A critical evaluation and extension of the application of quantitative techniques to the field of auditing

This dissertation constitutes, among other things, an inquiry into how various quantitative techniques are being used by members of the auditing profession. An extensive review of the literature is made to ascertain what quantitative techniques have been suggested for use by auditors. Such techniques include Markov chains, regression analysis, Bayesian statistics, simulation, and nonparametric statistics. An analysis is made of in-depth interviews conducted with personnel from the largest accounting firms. The results of the interviews indicate that in current practice very little use is made of quantitative techniques other than those associated with classical survey sampling and classical parametric univariate hypothesis testing. The reasons for this lack of use are reported and discussed. An investigation of how auditors, businessmen, and investors act in various uncertainty situations is discussed. Experimental results show that auditors tend to use weighted averages in preference to modes and medians. A quantitative approach to audit uncertainty is presented. This approach allows for formal recognition of the stochastic nature of many account items. Examples are provided which demonstrate how the approach might be used in practice.