We begin this section by describing each measure and the types of studies that use it. We then evaluate the measures
along several dimensions that determine how well each proxy captures audit quality. The principal dimensions we discuss
are directness, egregiousness, actual-versus-perceived, and several measurement-related dimensions. We define directness
as the extent to which the auditor influences, controls or is responsible for the output. Thus, this dimension only applies to
output measures. For example, the auditor has sole influence, control, and full responsibility for the type of opinion. We
define egregiousness as the severity of the misconduct implied by the measure. For example, auditor-related AAERs capture
relatively egregious misconduct. The actual-versus-perceived dimension describes whether the proxy attempts to measure
actual audit quality, such as restatements, or perceived audit quality, such as stock price reactions. Finally, we discuss a
variety of measurement-related dimensions, including discreteness, consensus on the measurement, and measurement
error. We discuss the output proxies in descending order of their directness, as summarized in