The Delphi method is an iterative process for consensus-building among a panel of experts who are anonymous to each other. Delphi is advocated as a means to achieve an optimally reliable expert consensus (Dalkey, 1963). Its primary original and current uses center on one of three objectives:
(1) forecasting future events ("exploratory Delphi");
(2) achieving policy consensus on goals and objectives within organizations or groups ( "normative Delphi"); and
(3) identifying diversity in and obtaining feedback from stakeholders in some policy outcome ("focus Delphi").
The term "policy Delphi" may include any of these three dimensions (Turoff, 1970 ; Adler & Ziglio, eds., 1996; Tapio, 2003).
In spite of its policy focus, the Delphi method also has applications in quantitative research and this, not policy Delphi, is the focus below. These applications include aiding in reaching consensus on the composition of survey or other data-collecting instruments, reaching consensus on the coding of phenomena, and reaching consensus on research agendas for given areas of study. Expert consensus may also be a form of validation for otherwise subjective judgments by the quantitative researcher, such as for determination of proper labels to infer from loadings in factor analysis, principal components analysis, structural equation modeling, partial least squares modeling, and other statistical procedures involving latent variables inferred from measured indicator variables.
In its traditional form, extensive questionnaires are distributed to the panel of experts, responses are synthesized and then used as feedback to the panel in the next round of questionnaires, for a series of rounds. Experts on the panel typically do not communicate directly with each other but rather only provide responses to the Delphi administrator.
The method was first developed in the 1950s by the RAND Corporation for the military, which wished to forecast the effects new military technology might have on the future of warfare. In early applications, Delphi was often accomplished through postal mail. Today, however, Delphi is independent of the mode of administration and examples exist using email, interactive online survey methods, and even face-to-face meetings.
The full content is now available from Statistical Associates Publishers. Click here.
Below is the unformatted table of contents.
THE DELPHI METHOD IN QUANTITATIVE RESEARCH Table of Contents Overview 4 Key Concepts and Terms 5 Appropriateness 5 Rounds 5 Forecasts vs. reasons 5 Legitimation 5 Anonymity 6 Examples of Delphi in quantitative research 6 Defining constructs 6 Developing and selecting indicators 7 Creating a multidimensional instrument 7 Creating a typological framework 8 Coding 9 Establishing the construct validity of factor weightings 9 Integrating factor eigenvalues and Delphi weights for purposes of optimal scoring for selection among alternatives 10 Adjusting statistical forecasts 11 Assumptions 12 Representative selection of experts 12 Avoidance of researcher bias 12 Intermediate level of predictability 12 Frequently Asked Questions 13 How many experts should serve on a Delphi panel? 13 How are experts selected? 13 How many rounds should there be in a Delphi process? 14 How long does the Delphi process take? 14 What if no consensus is reached? 15 Is the Delphi administrator always a neutral facilitator? 15 Bibliography 15 Pagecount: 20