Thread by @EvaWoodwardPhD

Overview thread of the Delphi technique, its methods and rigor, ethical concerns, and some darn good examples - I've been trying to learn and wanted to share. Please add resources as you wish 🧵 #AcademicTwitter #ImpSci #PhDChat

Delphi is a method of developing consensus on a topic or forecast a future pattern from experts on those topics // Multiple rounds of voting anonymously, feedback groups' response to all. Link to @RANDCorporation that developed it: rand.org/topics/delphi-…

Used in research and communities with controversy, some say it erroneously guides people to "group think" - some of this is discussed by Keeney et al: "consensus from a Delphi process does not mean that the correct answer has been found." pubmed.ncbi.nlm.nih.gov/11223060/

Uses quantitative ratings to determine consensus typically. Nice overview of methodological rigor, tradeoffs, and things to consider for: sampling experts, anonymity, how broad or narrow to be in each round of iterative feedback by Keeney et al: pubmed.ncbi.nlm.nih.gov/11223060/

Example of one group that used only qualitative consensus and shared a thoughtful study of Delphi with multilevel leaders in healthcare systems to protect confidentiality of people who don't want their opinions to be known or want added privacy journals.sagepub.com/doi/full/10.11…

Step by step of how another group did Delphi virtually, including a list of choice points to need to make if you are planning a Delphi method, like: what is consensus? who are really experts? how do we keep them engaged in multiple rounds of voting? pubmed.ncbi.nlm.nih.gov/32539717/

Other threads on Twitter have generated visual figures, strengths/challenges from @HPJournalClub

https://twitter.com/HPJournalClub/status/1245155769819029504

And for those of us in #ImpSci, an example of a modified Delphi from an article many have read - ERIC project to determine one taxonomy of implementation strategies …plementationscience.biomedcentral.com/articles/10.11…

Ironically, after this learning, I've realized this is not the method I need - ha! I share this for others looking for resources (I am not experienced, so can't field advanced questions). Many were sent to me by a fellow implementation scientist, @FemkeVanNassau - thank you!