ie=UTF8&qid=1332506416&sr=8-2

    ReplyDelete
  4. Forgot... speaking of the pedagogic importance of behavior analysis one should be aware of Project Follow Through:

    http://en.wikipedia.org/wiki/Project_Follow_Through

    The most successful pedagogical model used, Direct Instruction, is built on principles from applied behavior analysis.
    http://en.wikipedia.org/wiki/Direct_Instruction

    ReplyDelete
  5. Graham, if you're on Facebook then you might want to become a member of the Associative Learning Group:

    http://www.facebook.com/home.php#!/groups/110618798972939/

    139 members and counting, we're in rude health!

    ReplyDelete
  6. I would also like to add that John Pearce (FRS) et al have organized the Associative Learning Conference for the past 16 years in a delightful country house in the middle of rural Wales. Every year we enjoy the company of the big names in learning theory, and I am constantly delighted, also, at the number of young researchers from across the world, who are interested in associative learning, and who give great talks there (hangovers depending).

    I agree that there has been a contraction in the field of pure animal learning theory over the past 10 to 15 years, but there has also been an expansion of the field into other areas (for example, behavioural neuroscience is, from the looks of it, dominated by procedures and theories from learning theory).

    In 2013, I am taking over responsibility for conditioning/learning theory submissions to the QJEP and one of my priorities is to improve the profile of this field in the Journal.

    Personally I am optimistic about the future of learning theory.

    ReplyDelete
  7. I increasingly think that this is a story about the mishandling of first-year Intro Psych. The course does not talk about learning theory in a way that shows 1) that there are many of hard-won results in that field that are far more conclusive than in most of the rest of psychology, and 2) that those results are central to the field, and not a weird, quirky area that has something to do with rats and dogs. this problem is not unique to learning theory, there is a century of great work in all areas of psychology that is largely forgotten.

    I know few care, but if you want more details about how the field's problems start in intro courses, see here.

    P.S. I agree with many here. I also teach a lonely Psychology of Learning class. Though I am not much involved in the larger community, I know that there are more of us left than it seems.

    ReplyDelete
  8. Hi Graham.

    You really should get out more! Learning theory forms a central plank of the research ongoing at Sussex in Psychology, both in the human and animal labs, and some of us have offices within a couple of metres of yours! You’d be very welcome for a cup of coffee.

    You write that “In the 1980s I was researching performance models of Pavlovian conditioning – basically trying to define the mechanisms by which Pavlovian associations get translated into behaviour – particularly in unrestrained animals. Eventually it became clear to me that only me (grammar!) and maybe two or three other people worldwide shared this passion”.

    Well, several of us in the School of Psychology at Sussex are doing exactly this, some of us for nearly twenty years, largely in the context of models of drug and alcohol abuse, though there is also a good chunk of work on feeding behaviour. We’d be happy to supply you with reprints on our work (they are also available through the University website).

    Not surprisingly, given our research interests, we include many aspects of learning theory in the courses we teach (though we share your desire to see more included in the core courses). Thus aspects of learning theory are taught in the first year course Psychobiology, in the second year course on Brain and Behaviour, and in the final year options Psychobiology of Addiction, and Psychobiology of Learning and Memory (and, we suspect, others). In our experience, the students are really interested in concepts such as conditioned reinforcement, and pavlovian-instrumental transfer, when they are presented in a relevant way. And our problem lies not in attracting students to our courses, but in keeping a cap on class size! Not least, a number of undergraduate students carry out research projects in our labs in precisely these areas, and some even stay on to do PhDs.

    So, by all means, let’s have more learning theory in core courses, but just because you no longer read our literature, don’t make the mistake that we don’t exist. And some of the things that we and our colleagues elsewhere do may even be important!

    Dai Stephens, Dora Duka, Hans Crombag.

    ReplyDelete
  9. FYI: This post has lead to some discussion here and here.

    Eric

    ReplyDelete
  10. The cognitive theory is the best for solving various health issues and also for the mental developments.

    ReplyDelete
  11. Am loving this post and thread, which came up when I tried googling for the Wales AL conference - I also agree that there is lots of exciting work being published in this field e.g. PIT & addictive behaviors - particularly relevant to my research. I've found lots of handy links to conferences etc on this thread so will check them out.

    ReplyDelete