Evaluation of Nine Consensus Indices in Delphi Foresight Research and Their Dependency on Delphi Survey Characteristics: A Simulation Study and Debate on Delphi Design and Interpretation
Open accessThe DOI matches its Crossref record; 8 of the 8 references it cites were found in a registry or an index.
Abstract
The extent of consensus (or the lack thereof) among experts in emerging fields of innovation can serve as antecedents of scientific, societal, investor and stakeholder synergy or conflict. Naturally, how we measure consensus is of great importance to science and technology strategic foresight. The Delphi methodology is a widely used anonymous survey technique to evaluate consensus among a panel of experts. Surprisingly, there is little guidance on how indices of consensus can be influenced by parameters of the Delphi survey itself. We simulated a classic three-round Delphi survey building on the concept of clustered consensus/dissensus. We evaluated three study characteristics that are pertinent for design of Delphi foresight research: (1) the number of survey questions, (2) the sample size, and (3) the extent to which experts conform to group opinion (the Group Conformity Index) in a Delphi study. Their impacts on the following nine Delphi consensus indices were then examined in 1000 simulations: Clustered Mode, Clustered Pairwise Agreement, Conger's Kappa, De Moivre index, Extremities Version of the Clustered Pairwise Agreement, Fleiss' Kappa, Mode, the Interquartile Range and Pairwise Agreement. The dependency of a consensus index on the Delphi survey characteristics was expressed from 0.000 (no dependency) to 1.000 (full dependency). The number of questions (range: 6 to 40) in a survey did not have a notable impact whereby the dependency values remained below 0.030. The variation in sample size (range: 6 to 50) displayed the top three impacts for the Interquartile Range, the Clustered Mode and the Mode (dependency = 0.396, 0.130, 0.116, respectively). The Group Conformity Index, a construct akin to measuring stubbornness/flexibility of experts' opinions, greatly impacted all nine Delphi consensus indices (dependency = 0.200 to 0.504), except the Extremity CPWA and the Interquartile Range that were impacted only beyond the first decimal point (dependency = 0.087 and 0.083, respectively). Scholars in technology design, foresight research and future(s) studies might consider these new findings in strategic planning of Delphi studies, for example, in rational choice of consensus indices and sample size, or accounting for confounding factors such as experts' variable degrees of conformity (stubbornness/flexibility) in modifying their opinions.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Measuring nominal scale agreement among many raters.
Fleiss, J. L. (1971) · Psychological Bulletin · doi:10.1037/h0031619 · Cited by 8,733
High agreement but low Kappa: I. the problems of two paradoxes
Feinstein, A. R., & Cicchetti, D. V. (1990) · Journal of Clinical Epidemiology · doi:10.1016/0895-4356(90)90158-l · Cited by 3,034 · cited 1×
“First, there is debate and uncertainty regarding the calculation of kappa measures as a consensus measure [35, 36].”
BackgroundConsensus methods: characteristics and guidelines for use
Fink, A., Kosecoff, J., Chassin, M., & Brook, R. H. (1984) · American Journal of Public Health · doi:10.2105/ajph.74.9.979 · Cited by 2,089 · cited 1×
“The survey rounds iteratively ask the experts to prioritize the issues or rate them on implementation-related scales such as feasibility or desirability, providing controlled feedback of the previous round’s group results [1, 2].”
BackgroundConsensus measurement in Delphi studies
von der Gracht, H. A. (2012) · Technological Forecasting and Social Change · doi:10.1016/j.techfore.2012.04.013 · Cited by 1,831
Consulting the oracle: ten lessons from using the Delphi technique in nursing research
Keeney, S., Hasson, F., & McKenna, H. (2006) · Journal of Advanced Nursing · doi:10.1111/j.1365-2648.2006.03716.x · Cited by 1,511
Sinha, I. P., Smyth, R. L., & Williamson, P. R. (2011) · PLoS Medicine · doi:10.1371/journal.pmed.1000393 · Cited by 891
Integration and generalization of kappas for multiple raters
Conger, A. J. (1980) · Psychological Bulletin · doi:10.1037/0033-2909.88.2.322 · Cited by 545 · cited 1×
“The different ways of calculating kappa differ mainly in the value of the probability that there is chance agreement, due to different assumptions of distributions of object ratings among all experts [11, 12].”
BackgroundDiversity of decision-making models and the measurement of interrater agreement
Uebersax, J. S. (1987) · Psychological Bulletin · doi:10.1037/0033-2909.101.1.140 · Cited by 254 · cited 1×
“These “Kappa”multi-rater indices of consensus measure agreement between experts over that which would be expected by chance (although the way chance agreement is taken into account in the calculation of kappa indices has been criticized [10]).”
Background
Cited by
Jünger, S., Payne, S. A., Brine, J., Radbruch, L., & Brearley, S. G. (2017) · Palliative Medicine · doi:10.1177/0269216317690685 · Cited by 2,210
Delphi Technique in Health Sciences: A Map
Niederberger, M., & Spranger, J. (2020) · Frontiers in Public Health · doi:10.3389/fpubh.2020.00457 · Cited by 1,072
Use of Delphi in health sciences research: A narrative review
Shang, Z. (2023) · Medicine · doi:10.1097/md.0000000000032829 · Cited by 657
Consensus in the delphi method: What makes a decision change?
Barrios, M., Guilera, G., Nuño, L., & Gómez-Benito, J. (2021) · Technological Forecasting and Social Change · doi:10.1016/j.techfore.2020.120484 · Cited by 377
Artificial intelligence in medical education curriculum: An e-Delphi study for competencies
Çalışkan, S. A., Demir, K., & Karaca, O. (2022) · PLOS ONE · doi:10.1371/journal.pone.0271872 · Cited by 115
“Those with a median score less than or equal to 5, with a small IQR (less than or equal to 2), were considered to attain weak consensus [19].”
Metadata
- Authors
- Birko, S., Dove, E. S., & Özdemir, V.
- Venue
- PLOS ONE
- Publisher
- Public Library of Science (PLoS)
- Published
- 2015-08-13
- Volume
- 10
- Issue
- 8
- Pages
- e0135162
- ISSN
- 1932-6203 (ISSN-L)
- OpenAlex
- W1771332781
- Cited by
- 128
- References
- 8
- FWCI
- 11.07
- Type
- journal article
- Language
- en
- Open access
- gold
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.