Data are from the US FARS (Fatality Analysis Recording System) archive that is intended to include every accident in which there was at least one fatality. Data are limited to vehicles where the front seat passenger seat was occupied.
FARSA data frame with 153338 observations on the following 17 variables.
caseida character vector: identifies the vehicle
statea numeric vector. See the FARS website for details
agea numeric vector; 998=not reported; 999=not known
airbaga numeric vector
injurya numeric vector
restrainta numeric vector
sex1=male, 2=female, 9=unknown
inimpacta numeric vector
modelyra numeric vector
airbagAvaila factor with levels no yes NA-code
airbagDeploya factor with levels no yes NA-code
Restrainta factor with levels no yes NA-code
D_injurya numeric vector
D_airbagAvaila factor with levels no yes NA-code
D_airbagDeploya factor with levels no yes NA-code
D_Restrainta factor with levels no yes NA-code
yearyear of accident
Data is for automabiles where the right passenger seat was occupied, with one observation for each such passenger. Observations for vehicles where the most harmful event was a fire or explosion or immersion or gas inhalation, or where someone fell or jumped from the vehicle, are omitted. Data are limited to vehicle body types 1 to 19,48,49,61, or 62. This excludes large trucks, pickup trucks, vans and buses. The 2009 and 2010 data does not include information on whether airbags were installed.
http://maths-people.anu.edu.au/~johnm/nzsr/taws.html
Olson CM, Cummings P, Rivara FP. 2006. Association of first- and second-generation air bags with front occupant death in car crashes: a matched cohort study. Am J Epidemiol 164:161-169
Cummings, P; McKnight, B, 2010. Accounting for vehicle, crash, and occupant characteristics in traffic crash studies. Injury Prevention 16: 363-366
Braver, ER; Shardell, M; Teoh, ER, 2010. How have changes in air bag designs affected frontal crash mortality? Ann Epidemiol 20:499-510.