By Kajal Shah
On Thursday, Sept. 27, Gene Fiorini from Rutgers University presented “CrIME: Criminal Investigation through Math Examination,” a lecture introducing the Center for Discrete Mathematics and Theoretical Computer Science before venturing into some examples of how statistics have helped to solve mysteries recently in the Carlisle forensics center, ironically located in an old prison site.
Fiorini stated that forensics could be associated with a myriad of majors, ranging from medicine to botany to sociology. His path led him to forensics through statistics. Statistics can be applied to blood splatter patterns: by measuring their length and width and performing basic trigonometric functions, the forensics analysts can depict the angle of the blood projection to find if it was passive or projected. Such details are placed together to form a complete image of the crime scene. A simple calculation can connect the stride found in the crime scenes with the gender of that individual.
Statistics has the power to find someone guilty or to free him or her of an accusation. One crime investigation from the Carlisle forensics lab, “The Case of the Maggots on Cocaine,” shows the elegance of statistical analysis.
A woman was found dead at the base of her apartment and authorities could not figure out if she had just slipped down a precarious staircase or been murdered. The size of the maggots found growing in the crevice of her nose gave an estimated time of death.
When her ex-boyfriend came in for questioning, his story of when he had last seen her did not match with the time of death suspected. However, statistical analysis required that they take maggots from not only the nose, but also other parts of the body. It turned out that the woman had been sniffing cocaine so the maggots by her nose were the largest. In other parts of the body, the maggots were much smaller and the new estimated time intervals matched what her ex-boyfriend said. He was no longer a suspect.
Finally, Fiorini discussed the inaccuracies of crime investigations found on television shows. Fingerprint matches are rarely instantaneous matches, usually requiring comparisons with different sampling groups and analyzing variance in data sets. The algorithm used today requires weeks to identify prints, categorized into three patterns: the arch, loop and whorl. He closed by mentioning research opportunities for mathematics students through REU (Research Experiences for Undergraduates). This program will run tentatively from early June to late July and provides a stipend. For more information, visit the National Science Foundation website.