|
Activity Number:
|
124
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Government Statistics
|
| Abstract - #304221 |
|
Title:
|
Analysis of zero-inflated data from OSHA's worksite inspections
|
|
Author(s):
|
Jia Li*+ and Scott Henn and Aaron Sussell
|
|
Companies:
|
National Institute for Occupational Safety and Health and National Institute for Occupational Safety and Health and National Institute for Occupational Safety and Health
|
|
Address:
|
4676 Columbia Parkway, MS R17, Cincinnati, OH, 45226,
|
|
Keywords:
|
excess zero ; mixture-distribution model
|
|
Abstract:
|
In occupational exposure studies we sometimes observe non-negative outcome measures with excess zeros, which can cause lack of fit of usual parametric distributions. A common approach in such cases is to specify a mixture-distribution model. The Occupational Safety and Health Administration's (OSHA) Integrated Management Information System (IMIS) contains national data from worksite inspections that OSHA conducts during regulatory enforcement. In this study, we examined the change in the average number of violations and average amount of penalties per inspection issued over time from 1979 to 2002. The distributions for both outcomes are right-skewed with 20% and 34% zeros respectively. We compared several modeling strategies and chose zero-inflated negative binomial and lognormal hurdle models on the two outcomes respectively.
|
