The overriding objective in choosing audit quality proxies is to use measures that are most appropriate for the research
setting. In evaluating appropriateness, one broad consideration is whether the study examines the demand or supply of
audit quality. Output-based measures are usually best suited for tests that examine the supply of audit quality, while inputbased
measures are usually best suited for tests that examine the demand for audit quality. However, based on our
evaluation, it is obvious that no single proxy is capable of painting a complete picture of audit quality. Thus, for supply
studies, we recommend choosing measures across the four output measure categories when it is feasible. Comparing across
categories provides a more comprehensive understanding of the effect on audit quality than comparing within a given
category. We recognize that it is neither practical nor desirable to use all possible measures in a single study. However, this is
not likely to be a major constraint to our recommendation as there are only four output categories. We also note that this is a
practical strategy as evidenced by several studies that effectively adopt this approach (e.g. Lennox and Li, 2012; Dao et al.,
2012). Finally, we urge researchers to evaluate and carefully articulate the inferences that can and cannot be inferred from
the proxies based on their unique strengths and weaknesses. Currently, the literature often lacks such discussion.