Can we measure effort in cognitive tasks? Examining the application of Additive Conjoint Measurement and the Rasch model

‘Effort’ is a concept of interest in cognitive psychology and neuroscience where many theories include it as a postulate. Despite its intuitiveness it is difficult to define such that its operationalisation follows a logical derivation chain. Recently I have proposed conceptual definitions of both actual effort, and the perception of effort, as the ratio of task demands to capacity to meet task demands, both actual and perceived respectively. Clear conceptual definitions are key for determining whether a given operationalisation meets the necessary and sufficient conditions adequately. For physical tasks valid operationalisation, and indeed measurement, of actual effort is often trivial. But a problem arises for operationalisation of actual effort in cognitive tasks where the underlying capacity that disposes an individual to be able to attempt, and perhaps complete, the task is not directly observable nor is the demand the task presents. However, a solution may lie in applications of Additive Conjoint Measurement to determine conditions where classical measurement may be possible, and the Rasch model as a measurement operation of capacity and demands to derive effort. A key aspect of the Rasch model is that it posits and estimates from data two latent constructs that I accept here as conceptually equivalent to capacity and demands respectively in my definition of effort; first a characteristic of the individual (ability), and second a characteristic of the test or item (difficulty). As such, applications of these methods might provide a measurement operation of actual effort in cognitive tasks to enable more precise formulations and testing of theories that employ the concept. In this work I explore these ideas using simulation and analogical abduction of a task where effort is known and examine an empirical dataset. Finally, I discuss the conditions under which these methods may be suitable and their inherent limitations.