To draw conclusions about the information content of audit reports, one must isolate the effects of information conveyed specifically by audit-report components of aggregate signals. To do so, the information conveyed by other components of the aggregates must be controlled. This paper argues that, for most types of audit reports, the necessary control cannot be exercised with security-price research methods Accordingly, some inappropriate inferences have been drawn about the information content of certain types of opinions. Research strategies to provide and implement the necessary control are suggested.

The paper reports the results of a study of the information that four types of audit reports convey to Chartered Financial Analysts about the sources of both financial-statement and audit-report messages. Four hypotheses are tested, two of which concern the perceived identity of the source of specific financial-statement messages. Two others concern the effect of the audit reports on analysts' perceptions of the credibility of the source of financial-statement messages and of the auditor. The study indicates that auditors are perceived to be much more involved in the communication of financial-statement messages than many people realize. This perceived involvement and the information that audit reports convey about the auditor's credibility combine to produce some unsuspected credibility effects.