BETWEEN-AUDITOR DIFFERENCES IN THE INTERPRETATION OF PROBABILITY PHRASES

Auditors frequently interpret and use probability phrases. The focus o f this study is on between-auditor variance in the numerical interpret ation of probability phrases currently used in audit settings, as well as on the degree to which auditors are aware of this variance. Betwee n-auditor variance in numerical interpretation is an important measure because a goal of professional standards and audit engagement tools i s to increase the consistency of the auditing and accounting treatment of specific events (Ashton 1983; Joyce and Libby 1982; SFAC No. 2), a nd a goal of communication is to avoid misunderstanding. High between- auditor variance in the interpretation of probability phrases could re sult in miscommunication between auditors and the inconsistent applica tion of professional standards and audit engagement tools. Low varianc e awareness would suggest that auditors might not seek clarification o f communications or professional standards, and thus may be vulnerable to the costs associated with variance in interpretation. This study e licited from 49 audit managers their numerical interpretations of 17 p robability phrases currently used in audit practice, as well as their interpretations of six variations of those phrases. These interpretati ons were used to determine the relative likelihoods communicated by th e probability phrases, the between-auditor variance in interpretation of probability phrases, and the degree to which auditors are aware of between-auditor variance in interpretation. The results reveal substan tial between-auditor variance in interpretation and generally low leve ls of variance awareness. The data were also used to identify probabil ity phrases that have similar numerical interpretations, and to illust rate the use of the variance and variance awareness measures in select ing from among alternative probability phrases. These measures are use d to propose a five-level scale that contains those probability phrase s for which auditors display relatively low levels of between-auditor variance in interpretation and high levels of variance awareness in th e context employed in the experiment (the scale contains the phrases ' 'remote,'' ''low,'' 'as likely as not,'' ''more likely than not'' and 'highly probable''). The potential usefulness of probability phrases f or communicating the imprecision of probability estimates is also disc ussed.