A good management program provides the environment, housing, and care that . . . minimizes variations that can affect research' (National Research Council 1996). It would be naïve to rely on data collected from an animal who experiences depression and frustration resulting from an inability to show species-typical behaviours, or who experiences discomfort, pain, fear, anxiety and mental distress resulting from enforced bodily restraint. These experiences are re ected in the animal's physiological, psychological and behavioural responses to a test or to an experiment.The responses, however, differ from individual to individual because the experience itself is subjective. It is, at the least, questionable to do reliable 'scienti c' research under such methodological conditions because the data collected are bound to be in uenced by unaccounted-for extraneous variables such as distress, fear, anxiety, discomfort, depression and boredom. Sound scienti c methodology requires that research data should not be in uenced by uncontrolled variables and that the results can be replicated by other laboratories. 'Unless these requirements are met, an experiment is not considered scienti cally valid' (American Medical Association 1992).