Investigating Work Related Injuries/Fatalities in regions affected by Hurricane Katrina using Poisson Regression and Negative Binomial Regression Models

Data from the Census of Fatal Occupational Injuries (CFOI), collected by the Bureau of Labor Statistics (BLS), provide information on fatal occupational injuries that occur in the United States. Since 1992, when the BLS first introduced CFOI, BLS has annually reported data on fatal occupational injuries from all 50 states and the District of Columbia. In 2005, BLS reported a total of 5,702 work-related fatal injuries and a rate of 4.0 deaths per 100,000 workers (CDC MMWR, 2007). In this study data was collected at state level from the CDC Wonder, the U.S. Census Bureau and the Bureau of Labor Statistics from 2000 to 2007. The purpose of this study is to investigate work related injuries/fatalities pre and post Hurricane Katrina between the four major states affected compared to the rest of the United States. Variables such as state unemployment rate, state employment rate in all industries, state poverty rate, and population estimate occupation were entered into the Poisson Regression and Negative Binomial Regression Models. The literature on the Poisson regression model often discusses individually specification and prediction of the model and its variants and specification testing in the context of the basic model. The Poisson model arises in many contexts as the probability distribution for the distinct count of the number of occurrences of an event.