3. CORRELATION IS NOT CAUSATION
                                       
    Eight Rules

1. False Precision

2. Base Rate Fallacy

4. Single Variable

5. Substantial Differences

6. Longitudinal Study

7. Different From Chance

8. Dose-Response 		 
  • If A and B occur together (are correlated)

          —   A may cause B

          —   B may cause A

          —   C may cause both A and B

          —   Cannot automatically conclude any of these

    •  
     
    BROKEN HOMES CAUSE DELINQUENCY ?
  •   In 1950s researchers noticed a correlation
  •   Not necessarily cause-and-effect
          —   Conflict between spouses re childrearing
                  — May cause delinquency
                  — May cause divorce
                  — C causes both A and B
          —   Conflict over dealing with delinquent child
                  — May cause divorce
                  — B causes A
    		•  
      TOP
     
    EDUCATION
    SMOKERS
    LESS THAN HIGH SCHOOL
    24.7%
    GED
    41.9%
    HIGH SCHOOL GRADUATE
    23.1%
    SOME COLLEGE
    20.9%
    ASSOCIATE DEGREE
    17.9%
    UNDERGRADUATE DEGREE
    9.1%
    POSTGRADUATE DEGREE
    5.9%
          Reference
     
    POVERTY
    SMOKERS
    BELOW POVERTY LEVEL
    27.9%
    AT OR ABOVE POVERTY LEVEL
    17%
        TOP
     

  •   Does educational attainment affect smoking?

  •   Or does smoking affect education?
          —   Problems with authority/disdain for education
                  — Can lead to dropping out
                  — Can lead to smoking
                  — C causes both A and B

          —   Cost of smoking
                  — Might affect education
                  — B causes A

  •  Connection between poverty and smoking
          —   Cost of smoking (A causes B)

  •   Education (C causes both A and B)
          — Could affect smoking
          — Could affect financial status

  •   Could poverty cause smoking?

    •  
    LESSONS LEARNED WHEN COMPARING DIFFERENT COUNTRIES
  •   Highly unlikely just one (important) factor

  •   If there are multiple important factors
          — Easy to be misled if ignore them
          — How to know which are important?

  •  May have to examine other studies
          — To get an idea of which factors to examine
          — To avoid being fooled

    •   TOP
                                           
     
    COUNTRY
    SMOKERS(%)
    LIFE EXPECTANCY (YRS)
    RANK
    		  
     
    AUSTRALIA
    19
    82.07
    10
    		  
     
    AUSTRIA
    45
    80.17
    32
    		  
     
    BELGIUM
    22
    79.92
    37
    		  
     
    CANADA
    17
    81.67
    14
    		  
     
    DENMARK
    28
    79.09
    47
    		  
     
    FINLAND
    22
    79.69
    41
    		  
     
    FRANCE
    27
    81.66
    15
    		  
     
    GERMANY
    25
    80.44
    28
    		  
     
    ICELAND
    21
    81.22
    20
    		  
     
    ITALY
    19
    82.03
    11
    		  
     
    NETHERLANDS
    26
    81.12
    22
    		  
     
    NEW ZEALAND
    24
    80.93
    26
    		  
     
    NORWAY
    28
    81.60
    17
    		  
     
    SWITZERLAND
    19
    82.39
    8
    		  
     
    UK
    23
    80.42
    29
    		  
     
    USA
    25
    79.56
    42
    		  
      TOP Next