Identifying Correlates of Homicide Rates in Michoacán, Mexico

Author:
Kruger, Carl, Systems Engineering - School of Engineering and Applied Science, University of Virginia
Advisor:
Gerber, Matthew, Department of Systems and Information Engineering, University of Virginia
Abstract:

Violent crime rates in Mexico have reached pandemic levels in recent decades primarily due to corrupt and ineffective government and law enforcement officials as well as violent drug-trafficking organizations (DTOs) and other organized crime groups. The drug war is a large source of violent crimes in Mexico that has claimed over 100,000 lives, according to some estimates. This thesis will analyze social, economic, and demographic data, to determine what factors correlate with homicide rates. Previous research efforts by the Justice in Mexico Project (JMP) organization focused on understanding trends of homicide rates followed by subjective recommendations without any supporting empirical evidence. Using a large database of indicator variables, this thesis will provide a more objective explanation of high rates of homicides. The database consists of census-type data offered by the Information Bank on the Instituto Nacional De Estradistica Y Geografia (INEGI) website. This thesis found that, using linear regression modeling techniques, a stepwise regression model performed best when estimating the homicide rates in the data. The most statistically significant indicators substantiated previous research efforts reinforcing the importance of economic and education reform as well as identifying the importance of social reform through organizations such as Oportunidades. The results developed in this thesis will include the identification of significant correlates of homicide rates in Michoacan, Mexico, which could assist the Mexican government in drafting and implementing reforms in education and economics aiming to decrease homicide rates.

Degree:
MS (Master of Science)
Keywords:
Homicide, Principal Component Regression, Mexico, Linear Regression
Language:
English
Rights:
All rights reserved (no additional license for public reuse)
Issued Date:
2015/04/27