Both variations are insightful; I like using Wolfinger's because it is an accurate observation that all data start out as anecdotes, and are influenced/distorted by the collection and filtering process. Still, this also implies that one outlying anecdote is not enough evidence alone to dismiss the aggregate findings across all of the other anecdotes.
http://blog.revolutionanalytics.com/2011/04/the-plural-of-an...
https://web.archive.org/web/20080523225000/http://listserv.l...
Both variations are insightful; I like using Wolfinger's because it is an accurate observation that all data start out as anecdotes, and are influenced/distorted by the collection and filtering process. Still, this also implies that one outlying anecdote is not enough evidence alone to dismiss the aggregate findings across all of the other anecdotes.