In ‘Two Surgeon General's reports on smoking and cancer: a historical investigation of the practice of causal inference’ Emerging Themes in Epidemiology 2006 3, 1-11  
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1343554
Mark Parascandola, Douglas L Weed and Abhijit Dasgupta found that application of the 5 specified ‘criteria for causality’ had been variable and non-transparent – and, more important, different weights were implicitly assigned. The snapshot attempts to capture the authors’ interpretation. Note however that multi-criteria weighting is a matter of value judgments. The Ratings in the snapshot are purely conjectural and an invitation to others to supply evidence-based ones using the alt file here .