(5 reader(s) have rated this comic 4.8 out of 5)
Loading ...
Playing the Percentages
Friday, January 12th, 2007 by Admiral_CoeymanIf I came to you with an investment that had a 30% chance of making a profit, would you bet your mortgage on it? Could you spare a kidney for a 70% chance of failure? While I understand that 30% is statistically significant, it is not a sure thing. A sure thing requires an excuse when it doesn’t happen.
Why, then do we accept that something which increases risk by less than the odds you get flipping a coin proves that it causes a disease? I would want odds of at least 96% before I stopped looking for the actual cause. Less than 50% and you are at the beginning of the right track. You have a long way to go from there.
I will not argue that a statistically significant increase in risk is nothing. Okay, there is something to it. What I am writing about is the implication that statistiacally significant is the same thing as definite. Why do we accept this causal relationship on so little evidence?
If something that I do can increase my risk of getting one of a thousand diseases, then you have not found the actual cause for any of these links. A causal relationship is one on one. I do this and this is the result. Keep looking until you find this relationship. It is important. Blocking the action that can lead to 1000 things only masks the actual causes of those thousand things.
Stand Up philosophers are not afraid to state the unpopular truth that everybody else is happy not to look at. That is not where I am sitting right now. I honestly want to know why we play such low odds as definitive.
