• No results found

Chapter 5 COMPLEXITY AND MANAGEMENT

6.7.7. Quality of the Quantitative Research

The principles of reliability and validity are widely used by positivists in their scientific methods. These illustrate the quality of a quantitative study according to the accuracy of measurements. Whenever a measuring device in any form is utilised as a part of the data collection process, the validity and reliability of the measurement is important. This section evaluates the reliability and validity of the result of the quantitative inquiry.

6.7.7.1. Reliability

Reliability refers to ―the consistency of a measure of a concept‖ (Bryman & Bell, 2003, p. 76). The consistency of findings enables them to be replicated (Burns &

Burns, 2008). A postal questionnaire is a reliable method, because it is easy to standardise questions and structure the questionnaire in a way that all respondents answer exactly the same questions (Collis & Hussey, 2009). The questionnaire used in this study consisted of closed questions which reduces the risk of misinterpreting of answers by the researcher. Furthermore, its standard structure makes it replicable and helps other researchers to repeat the study.

To test the reliability of the questionnaire, Cronbach's α (alpha) method is used. It is a test commonly utilised by the quantitative researcher to figure out internal reliability. The concern of internal reliability is ―whether respondents' scores on any one indicator tend to be related to their scores on the other indicators‖ (Bryman &

Bell, 2003, p. 76). Cronbach's α (alpha) method ―calculates the average of all possible split-half reliability coefficients‖ (Bryman & Bell, 2003). The value of 0.80 for alpha is typically considered as an acceptable level of reliability (Bryman & Bell, 2003).

To calculate alpha, the reliability analysis feature of SPSS is used. All Likert rating and multiple choice questions are considered for the analysis. The calculated

Chapter 6

METHODOLOGY

155 Cronbach's α is 0.805. George and Mallery (2003, p. 231) provided the following rule of thumb for Cronbach's alpha: α > .9: Excellent, α > .8: Good, α > .7:

Acceptable, α > .6: Questionable, α > .5: Poor, and α < .5: Unacceptable. Based on George and Mallery‘s rule of thumb, the reliability of the questionnaire is good.

6.7.7.2. Validity

The validity of the results can be discussed only when the results are reliable.

Validity distinguishes ―whether a measure of a concept really measures that concept‖

(Bryman & Bell, 2003, p. 77). There are two types of validity: (a) internal and (b) external. Internal validity is ―the degree to which the conditions within the experiment are controlled, so that any differences or relationships can be ascribed to the independent variable, and not other factors‖ (Burns & Burns, 2008, p. 427).

External validity is ―the extent to which the results of a sample are transferable to a population‖ (Burns & Burns, 2008, p. 426). In contrast to reliability, validity is not evaluated numerically and should be explained based on the way the research is carried out.

The research has internal validity because the sample size selected is large enough to study the population. Furthermore, the selection of measurement items was based on the qualitative data analysis (QDA) and this ensures a high level of congruence between concepts and measurements. Moreover, the positive feedback received from the two-steps pilot test of the questionnaire ensures that the measures can assess the concept that is focus of attention.

To make the research externally valid, the methods used were described comprehensively. Furthermore, the purpose of each question was explained and the variables associated with that question were defined (Appendix five). In terms of selected sample, the process of random sampling was expressed in detail. This sampling method ensures that the sample is a representative of the target population.

Moreover, the high response rate achieved minimises the risk of bias in the sample.

Also, the demographical information of the sample is provided which illustrates the characteristics of the sample and the extent to which the results can be generalised to wider contexts.

156 6. 8. Conclusion

This chapter described the way that the research was conducted. First, the philosophical position of the research clarified that it has a constructivist ontology and interpretivist epistemology. Then, qualitative and quantitative inquiries were expressed as two approaches to the research. It was also indicated that this research intended to mix qualitative and quantitative methods and take advantage of methodological pluralism. Hence, the complementarity approach was used, which allows the quantitative methods to complement the results of the qualitative investigation.

The next section of the chapter was assigned to qualitative inquiry. All theoretical and practical matters of the research design, data collection and analysis were elucidated. It was explained that twenty four interviews were conducted with experienced Iranian construction practitioners. The interview information was transcribed into NVivo for qualitative data analysis. The QDA included coding the interview transcriptions, identifying categories and subcategories (nodes), seeking patterns, and finding relationships among the logistics system‘s agents. At the end, the quality of the qualitative inquiry was evaluated and the advantages and disadvantages of qualitative methods were discussed. The results of the QDA are presented in the analysis chapter and are used to create a model for managing construction logistics in projects.

The final section of this chapter described the process of quantitative inquiry. In this section, how quantitative methods were utilised in the research was stated. The emerging variables from the QDA were introduced and the methods measuring these variables were clarified. Practical matters, including questionnaire design, questions‘

logic, pilot testing and sampling, were considered in detail. Then, the process of distributing the questionnaires and inputting the data into SPSS was explained.

Among the population of 794 Iranian building contractors (grades one to three), 287 companies were selected randomly for completing the questionnaire. Among those, 135 questionnaires were returned, which indicates an acceptable response rate of 47.03 per cent. At the end, the reliability and validity of the quantitative inquiry was discussed. The result of the quantitative study is presented in the analysis chapters, wherever appropriate to complement the results of the QDA.

157

Chapter 7