Information compiled
- Historic data: The historic data on the running of the bulls from 1980 to 2005 was compiled from newspaper libraries (Diario de Noticias, Diario de Navarra, Navarra Hoy); from the Government of Navarra press office, from SOS Navarra and from The Red Cross. These all have information related to the different kinds of mishaps which have occurred in each running of each day, with information on dates, bull-breeds, the particular stretch of the course, the age of the runner in serious accidents, etc. With the help of this data it has been possible to identity some of the risk factors, such as over-crowding, the breed of the bull, etc.
- Opinion of experts: A panel of experts was selected, among them, runners, first-aid workers, experts in safety and citizen security, from the world of bullfights, etc. A questionaire was drawn up using, among other factors, the methodology of ortogonal experimental design. In this way, on the one hand, a comparison of distinct risk factors was obtained and, on the other hand, a numeric valuation in different situations and settings...
Design of the statistical model and risk estimation
The risk factor was designed by taking the information provided by the experts as a base and by using methodology from the multi-variant statistics. In this design the influence of each factor on the risk of taking part in the running of the bulls is established. To do so, regression models were used with logistical functions which have given reliable estimations for all parameters. The model was checked and it provides satisfactory levels with regard to its prediction capacity and strictness.
Algorithm programming
The prototype of the "Running-of-the-Bulls Test" was programmed in Java language with the JDK 5.0(J2SE 5.0) development kit. Application constitutes an extension of the JApplet class and it uses standard methods of language so that it can work on any outlet with a recent version of JRE (Java run time) from Java.