Menu Close
Reader of Clinical Psychology, University of Manchester

My research takes many forms, but I now subscribe to an ideal for research in the behavioural sciences influenced by the work of William T. Powers (1926-2013) and his collaborators on Perceptual Control Theory (PCT; Powers et al., 1960a, b; Powers, 1973, 2005, 2008). Below are a number of propositions of PCT that inform theory, research and practice in the behavioural, social and life sciences:

1) Humans and other animals purposefully seek out, select, control, manipulate and influence their own perceptual input from the environment. This is the function of what we call 'behaviour'. Behaviours are not triggered, learned or planned, but emerge dynamically through constraints within the current moment in an attempt to control current perception.

2) The default functioning of humans and other animals is an unbroken system of parallel, hierarchically organised, continuous, 'closed' loops between perception, comparison with reference values for that perception, action, the environment, and (again) perception. Therefore, it is impossible to 'independently manipulate' thoughts, feelings, behaviours, physiology or the environment.

3) Despite all behaviour being purposeful, control can be executed automatically (i.e. outside 'conscious awareness'), and this process can be understood in quantitative terms that facilitate computer modelling and prototype testing of models of an individual to compare with the real system. The functioning of an accurate model of an individual, generated by a correct theory, would have a near perfect match with the real system.

4) Given the above propositions, empirical research that studies groups of individuals and aims to examine the statistical relationship between 'Independent Variables (IVs)' in the environment and 'Dependent Variables (DVs)' - measured as behaviour or other observable properties of individuals - has limited validity. This linear model is ideal for physical systems that do not control (such as a moving object), but do not apply to living systems (or control systems designed by humans), because of their intrinsically purposeful nature.

5) Researchers can use an alternative research paradigm - the 'Test for the Controlled Variable', which examines the perceptual variables that an individual is controlling using their behaviour through its manipulation of the environment, despite other aspects of the environment that disturb these attempts at control. There are many published papers demonstrating this approach, even though the 'IV-DV' approach currently dominates. The aim of this manifesto is the adjust this balance in order to raise the scientific credibility of the behavioural sciences.

For more information, see, which provides links to the key papers supporting this approach. Please feel free to contact me concerning this manifesto.


  • 2011–present
    Clinical professor, University of Manchester