This study investigates the effects of long-term use of Intelligent Decision Aids (IDAs) on auditors' decision-making skills. The research therefore addresses the questions of 'Does the long-term use of IDAs reduce auditors' professional knowledge?' and 'By what means might the long-term use of IDAs reduce auditors' professional knowledge?' In addressing these research questions, the study develops a theory of Technology Impeded Knowledge Acquisition and Retention (TIKAR) by drawing upon perspectives suggested by the work of prior researchers examining the de-skilling proposition of the Theory of Technology Dominance (Arnold & Sutton, 1998).
The theory of TIKAR is complementary to the de-skilling proposition of the TTD in that TIKAR identifies potential causal mechanisms that underpin the de-skilling proposition. As with the de-skilling proposition, this theory of TIKAR recognises that auditors may de-skill over the long term. This conceptualisation of de-skilling recognises that auditors may be impeded from acquiring skills as well as from retaining the skills they already hold. The general theory of TIKAR is concerned with the relationship between the long-term use of an IDA and the auditors' professional knowledge.
This study adopts a post-positivist ontological perspective and sets out a research model for the theory of TIKAR. The propositions of the research model identify relationships between the constructs of ability, experience, and the tool's declarative and procedural knowledge upon the auditor's levels of both declarative and procedural knowledge. The research model also proposes that the auditor's time with the tool positively moderates TIKAR effects, whilst the auditor's motivation, or 'passion', for the task of audit lessens the TIKAR effects of long-term use of intelligent technologies. The final proposition identifies that declarative knowledge is necessary (but not sufficient) for the auditor to develop procedural knowledge.
This study investigates these propositions through open-ended semi-structured interviews with 55 public sector auditors in ten different public sector audit offices. These auditors were members of a target population of senior and highly experienced auditors in the field across multiple regulatory settings and organisational contexts, and were drawn from Australia, Canada, New Zealand, and the United Kingdom. Qualitative research methods based upon the manual coding of interview transcripts, supplemented through the use of the automated text analysis tool 'Leximancer', were adopted. As the interview questions were open-ended and exploratory, an initial manual deductive analysis assessed the level of 'apparent support' for each proposition. A proposition was considered to have 'apparent support' where more than half of the research participants identifying the relationship also supported the proposition. This analysis indicated apparent support for all concepts and relationships identified in the research model, although the level of this apparent support ranged from 'limited' to 'strong' apparent support.
After considering the research model, a post-hoc machine-guided dual analysis was undertaken to identify prospective new concepts and relationships not set out in the theory of TIKAR. This post-hoc dual analysis identified several implications and insights for the refinement and future direction of the theory. Key amongst these refinements was a reconceptualisation of the theory to focus upon the many transactive memory systems used by the auditor rather than a single IDA. The analysis also identified that the structural restrictiveness of the IDA, and environmental attributes in the auditor's working environment (process restrictiveness, mentoring by senior staff and the availability of training) were also avenues of future research.
Overall, this research contributes to theory by furthering the understanding of why TIKAR effects might arise. The research also expands the range of application of TIKAR-related effects in relation to auditor de-skilling and auditor procedural knowledge, which has not previously been investigated. A further major theoretical contribution of this study arises from the extensive multi-national sample of highly experienced auditors participating in this research. Finally, this research presents a rigorous qualitative approach for the analysis of large sets of qualitative data in text form that incorporates manual and machine coding into a cohesive and complementary research method.
In the context of practice, the research contributes by developing the audit profession's understanding of the relationship between the long-term use of IDAs and auditor skill levels. The study also identifies specific interventions that audit offices adopt to mitigate TIKAR effects. Finally, the research also contributes to the debate on the need for auditors to maintain an attitude of 'professional scepticism' in an increasingly complex audit environment. The adoption of more intelligent technologies to cope with audit complexity may lead to less expert auditors in the future, and auditors' development of professional knowledge and attitude of 'professional scepticism' may be impeded. These three contributions have practical benefits for the audit profession.
This work develops and refines a theory of TIKAR. Overall, this exploratory research provides a solid foundation for a program of future research. An understanding of the issue of whether audit professionals using intelligent technologies are impeded from acquiring professional knowledge, and the circumstances under which it may arise, is important to the audit profession. This study provides a basis for refining and maturing the understanding of the TIKAR phenomenon.