Media local and abroad have gone to town slamming countries that are regarded as most corrupt in recent survey by Ernst & Young reported in Asia-Pacific Fraud Survey Report Series 2013.
Most disheartening ly, local media does not seem to have done their homework before writing those reports, some of which were headlined Malaysia one of the most corrupt nations, survey shows or Asia weakest against business fraud,
Malaysian firms ready to cut corners. Realistically, the survey is highly flawed and it is shocking those governments and citizens of countries implicated have not taken umbrage to this highly flawed unreliable and unrepresentative report. Just because an accounting firm released the report does not mean the results are conclusive or valid. Unfortunately, this is not the case.
Researchers who conduct any research are often motivated by external factors, such as the desire to be published, advance their careers, receive funding, or seek certain results.
Questionable Research Methodology
Commissioned by EY and carried out by Asia Risk, The Asia-Pacific Fraud Survey only polled 681 executives, senior managers and working level employees from March to May 2013 across the Asia-Pacific area including Australia, China, Indonesia, Malaysia, New Zealand, Singapore, South Korea and Vietnam.
The sample size for Malaysia (page 25) was only 103 executives. Can the views of 103 executives be representative of the integrity level of the business environment? How powerful, accurate, and reliable are the views of these elite 103 executives for the report to allege:
… bribery and corrupt practices are prevalent in Malaysia. Thirty-nine percent of respondents say that bribery or corrupt practices happen widely in Malaysia, which is nearly double the Asia-Pacific average of 21%. In addition, 29% of respondents say that bribery or corrupt practices have increased due to tough economic times and increased competition, which is the third highest amongst the countries surveyed.
How can such a minute sample of 103 out of a total population of 29.24 million is only 0.0003522571% of the population give Ernst and Young the audacious right to claim the right and authority to sermonize about corruption levels in Malaysia?
The report makes no mention of the sampling method used to obtain the size of 108 executives. How or why did they select those 103 executives?
All that is disclosed is that “the polling sample was designed to elicit the views of staff with responsibility for tackling fraud, bribery and compliance matters at multinational corporations, domestic companies and state-owned enterprises across sectors including Oil & Gas, Financial Services, Technology, Private Equity, Retail, Hospitality, and Mining & Minerals.
The phrase “the polling sample was designed to elicit the views of staff” undermines the reliability of the miniscule survey because it is a clear declaration that the sampling is flawed from the onset because it was highly probable that it was purposeful sampling designed to reap the intended findings.
Whenever possible, reliable studies MUST use random samples, utilize appropriate sample sizes, avoid biases, and must be conducted by researchers who remain objective and uninfluenced by funding or the desire to seek certain results. Randomization in sampling is critical to ensuring the validity of research. Is this evident in this highly biased research?
Who cares what Ernst and Young says when the report cited:
21% of respondents indicate they are not confident with the company’s existing internal procedures to detect fraud. Only a small portion of those polled considers tools other than the established methods. For example, only 12% of the respondents said that the use of technology, such as forensic technology, is a good method of detecting fraud.
Who doesn’t know a research finding is less likely to be true when the studies conducted in a field are smaller? Can there be representativeness when unreliability implies greater financial, other interest and prejudice or when the researchers are deliberately out on a chase of statistical significance as in this case?
This Asia-Pacific Fraud Survey Report Series 2013 oozes with research bias from start to finish. We can see how the combination of various design, data, analysis, and presentation factors have been manipulated to produce research findings when they should not be produced in the first place. Look at the methodology and analysis or reporting of findings. Clearly, the report displays selective or distorted reporting which is a typical form of such bias.
The report does not disclose the research instrument used. All we can deduce is that the researcher asked the respondents for their “opinions on what efforts are being taken successfully to mitigate those risks. All participants of our survey are employed at corporations with a turnover in excess of USD500 million equivalent and the industries covered range from industrial to financial services, from retail to natural resources.”
Was this done via administered or unadministered questionnaires? Alternatively, was it via structured or unstructured interviews? Were the questions open-ended or closed-ended? Each type has its own advantages and limitations. As it stands, we have no idea whatsoever as to how the survey was conducted.
As if to insult our intelligence, all that is disclosed is the following information (page 25) IF it is meaningful to anyone: