8
edits
Changes
Jump to navigation
Jump to search
Line 742:
Line 742: + +
Talk:Examples of Bias in Wikipedia (view source)
Revision as of 01:46, March 16, 2007
, 01:46, March 16, 2007→Six times? Not even close.
Regardless of the math, we must remember the stat teacher's old chestnut, "Garbage in, garbage out." Bad data gives bad results. The data from Wikipedia is based on voluntary response samples, which are notoriously prone to bias. (Remember Alf Landon and the ''Literary Digest'' poll?) The six times figure can only be attributed to serious mathematical error or intellectual dishonesty. [[User:John Brown|John Brown]] 21:17, 15 March 2007 (EDT)
Regardless of the math, we must remember the stat teacher's old chestnut, "Garbage in, garbage out." Bad data gives bad results. The data from Wikipedia is based on voluntary response samples, which are notoriously prone to bias. (Remember Alf Landon and the ''Literary Digest'' poll?) The six times figure can only be attributed to serious mathematical error or intellectual dishonesty. [[User:John Brown|John Brown]] 21:17, 15 March 2007 (EDT)
In fact, with the data provided, the correct suggested figure would be '''2.25 times as liberal''', not 6 times. Let (AC,AL) be the percentage of conservatives and liberals respectively of all Americans and (WC,WL) the and = ratio for Wikipedia editors. Since these are probabilities, AC + AL = WC + WL = 1. Then the ratio of liberal Wikipedia editors to liberal Americans is WL/AL = ((WL/(WC + WL))/(AL/(AC + AL))) = (1/(WC/WL + 1))/(1/(AC/AL + 1)) = (AC/AL + 1)/(WC/WL + 1). Since the suggested ratios are AC/AL = 2 and WC/WL = 1/3, then WL/AL = (2 + 1)/(1/3 + 1) = 3/(4/3) = 9/4 = 2.25. [[User:Hdelbrook|Hdelbrook]] 21:46, 15 March 2007 (EDT)