Performance of Safety Incidents Statistical Analysis of Safety Incident Rates Table of Contents Introduction3 Part I. Graphical Descriptive Statistics3 Part II. Binomial Probability Distribution4 Part III. Inferential Statistics5 Part IV. One Sample Hypothesis T-test5 Part V. Two Sample Hypothesis T-test6 Part VI. Paired (matched) Observation – Two Populations Hypothesis6 Part VII. Linear Regression and Correlation Study7 Part VIII. ANOVA – One-Way Test of Variance7 Part XI. Chi-square Goodness-of-Fit Test7
Part X. Chi-square Independence Test7 Conclusion7 Appendix A:9 Introduction The measurement of safety performance is a hot topic in most companies today. Safety performance measurements seek to answer such questions as how do we compare to others? Are we getting better or worse over time? Is our management of safety effective (doing the right things)? Safety differs from many areas measured by managers because success results in the absence of an outcome (injuries or ill health) rather than a presence.
As the scope of safety is vast, many stakeholders who are not safety professionals do not see safety as in such a broad way. They see “no injury” as good and “injury” as bad and that where it ends. This is one reason why incident rates and worker’s compensations costs have moved their way to the forefront of safety metrics. As an employee of a Facilities Engineering Command (a federal agency), working in the Operations Division, my role as an analyst requires me to gather and review metrics.
Our safety metrics are very important to our organization as the safety and health program is dedicated to establishing a safety conscious culture throughout the command, ensuring that safe facilities are designed and constructed, ensuring that facilities and equipment are maintained in a safe, operating condition, eliminating facility, public works, and contact-related injuries and illnesses. In this paper will provide a statistical analysis of safety performance incident rates at my organization. These data will describe the number of major incident rates, i. . material movers/laborers incidents, and truck drivers’ incidents. The data in appendix A represents the number(s) of major incidents for 2008 for an organization of 325 employees. Part I. Graphical Descriptive Statistics Table 1 |Sample Size |30 | |Mean |29. 96 | |Median |26. | |Mode |21 | |Standard Deviation |15. 921 | |Variance |253. 81 | |Range |53 (63-10) | |Min |10 | |Max |63 | |25th Percentile |7. 5 (between 15 and 16 [15. 75]) | |50th Percentile |15. 5 (between 26 and 27 [26. 5]) | |75th Percentile |23. 25 (between 41 and 46 [43. 25]) | Figure 1: [pic] Table 2: Stem and Leaf 1 |001224569 | |2 |111466799 | |3 |2467 | |4 |167 | |5 |126 | |6 |13 | Figure 1 shows that the data is skewed to the right.
Distribution is greatest in the 10-29 incident range and at its lowest in the 59-69 range thus indicating that there are more incidents at the lower safety incident range. Part II. Binomial Probability Distribution According to the Bureau of Labor Statistics, occupational injuries and illnesses requiring days away from work for state and local government, there were State government workers sustained occupation injuries and illnesses at an incident rate of 170 cases per 10, 000 full-time workers and required a median of 8 days away from work to recuperate.
Suppose the probability of having fewer than 10 incidents in 30 days were . 33. Solution: There is a 99% probability that fewer than 10 incidents will take place in 30 days. | # of Trials (sample sz) |10 | |Binomial “successes” |9 | |Probability |0. 5 | Table 3 Part III. Inferential Statistics
Occupational Safety and Health Administration (OSHA) has published the population mean for incident rates. The 95% level of confidence, the true population mean number of incident rates lies between 24. 263 and 35. 657. The 90% level of confidence shows the population mean number of incident rates lies between 25. 164 and 34. 756. The supporting statistics are presented below. Sample Size: 30df = 30-1=29 Sample Mean: 29. 96 Sample SD: 15. 921 Confidence level P
