causation and averages: can't find any reference online

futurewear

New member
Joined
Jul 3, 2007
Messages
9
Can anyone explian what “causation” in statistical analysis is? I have looked online and been unsuccessful.. I need to find what it is so I can describe at least two factors that influence relationships between two variables and can lead to misinterpretation of data analysis.

Then I need to explain when it is appropriate to use averages when computing correlations. and what statisticians should be aware of when doing this. I am just having no luck online.
 
Commonly Confused: Causation and Correlation

Causation implies Correlation

Correlation does NOT imply Causation.

Classic Example: Netherlands Human Births and Number of Storks are correlated. Storks don't cause human births. I hope you are not shocked by that. On the other hand, an increase of human births may require additional housing and this may make available additional warm chimneys which would create more nesting places for storks. The two ARE correlated, and Human births MAY lead to more storks.

Modern Example: Building a casino and local gambling activity are correlated. There is not necessarily causation because of at least two factors. 1) Some people gamble whether or not there is a casino handy. 2) People can choose not to gamble.

Particularly Pertinent Example: U.S. Congress - but probably any sort of pluralist legislative body. Anything that goes wrong is the other guys' fault. Anything that goes well is my fault. This is an excellent example of assumed causation, really, whether or not there is any evidence in either direction.

Odd Example: Stock Markets - Have you EVER heard a market analyst say, "I do not know why the market went up 2 points, today. Sometimes it just does that." There is ALWAYS some reason why the market does ANYthing. If this is so, why can the market not be exploited simply by watching these factors? The factors mentioned by the analysts have some correlation and sometimes make logical sense. Do they always CAUSE the behavior of the market? No.
 
Thanks TKHunny! Do you know what 2 factors I can use? And when it is appropriate to use averages when computing correlations. and what statisticians should be aware of when doing this?
 
Top