Low response rate may introduce bias into the survey results. To investigate how frequently clients manage earnings and how auditors adjust managers' earnings management attempts, Nelson et al. (2002) survey auditors from a Big 5 firm. They report that the response rate of the questionnaires sent to auditors is around 16%. In evaluating the potential bias arising from the low response rate, the authors report the followings:

In our study, 52 ''late'' responses, received following a reminder, are not significantly different from other responses in any of the analyses reported in the results section. Likewise, there are no significant differences between the first 50 and last 50 responses received, or between the first half and last half of the responses received.

We also tested for similarity between respondents and the population from which they were drawn. Although confidentiality restrictions prevented us from getting some forms of demographic data, respondents did provide the percentage of their experience that involved clients in each of the many industry classifications used by their firm. The average proportion of auditors' experience in each industry is very similar to the proportion of audit clients the firm has in that industry, suggesting that respondents' backgrounds reflect the general practice of their firm.

Required:

1.Explain why "non-response" may bias the research results in survey research.[4 marks]

2.One purpose of Nelson et al. (2002) is to investigate how frequently auditors would make audit adjustments if client managers attempt to manage earnings. Analyze whether the non-response bias would understate or overstate this frequency.[4 marks]

3.Discuss the rationale of the tests conducted by the authors in evaluating the potential bias. Do you think the results indicate the absence or presence of non-response bias? [12 marks]