The questionnaire comprises two major sections. The first captures the main
independent variables that examine the factors potentially associated with the
non-occurrence of fraud in organizations. The study’s dependent variable was a
dichotomous variable that took a value of 1 if an organization had never experienced
fraud and 0 otherwise. Factor analysis with principal component extraction using
varimax rotation was first conducted to confirm whether the questions under each
of the four independent variables, namely, “audit committee effectiveness,”
“internal audit effectiveness,” “tone at the top,” and “receipt of qualified audit reports,”
actually measured each of the underlying constructs (Table I) (Field, 2005).
Audit committee effectiveness and internal audit effectiveness were each measured
with five items modified from Leblanc (2007) (Cronbach’s a ¼ 0.93 and 0.90,
respectively). Tone at the top was measured with five items (Cronbach’s a ¼ 0.92) and
receipt of a qualified audit report with two (Cronbach’s a ¼ 0.84). All of the items in
these four variables were ranked on a Likert scale that ranged from 1 (“strongly
disagree”) to 5 (“strongly agree”). Ethical guidelines/policy was measured on a category
scale of 1 (“adhered to by employees”) and 0 (otherwise). Organization employs auditors
with prior success in fraud detection was also measured on a category scale of 1 (“Yes”)
and 0 (“No”), as was Auditor type (1 ¼ Big4 auditor) (0 ¼ Non-Big4 auditor).The second
part of the questionnaire asked the respondents for basic demographic information.
A number of pre-tests were carried out on the questionnaire, and as the results were
found to be satisfactory, ambiguity was not a concern. Hence, the questionnaire is
considered to have a high degree of content validity (Cavana et al., 2001).