CHAPTER 5 RESEARCH METHODOLOGY
5.4 Research Procedure
The two firms selected to participate in this study were Jeffares and Green
Consulting Engineers (Pty) Ltd and Mouchel plc, as discussed in Chapter 2:
Industry Analysis. The two firms were selected based upon a number of factors. Firstly, they are both multidimensional consultancies with interests and operations within the geographies under study. Mouchel plc has wide operational experience in the Middle East, whilst Jeffares & Green is heavily involved with developments in Africa. Mouchel plc is based in the UK and thus represents a developed world perspective, whilst Jeffares & Green is a South African firm and therefore provides information from an emerging market point of view. Lastly, the researcher has been an employee of both firms in the past and is thus intimately knowledgeable of the applicable divisions and staff to be approached in the interests of data retrieval.
The form of research model utilised in this study was defined by Yin (2003) as a “multiple-case replication” case study format. This replication approach is graphically illustrated in Figure 5.4 overleaf. As the model suggests, the theory was first evolved based upon the extensive literature review conducted on the subject, as described in chapters two, three and four. The two participative firms were then selected, based upon their exposure to the target markets, and case studies addressed on an individual basis. Each case study consists of a “whole” study, in which convergent evidence was sought regarding the facts and conclusions for the case, with each case’s conclusion then providing the information requiring replication by other individual cases. The “feedback loop” shown in Yin’s model represents the redesign potential required when unexpected elements emerge
from cases that require adjustment to the original theory, or potential inclusion of additional case studies (Yin, 2003: 49-50).
Figure 5.4: Case Study Method utilised by the Cosmos Corporation.
(Source: Yin, 2003: 50)
5.4.1 The Questionnaire
According to Denscombe (2003), there is no “golden formula” for questionnaire design which, if slavishly adhered to, will ensure success and fend off all potential criticisms. Almost inevitably, the researcher will need to apply discretion, make trade-offs and exercise judgement when producing and implementing a questionnaire. One of the key issues in designing the questionnaire is that it is required to be as crisp and concise as possible, asking just those questions which
are critical to the research. The following principles were recommended by Denscombe (2003: 154), and utilised in formulating the most suitable questions:
− Avoid the use of “leading” questions that suggest an answer or prompt the recipient towards a certain kind of answer;
− Avoid asking the same question twice in different fashions; − Make sure the wording is completely unambiguous;
− Avoid vague questions;
− Give sufficient options in the answer;
− Use only a minimum amount of technical jargon;
− Keep the questions as short and straightforward as possible; − include only those questions vital to the research;
− Co-ordinate numbering of questions and sections in a logical format; − Do not make unwarranted presumptions in the questions; and − Avoid any words or phrases that cause offence.
Two forms of questions prevail in questionnaires, these being “open” and “closed” questions. Open questions allow the respondent to phrase an answer in their own words, deciding on the wording, the length of the answer and the matters to be raised in the answer. Generally these forms of questions are relatively short, whilst the corresponding answers are relatively long. The advantage of open questions is that the information gathered by way of the responses is more likely to reflect the full opinion of the views held by the respondent. These questions do, however,
require more effort on the part of the respondent and provide the researcher with large quantities of raw data requiring analysis (Denscombe, 2003: 155-156).
Closed questions structure the answers by allowing only answers that fit into categories that have been established in advance by the researcher. A form of closed question, known as a Likert scale, allows participants to give more discriminating responses, as well as stating if they have no opinion, by providing them with some form of a rating scale. This turns the question into a statement and asks the respondent to indicate their level of agreement with the statement by ticking a box (Collis & Hussey, 2003: 184).
5.4.2 Testing the Questionnaire
The questionnaire was first tested prior to being released to respondents. This testing was conducted by the researcher’s promoter, who is in possession of a Doctorate degree in Business Administration. Required corrections and additions were made by the researcher, after which final testing was conducted and approval for release obtained.
5.4.3 Layout and Content of the Questionnaire
The questionnaire consisted of two sections, a Demographic Information section requiring input of background data on the respondent, and then a Project Information section which consisted of forty-five “open” as well as “closed” questions. Whilst “closed” questions lend themselves towards more quantitative than qualitative data, their inclusion served to ensure variety in the questioning methodology and thus maintain the interest of the respondents, as well as providing critical data for quantitative analysis.
The Project Information section was divided into five categories, each reflective of the five primary issues emerging from the literature review. Category A consisted of six questions revolving around the importance and impact of energy considerations to projects in Africa and the Middle East. Category B comprised of nine questions, targeting issues surrounding the remaining forms of built infrastructure, such as roads, harbours and airports. Category C had eight questions that queried respondents experience with tax rates and the cost of finance in Africa and the Middle East. Category D consisted of eight questions that evaluated the frequency and extent of corruption in the targeted regions, whilst the final category, Category E’s ten questions, reviewed conflict and political instability. A covering letter outlined a brief summary of the need for the questionnaire, its layout and instructions for its completion.
Initially two respondents were contacted within each of the two target companies, some months prior to the questionnaire being produced, and yielding four respondents in total. When respondents were issued with the questionnaire it transpired that one of the original respondents had left the employ of the company, and thus a different respondent within the same firm was identified as being suitable, contacted by the researcher and agreed to participate in the survey.
The covering letter was emailed to all respondents in .pdf format, whilst the questionnaire itself was emailed in Microsoft Word 2003 format. As mentioned earlier in Section 5.3, a copy of the covering letter is attached in Appendix A, whilst the research questionnaire is attached in Appendix B.