Analyze the research methodology

Introduction

This work has chosen two research papers to critically analyze the research methodology and method used, as well as to compare and contrast the approaches of the two papers.

The first paper (refer to paper one hereafter) is to contribute the understanding of organizational culture in Turkish construction industry. A case study (Oney-Yazici E., et al, 2007) about “Organizational culture: the case of Turkish construction industry” was conducted to examine the cultural profile of construction organization in terms of firm type, size and age within the context of Turkish construction industry in 15 countries.

The second paper (refer to paper two hereafter) is to demonstrate the situation of claims management in construction section in Egypt from contractor’s perspective. A questionnaire survey (Hassanein A and Nemr W, 2008) of “Claims management in the Egyptian industrial construction sector: a contractor’s perspective” was taken to discuss the issues of claim management, change of order in particular, conducting on a sample collection in construction industry companies.

Research design and research question

Two basic types of research methodologies are divided into quantitative and qualitative methods (Kumar, 2005). Creswell (2009) further advanced three types of designs as qualitative, quantitative and mixed methods.

In order to overcome the disadvantage of qualitative and quantitative research, mixed methods research emerged trying to combine or associate both forms. It involves the use of both qualitative and quantitative methods to achieve greater strength of a study than using either alone (Creswell 2009). A mix method is defined by Crewell (2009) as “an approach to inquiry that combines or associates both qualitative and quantitative forms”. Thus, this method can synthesize either strength or weakness of both quantitative method and qualitative method (Fellows and Liu, 2008).

The research of organizational culture in paper one was conducted with questionnaire, which has Part I and Part II. Part I is mainly categorical or nominal variables (age, gender and size etc) to find out the demographic characteristics of respondents and profile of their firms. Part II comprises of questions with 1-5 Likert scale to measure the organizational culture. Paper one is therefore a straight-forward quantitative research paper, which emphasizes on the quantification the data collection and the subsequent analysis of data (Brymen 2009). The quantitative research was also defined as “a means for testing objective theories by examining the relationship among variables. These variables, in turn, can be measured on instrument, so that numbered data can be analyzed using statistical procedures” (Creswell, 2009).

The aim of paper one was to examine existing culture profile in the Turkish construction industry. Quantitative research is suitable for fact-finding based on evidence or records (Table 1, Naoum 1998). Therefore, the objective of paper one to find out the culture profile can be achieved by quantitative research. Thus it is appropriate in this case. However, if the objective is extended to further understanding as to why such a culture profile existed in Turkish construction industry, a qualitative research could be conducted with methods like interview to gather more data to comprehend the mentality behind the organizational culture, thus to supplement the finding from quantitative research since it has rich and deep data (Naoum 1998), and it places greater emphasis on understanding, rather than merely testing and verification (Ghauri & Gronhaug, 2005).

The objective of the paper two is to finding and improving the status of claims management in Egypt, particularly the change order claims. The research of claims management in paper two is also conducted by questionnaire but it is not the same type of quantitative questionnaire as in paper one. Rather, it has both closed-ended and open questions, hence with combination of quantitative and qualitative questions.. It was actually conducted by researcher in person as an interview, which the researcher admitted it last longer than expected due to gathering of the vast amount of information. A semi-structured interview is believed to be used in this case (Bryman 2008).

The questionnaire in paper two has 7 questions. Questions 1-3 are all closed-ended and are concerned with categorical (Q1) or nominal (Q2 & Q3) variables with purpose to determine the profile claims statement status in Egypt. Questions 4-7, however, are all open questions. Interestingly, question 4 in fact has three questions rather than one question for the remaining part of questionnaire. It therefore can be concluded that the research in paper two is using mixed method research because both quantitative and qualitative data are gathered concurrently and analyzed separately. It occurred to me that the concurrent embedded strategy (Creswell 2009) is used in paper two, which will be discussed later in this article.

In my opinion, mixed method is acceptable to use for paper but I do not like the combination of three different questions as Question 4. These questions are concerned with procedure, communication and documentation. They are interlinked but they could be separately asked. Moreover, questions 5 is asking respondents that “does the contract … stipulate that all change orders must be written?” It is a missed opportunity as it could ask more information from the respondents such as which version/edition of the contract they used, what are the terms and conditions related to the change orders, whether these terms and conditions require all orders to be in written form only.

Data collection and sampling

According to Kumar (2005), the two main types of data collection are primary data and secondary data. Literately, primary data is collected by researchers themselves and secondary data are available data from another source. In both paper one and paper two, primary data were gathered. However, data collection in paper one is purely quantitative with Likert-scale questionnaire while the data collection in paper two are for both quantitative and qualitative data with closed-end questions and semi-structured interview.

In order to gather primary research data, quantitative method combining with structured questionnaire was conducted in both of the two reports. A questionnaire is defined by Kumar (2005) as “a written list of questions, the answers to which are recorded by respondents”. In a questionnaire, respondents read the question and then provide the answers according to the questions. It is clear that a questionnaire is less expensive and easy to understand for the respondents giving the answers. Sometimes, sensitive questions can list on the questionnaire as the respondents can answer in anonymity in less distressed way (Kumar, 2005). However, questionnaires have disadvantages including low response rate, limited to literate population only, biased self-selection of respondents, lack of opportunity to clarify questions to respondents and no allowance for spontaneous responses.

As Bryman (2008) asserted that, the reason why a questionnaire belongs to quantitative method is that most of questions are structured and closed, though it can combine a few open questions. Closed-ended questions refer to “possible answers are set out on the questionnaire, and the respondents ticks the best category the best describes the respondent’s answer”, concurrently open-ended questions refers to “possible answers are not given. In the case of a questionnaire, respondents write down the answers in their own words” (Kumar, 2005).

It is apparent that closed-ended and open-ended questions are presented in paper two of claims management. The questionnaire was formulated based on an extensive literature review of issues on claim management, change order claims in particular. Meanwhile, paper one designed a questionnaire which based on Organizational Culture Assessment Instrument (OCAI) with two major dimensions in which the first dimension emphasizes the organizational focus, whereas the second one distinguishes between the stability and control and the flexibility and discretion, which formulated by clan, adhocracy, market and hierarchy.

All questions in questionnaire of paper one are closed-ended questions which are easier and faster to respond because all alternative answers have been given. However, lack of depth and variety could be a main disadvantage as answers were restricted in a given scope (Kumar 2005).

In paper two, a semi-structured interview in person was conducted with engineers and managers who have more than 15 years working experiences. This interview was based on previous literature in order to get more evidence to examine the claim management in Egypt, change order claims in particular.

Open questions in paper two, with contrast of closed-ended questions used in paper one, could get more in-depth information, and allows respondents to freely express what they intended to say (Kumar 2005, Fellows and Liu 2008). Even closed-ended questions 1-3 in paper two have provided opportunity for respondents to write down their own answers (if other – specify). However, the analysis of data from open questions could be more complex and difficult because respondents can say anything. Meanwhile, the chances are that these questions could have biases from interviewees.

According to Kumar (2005), sampling can be divided into probability random sampling, non-probability non-random sampling and mixed sampling design. Stratified random sampling collection was used in in paper two, and non-random judgmental sampling was employed in paper one. As Kumar (2005) and Fellows and Liu (2008) asserted that there is an equal chance of selection for each member of the population in random sampling. Stratified sampling is one of the probability random sampling and appropriate where the population occurs in ‘distinct’, groups or strata (Fellows and Liu, 2008). In CM, the sample was selected among major companies in construction industry in Egypt and seven of the main companies were collected in this research. Using stratified random probability sampling method, the inferences drawn from the samples can be generalized to the whole population (Kumar, 2005). Judgmental sampling that called Purposive sampling as well, was conducted in Organizational Culture with the criteria of nationality, firm size and market share. Thus, 134 in 351 firms participated in the research study. In judgmental sampling, researcher can gather case and information in those people who meet requirement of the sample chosen criteria, and this sampling method is extremely useful when the researcher want to describe a phenomenon and develop something about which only is a little known (Kumar, 2005 and Bryman, 2008).

However, there can be bias when the sampling was chosen. The random sampling can be influenced by human preference unconsciously. On the other hand, the sample frame may not cover all the features of the whole population. In addition, the sample can not represent all the population, so that the conclusion can only be inferred from the samplings (Kumar, 2005).

In paper one, the population is the whole Turkish construction industry. The stratified random sampling were used to gathered from 826 respondents, who are working for 107 contracting and 27 architectural firms, with male/female ratio of 74.9%/25.1%. The absolute size of sample of 826 is quite impressive and is significant in terms of sample size. The male/female ratio is not balanced, however, it may be the fact that most people working in architectural/contracting industry are dominantly male. The response rate is 38.18%, with 134 participated from 351 firms contacted.

The sample selection in paper two are targeting at the middle to large size contracting firms that are more then LE 50 million in industrial work turnover and LE 100 million in cumulative construction work turnover. Due to limited number of firms fall into the category, seven major companies in Egypt were chosen and 21 industrial projects from them were the sample for the research. However, the actual data collection was conducted by the interview to the senior managers in these firms, but exact number of interviewee was not revealed in the paper. Therefore, when the data analysis present the data as percentage, there is no way for reader to know exactly how many respondents chose a particular answer. My guess is there might be at least 21 respondents since there are 21 industrial projects. If the respondents is only 7 (there are 7 companies chosen), the sample would be questionable due to limited sample size. The research in paper two have omitted small contracting firms and inexperienced site manager/site engineers. Only medium and large contracting firms and hugely experienced personnel are invited for the interview. This may “paint a rosier picture” than the reality. Inexperienced site manager/site engineers tend to be worse in terms of contract administration, understanding procedures, using documentation.

The main aims to select sample are to “achieve maximum precision in the estimates within a sample size and avoid the bias in the selection of sample” (Kumar, 2005). In this case, I believe that paper one has done pretty well to using a large representative sample, while paper two, on the other hand, use a relative smaller number of interviewee and the exact number of interviewee is unknown or unspecified.

Reliability and Validity

Another important process in a social research is the issues of reliability and validity. Reliability means concern with the results can be retested and can get the same result, time after, time in different conditions (Fellows and Liu, 2008). In Organizational culture paper, reliability coefficients (Cronbach alpha) were calculated with 0.89 for the clan and adhocracy cultures, and 0.86 for the market and hierarchy cultures. Both values indicate good internal consistency reliability for the fairness of all culture types because values above 0.7 are considered acceptable and above 0.8 are preferable (Pallant 2007). Meanwhile, despite of the statistical data display, there is no statistical test in the survey of Claims Management paper, hence the reliability of the data is not mentioned.

Validity means “the integrity of the conclusions that are generated from a piece of research”. It is to determine if the research question is properly answered by the research. Internal validity refers to the causality relationship and external validty criterion refers to the generalization from research findings. (Yin, 1994; Amaratunga, 2002). Both papers are trying to generalize the situation. It can be concluded that both research are valid.

Data Analysis

In paper one, the questionnaire consists of 6 relevant questions to key dimensions of organizational culture and each statement has four alternative statements, therefore representing 24 questions. All questions are 1-5 Likert scale, with 5 as “completely true”. The average score was calculated to determine the overall cultural profile of a particular firm. The results of the questionnaire carried out have been presented in table II in paper one.

Independent sample t-tests and Analysis of variance (ANOVA) test was used to examine the differences of organizational culture by firm size and age. A post hoc Scheffe test was used to explore difference among subgroups. The results have been presented in table III.

When it comes to firm type, independent sample t-tests was used on the overall scores of each culture types to compare architectural and contracting firms. in order to have the equal sample sizes (107 contracting firms vesus 27 architectural firms), items were randomly selected from them. An Independent sample t-test is used to compare the mean score for two different groups on same continuous variable (Pallant 2007). The research suggest that “Market” has more influence to Contracting firm than Architectural firm. The average score of randomly selected 32 Contracting firms is 3.37 while the 27 Architectural firms is 2.93, with t=3.849, p

Comparing with Independent-sample T-test which use to compare in two groups and conditions, One-way ANOVA is used to determine whether there are significant differences in the mean scores on the dependent variables between more than two groups (Pallant, 2007). ANOVA test is based on the statistical F-test which is the ratio of the between-groups variance to the within-groups or error variance (Fellows and Liu, 2008). According to Pallant (2007), the significance of F-value indicates the evidence to reject the null hypothesis, which states that the population means are equal. In other words, the larger the F-value is, the more the independent variables influence on dependent variables.

In the Table II of paper one under firm size, the value under four culture types for small and medium firms are all greater than those for large firms. Our of four culture types, three culture types have found that the difference is significant to reject the null hypnoses (p

Paper one further had with k-means cluster analysis to group firms with similar cultural characteristics, together with Ward’s method and squared Euclidean distance. Three underlying patterns of cultural types among sampled firms were found from the results of hierarchy cluster analysis, and this formed as basis for the non-hierarchical k-means analysis. Hierarchical Data clustering algorithms care to find successive clusters with already-established clusters. To select a distance measurement is a vital step in clustering, from which how the similarity of two elements is calculated. The shape of the clusters will be influenced as “as some elements may be close to one another according to one distance and farther away according to another” (Huang 1998, Lu 2004).

In paper two, the data from question 1 about the causes of claims are presented in Figure 1, which is a pie chart showing the percentage of various causes. The result validated with other research which indicated that change order claims is the main reason of claims in most projects while the owner factor ranked second. The result confirmed the necessity to evaluate change order claims in particular (this suggest that the research is valid in terms of the research purpose).

In claims notification status related question, the results showed that notifications were tied to a contract clause in 67% of projects. But the research did admit the interviewee bias and its limitation because respondents tend to answer this question optimistically.

In general claims documentation status related questions, the researcher undertook to categorize the answers into seven groups by using pattern matching for qualitative data analysis, which is one of the most desirable strategies (Yin, 1994). However, this method is under criticism for the subjective risk of interpretive discretion of the researchers (Yin 1994, Amaratunga 2002). This suggests that different researcher may group the data and interpret them differently.

In change order document questions, pie chart is used again in figure 2 to highlight that documentation should be improved to better records management of the project.

The research used explanation building strategy to explain and discuss the results of remaining questions.

The researcher went further from these results to suggest solutions to improve change orders. In the conclusion paper, the research summarized the findings as well as the recommendations based on these findings.

REFERENCE:

• Amaratunga D. et al, (2002) Quantitative and qualitative research in the built environment: application of “mixed” research approach, Work Study Vol. 51 No. 1, pp.17-31
• Bryman, A. (1998) Quantity and Quality in Social Research. Unwin Hyman.
• Bryman, A. (2008) Social Research Methods, 3rd edition, Oxford University Press, Oxford
• Creswell, J. (2009) Research Design: Qualitative, Quantitative and Mixed Methods Approaches, Sage.
• Hassanein A and Nemr W, (2008) Claims Management in the Egyptian industrial construction sector: a contractor’s perspective, Engineering, Construction and Architectural Management Vol. 15 No.5, pp. 456-469
• Huang, Z. (1998). Extensions to the K-means Algorithm for Clustering Large Datasets with Categorical Values. Data Mining and Knowledge Discovery, 2, p. 283-304.
• Kumar R. (2005) Research Methodology: A Step-by-Step Guide for Beginners, SAGE, London
• Lu Y, et al (2004), FGKA: A Fast Genetic K-means Algorithm, in Proc. of the 19th ACM Symposium on Applied Computing, pp. 162-163, Nicosia, Cyprus
• Fellows R and Liu A, (2008), Research Methods for Construction 3rd edition, Wiley-Blackwell, Oxford
• Oney-Yazici E, et al (2007), Organizational culture: the case of Turkish construction industry, Engineering, Construction and Architectural Management Vol.14 No.6, pp.519-531
• Pallant J, (2007), A step-by-step guide to data analysis using SPSS version 15 SPSS Survival Manual Third Edition, Open University Press
• Yin, K. (1994), Cast Study Research: Design and Methods, Sage Publications, Newbury Park, CA.