2,002
edits
Changes
Talk:Examples of Bias in Wikipedia (view source)
Revision as of 04:13, April 5, 2007
, 04:13, April 5, 2007Bad data
Forgive me if I am disinclined to trust the personal opinion of a man who's publically gloats about his utter inability to understand basic mathematics. --[[User:RexMundane|RexMundane]] 23:15, 4 April 2007 (EDT)
Forgive me if I am disinclined to trust the personal opinion of a man who's publically gloats about his utter inability to understand basic mathematics. --[[User:RexMundane|RexMundane]] 23:15, 4 April 2007 (EDT)
== Bad data ==
The discussion about the six times more liberal is going strong again. But lets organize the discussion. There are two problems. First the data that is used is corrupt. Second, the definition of Liberal bias is nonsensical.
First "Bad Data"
The Wikipedia bias is computed with data, obtained from this page:[http://en.wikipedia.org/wiki/Category:Wikipedians_by_politics | Wikipedians by politics].
::* It relies on raw data from a self-selected sample.
::* You compare the outcome of a poll with 3 possible answers, to one with 131 different answers.
::* You compare the outcome of a poll where you can only choose one option, with one where you choose as many options as you like.
::* Wikipedia is not measured by the Wikipedians, but by the articles that Wikipedians write.
Each of the first three issues make any conclusion drawn from this data set useless. As someone said, counting bumper stickers on your way to work would be more reliable. The last mistake to measure Wikipedia by its editors is easily made, but also easily avoided. Wikipedia gives you a tool to select a random sample of articles, which you could use to determine bias. [[User:Order]] 5 April 2:15 (AEST)