Page 1,  Page 2, Page 3, Page 4, Page 5, Page 6, Page 7, Page 8, Page 9, Page 10, Page 11, Page 12, Page 13

HOME

Meta-Analysis

2 major caveats: (1) Different studies use different methodologies and different populations. Therefore, combining these results while controlling only for one variable (i.e. ETS-related mortality), will bias the results. (2) Because meta-analysis are addressing a set of related hypothesis, there is a tendency to ignore studies that don’t fit the outcome (publication bias)

The Surgeon General’s 2006 report on ETS was a meta analysis that ignored the largest study on ETS related mortality (Enstrom and Kabat (2006).

List of all studies on ETS through 2006:

Number of spousal studies: 81.    

Not statistically significant risk elevation: 69

Not statistically significant risk reduction (protection): 10 

Statistically significant risk elevation: 1 

Statistically significant risk reduction (protection): 1

Only 9 spousal studies had a sample base of over 200 people,  the others ranged from 8 to 191 people.

Number of workplace studies: 31

Not statistically significant risk elevation: 24

Not statistically significant risk reduction (protection): 7

Statistically significant risk elevation: 0

Number of childhood studies: 37.  

Not statistically significant risk elevation: 25 

Not statistically significant risk reduction (protection): 10 

Statistically significant risk elevation: 1

Statistically significant risk reduction (protection): 1

All studies with a 95% confidence level have a 3% margin of error.  Therefore only studies that show an elevation of 3.0 or above are of statistical significance.