Differential Evaluation of Audit Evidence from Fixed versus Sequential Sampling

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Journal: Differential Evaluation of Audit Evidence from Fixed versus Sequential



Perceived strength of evidence will be less when evidence is obtained using a sequential sampling plan that when it is obtained using a fixed sampling plan.

Evaluating the strength of audit evidence obtained is important for evaluating the effectiveness of specific audit test. In accordance with the requirement of professional standards, auditors aim to obtain sufficient competent evidence while minimizing the cost of evidence. As for the process of audit review, it involves examination of evidence obtained by another auditor and making a judgment about how strongly the evidence obtained by the auditor supports or disconfirms the auditor's conclusion.

This journal studies the literature on probabilistic inference in auditing by examining empirically how the choice of sampling plan used by the auditor influences the subsequent evaluation of the strength of audit evidence. There are differential evaluation of audit evidence from fixed versus sequential sampling. Sequential sampling also known as stop-or-go sampling. As we know, auditors can use either fixed or sequential sampling plans. As for a fixed sampling plan, it determines in advance the sample size that should be obtained by the auditor. On the other hand, sequential sampling allows for optional early stopping of sampling when support for one hypothesis over another is observed.

According to audit manual, it stresses the importance of achieving efficiency in sampling. After overall discussion, it concludes that sequential sampling is an approach that maximizes sampling efficiency. This is because it allows for sampling to be stopped at any time. However, auditors are advised to use sequential sampling only if those who subsequently assess the sufficiency of their work reason in accordance with the fixed sampling plan. Fixed sampling plans are primarily used in audit practice. By using fixed sampling plans, it allows auditors to collect more evidence. In this case, the frequentist approach may dominate the judgments of jurors and other subsequent reviewers of the auditor's work.

Audit research are carried out for the purpose of testing whether the type of sampling plan used influences evaluator's judgment about the sufficiency of audit evidence. Its practical implications with respect to audit efficiency and evidence evaluation in audit review and subsequent inspections of an auditor's work is important.


There are conduct two experiments to investigate the Fixed sampling plan and Sequential sampling plan.

Experiment 1: Sampling Plan

One of the methodologies used is participants. In experiment 1, there were 32 students as participants that consists of 18 males and 14 females in Accounting program. They had an average of 2.9 years of work experience. The purpose of this program is to prepare future auditor for the CPA examination. Peecher and Solomon (2001) argue that absent a theory specifying an expected interaction between level of audit experiment and the variable of interest in the study, the use of students is theoretically and practically justified. In the other hand, in the absence of any theory suggesting that students would assess evidence from fixed and sequential sampling plans differently than more experienced decision makers, use students as participants in Experiment 1. (Peecher and Piercey, 2008). Besides that, experimenting with a group that is less influenced by experience with fixed sampling provides for cleaner tests of hypothesis since fixed sampling plans are prevalent in audit practice (Wilburn 1984, Arens 2007, Rittenberg 2008).

The second one of the methodologies used is research design and experimental procedure. The participants were given scenarios that describing auditor’s tests of the effectiveness of internal controls. Using a 2 x 2 x 2 mixed between- and within-subjects design to test the hypothesis. The control variables EVIDENCE DIRECTION and SAMPLERS were manipulated within subjects. We used a diagram-balanced Latin Square design to assign the order of within-subjects treatments. This design ensured that each of the treatments occupies every position in the sequence of cases presented to participants the same number of times as each of the other treatments, and is followed and preceded by every other treatment an equal number of times (Keppel and Wickens, 2004).

The following methodology is independent variables. There are separate into two variables; Between-subjects variable and Within-subjects variable. In the Between-subjects variable, we manipulated the SAMPLING PLAN at two levels; Fixed and Sequential. Participants in Fixed condition read that the randomly sampled evidence. Participants in Sequential condition read that evidence was sampled and examined sequentially, and the plan allowed for optional early stopping of sampling. In the Within-subjects variable, EVIDENCE DIRECTION was manipulated at two levels; For and Against. Narrative in the For is presented the evidence confirming the hypothesis that controls failure does not exceed the specified threshold. Conversely, narrative in the Against is presented the evidence disconfirming audit hypothesis. Furthermore, the independent variable SAMPLER was manipulated also at two levels; Self and Another. The narrative in the Self condition specifying someone also has chosen the sampling plan, but the decision makers has obtained the evidence by himself/herself. In the Another condition, the person who has chosen the sampling plan also obtains the evidence at the same time.

Dependent variable is also used in Experiment 1. The dependent variable was PERCEIVED STRENGTH OF THE OBDERVED EVIDENCE. Suppose that you are reviewing the work of another team member who sampled invoice such as sales invoices and shipping documents for proper supporting documentation and you needs to assess how strongly the evidence confirms the audit hypothesis that sample error rate in the population of the invoices is 10 or less. There is excerpt from the instruction for accompanying the scenario. The PERCEIVED STRENGTH OF THE OBSERVED EVIDENCE was measured by six-point Likert scale. This scale was consists of six choices. There were not at all, weakly, moderately, fairly strongly, very strongly and totally.

In this experiment, we study to showing how sampling plan to affect the assessment of the strength of audit evidence. Independent variable and dependent variable is related. Difference evidence direction was effect the sampling plan under confirms or disconfirms the audit hypothesis. Evidence was perceived based on the sample error rate in the population of invoices. The sample error rate of either 4 percent or 8 percent is consider as For while sample error rate of either 12 percent or 16 percent is consider as Against.

Experiment 2: Sampling Plan Test with Auditors

Experiment 2 is test research hypothesis with sample of practicing auditors to enhance the external validity of results documented in Experiment 1.

In the Experiment 2, we are also using participants as our methodology. 23 auditors from an international PCAOB-registered were selected as participants.

Research design and experimental procedure is used in Experiment 2 too. This experiment was designed as a within-subjects study is due to small sample size of practicing auditors. The experimental narrative was similar to that used in Experiment 1. The sample size of audit evidence was the same between the fixed and sequential sampling plans. The evidence of the two sampling plans was also almost same.


The author is using a small sample practicing auditors in Experiment 2 to test the robustness of research results to differences in audits experience. In this test, the author finds that both students and practicing auditors are rates the greater on the strength of evidence from fixed sampling plans. This is clearly show that the results are not driven by the level of experience of the participants with fixed sampling plans, but it is depend on their frequentist thinking on the obtaining evidence from the sampling plans. In 2 x 2 x 2 experiment conducted by Masters in Accounting students, evidence from sequential sampling plans is invariably assessed as weaker compared with the fixed sampling plans, at every level of the two within-subjects variable. The result suggest the presence of some confirmation bias, the participants assess the evidence confirming the maintained audit hypothesis as stronger. Additional results show that participants assign less strength to evidence obtained by someone else when a fixed sampling plan has been used. This finding helps to drive away the alternative explanation that concern about the integrity of the random sampling random sampling process instead of concern about the adequacy of sequential sampling plans.

Bayesians and Frequentists, the two main schools of statistical inference hold opposing views on the differential evaluation of audit evidence from fixed versus sequential sampling. As for Bayesians, it suggests that the choice of sampling plan should be irrelevant to judgments about audit evidence. In contrast, Frequentists deems knowledge of the sampling plan essential to evidence evaluation. It assigns a greater weight to evidence obtained using a fixed sampling plan, as opposed to evidence obtained using a sequential sampling, on the same number of items.

Ultimately, it leads a reviewer of an auditor's work to take a different stand whether sufficient audit evidence has been obtained. According to Bayesian theory, it leads a reviewer of an auditor's work to be indifferent between the two plans. According to Frequentist theory, it leads a reviewer of an auditor's work to conclude that a larger sample size should have been obtained for a specified level of confidence under sequential sampling than under fixed sampling.