Ethics and experiments in accounting : A contribution to the debate on measuring ethical behaviour

Purpose – This research aims to measure compliance in a tax experiment among students. The aim of the study is to investigate relationships between claimed behaviour in a questionnaire and actual behaviour in an experimental environment, together with different behaviours between males and females, and different age cohorts. Design/methodology/approach – A total of 630 undergraduate Commerce students at a New Zealand university completed a questionnaire on attitudes towards the tax system. The students subsequently participated in a simulation experiment requiring responses to hypothetical tax evasion decisions. Individual reward payments were contingent on the outcome of these tax evasion decisions. Questionnaire responses, which captured intended behaviour, were compared with actual behaviour in the experiment. Findings – The study finds more compliant behaviour among older students and students who have been at university longer. It also finds female students demonstrate more ethical responses in their behaviour than male students. In contrast to extant literature, it finds a positive relationship between students indicating a preference for compliant behaviour in the questionnaire, and behaviour in the experiment. This leads support for the use of Defining Issues Tests (or similar instruments that capture moral development intentions) in ethics education research, and challenges recent studies that find a gap between intended and actual behaviour. Research limitations/implications – As with all experimental research, the design is necessarily an artificial representation of the real world. Thus, the ability to generalise from this research is restricted. Originality/value – Much of the research into the influence of ethics education on accounting students focuses on student claims of how they would respond in a hypothetical situation as measured by a Defining Issues Test or similar instrument, in order to provide a measure of ethical development. In contrast, this study adopts a behavioural approach. The findings indicate that Defining Issues Tests are likely to be an appropriate tool for ethics education research.