Reflections on Task Analysis; Psychological Sciences; and Clinical Psychology by John Annett

‘Reflections on Task Analysis: the origins of HTA’ by John Annett

An approach (let’s not say ‘method’) to the analysis of complex tasks was dubbed Hierarchical Task Analysis (HTA) by my old friend and colleague Keith Duncan. It has been successfully  applied to the description of both  routine tasks such as typewriter assembly, and more complex tasks such as target identification by naval combat teams and the control of nuclear power plant. Its origins go back to 1951 when as an undergraduate at Oxford I first encountered  the work of Clark Hull. Not only did I find Hull’s pseudo-mathematical behaviourism incomprehensible, but the ideas of Shannon and Weaver on information processing and the building of the first commercial  computers were beginning to influence thinking in experimental psychology.

In 1960 I was invited to spend a summer at the US Naval  Training Device Center  just outside New York. There was a lot of interest in ways of specifying training methods for military tasks and I was impressed by the work of R.B. Miller (for  USAF) who stressed the importance of specifying sensory feedback in acquiring a skill. I also met B.F. Skinner and, whilst being interested in his ideas on ‘teaching machines’, was unconvinced by his too-easy extrapolation from rewarding half-starved pigeons (actually only 2/3^rd. starved)  to the acquisition of mathematical skills  in children by small ‘positively reinforced’ steps.

Returning to the UK and a project on teaching machines in industrial training under Harry Kay at Sheffield in 1961 I was faced with the problem of specifying the re-training of ‘open-hearth’ steel makers to operate the new electric arc furnaces. The team responsible for an open-hearth furnace melt over 100 tons of ore and scrap metal to a very high temperature deciding how long to apply fuel and how much  carbon, nickel, oxygen or other components to add before that steel is ready to pour from the furnace into moulds  This judgement is made by checking the temperature and inspecting the flame. It became obvious that the skill  depended on the assessment of feedback concerning the state of the hot metal rather than a sequence of routine actions and that human skilled activities are governed by the checking the current state of a hierarchy of  goal conditions. In other words HTA represented not just a technique for listing a set of duties but a way of understanding intelligent, and unintelligent, human behaviour in a complex environment.

We had many visitors to Sheffield intrigued by the idea of teaching machines, among them a somewhat sinister delegation from the USSR who photographed everything but also Keith Duncan from the Army Operational Research Unit who showed especial interest in my ideas on task analysis.

After a period lecturing at Aberdeen I moved to Hull in 1965 and there set up the rese\arch project on Task Analysis funded by the Ministry of Labour. I recruited Keith Duncan and two of my students Rob Stammers and Mike Gray to the research team. The project involved work in a variety of local industries including chemical process control with BP Chemicals and the construction of a computer controlled simulator, the assembly of typewtiters and the manufacture of sticky dressings (band-aids). John Patrick  and Andy Shepherd, also students at Hull along with Rob Stammers went on to spread these basic ideas at Aston, and so did Keith Duncan at Cardiff and elsewhere.

I left Hull in 1972 to try my luck at at preparing the first degree course in Psychology at the Open University. At the OU and subsequently on appointment to the new Chair of Psychology at Warwick in 1974  my interests were, of necessity, much wider and it was not until I joined the MRC  RNPRC under Bob Audley’s chairmanship that I became involved once more in military applications, This led to work with Dave Cunningham, another former Hull student, and became involved in analysing teamwork interactions in Naval Command and Control systems using the basic principles of HTA

In summary, HTA is not just an ergonomic method but a general approach to understanding how people acquire, succeed or fail in complex control skills.

Principal Publications on HTA

1. ‘Task Analysis.  (1971) J. Annett, K.D. Duncan, R.B. Stammers & M.J. Gray HMSO.
2. Task Analysis (2000)  J. Annett & N.A. Stanton. Taylor & Francis.
3. Task Analysis (2006) J. Annett & N.A. Stanton.. Pp. 45-78 in G.P Hodgkinson and J. K. Ford (eds)  International Review of Industrial and Organizational Psychology, Vol. 21. Wiley.
4. Hierarchical Task Analysis (2005) J. Annett  Ch. 33 in N. Stanton, A. Hedge, K. Brookhuis, E. Salas & H. Hendrik (eds)  Handbook of Ergonomic Methods.  CRC Press.
5. Hierarchical Task Analysis. (2001) A Shepherd. Taylor & Francis
6. Ergonomics Vo.l 41, No.11, Nov. 1998. Special Issue on Task Analysis, eds. J. Annett & N. Stanton.

Supplementary  Comments on HTA

There are two important features of HTA that are sometimes either missed or misunderstood. Both are explained in the original research report  (Annett et al. 1971. Training information Paper number 6, Ministry of Labour, HMSO).

First, HTA has a unit of analysis called an operation. This has three elements: an input or the current state of affairs available to the operator, an action or strategy that is what one does to change to am more desirable state of affairs,  and feedback,  namely the new input resulting from the action, ideally an improved state of affairs, that is something closer to the overall goal. These units are not are not fixed in size or complexity like, Gilbreth’s ‘therbligs’, but are essentially goals. They can be as comprehensive as ‘build a house’ or as detailed as ‘lay the next brick’. Put another way, the process of analysis is to determine a goal hierarchy.

This leads to something which has puzzled those who have failed to grasp the essence of the  hierarchical principle, namely the p x c criterion. When this criterion is reached then no further analysis (or decomposition of the goal hierarchy) is worthwhile. If, given a particular task and a particular learner it seems unlikely either that the probability of failure (p) or that  the cost of failure (c) is acceptably low such that their their product is low, then no further detailed analysis is likely to be of benefit. Thus, if the learner (perhaps a population of learners) is especially naïve or the cost to the overall goal of the task is high then further detailed analysis is required. The p x c criterion is thus specific to the nature of the system and the calibre of the trainee. It is not a single figure that can be estimated to fit all circumstances nor a number that is a permanent property of ‘the task’.

John  Annett

10^th April 2014

‘Reflections on Psychological Sciences’ by John Annett

Fashions in Psychology change. In the 1950s I read Psychology & Philosophy at the Oxford Institute of Experimental Psychology. By the 1960s I thought of myself as a Cognitive Psychologist. Currently the OK term for a truly scientific approach to understanding the human mind is Cognitive Neuropsychology. The Oxford Institute had been founded in 1936 by Kathleen Watts. McDougal, Wilde reader in Psychology had been influential in the current ‘take’ on Psychology, but  W. Brown was the first director of the Institute which by the time I arrived, under the auspices of George Humphrey, the Institute had adopted the general principles established by Hemlholtz and Wundt which were in turn based on a physiological view of psychology

Mrs Watts had lost her husband in the first World War and  may well, like others,  have turned to the nascent ‘science’ of psychology for illumination of possible modes of access to the ‘spirits’ of those lost.  In the age of Gilbert Ryle (I attended his lectures on ‘the concept of mind’) it was no longer thought appropriate to pursue a category error. By 1950 Psychology at Oxford, Cambridge (and Reading) was firmly based on the pioneering work of German psychologists such Fechner and Wundt and the Dutchman Donders  London University, mainly influenced by Galton and Spearman and subsequently Burt was primarily concerned with the measurement of personality and intelligence. The Experimental Psychology Society was formed by the  Oxford, Cambridge, Reading group and explicitly excluded Eysenck from membership, presumably because of his theory of personality based on psychometrics..

In our first lab class my partner, Roy Davis and I, used an electronic device, a Chronotron, to measure each other’s reaction time..  Helmholtz had  demonstrated that the mental processes underlying thought take a measurable amount of time[i].  Donders published his first paper on reactiokn time in 1868. using reaction time  measurement to explore choice under a variety of conditions. In my study I have a poster celebrating the centenary of  the founding of Wundt’s ‘reaction time’ laboratory at Leipzig and 1879. I also have a Kymograph, rescued from a purge of outdated laboratory equipment and popularly known as a Palmer Drum designed to allow biologists to measure time, usually against the vibrations of a tuning fork.

But Wundtian experimental psychology was already being challenged  by the Shannon-Weaver Theory of Information and its application to the interpretation of reaction time data, for example by Hick, 1952 and Fitts 1954 (followed by me in 1957 and 1958). Other experimentalists,  especially Donald Broadbent in Cambridge with whom I shared a platform at the 1957 British Association for the Advancement of Science, focused on perception and attention as essential to the new cognitive psychology.

In those days knowledge of brain function was limited and EEG the only means of recording brain activity in the live individual. In my first job at the Burden in Bristol I had been impressed with  both the recording of brain activity and the construction of working models such  as Grey Walter’s ‘tortoises’. The latter was part of the beginnings of Cognitive Psychology which by the 1960’s had become firmly entrenched as the new approach to understanding the workings of ‘the mind’, inspired in part by the work of linguists such as Colin Cherry in London and Noam Chomsky in the USA.

By the 1970s I was busy setting up a new department at Warwick and decided on computing and EEG as the way forward, rejecting the classical requirement for animals in a psychological laboratory. By the 1980s new technology was introducing new methods of brain scanning, a little too late (and too expensive) for me.

Nowadays the idea of  ‘cognitive neuropsychology’ has been adopted by the press, whilst classical psychology, (now taught in schools) has gone back to more popular models, for example, the ‘mindfulness’ movement, and business schools promote the use of poorly validated tests of management skills such as ‘leadership’. I almost despair!

I am not entirely happy with the way in which the popular press appears to deal with what it understands to be ‘psychology’ as distinct from ‘cognitive neuropsychology’. As I write this a well-known BBC journalist is announced as soon to present a series on ‘psychology’ which, from the BBCs description is interpreted as psychoanalysis. Nor am I entirely happy with the manner in which the press is captivated by the technology of brain scanning techniques whilst ignoring the necessity to relate functional analysis of behaviour to the detailed mechanisms of the nervous system. In the early 1980s I proposed a theory of observational learning, including ‘imaginary prectice’ that required a mechanism shared between perception and action. Only later I  learned that Rizzolatti had shown that mirror neurons did in fact provide a plausible neural mechanism[ii]. Such links between function and mechanism are important to the development of the science of psychology.

John Annett

April 2014

‘Reflections on Clinical Psychology’ by John Annett

Just at the end of  WW2 at age of about 15 I was doing well at school but suffering from an undiagnosed set of symptoms and, as a result, was sent to the local ‘Child Guidance Clinic’ where I met a very charming Viennese lady, a Dr. Grunbaum. At her behest I produced some highly spectacular dreams although I have no recollections of any attempt at interpretation. I had only a few sessions but I decided to look up Freud in the local library and read several of his books, especially, since I was interested in religion, Moses and Monotheism followed by Totem and Taboo. I inflicted a graphic account of the incest taboo on an alarmed RE teacher in a sixth form discussion. Later I read other sources on Psychology, notably Woodworth’s Schools of Psychology and was then able to distinguish between psychoanalytic psychiatry and lab-based general experimental psychology.

In 1948 I began my National Service in the Navy as a Sick Berth Attendant (essentially a male nurse) and was sent for basic medical training at Stonehouse  Naval Hospital[i], Plymouth where I learned hospital routines and medical procedures which ranged from attending post-mortems, to assisting in the operating theatre and general ward duties, such as emptying bed pans. As a qualified SBA at Chatham Naval Hospital I assisted the medical consultant and progressed to more technical procedures such as using the new ECG machine and was given charge of a patient paralysed by polio (in pre-vaccination days) and in a ‘Both’  respirator or ‘Iron Lung’. On discharge from National Service in May 1950 and due to go up to Oxford in October I obtained a temporary nursing post at the aptly named Barming Hospital, the Kent County Asylum, and here was my first experience of  psychiatry as then practised.

I was assigned to the male admissions ward. The Hospital had been built in the 19^th Century, about the same time, and probably by the same architect who designed Maidstone Jail. Surrounded by a formidable wall the hospital accommodated 2000 male and female patients, many suffering from GPI, often due to syphilis acquired during WW1, along with other forms of insanity and dementia. The admissions ward, accessed through several locked doors, comprised first, a corridor with padded cells for the more violent patients, and then a long ward with up to 50 beds. A small team of male nurses, myself among them, had to see each patient washed, shaved and fed and all the beds made and tidy before doctor’s rounds and treatments. In those days treatment was mainly by electroconvulsive therapy. I well remember a senior male nurse telling me ‘when I started there was two treatments, hot water and cold water. Now it’s all done by electricity!’ Patients for ECT were laid on a stretcher and given an injection of curare to prevent violent muscular spasms when electric shocks were administered to their temples. Patients were usually rendered unconscious by a few shocks and it was often my task to monitor up to a dozen laid out unconscious on stretchers and to apply artificial respiration to any who stopped breathing while in ‘recovery’. I had no indication as to the effectiveness of this ECT, or its companion technique based on an induced insulin coma. I do not recall any drugs being administered nor was I aware of any surgical treatment, such as pre-frontal lobotomy. Although there was ‘occupational therapy’ I never met a clinical psychologist during my few months at this hospital.

At Oxford where I switched from Theology to Psychology with Philosophy in 1951 I was introduced to the measurement of reaction time using a chronotron in the lab and later to the study of perception by Jimmy Gibson who was spending a sabbatical year with the Institute. We were taught by Oliver Zangwill, later the successor to Bartlett at Cambridge, and were also introduced to the care of badly disturbed boys and elderly dementia suffers at different institutions in Oxfordshire. By my final year my psychology tutor was Harry Kay, newly arrived from Cambridge where he had been working in Alan Welford’s unit on ‘Ageing’. Harry introduced me to the ideas of information theory [ii]. I also made the acquaintance of May Davidson clinical Psychologist at the Warneford Hospital and it was she who recommended me to my first job in clinical research at the Burden Mental Research Department, Stoke Park, Bristol.

Stoke Park Hospital was an institution for the ‘mentally deficient’ then classified as ‘idiots’ and ‘imbeciles’. The research department provided assessment services to the hospital as well as following a research programme on simplified methods of measuring intelligence. The hospital also housed the Burden Institute under Professor Golla who was a pioneer of a new surgical treatment, pre-frontal lobotomy. Also at the Burden was Dr. Grey Walter a pioneer in the technique of EEG developed just before the war. Grey Walter and was also an  early advocate of artificial intelligence. He constructed semi-intelligent ‘tortoises’ able to move about a room and return to a docking station when the battery was running low.[iii]

Stoke Park Hospital held, like Barming, about 2,000 inmates. I learned these were of two types: first, those born with serious brain defects, the ’idiots’, and second those who came to public notice, for example by the police, and whose intelligence was suspect as established by a test IQ of 70 or less, the ‘imbeciles’. In either case they were committed to the ‘hospital’, essentially for life, since low intelligence did not appear to benefit from any form of ‘treatment’. Sometimes the borderline mentally subnormal would escape and be brought back into hospital after a few days where they were indispensable as helpers for the severely brain damaged. The females worked as helpers on the wards and in the laundry whilst the male imbeciles worked on the farm attached to the hospital estate providing vegetables for the kitchen.

As a psychologist I and two others[iv] joined the medical and administrative staff in the dining room for a daily lunch, presided over by the medical superintendent in the Dower House, the original building still visible from the M32 on the approach to Bristol. The MS was a towering 6ft 7 but, perversely, drove a mini. Amongst the medical staff was a distant relative of Lewis Caroll who had his own pathology lab and from time to time took specimens home for study. At lunch one day he admitted to having left a brain in a biscuit tin on the bus home and was worried about approaching the lost property office to ask if ‘his’ brain had been found.

At this time  I was hovering between my initial interest in clinical work and the new approach to experimental psychology so I began working on new ways of assessing, and perhaps improving the physical capabilities of the mentally handicapped by designing motor tasks with simplified displays[v]. My first published paper indicated that the speed of motor performance depended on the (measurable) amount of information in the display rather than muscular weakness.[vi]

My work on information processing led to a recall to a research post at the Institute of Experimental Psychology,  Oxford assessing Fitts’ Law. In the meantime my new wife Marian took an appointment as clinical psychologist with May Davidson at the Warneford Hospital whilst I became more involved with ideas about information theory and motor skills[vii]. The idea of the peg-board as a convenient measure of manual skill was later adopted by Marian as a measure of handedness and cerebral dominance[viii]

I next became involved in clinical psychology in the 1970s through my role as head of the newly-found Department of Psychology at Warwick. After the principal undergraduate courses had been established I appointed John and Marcia Davis (who was also part time at the local mental health clinic) to teach clinical psychology at undergraduate and postgraduate level. Their line was in ‘cognitive behavioural therapy’ which has become generally popular, Like many clinicians, and despite my anxiety as head of department facing assessment, they published little research. After my retirement from Warwick a psychiatrist was appointed to teach part time on the postgraduate clinical course. I attended one lecture by an incumbent of this post on ‘mindfulness’, which appears to be strongly influenced by Buddhist philosophy. I concluded this approach may well be of benefit to patients suffering from mild anxiety and depression. I regret that this appears to be only the most recent of a long line of ‘soft’ treatments that may supplement the currently widespread use of drugs. The technical advances in brain studies (such as functional Brain Imagery) have now focused clinicians and researchers (not to say journalists)  on ‘cognitive neuroscience’[ix] as a scientific approach to the many unsolved problems of psychiatry.

I am far from satisfied with progress in clinical psychology since my initial interest some 70 years ago. Mental illness provides a set of serious problems which neither  mesmerism, analysis or physical treatments and even the anti-psychiatry movement of the 1960s have been able to solve. There have even recently been suggestions that the DSM classification of mental disorders be abandoned. The large mental hospitals have now been closed and their estates sold off, often for housing.  There is still a long way to go before we can solve the problems of mental disorder and in the long term one is left to wonder if natural variation, including some degree of mental abnormality, is not essential to the future development of human society.

John Annett 27^th. May 2014.