From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.5-pre1 (2020-06-20) on ip-172-31-74-118.ec2.internal X-Spam-Level: X-Spam-Status: No, score=1.0 required=3.0 tests=BAYES_40,TO_NO_BRKTS_PCNT autolearn=no autolearn_force=no version=3.4.5-pre1 Date: 2 Aug 93 13:32:11 GMT From: tis.com!mjr@uunet.uu.net (Marcus J. Ranum) Subject: Re: Gauntlet gathering rust and dust Message-ID: <23j50r$d9b@sol.TIS.COM> List-Id: stt@spock.camb.inmet.com (Tucker Taft) writes: >It is generally a mistake to use responses in a network bulletin board >as any statistical indication of anything. There are many people who >simply don't read comp.lang.. [...other reasons...] The best reason not to take responses on USENET as statistical indicators of anything is because they're a self-selected sample. Such samples have virtually no statistical relevance, because they often amount to polling the choir. For example, what is the statistical validity of a self-selected sample of the NRA's voting membership, polled on their feelings about gun control? (both NRA and HCI play this game, but mostly as a fund-raiser: "Surprisingly, 95% of our membership agreed with our basic organizational purpose!!!") More often, they're skewed because the only folks who respond are the folks who feel strongly about the issue in the first place. This is basic stats 101 stuff. mjr.