The California Occupational Mortality System (COMS) database was used to address the research question of whether different case definitions would result in different descriptive profiles of agricultural occupational deaths. If different definitions would identify different groups of deaths, then this would have significant implications for targeting prevention strategies. The three definitions used were: the place of injury was a farm and the external cause International Classification for Diseases (ICD) code indicated an externally caused death; the injury at work box was checked yes or unknown and an agricultural industry or occupation was listed in addition to an external cause ICD code; or death was caused by an agricultural machine. Of 504 total occupational agricultural deaths identified using the three different definitions, only 36 cases were captured by all three. Farm location identified 258 deaths, agricultural industry or occupation identified 306 deaths, and 104 deaths were identified using agricultural machinery. Analysis of the findings suggested that it may be more important for surveillance purposes to identify representative cases than to identify all cases. If the goal is to target prevention strategies towards the most frequent cause of death, and if only deaths resulting from injuries on a farm are included, then deaths caused by motor vehicles would be ignored. If the goal is to prevent all injury deaths related to agriculture, it may be useful to consider a broad definition of occupational injury when analyzing death certificate data, requiring going beyond using a single algorithm such as the injury at work box, to identify deaths. The authors indicate that their findings illustrate why researchers using death certificates must consider what their research questions are in relation to how they are defining occupational agricultural deaths.