Reliability of Data
The initial aim of this study on completed suicides in West Berlin was to check or replicate research findings published mainly in American literature about suicides of physicians, who are supposed to have a higher suicide rate than the average population or other related professions ( Steppacher and Mausner, 1974; Rose and Rosow, 1982). The fact that professional or occupational variables are not included in official statistics made it necessary to use another data reference system, i.e., access to police or prosecution files was needed. Among other variables, these documents include professional or occupational data for coding, whatever the validity and reliability of these data may be. The results reported here are based on differences that were found between official statistics and the prosecutional suicide data collected.
The data pool contained files from the police which were turned over to the prosecution for final legal evaluation, i.e., to exclude the possibility of a crime. These files--ranging from undetermined causes of disease or death; evident suicides and cases officially relabelled later on as suicides; to undetermined homicide; and accidents in the household and on the job--were analysed by three sociologists with the aid of a coding list comprising eighty variables. The research period covered five years ( 1981-5). The average number of analysed files per year was approximately 3800.