Application of Delphi Method in Construction Engineering and Management Research: A Quantitative Perspective

Abstract

Abstract. The Delphi method has been used as a main research method by a growing number of researchers in the Construction Engineering and Management (CEM) field in the past two decades. Although a number of studies are available on the use of Delphi, few researchers fully examine the potential of the Delphi method in the combined use of statistical techniques, which is an inevitable trend for future Delphi research. This paper aims to review the combined use of Delphi and other quantitative methods in the CEM field based on a structured literature review of 88 relevant papers. All of the 88 papers are systematically identified from ten well-known peer-reviewed CEM journals published in the period of 1990-2012. Topic coverage, application requirements, and statistical techniques in the 88 Delphi papers are reviewed. The mix use of the Delphi method with three advanced modelling methods, such as Fuzzy Sets, Analytical Hierarchy Process, and Analytical Network Process is also examined. These review results provide practical references for researchers having interests in applying Delphi method in CEM research.Keywords: Delphi method; Construction engineering and management research; Quantitative perspective; Review IntroductionThe Delphi method is a structured communication and consensus building approach amongst a group of experts on a complex problem (Chan et al. 2001). This method refers to an iterative process where consensus is often reached through rounds of feedbacks of experts' opinion and judgment on a particular subject (Hallowell and Gambatese 2010b). Although sometimes the reliability of the findings derived from a Delphi study may raise some controversy because of the inappropriate design and execution of the Delphi study, such as shortcomings of the survey instrument, poor choice of experts, weak bias control, unreliable analyses, and limited feedback during the study (Gupta and Clarke 1996;Keeney et al. 2001), the Delphi method remains a particularly useful alternative for the situation when objective data are unattainable, there is a lack of empirical evidence, or experimental research is unrealistic or unethical (Hallowell and Gambatese 2010b).As an established profession in the construction industry, CEM is a practice-driven field in nature. Many research questions in this field need to address the impacts of individuals, organizations and the society on construction management activities, particularly those on planning, forecasting and decision making activities (Fellows and Liu 2009). The key to resolve these problems should draw upon the collective knowledge and experience of selected experts in a given area. By contrast to other methods such as interviews, Delphi provides a more reliable and efficient alternative for solving these problems with high uncertainty (Chan et al. 2001). Therefore, a growing number of researchers have adopted Delphi method in CEM research since the early 1990s (Hallowell and Gambatese 2010b).Although many researchers regard Delphi as a qualitative method (Hasson et ...

An Exploratory Analysis of Risks in Green Residential Building Construction Projects: The Case of Singapore

“…To look into possible differences in RC values between green and traditional residential building construction projects, the Wilcoxon signed-rank test was conducted. This method is a non-parametric statistical test to compare two sets of scores that come from the same participants, without requiring the data must be normally distributed [51]. In this study, test results showed that the p-values of 35 risks were less than 0.05, suggesting there were significant differences in RC values of the most risks between green and traditional residential building construction projects.…”

Consensus Measure with Multi-stage Fluctuation Utility Based on China’s Urban Demolition Negotiation

“…However, the key concerns for the moderator focus on two aspects: (1) what extent the consensus can really reach up to, namely the consensus level is the actual issue the moderator cares about; and (2) how much the moderator needs to cost for reaching a consensus, i.e., the negotiation cost. Available research about consensus measure, including qualitative judgment methods (Skinner et Akiyama et al 2016), mainly focused on a single distance-based measurement of the consensus level.…”

Demand-Oriented Design Strategies for Low Environmental Impact Housing in the Tropics

“…Each item of the questionnaire is assessed by an evaluator team with a minimum value representing the evaluators' most "conservative cognitive value" (C i ) of the item, and a maximum representing the most "optimistic cognitive value" (O i ). [63,64], is used in FDM, which is an attractive characteristic of the method. With a limited sample size, FDM can build consensus despite different perspectives from the participants while obtaining a reliable and objective opinion.…”

“…Repeat the calculation for the tree scores of ", i.e., minimum ( ), geometric mean ( ), and maximum scores ( ). (3) Examine whether evaluators' judgments converge using A small sample size, mostly between 15 to 20 [63,64], is used in FDM, which is an attractive characteristic of the method. With a limited sample size, FDM can build consensus despite different perspectives from the participants while obtaining a reliable and objective opinion.…”

“…The questionnaire aimed at nuclear and two-generation families in northern, central, and southern Taiwan, where 17, 10, and 15 questionnaires were distributed, respectively. Multi-criteria decision analysis (MCDA), such as the Fuzzy Delphi method and ANP, require a smaller sample size, which is typically between 15 and 20 respondents [63,64], while still deriving objective and reasonable results. This method is not only time-and cost-saving for collecting expert opinions but also sufficient for obtaining experts opinions without distortion [61,65,66].…”