Google Search
A more traditional study method was used by the Société de l'assurance automobile du Québec (SAAQ) for their 2003 study. Questionnaires were sent to 175,000 drivers and analysis was done on the 36,078 who responded. The questionnaire asked about driving habits, risk exposure, collisions over the past 24 months, socio-demographic information, and cell phone use. Questionnaires were supported with data from cell phone companies and police crash records. The study found that the overall relative risk (RR) of having an accident for cell phone users when compared to non-cell phone users averaged 1.38 across all groups. When adjusted for kilometers driven per year and other crash risk exposures, RR was 1.11 for men and 1.21 for women. They also found that increased cell phone use correlated with an increase in RR. When the same data were reanalyzed using a Bayesian approach, the calculated RR of 0.78 for those making less than 1 call/day and 2.27 for those with more than 7 calls/day was similar to cohort analysis. When the data were reanalyzed using case-crossover analysis, RR was calculated at a much higher 5.13. The authors expressed concern that misclassification of phone calls due to reporting errors of the exact time of the collisions was a major source of bias with all case-crossover analysis of this issue.

