*Specificity of Association

Since you asked...

There are a few rules to follow when trying to prove causality.  Say, for example, I observe that it rains every time I wear my pink suede shoes. 

Is there a causal link between wearing pink suede shoes and the weather?  

This guy, Sir Austin Bradford Hill, put together a list of associations that should be checked when trying to prove causality.   These rules (they're called Bradford Hills Criteria, btw) are:

1. Temporality
• how quickly after I put the shoes on does it start to cloud up?
• This is not a big requirement so much anymore, either, depending on what is under discussion.  The temporal association between smoking and lung cancer, for example, is so disparate to be almost nil.
2. Strength
• a mathematical measurement of correlation
3. Biologic gradient (aka dose-response relationship)
• does it rain all day if I wear the shoes all day?  does it only rain for half the day if I wear one pink suede shoe and one sneaker?
4. Consistency
• does it happen every time or only some times?
5. Biologic plausibility
• This would be something like "the color pink is on a light frequency that attracts water droplets and suede makes water droplets condense," e.g., a scientific reason for why pink shoes make it rain
6. Specificity
• does it always rain or does it sometimes get foggy or even snow?
• This is considered the weakest criteria b/c there are lots of cause and effects that don't seem to make any sense, especially in biology.  For example, a concussion will make you cranky (i.e., physical injury results in behavioral symptoms)
7. Experimental evidence (aka effect of removing the exposure):
• does it stop raining if I take them off?
8. Alternate hypothesis or analogy
• is there any other possible explanation?
9. Coherence
• is there anything else like this in the known universe that operates similarly?