Jay Smith, CPA, is doing the audit of accounts receivable ofJones Wholesale Stationery and Office Supplies. There are 820 accounts receivable, totaling \($887,500\). Total current assets are \($1,450,000\), and total assets are \($2,200,000\).

Jay makes no effort to determine the appropriate sample size statistically. Owing to the high expectation of errors in accounts receivable, he decides to select a random sample of 200 accounts receivable for positive confirmation. He plans to evaluate the results statistically, using several different statistical methods.

Required :

a. Is it acceptable to statistically calculate confidence limits from a random sample without having determined the appropriate sample size with the use of appropriate formulas ? Explain. Why should anyone ever use statistical formulas in determining the sample size ?

b. Explain how Smith should handle nonresponses to first and second requests for confirmation. What can he do to minimize the likelihood of including a confirmation response value at the wrong amount due if a respondent states that a confirmation is correct when it is actually wrong ?

c. Assume that all 200 confirmations were returned by customers or alternate procedures were followed by Smith to determine the audited value of each account. The audited value of the 200 confirmations was \($216,512\), and the client’s recorded value was \($237,416\). There were 34 errors discovered. An analysis of the 34 errors indicated that 22 were overstatements and 12 were understatements. The size of individual errors appeared closely related to the size of the recorded balance. Calculate the point estimate of the audited value of the population using mean-per-unit, difference, and ratio estimation. Explain the reason for the differences in the three answers.

d. Which of the three methods is likely to result in the smallest computed precision interval at a given confidence level ? Which method is likely to result in the largest precision interval ? Explain why you answered as you did.