|
Eight Rules
1. False Precision 2. Base Rate Fallacy 3. Correlation Not Causation 4. Single Variable 6. Longitudinal Study 7. Different From Chance 8. Dose-Response | |||||||||||||||||||
| -------------> <------------- -------------> |
In the diagram to the left, the arrows originate from, or point to, the brown bar. The purple arrow pointing from Education to the brown bar, then from the brown bar to Life Expectancy, indicates that there is a correlation between Education and Life Expectancy. (If you have a high income you can afford more education; if you have more education you tend to have a higher income.) |
||||||||||||||||||
| <------------- -------------> -------------> | |||||||||||||||||||
| <------------- -------------> -------------> | |||||||||||||||||||
| <------------- <------------- <------------- | |||||||||||||||||||
| <------------- <------------- <------------- | |||||||||||||||||||
There are several other correlations, and all of the factors (access to health care, income, education, and smoking) affect life expectancy. And none of the factors are independent. For example, more income means more access to healthcare. And more income may result from more education, which also tends to mean more access to healthcare, independent of income. These are some of the reasons that precisely determining the magnitude of the effect on smoking on life expectancy is so difficult. It is for these reasons that it is hard to say "smoking cuts life expectancy by 13 years" or a similar statement; the numbers cannot be that precise. |
|||||||||||||||||||
| TOP | Next | ||||||||||||||||||