4.2. Procedure
The logistics of this experiment required careful implementation of procedures in order to
reduce threats from internal validity problems such as history and maturation. A preliminary
questionnaire was given to the partners immediately after field interviews, for their com-
ments and suggestions. The experimental instrument was amended to incorporate some of
the partners’ comments. A further interview was conducted with the partners (and their sec-
retaries) to co-ordinate the administration, distribution and collection of the questionnaire.
The coordinator, generally the partner’s secretary, was given an explanatory letter on the
procedures and subjects required for the experiment. For example, the researchers requested
a representative number of subjects from the five hierarchical levels and the different divi-
sions of their firms. The completed questionnaires were collected by the co-ordinator shortly
after distribution. Partners involved in coordinating the questionnaires were requested not
to participate in the study nor advise subjects about the contents of the instrument.
Controls for internal and external validity threats were included in the instructions. For
example, the participant was requested to complete the questionnaire within ten minutes.
To ensure anonymity and confidentiality, each respondent was provided with a self-sealing
envelope and requested to seal it upon completion. The completed questionnaires were
personally picked up within three weeks of delivery to the firms. An overall response rate
of 93% was achieved with 183 useable questionnaires.
4.3. Research instrument
The questionnaire was adjusted after qualitative interviews with male and female
accounting partners in Singapore and Australia. It comprised of demographic information
of the respondents: age, education, marital status, gender, dependent children, the numbers
of years for the respondent to attain the current organizational level, and the number of years
as a professional member. Further, organizational culture values (O’Reilly et al., 1991) of
the three participating multinational accounting firms were measured to ensure cultural
homogeneity in relation to gender oriented hypotheses. We intend to diffuse the effect of
organizational culture through the principle of matching (Iselin, 1994). Chatman and Jehn
(1994) found that the then Big 8 accounting firms had no significant differences based on
the O’Reilly et al. (1991) measure (see Appendix A).