Access Type

Open Access Dissertation

Date of Award

January 2012

Degree Type


Degree Name



Civil and Environmental Engineering

First Advisor

Mumtaz Usmen






" Hulya Cakan"

" August 2012"

" Advisor: " Dr. Mumtaz Usmen

" Major: " Civil and Environmental Engineering

" Degree: " Doctor of Philosophy

There are more than nine million construction workers in the US. Roofers and steel workers are the highest risk construction trades according to BLS, and fall from elevation accounts for a large percentage of fatalities and injuries among the construction trades.

In this study, 2114 OSHA accident case reports involving roofers and steel workers were reviewed to identify and analyze the factors contributing to construction fall accidents. Using data for the years between 1994 and 2008, the relationships between these factors were determined and further studied to develop predictive models. Univariate frequency, cross tabulation and logistic regression analyses were used to estimate the effect of the statistically significant factors on the degree of injury (fatality vs. nonfatality)

Chi square tests on the entire data showed that there is a significant relationship between the degree of injury and union status, SIC code, construction operation prompting fall, environmental factor, human factor, project type, construction end use, safety protective system provision, safety protective system usage, fall distance, and fatality/injury cause.

Logistic regression model created for the combined SIC Codes of 1761 and 1791 showed that among the six independent dichotomous variables only four were significantly associated with the degree of injury. These factors were project type, SIC code, safety training and safety protection system usage.

Two separate logistic regression models, one for roofers and another for steel workers were also developed. The roofers' model showed that among the five independent categorical dichotomous variables only three showed significant association with injury severity. These were project type, safety training, and safety protection system usage. The steel worker model showed that only two independent variables had significant association with the degree of injury, and they were union status and project type.

The study showed that cross tabulation analysis and logistic regression modeling can be used for analyzing data on construction fall accidents in a meaningful way, producing useful results.