Auditing Section Research Summaries Space

A Database of Auditing Research - Building Bridges with Practice

This is a public Custom Hive  public

research summary

    Covariation Assessments with Costly Information Collection...
    research summary posted April 16, 2012 by The Auditing Section, last edited May 25, 2012, tagged 09.0 Auditor Judgment, 09.03 Adequacy of Evidence 
    Covariation Assessments with Costly Information Collection in Audit Planning: An Experimental Study
    Practical Implications:

    The study has significant implications for situations where auditors must make decisions based on preliminary analyses (i.e. analytical procedures). From this analysis they either pass or extend testing on the accounts under review. In the absence of guidance on the importance of subjecting all accounts and transactions to testing, auditors may tend to over (under) test in response to greater (lesser) sensitivity of an account to error (based on the preliminary testing). The ineffectiveness and inefficiencies that may result are potentially greater when transactions to be tested are chosen judgmentally. 


    Ganguly, A.R. and J. S. Hammersley. 2009. Covariation Assessments with Costly Information Collection in Audit Planning: An Experimental Study. Auditing: A Journal of Practice and Theory 28(1):1-27.

    auditing; audit planning; covariation; correlation; costly information.
    Purpose of the Study:

    Auditors often must estimate the potential that the presence of observable clues and could lead to or provide some insight into the actual presence of material misstatements in financial reporting. For example, while studying the internal control system during audit planning, the auditor may discover a weakness in internal control (clue). The auditor must then assess how this weakness may affect the risk of material misstatement and assign resources to testing accordingly. If the assessment is too high (or too low) the auditor will likely overemphasize (or underemphasize) the necessary substantive testing. 

    Prior auditing research has consistently demonstrated that people tend to place much more emphasis on presence, rather than absence, of clues. In this study, the authors study covariation and define it loosely as a measure of how much two variables change together (e.g. an internal control clue and an actual material misstatement). The purpose of this study is to explore how auditors assess the relationship between clues (specifically internal control clues) and resulting conditions (e.g. a material misstatement). In this setting the auditors have incentive to make accurate assessments; however, the cost of obtaining sufficient evidence to support the assessment is high.

    Design/Method/ Approach:

    The authors conducted two experiments during which students (surrogates for auditors) chose which costly information to obtain in order to assess the relationship between an observable clue and its associated condition. The experiments were conducted in the early 2000’s time period.

    • The authors find evidence suggesting that when people are simultaneously incentivized to be both effective and efficient, they collect insufficient information in specifically predictable patterns.
    • The authors find evidence suggesting that when participants were biased in their estimation assessments, the biases occurred more due to the type of information collected rather than improper processing of collected information.
    • The authors find evidence suggesting that collecting information randomly from all available populations can potentially      improve understanding of the relationship between clues and resulting conditions. However, it can lead to a stronger revision of initial beliefs based upon some perceived or real standard rather than collecting information only from judgmentally-selected populations.  
    • The authors find that when collecting information is costly and auditors are allowed to select populations from which to sample they are more likely to collect sufficient information if they have been instructed that sampling from all populations of interest is important.
    Auditor Judgment
    Adequacy of Evidence
    home button