Art of Deception

Art of Deception (continued)
"Refutation of Statistics" p 128


Evidence may be false
Use of statistics means that you cannot get at something directly.

For statistics to mean anything they must be representative and random.  One of the worst polls ever done was the presidential election of 1936, when a sampling from Oshkosh predicted that FDR would lose massively.  This was absolutely wrong because of the way the poll occurred.

Declare the other guy's stats wrong.
Question the methods and strategies used to get the information.
Use other statistics that conflict.
Do they measure what the speaker says?
Use the stats with an ad hominem attack.
Attack stats altogether.
Can't predict the unique from the general.
Extrapolationis too risky to make conclusions
Gambler's fallacy.

The thinking skews the math.
Who is more likely to win?

Statistics are too easily twisted.  They begin to collect all kinds of emotional elements that do not believe, and they are too easily applied to unique cases.  In fact their usefulness is hooked to science.  Because there is a 50% chance of something will happen does not equate to the possibility of its occurring in real life.