<%BANNER%>

The Owner's Role in Construction Safety


PAGE 1

THE OWNER’S ROLE IN CONSTRUCTION SAFETY By XINYU HUANG A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2003

PAGE 2

Copyright 2003 by Xinyu Huang

PAGE 3

This dissertation is dedicated to my wife, Ji e Bai; my parents, Xiaoyi Huang and Guizhen Zhang; my family and everyone engaged daily in the battle against the poor safety and health performance of the construction industry.

PAGE 4

iv ACKNOWLEDGMENTS First I wish to thank my wife and my pare nts for their enduring support, trust, and encouragement for my study and research. Their unquestioning belief in my ability to complete my study was often the only inspirat ion and motivation I needed to keep me from succumbing to frequent feelings of inadequacy and ineptitude. I am indebted to them all for not demanding too much. I wish to acknowledge my supervisor for PhD study, Professor Jimmie Hinze, for his patience, generosity, encouragement, a nd guidance during my st udy at the University of Florida. The invaluable assistance and guidance of him in the course of completing this project were essential for me to complete the study. The academic and profession spirit he demonstrates will inspire me in my future life and career. I am also indebted to all the safety pe rsonnel I interviewed or surveyed. The study could not have been completed without their enthusiasm for construc tion safety and their experience in safety management. I want to thank Professor Jo Hasell, w ho guided and helped me so warmly ever since the first day I arrived at the University of Florida. I want to thank Drs. Dennis Fukai, Charles Glagola, Leon Wetherington, and Kwaku Tenah, for their help with my research. I want to thank Dr. Richard C oble for encouraging and recommending me to study for my PhD at the University of Florida.

PAGE 5

v The statistical courses I took with Drs. Alan Agristi, Bhramar Mukherjee, and Ron Marks at the Department of Statistics help ed me to conduct research design and data analysis. Their guid ance is appreciated. This research was supported primarily by funds provided by the National Institute of Occupational Safety and Health (NIOSH ). Support was also provided through the cooperative efforts of the Constr uction Industry Institute (CII). This research could not have been successfully undertaken without th e assistance of both organizations. Their contributions of funds and effort are gratefully acknowledged. Finally, I owe an unquantifiable debt to ma ny of my friends and teachers who have encouraged and helped me to achieve my degree.

PAGE 6

vi TABLE OF CONTENTS Page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................ix LIST OF FIGURES.........................................................................................................xiii ABSTRACT....................................................................................................................... xv 1 INTRODUCTION........................................................................................................1 2 LITERATURE REVIEW.............................................................................................5 Accident Facts..............................................................................................................5 Owners' More Active Involvement In Safety...............................................................7 How Owners Take Their Role In Safety....................................................................12 Establish And Communicate Attitudes Towards Safety.....................................14 Consider Safety In Contractor Selection.............................................................14 Contractual Safety Arrangements........................................................................16 Address Safety During Design And Constructability Review............................18 Participate In Safety During Construction..........................................................21 Total Safety Culture And Behavior-Based Safety...............................................24 Summary.....................................................................................................................29 3 RESEARCH MOTHODOLOGY...............................................................................30 Research Design.........................................................................................................30 Research Components.........................................................................................30 Method Design....................................................................................................34 Sample Size Determination.................................................................................35 Sample Selection.................................................................................................37 Data Collection...........................................................................................................37 Mailed Survey Study (Phase 1)...........................................................................38 Project Interview (Phase 2):................................................................................39 Follow-Up Survey To Contractors (Phase 3)......................................................41 Data Analysis..............................................................................................................41

PAGE 7

vii 4 QUESTIONNAIRE SURVEY S FOR LARGE OWNERS........................................44 The Data Set...............................................................................................................44 Analysis Of The Mailed Survey Data.........................................................................46 The Projects Context...........................................................................................46 Commitment Of The Owner To Construction Safety.........................................48 Selection Of Safe Contractors.............................................................................49 Safety Management Practices Of The Owners....................................................51 Summary.....................................................................................................................57 5 RESULTS: OWNER'S RO LE IN PROJECT SAFETY.............................................59 The Interviews............................................................................................................59 Projects Interviewed In The Study.............................................................................62 The Most Common Safety Practices..........................................................................64 Project Descriptions And Safety Performance...........................................................66 Shutdown Projects...............................................................................................67 Public Or Private Project.....................................................................................67 Petrochemical Projects A nd Manufacturing Projects..........................................68 Union Projects Or Open Shop Projects...............................................................70 Type Of Contract.................................................................................................71 Size Of The Projects............................................................................................73 Work Shift And Workdays..................................................................................74 Owner’s Selection Of The Contractor........................................................................75 Preferred Contractors List...................................................................................75 Importance Of Safety During Selection Of Contractors.....................................76 Criteria Used To Evaluate Safety Performances Of Contractors........................76 Owner’s Contractual Arrangement.............................................................................80 Owner Involvement During Project Execution..........................................................84 Safety Program Of The Contractor......................................................................85 Owner Monitors Near Misses On The Project....................................................87 Accident/Incident Investigations.........................................................................88 Safety Recognition Program................................................................................89 Safety Education And Training...........................................................................91 Responsibilities Of Owner’s Site Representative................................................93 Setting Zero-Injury Objectives............................................................................95 Additional Findings....................................................................................................95 Safety Recognition Program................................................................................95 Safety Observation Program................................................................................96 Collaborative Efforts On Safety With Contractors And Designers.....................98 Answers To Open-Ended Questions.................................................................101 Industrial Versus Commercial Construction (A Case Summary).............................102 Owner Safety Model Building..................................................................................104 Quantification Of The Index.............................................................................105 Multi-Linear Regression Between TRIR And The Indices...............................110 Explanation And Discussion Of The Regression Results.................................115 Safety Performance Model: How To Ac hieve Better Safety Performance.......117

PAGE 8

viii Follow-up Survey: Contractors' Responses..............................................................121 Summary...................................................................................................................122 6 HOW CAN OWNER ACHIEVE PROJECT SAFETY...........................................124 Case 1: Owner R.......................................................................................................124 Case 2: Owner S.......................................................................................................130 Project Description............................................................................................130 Analysis And Diagnosis....................................................................................131 Summary...................................................................................................................135 7 CONCLUSIONS, DISCUSSI ONS AND RECOMMENDATIONS.......................136 Conclusions...............................................................................................................136 Discussions...............................................................................................................138 Recommendations.....................................................................................................140 APPENDIX A OWNER SURVEY QUESTIONNAIRE.................................................................145 B OWNER INTERVIEW QUESTIONNAIRE...........................................................149 C CONTRACTOR SURVEY QUESTIONNAIRE.....................................................155 D ANALYSIS OF THE C ONTRACTOR SURVEY...................................................156 The Contractors Interviewed....................................................................................156 Project Content And Safety Performance.................................................................157 The Owner's Involvement In Project Safety Management.......................................159 What Percentage Of The Safety Res ponsibility Is Assumed By The Owner....159 Level Of Importance Owners Place On Safety When Awarding Contracts.....161 Does The Owner Provide Suffici ent Support For Project Safety......................161 How Can Owners Facilitate Contractor s In Project Safety Management................162 In What Area Can The Owner Do The Most To Promote Safety.....................162 What Is The Best Way That Owners Ca n Address Their Concern For Safety.163 To What Extent Do You Prefer Owner s To Be Involved In Project Safety......165 Summary...................................................................................................................166 LIST OF REFERENCES.................................................................................................167 BIOGRAPHICAL SKETCH...........................................................................................172

PAGE 9

ix LIST OF TABLES Table page 2-1. Average cost of construction site injuries...................................................................10 4-1. Importance of safety during selection of c ontractors to owners of different types of projects (1 to 10 with 10 bei ng the highest possible rating).....................................47 4-2. Importance of safety when selecti ng contractors to owners with differing construction budgets.................................................................................................47 4-3. Earliest stage to emphasize sa fety vs. annual construction budget............................49 4-4. Earliest stage to emphasize safe ty vs. major type of projects....................................49 4-5. Number of safety criteria vs. annual construction budget..........................................50 4-6. Number of safety criteria vs. major type of projects..................................................50 4-7. Results of Mann-Whitney test: safety sele ction criterion vs. importance of safety in the selection of contractors.......................................................................................50 4-8. Number of contractual safety requirements vs. major type of projects......................51 4-9. Number of contractual safety requirements vs. annual construction budget..............52 4-10. Total number of safety practices im plemented vs. major type of projects...............52 4-11. Total number of safety practices im plemented vs. annual construction budget.......53 4-12. Importance of safety vs. safety approaches taken....................................................53 4-13. Non-Parametric correlation matrix be tween importance of safety, number of contractual requirements and sa fety involvement approaches.................................54 4-14. Mann-Whitney test: Importance of safety vs. safety training inputs........................55 4-15. Total amount of safety trai ning input vs. project type..............................................55 4-16. Total amount of safety input vs. annual construction budget...................................55

PAGE 10

x 4-17. Relationship between the importance of safety and whether or not safety is addressed in the design phase...................................................................................56 4-18. Priority of different approaches preferred by owners (1 to 5 with 1 being the highest possible priority)......................................................................................................56 4-19. Comparison of the ranks of selecting sa fe contractor (Q232) v.s. participating in and monitoring safety (Q233)..................................................................................57 5-1. Safety performances and sizes of th e projects included in the research.....................63 5-2. Injury rates of shutdown projects and all other projects.............................................67 5-3. Type of facility owner and safety performance (excludes shutdown projects)..........68 5-4. Type of project facility (including shutdown projects) and injury rates....................69 5-5. Type of project facility (excluding shutdown projects) and injury rates....................69 5-6. Type of labor and safety performance (US projects only).........................................70 5-7. Type of contract used on the project..........................................................................73 5-8. Worker hours expended (in thousands of hours) and safety performance.................74 5-9. Number of shifts worked and safety performance......................................................75 5-10. Number of days worked per week and safety performance.....................................75 5-11. Emphasis placed on safety in the overall review of contractors (rating from 1 to 7 with 7 being the most important).............................................................................76 5-12. Is the EMR used to evaluate safety pe rformance of contract ors? (No significant differences)...............................................................................................................77 5-13. Is the TRIR used to evaluate safety performance of contractors?............................78 5-14. Threshold value of TRIR set fo r contractor safety performance..............................78 5-15. Are qualifications of safety staffs reviewed wh en evaluating contractors?.............79 5-16. Are qualifications of th e project team reviewed wh en evaluating contractors?.......79 5-17. Number of proactive criteria utili zed for evaluating contractors on safety..............80 5-18. Contract requires the contractor to place at least one full-time safety representative on site?......................................................................................................................8 2

PAGE 11

xi 5-19. Contract requires the contractor to submit the rsums of key safety personnel for the owner’s approval?.......................................................................................................83 5-20. Relationship between TRIR and total num ber of contractual safety requirements..83 5-21. Number of leading indicator safety requirements included in the contract..............84 5-22. Owner requires the emergency plan (medical and hazardous materials) to be included in the contractor's safety program?...............................................................................85 5-23. Owner requires the daily JSA (job safety analysis) to be included in the contractor's safety program?..........................................................................................................86 5-24. Owner requires a substance abuse program to be included in the contractor's safety program?....................................................................................................................86 5-25. Does the owner’s representative monitor near misses on the project?.....................88 5-26. Does the owner maintain inju ry statistics by contractor?.........................................89 5-27. Are the contractor’s safety performance statistics included in the owner’s safety performance statistics?.............................................................................................89 5-28. Are some funds provided to the contr actor, above and beyond th e contract amount, to promote project safety?........................................................................................90 5-29. The owner participates in the safety recognition program.......................................90 5-30. Safety training methods used on the project.............................................................92 5-31. Hours of monthly refresher safe ty training received by the workers.......................92 5-32. Is there any means of verifying the comprehension of safety orientation?..............93 5-33. The owner participates in safe ty meetings and toolbox meetings............................93 5-34. Total number of key activ ities performed by the owne r’s safety representatives....94 5-35. The owner’s representative monitors proj ect safety inspection records on a regular basis.......................................................................................................................... 94 5-36. The owner’s representative monitors the project near mi ss rate and project inspection records on a regular basis........................................................................95 5-37. Is zero TRIR set as a safety objective by the owner before project commencement?95 5-38. Is there a safety recognition program on the project?..............................................95 5-39. Is safety dinner held on the project?.........................................................................96

PAGE 12

xii 5-40. Are safety observers used on the project?................................................................97 5-41. Does the contractor repor t near misses to the owner?..............................................98 5-42. Comparison between number of OSHA record able injuries recorded on the project with number of near misses recorded.......................................................................98 5-43. How does the owner rate the prime contr actor’s commitment to safety? (rating from 1 to 7, with 7 being the best)....................................................................................99 5-44. How would the owner rate the coopera tion and communication between the owner and contractor? (rating from 1 to 7, with 7 being the best)......................................99 5-45. Distribution of the indices......................................................................................108 5-46. Relationship between TRIR and the project index (P)...........................................108 5-47. Relationship between TRIR and the selection index (S)........................................108 5-48. Relationship between TRIR and the contract index (C).........................................109 5-49. Relationship between TRIR and the owner’s index (O).........................................109 5-50. ANOVA for between-subject effects.....................................................................110 5-51. Spearman’s Correlation between TRIR and the index...........................................111 5-52. Model Summary of the multi-variable regression..................................................112 5-53. ANOVA of the multi-variable regression..............................................................112 5-54. Coefficients of the multi-variable regression.........................................................112 5-55. The safety performances of proj ects with highest and lowest scores.....................121 6-1. Comparison of injury rate and scorecard evaluation................................................125 6-2. Comparison between TRIR and the scores of the scorecard....................................125 7-1. Comparison of the three levels of safety management methods..............................143 7-2. Comparison of findings in different studies.............................................................144 A-1. Characteristics of the respondents...........................................................................156 A-2. What percent of safety respon sibility is assumed by the owner..............................159 A-3. Level of importance of the fact ors when owners award contracts..........................161

PAGE 13

xiii LIST OF FIGURES Figure page 1-1. The organizational chart of a ty pical general contract agreement................................2 2-1. Injury rate of constr uction and all private indust ry (data source: www.bls.gov).........5 2-2. Fatalities in construction industry and all industries....................................................6 2-3. Causes of construction fall acciden ts investigated by OSHA (01/90-10/01)...............6 2-4. Breakdown of OSHA investigated accidents in construction (01/90-10/01)...............7 2-5. Construction Accident Causatio n Model: Constraints and Responses.........................9 2-6. Relationship between Desi gn Effort and Project Cost...............................................19 2-7. Domino accident causation model suggested by Heinrich (1959).............................25 5-1. Locations of the project s included in the research.....................................................63 5-2. Injury pyramid for the projects...................................................................................64 5-3. Other leading indicators us ed in the project contract (n ot statistically significant)...83 5-4. Corporate history of safety performance..................................................................103 5-5. Histogram of the regre ssion standardized residual...................................................113 5-6. P-P plot of the regression..........................................................................................114 5-7. The predicted value vs. standardized residual plot...................................................114 5-8. Project safety model: owner’s impact......................................................................119 5-9. Owner's influence on construction safety scorecard................................................120 6-1. Owner's influence on construction safety scorecard................................................133 7-1. Owner's involvement in Safety management...........................................................142 7-2. Focus of different levels of safety management methods........................................143

PAGE 14

xiv A-1. What type of project has the best safety performance? (type of contract)..............157 A-2. What type of project has the best safety performance? (union or open shop).........158 A-3. What type of project has the best safety performance? (type of project)................158 A-4. Change of owner involvement in projec t construction safety in the past decade....160 A-5. Does the owner provide sufficien t support for safety on the projects.....................162

PAGE 15

xv Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE OWNER’S ROLE IN CONSTRUCTION SAFETY By Xinyu Huang December 2003 Chair: Jimmie Hinze Major Department: Building Construction The construction industry has long been re garded as one of the most dangerous industries. The construction industry has a hi story of poor safety pe rformance. This is despite the fact that dramatic improvements ha ve been made in the safety performance of the construction industry in th e past decade. The improvements are due, in part, to the concerted efforts of owners, contractors, subcontractors, and de signers. While past studies have investigated the safety roles of contractors, subcontra ctors, and designers, the owner's impact on construction safe ty has not been thoroughly studied. The owners of projects are the primary consumers of constr uction services, the sources of project finances, and, in many cases, the end users of the facilities. They are often at the pivotal position of the projects. Traditionally, owners have not been directly involved in construction safe ty, often in order to avoi d economic losses and legal entanglements resulting from injuries. With the increase of accident costs and legal cases

PAGE 16

xvi involving owners as the thir d-party defendants, owners have come to realize the importance of safety. In this study, the owner's role in cons truction safety was investigated. The relationship between project sa fety performance and the owner's influence was examined, with particular focus on the project context, selection of safe cont ractors, contractual safety requirements, and the owner's proact ive involvement in safety management. A questionnaire survey sent to owners with large construction budge ts and interviews conducted on projects with large worker hour exposures were the primary means of data collection for the study. Statis tical analysis shows that th ere is a strong relationship between project safety performance and ow ner involvement. Practices of owners significantly associated with project safety performances were identified. A model to demonstrate how owners can help improve safe ty was established. A scorecard that can be used to assess the owner's involvement in safety was developed and tested. It was concluded that the owner's involve ment can significantly influence project safety performance. Owners can achieve be tter project safety performances by setting safety objectives, selecting sa fe contractors, and particip ating in safety management during construction.

PAGE 17

1 CHAPTER 1 INTRODUCTION With peak levels of employment reaching eight million workers, the United States construction industry has the dubious distin ction of perpetually being one of the industries with the worst injury and fatal ity records. The mining industry and the agriculture industry are the only industries that have worse records than the construction industry. While the actual inci dence of construction worker in juries has declined over the past three decades, the number of injuries a nd fatalities is still at an unacceptable level (Hinze, 1997). According to the Census of Fatal Occupational Injuries (CFOI, 2003), an average of 1,115 workers were killed annua lly on construction sites from 1995 to 2000, which accounted for nearly 20% of all industria l fatalities in the United States. This number is a disproportionate distribution of constructi on worker fatalities since construction workers account for only 7% of the industrial workforce. While many strides have been made to reduce the incidence of injuries, there is considerable room for improvement. The industry is trying to find ne w ways to improve safety performance. In the construction industry, most construc tion projects involve th e participation of owners, designers, and contractors. The owners (also called facility clients, or project buyers) of projects are the primary consumer s of construction services, the sources of project finances, and, in some cases, the end us ers of the facilities. Owners include both public and private entities. Some owners obt ain the services of a construction firm to build a facility with virtually no interaction between the ow ner and the cons truction firm during the construction process. Other owners play varying roles during the construction

PAGE 18

2 Designer Workers Workers Worker Worker Subcontractor Worker Worker Subcontractor General Contractor Owner phase, with some being closely involved with the construction effo rt during every stage of project execution. Generally, an owner of a project will requi re various objectives to be satisfied by the designer and contractor (Hinze, 2001). Th e owner defines the scope of a project to become a reality, which requires the services of several parties. Fi rst, the designer will develop the guidance document for building the project. The contractor will then follow by executing the construction of the project. Typical objectives required by the owne rs include, but are not limited to the following: when a particular project should be finished; what quality requirements the project must satisfy; the owner’s cost of the project; and possibly the safety standards that must be met during project construction. The roles of some of the parties are shown in the organizational chart for a ge neral contract agreement in Figure 1-1. In all types of contract agreements, the owne r is at the top position of the project organization. Essentially, the owner has the overa ll authority on the project. Figure 1-1. The organizational chart of a typical general contract agreement

PAGE 19

3 Traditionally, safety responsibilities ha ve rested solely on the shoulders of contractors. Owners and designe rs have held the view that they should no t get involved in construction worker safety for fear of incurring increased liability exposure. However, in the past decades more and more parties to the construction process have come to realize that zero accidents is an attainable objective, but only through the concerted efforts of all parties involved in the constr uction process. Owners, A/E firms (also called designers), and contractors/subcontractors ha ve different roles in preventing accidents to achieve an injury-free worksite (Hinze, 1997; Gambatese, 1996; Toole, 2002). The contractor is undoubtedly the pi votal party to control jobsit e safety. Strategies and approaches taken by contractors to improve project safety have been thoroughly investigated in past research studies (L evitt and Samelson, 1993; Hinze, 1997; Hinze, 2002). Subcontractor safety influenced by th e general contractor in various sizes of projects was investigated by Hinze and Fi gone (1988, for small and medium projects), and Hinze and Talley (1988, for large projects). Designers can reduce safety hazards in the working environment by considering worker safety issues in their design decisions. Hinze and Gambatese (1996) gathered various “best practices” for designers to address safety issues in their designs and develope d a safety design tool to help designers eliminate hazards when making decisions. T oday, more and more owners have come to realize that the costs of c onstruction accidents are ultimate ly their own financial burden, and that they cannot with certain ty shield themselves from th e legal liabilities associated with worker injuries. As a result, ma ny owners are taking more active roles in construction safety. This begins with the selection of safer construction firms and can also be observed in the safety provisions in cluded in the construction contract. Owners

PAGE 20

4 will continue to influence safety performan ce by the nature of their involvement during the construction process. Despite all the research th at has been conducted with the various parties to the construction process, little prior work has examined the specific role played by project owners in promoting safety performance. Ho w owners regard their role in safety and how they attempt to exert their influence in the construction process have not been previously examined in detail. This study is focused on the investiga tion of the impact of owners on construction safety.

PAGE 21

5 CHAPTER 2 LITERATURE REVIEW Accident Facts Accident data prepared by the Bureau of Labor Statistics (200 3) show that the construction industry has performed much wors e than the average of all industries (see Figure 2-1). Although the statis tics have improved dramatically in the 1990's, accident rates in the constructi on industry are still 50% higher than that of all i ndustries, lagging all industries by about 10 years. An alarming fact is that the number of fatalities in the construction industry has increased in the past decade (see Figure 2-2). 5.8 6.8 6.3 6.7 6.6 6 5.7 5.45.4 4.8 4.44.4 4 4.1 4 3.9 3.8 3.93.9 3.7 3.6 3.53.5 3.4 3.13.1 2.9 2.82.8 2.6 15.2 16 14.9 14.5 15.4 14.4 14.2 14.1 12.8 12 11.5 10.4 9.7 9.3 8.7 8.4 8.2 7.8 8.9 9.2 7.6 8.3 8.2 8.3 7.9 8.3 7.9 7.7 7.5 6.9 6.6 6.2 5.9 5.8 5.4 5.4 6.3 6.5 6 6.2 6.9 6.86.8 6.7 6.8 3.7 3.6 3.6 3.6 3.4 3.7 3.9 4.2 4 3.7 3.43.4 12.9 14.5 15.1 14.7 15.5 15.8 14.9 15 8.1 9.2 7.7 7.7 7.5 9 8.5 7.8 8 0 2 4 6 8 10 12 14 16 18 19751977197919811983198519871989199119931995199719992001 YearInjury Rate (per 100 full-time workers) Lost Workday Injury Rate of construction industry Lost Workday Injury Rate of all private industry OSHA Recordable Injury Rate of construction industry OSHA Recordable Injury Rate of all private industry Figure 2-1. Injury rate of constr uction and all private industry

PAGE 22

6 6217 6331 6632 6275 6202 6238 60556054 5915 963 971 10771083 1086 1123 1198 1210 1159 0 1000 2000 3000 4000 5000 6000 7000 19911992199319941995199619971998199920002001 YearNumber of fatalities Fatalities in all industries Fatalities in construction industry Figure 2-2. Fatalities in constr uction industry and all industries Figure 2-3 shows the major types of construction accidents that have occurred in the past decade, based on the analysis of all serious injuries and fatalities investigated by OSHA. Falls, struck-bys, electric shocks, ca ught-in-or-betweens, and other are the major causes of injuries, although the relative proporti on of each type has changed somewhat in recent years (see Figure 2-4). Falls 36.3% Struck By 24.3% Shock 13.4% Caught in or Between 11.6% Other 14.5% Others 5.7% CardioVascular/Respirator y System Failure 4.2% Struck Against 2.3% Repeated Motion/Pressure 0.1% Bite/Sting/Scratch 0.1% Rubbed/Abraded 0.2% Inhalation 1.6% Ingestion 0.1% Absorption 0.2% Figure 2-3. Causes of construction fall acci dents investigated by OSHA (01/90-10/01)

PAGE 23

7 0% 20% 40% 60% 80% 100% 199019911992199319941995199619971998199920002001 Year Falls Struck By Shock Caught in or Between Other Figure 2-4. Breakdown of OSHA investigated accidents in construction (01/90-10/01) Owners' More Active Involvement In Safety With a higher proportion of injuries, th e construction industry has long been regarded as a dangerous industry. With an average employment of approximately 7% of the industrial workforce, th e construction industry has ge nerally accounted for nearly 20% of all industrial worker fatalities (CFOI, 2003). The re search conducted by Everett and Frank (1996) concluded that the total costs of construction accidents accounted for 7.9-15.0% of the total costs of new, non-resi dential projects. A more recent but unpublished research by Coble and Hinze (2000 ) showed that the average worker's compensation insurance costs could be conser vatively estimated as constituting 3.5% of the total project cost. Safety performance of the construction i ndustry did improve during the 1990’s (see Figure 2-1). As suggested by Hislop (1999), perhaps th e most significant factor attributed to this improvement is the in creased management commitment to safety. Management has been motivated to make a greater commitment to safety, based on its

PAGE 24

8 increased awareness of the impact of the hi gh costs of worker’s compensation payments, the higher dollar value settlements in lawsu its, the increased amount s of OSHA fines for safety violations, and the adverse impacts of poor safety performance on the corporate image. In order to reduce and eventually eliminat e construction accidents, researchers have explored techniques to fulfill the “zero inju ry objective” through the concerted efforts of all parties involved. The involvement of ow ners has been regarded as an essential requirement for this objective. For example, in the research conducted by Liska et al. (1993), it was found that an important prere quisite attributed to excellent safety performance was the involvement of the owne r not only in pre-project planning including financially supporting the contract or’s safety program, but also in the day-to-day project safety activities. In the c onstruction accident causation m odel developed by Suraji et al. (2001, see Figure 2-5), construction accidents ar e caused by different responses to certain constraints and the environment on construction projects. In the model, owner (client) responses are the actions (or in actions) of the owner in resp onse to constraints during the development of a project brief (Duff, 2000). These include, for example, reducing the project budget, adding new project criteria, changing project objectiv es, and accelerating the design or construction efforts of the project. All these elements will play an essential role in accident occurrences.

PAGE 25

9 Figure 2-5. Construction A ccident Causation Model: Co nstraints and Responses In the past, there was a reluctance of owne rs to become involved in matters related to construction safety issues for the fear of added liability exposure (Sikes et al., 2000). However, since the 1980’s more and more owners, especially owners with larger construction budgets, have voluntarily expanded th eir role in ensuring worker safety. A series of studies conducted at the Univ ersity of Washington in the early 1990’s

PAGE 26

10 demonstrated that owners’ concern for constr uction safety was increasing (Hinze, 1997). The major reasons include the following: The rising costs of health care and worker s’ compensation are not being ignored by owners (see Table 2-1). Table 2-1. Average cost of c onstruction site injuries Job Costs Type of injury Direct Indirect Estimated Liability Costs Total Cost to Employer Medical Only $520 $440 $240 $1,200 Lost Work Day $6,900 $1,600 $16,500 $25,000 (Hinze and Appelgate, 1991) Owners realize that the costs of injuries are ultimately reflected in the cost of construction (Gambatese, 2000b). Litigation involving third parties has escal ated in the past three decades. For example, in the case of Phillips v. Unite d Engineers & Constructors, Inc., and Plasteel Products Corp., 500 N.E. 2d 1265 ( 1986), owner was sued but was not held responsible for a worker's fall from catwal k during steel erection. In another case Rigatti v. Reddy, 723 A.2d 1283 (1999), the owne r was similarly cleared of being responsible for a roofer's fall. In the case of Stark v. Rotterdam Square, 603 N.Y.S.2d 347 (1993), the owner of a mall was held liable for injuries suffered by a roofer when he fell throu gh a hole cut into the roof. Because of these types of lawsuits, many ow ners have come to realize that reducing the frequency and severity of construction inju ries is the only sure way of reducing their potential liability for worker in juries (Levitt and Samelson, 1993). Since the early 1980s, several efforts have been undertaken to formally require owners to participate in cons truction site safety. A majo r effort to expand safety legislation resulted after 28 workers died in the tragic collapse of the L’Ambiance Plaza Building in Bridgeport, Connecticut (Godfrey, 1988). This accident became the driving force behind U.S. Senate Bill 2581 to amend the Occupational Safety and Health Act to “require all construction projects to be supervised by . a professional engineer-architect designated by the owner and registered in the state where the construction is to be

PAGE 27

11 performed” (ASCE, 1988). Opposition to th is bill from a large segment of the construction industry ultimatel y resulted in its failure. Although these legislation efforts failed in th e U.S., the liability of owners in the United Kingdom was specified by the C onstruction (Design and Management) Regulations 1994 (CDM), in which the owner’s main duties are contained in Regulations 6, 10, 11 and 12 (Holt, 2001). In CDM, a ow ner is defined as “any person for whom a project is carried out, whet her carried out by another pers on or in-house.” CDM imposed criminal liability onto the owne r who ignores construction safety and where this results in an accident. Each project owner has a responsib ility for safety on the project. In cases where there are multiple owners, the owners can appoint an agent or another owner to carry out the owner’s duties, and then have to make a d eclaration to the enforcing authority (the Health and Safety Executive) that the transfer of duties has been made. Under CDM, the owner must do the following: Appoint a Planning Supervisor and a Prin cipal Contractor for each project, being satisfied that these "duty holders" are comp etent and have the resources to perform their duties adequately. Not permit the construction work to start unless a health and safety plan, which complies with the safety regulations, is in place for that project. Provide the planning supervisor with inform ation about the state or condition of the premises where the work is to be carried out This is information which is relevant, and which the owner either has or coul d get after making reasonable inquiries. Verify that any designer or c ontractor that is appointed directly is competent for the task and has allocated sufficient resources to it. Make the health and safety file availa ble for inspection by anyone who may need information to comply with legal requirements. The owner will sell or pass on the file to a future owner or a person acquir ing the interest in the property of the structure to which it refers.

PAGE 28

12 The trend is clear that owners are getting more concerned about construction safety, but many owners are confronted with the prob lem of how to effectively influence project safety. For example, a safety campaign supported by the owner was conducted during the construction of the resund Link between Denmark and Sweden. Primarily, the safety campaign consisted of a comprehensive information campaign aimed at promoting positive attitudes towards safety among the employees, and a specific campaign aimed at the behavioral aspects to increa se the level of safety when performing routine work at the construction site (Spangerberg, 2002). In th e end, the results of the campaign did not satisfy the owner’s expectations of low inju ry rate, which led people to question what techniques by the owner can effec tively improve construction safety. Therefore, it would be essential to inve stigate the current techniques employed by owners to achieve good safety performances on construction sites, and if possible establish the effect the techniques on the resultant safety performance. How Owners Take Their Role In Safety Owners can actively impact construction sa fety by selecting safe contractors, addressing safety issues in design, and participating in safety management during construction (Hinze, 1997). To the extent possible, the owners, through their project representatives, should particip ate with the contractors in al l project safety activities, including but not limited to new employee orientation, safety meetings, audits and accident investigations, training, incentive programs and other safety related programs (Gambatese, 2000b). One of the earlier studies on the owner’s ro le in construction safety by Levitt et al. (1981) reached the conclusion that construc tion owners who selected or prequalified contractors based on safety performance, and/ or who got involved in construction safety

PAGE 29

13 management, had fewer accidents on their projects. The owner's involvement in construction safety management not only reduc es the number and severity of accidents but probably also reduces the ow ner’s total liability exposure. The American Society of Civil Engineers (ASCE) moved to the forefront in the trend to involve owners in safety followi ng the issuance of ASCE’s Policy Statement 350 on construction site safety in 1998. The statement outlines ASCE’s belief that “improving construction site safety requires attention and commitment from all parties involved.” The policy states that safety should be addressed “for each project on a project specific basis,” and th at owners should “take an activ e role in project safety.” Various ways owners can actively address safe ty were given in the policy, including the following: Assigning overall project safety respons ibility and authority to a specific organization or individual (or specifically retaining that responsibility) that is qualified in construction safety principles, rules, and practices appropriate for the particular project. Including prior safety performance as a criterion for cont ractor selection. Designating an individual or organization to monitor safety performance during construction. Designating in the contract documents t hose parties responsible for the final approval of shop draw ings and details. Gambatese (2000b) summarized various ways in which owners can actively address safety, including the following: Establish a clear position on safety. Ensure that safety is addressed in project planning and design. Consider safety performance when selecting a contractor. Address safety in the construction contract. Assign safety responsibility during construction. Participate in project sa fety during construction.

PAGE 30

14 Establish And Communicate Attitudes Towards Safety All owners have a legal and moral responsib ility to insist on the safe performance of their construction contractors and to use r easonable care to prev ent contractors from injuring others on the site. Owners shoul d understand that their involvement in construction safety management not only reduces the number and severity of injuries, but is also likely to reduce the ow ners’ total liability. In a ddition, the owner should realize that safe performance is generally related to lower costs, better quality work, improved productivity, adherence to schedule, reduced exposure to bad publicity, and minimal disruption of the work being performed. As Hinze (1997) stated, owners whose objectives are to avoid injuries are likely to have a proactiv e and more dire ct involvement in construction operations. Therefore, the firs t step of the owner’s involvement in safety is commitment, especially management commitm ent, to create an injury-free project. Each owner in the survey conducted by Levitt and Samelson (1993) stated the belief that active involvement in construction safety, done properly, served to reduce rather than increase the firm’s potential liabi lity for construction accidents. The owner’s position on safety should be cl early communicated to the project team at the beginning of the project and to all team members joining the project during the construction phase (Gambatese, 2000b). Du ring project development, owners should convey their commitment to safety to the contractors through vari ous means, including devising incentives for safety and by impl ementing sanctions for failing to support the owners' safety initiatives. Consider Safety In Contractor Selection Traditionally, selecting safe contractor s was recognized by many owners as the most effective way to guarantee safety pe rformance on their projects. Criteria for

PAGE 31

15 selecting safe contractors have been extens ively investigated in the past. Levitt and Samelson (1993) stated that screening contra ctors in terms of their expected safety performance is an easy and effective wa y for construction owners to enhance construction safety. Levitt and Samelson (1981) suggested that the criteria must be (1) predictive of safe project performance; (2) equally applicable to different construction firms; and (3) objective. Possible selecti on measures were EMRs, OSHA recordable injury rates, and management safety accountab ility. Findings from the research that are indicative of safety accountability include the following: 1. Who in the organization receives and re views accident reports, and what is the frequency of distribution of these reports? (Frequent re ports, detailed by operating units, and reviewed by the pr esident of the firm would be an indication of high company accountability for safety.) 2. Frequency of safety meetings for field supervisors. 3. Compilation of accident records by foreme n and superintendents and the frequency of reporting. (Contractors w ho categorize their accident statistics by superintendent and foremen, rather than by larger uni ts, have a more detailed accountability system.) 4. Frequency of project safety inspections and the degree to which they involve project managers and field superintendents. 5. Use of an accident cost system allocati ng or assigning respons ibility to individual foremen and superintendents, as well as project managers. Certain safety criteria during the sele ction of contractors suggested by Hinze (1997), and Hinze and Godfrey (2003) include the following: Injury incident rates (including lost wor kday injury rates, OSHA recordable injury rates, first aid injury rates, and etc.) Job site safety inspections Behavior based worker observations Experience modification rates (EMR) Loss ratios of workers’ compensation Records of OSHA citations and fines Litigation related to injuries Performance records of key personnel

PAGE 32

16 Project safety plans Contractor qualification safety surveys Worker's conception surveys Contractual Safety Arrangements Owners must also make sure that co ntractors recognize their contractual responsibility to perform safely (Levitt a nd Samelson, 1993). Contractually, most owners have the contractor indemnify them from a ny losses or liabilities resulting from injuries, but it is also essential to include specif ic and thorough safety requirements in the construction contract. The requirements mi ght include the following(Gambatese, 2000b): The requirement that the contractor ab ide by all applicable safety laws and regulations; A delineation of the responsibility for safety on the jobsite; The submission of a written contractor safety program before work begins; A requirement for implementing a substance abuse program; and The submission of an emergency plan and accident reporting procedure. Questionnaires sent to owners and cont ractors (Business Roundtable, 1982) were used by Stanford University to identify safety requirements owners placed on construction contractors and sp ecific practices of owners to emphasize safety with contractors. The responses are arranged in decreasing or der of use by the respondents. Require use of a system of permits for potentially hazardous activities. 1. Require the contractor to designate a responsible superv isor for safety coordination on the job site. 2. Provide the contractor with safety guidelines that mu st be followed. 3. Discuss safety at owner-contractor meetings. 4. Periodically discuss safety aud its of the contractor operations. 5. Require immediate reporting of contractor accidents.

PAGE 33

17 6. Stress safety as part of the c ontract during pre-bid walk-arounds. 7. Investigate contractor accidents. 8. Maintain statistics of contractor accidents. 9. Conduct periodic safety inspections. 10. Set goals for construction safety. 11. Consider safety in prequalifying contractors for bidding on projects. 12. Set up a construction safety departme nt to monitor contractor safety. 13. Set safety guidelines in the body of the contract. 14. Be involved in orientatio n sessions alerting workers to safety hazards on the job. An owner, to be actively involved in c onstruction safety, might consider several contractual issues. Many of the issues re late to the safety obligations placed on the contractor. Hinze (1997) suggested that c ontract provisions may include the following requirements: Submittal of a project-specific safety plan Job hazard analysis Regular safety meetings with supervisory personnel A designated project safety coordinator Mandatory reporting of accidents, safety inspections, and safety meetings Inclusion of subcontractor s in the safety program Compliance with the owner’s safety guidelines Establishment of an effectiv e worker orientation program As the major binding document between th e owner and contractor, the contractual arrangement (lump-sum or cost reimbursabl e) will also impact how safety will be addressed on the project. One reason that th e contract type should be considered when addressing project safety is that the contra ct essentially determines how the owner will make payment to contractors to compensate them for their services on the project, and therefore, can impact the relationship betw een owner and contractor (Dagostino, 2002). As safety can be obtained only through the c oncerted efforts of diffe rent parties (owner,

PAGE 34

18 contractor, and designer), the na ture of their relationships can be expected to impact safety. For example, if the contractor has a close, long-term relationship with the owner, they will generally use cost reimbursable cont racts (cost plus). Thus, safety investments will be reimbursed by the owner to support the contractor’s efforts. This arrangement encourages the contractor to sp end the necessary funds on safety. Address Safety During Design And Constructability Review Design decisions can dictate the structur e, elements, materials and even the construction methods to be used on the projec t. For example, the selection of forming systems, cycle times, sequences, equipment, and design of temporary structures are significantly impacted by the de sign of the permanent construc tion. Therefore, designers are very influential, perhaps unknowingly, in impacting the way safety issues will be addressed during construction. Consideration of construction worker safety by designers (hereinafter also called A/E) can lead to a reduction in injuries and associated costs during the construction phase (Hinze and Wiegand, 1992; Gambatese, 1998; Gambatese 2000a). Toole (2002) stressed that the A/Es are in the best po sition to implement specific safety design recommendations, thereby preven ting less safe conditions on the site. Also, A/Es may be best able to identify questiona ble structural situati ons such as temporary loadings on the permanent structure or tem porary work platforms, provided they are explicitly requested to do so and possess all of the data necessary to perform the analysis. One reason why the Corps of Engineers has ha d particularly low injury rates on their construction projects is that it coordina tes the design with erection procedures and coordinates their efforts on safety (MacCollum, 1995). Although the involvement of design professionals in construction site sa fety has been minimal to nonexistent on many projects, when they are involved their influen ce can be significant, and can help reduce

PAGE 35

19 accident occurrence. This influence has favor able cost implications for the projects (Gambatese, 2000a. see Figure 2-6). Figure 2-6. Relationship between De sign Effort and Project Cost In the past, design professionals typica lly would distance themselves from the responsibility for the safety of the construction workers. Th e reasons include their lack of safety education and training, the lack of safety design tools, their restricted role on the project team, and an attemp t to limit their li ability exposure (Gambatese, 1996). However, with the unacceptable injury and fa tality rates, the construction industry has concluded, “The need for changes in attitude does not stop at educati ng erectors to work more safely. It has to go back to the architect and engineer who should not only ask themselves if it can be built, but can it be built safely.” (Baggs, and Cunningham, 1988) In the UK, the Construction (Design and Management) Regulations 1994 place duties on designers and owners to make assessments of the impact of the proposed design on the life cycle safety of the facility, includ ing the construction workers, maintenance personnel, and the facility occupants (Holt, 2001). Although a similar U.S. bill to expand Additional costs: Delays, design changes, CM efforts, etc. Additional costs: Plan review, drafting, plan coordination, etc. Low Low High Optimal High Optimal costs Project costs Extent of Desi g n Efforts

PAGE 36

20 the designer and owner responsibilities on c onstruction safety did not pass, the OSHA standards are certainly not fore ign to designers, which is a potential source of designer liability. This occurs not solely by assuming the contractor’s responsibilities for the means or methods of the construction process, but through the understa nding of their duty as designers to help provide a safe workpl ace which is basic in their review of the contractor’s work product (Cobl e and Blatter, 1999). In re sponse to the needs of the industry, some architects starte d to work as construction ma nagers on small projects and thus coordinated the construction and de sign efforts to reduce accidents and costs (McKee, 1994). MacCollum ( 1995) ranked the methods used for reducing safety risks according to their priority and effectiveness, including the following: Design to eliminate or minimize the hazards; Guard the hazards; Give warnings; Provide special procedur es and training; and Provide personal protective equipment. He further suggested that the design t eam should include at least one safety professional, and should develop and impl ement a hazard prevention plan for each project. Owners can impact designers and contract ors through their pro active participation in construction safety issues. In the c onstruction accident causation model developed by Suraji et al. (2001, see Figure 2-5), designer responses to project design constraints (which may arise from the owner) can in turn place constraints on the construction process. This may result in a chain reacti on effect on accident causation. For example, the constraints imposed by the owner may include the following: Accelerated design program Inadequate design budget Conflict of objectives or de mands of other projects

PAGE 37

21 Designer response is the action or inaction by designers to respond or react to the constraints existing during the project de sign stage. These include the following: Increased design complexity Sublet part of design process Reduction in design resources Reduction in quality of components Ignore legal liabilities Therefore, owners can take an active role in design for safety through: Setting their expectations in the design phase that construction safety is one of their major concerns and is to be built into the project design; Addressing safety issues as early as th e feasibility study and conceptual design phases and integrate safety into the objectives of the project; Actively participating and coordinating the efforts of the designer and the contractor through regular safety/constructa bility reviews of the project design; and If possible, awarding the c ontract to an engineering and construction company to help promote safety performance. Participate In Safety During Construction In the research conducted by Levitt and Samelson (1993), it was found that the owners with the safest c onstruction projects tended to use many of the following strategies with their contractors: Stress safety as part of the contract during the pre-job walk-around. This gives contractors the opportunity to include all safe ty-related items in their bids and lets contractors know early that the owner is serious about safety. Require short-term permits, rather than ongoing permits, for hazardous activities. This means that contractors must check daily or more frequently to ensure that any planned hazardous activities are coordinate d with other contractors and with the owner’s own plant work force. Conduct safety audits of the contractor during construction. The owner’s or construction manager’s safety staff conducts these audits to ensure compliance with the owner’s safety requirements and with all state and federal safety regulations. They are aimed at systems and procedures rather than at specific hazards. Conduct periodic safety inspections.

PAGE 38

22 Require safety training of all project employees. Maintain statistics on the cont ractor’s safety performance. Set goals for construction safety. Project -wide safety goals should be set, along with specific goals for contractors who n eed special attention because of past poor performance or particularly hazardous work operations. Include general safety guidelines in the body of the contract. Set up a construction safety departme nt to monitor contractor safety. Require immediate reporting of all work er accidents. Immediate rather than periodic reporting gives the owner more time to intervene and ensure that the contractor has corrected any identified hazards before others can be injured by them. Such notification can also serve to initiate consultations w ith the contractor’s senior management, if needed. Investigate the contractors’ accidents. The owner’s involvement in investigating the contractors’ accidents gives the owne r valuable insights about generic safety hazards on the project, as well as add itional insights about the contractors’ organization and capabilities in the area of safety. Always include safety on the age nda at owner-contractor meetings. Provide contractors with special sa fety guidelines they must follow. Require the contractor to assign safety coordination responsibility to someone on site. Reimburse the contractor’s safety costs in full. The following is a sample of best practi ces suggested by Gambatese (2000b) that an owner can implement: Adjust the scheduling of different activit ies or construction phases that would otherwise occur at the same locati on and be performed simultaneously. Provide a list and location of toxic substances and othe r hazardous materials that are located on the site. Do not allow schedules that contain sustained overtime or night work. Impose a ceiling on the number of worker s on site or in a particular area.

PAGE 39

23 Confirm that the contractor knows of th e potential hazards of all construction materials and their prope r storage and disposal. On renovation or retrofit projects, provi de the contractor with complete and updated as-built drawings of the existing structure. Conduct a pre-construction meeting w ith the contractor (including all subcontractors) to discuss safety issues. Consider involving OSHA in planning fo r project safety prior to starting construction. Gambatese (2000b) also suggest ed that owners should: Establish a clear position on safety. The owner’s position on safety should be clearly communicated to the project team at the beginning of the project and to all team members joining the project as part of the construction phase. The position can be written in the project documents and contracts, and verbally communicated in project team meetings during desi gn and construction. The actions of all members of the owner’s organization during the course of the pr oject must reflect and reinforce the established position. Ensure that safety is addressed in project planning and design. Thus, owners must provide the initial impetus, by requesting or even requiring, by cont ract terms, that designers consider construction site safety in their designs. Consider safety performance when selecting a contractor. Address safety in the construction contract There are numerous clauses that could be included in a contract to promote safety. Assign safety responsibility during constr uction. The responsibility for overseeing safety on a construction proj ect should be held by a competent organization or individual. Participate in project safety during constr uction. Perhaps the most effective way an owner can influence safety is through jobsite participati on. Another common means by which owners can be involved in safety is by requiri ng the submittal of regular safety reports. Reports can be requested that provide information on the results of jobsite inspections a listing of all injuries, safety meeting minutes, and investigations of major accidents. Hinze (1997) summarized some unique owner practices as the following: Placing company representatives on every construction project Conducting safety meetings with the contractors Requiring the contractors to adhere to owner-developed safety practices

PAGE 40

24 Providing specialized sa fety training for the contractor personnel Requiring all workers to go through safety orientation Reviewing each contractor’s safety program Conducting regular audits of c ontractor safety performance Implementing safety incentive programs on all construction projects The participation of owners in safety duri ng construction is an in tegral part of the efforts of many owners, but it is far from the sole participation of owners in safety. Safety involvement of owners often starts at the very beginning of the project, and lasts throughout the life cycle of the facility. Owner involvement in safety includes selecting safe contractors, addressing safety in de sign, including safety requirements in the construction contract, and being actively involved in project safety management. Total Safety Culture And Behavior-Based Safety When participating in cons truction safety management, proactive owners generally implement initiatives that promote safety. The concept of total safety culture and behavior-based safety management are among the latest safety ma nagement initiatives that will be introduced. Originally, people believed that accident s were a result of pure chance, and could happen to anyone at any time, and some people were more likely to suffer an accident than others. This was referred to as the accident proneness model (Cooper, 1998). In the past, the prevailing influence of this approach meant that most accidents were blamed solely on employees rather th an on the work processes, poor management practices or a combination of these. Recent studies have found no solid evidence to show that certain people were more likely to be involved in acc idents, and thus concluded that accident proneness was essentially relate d the individual's propensity to take risks (Hinze, 1997). The most famous accident causation m odel is the Domino model suggested by Heinrich (1959, see Figure 2-7). The model shows that an acciden t is the result of a

PAGE 41

25 series of events related to the social e nvironment and heredity, personal failings or mistakes, physical hazards and unsafe behavior These events, if permitted to occur in continuity, result in an accident, which may re sult in injuries and damages. Any accident is prevented by breaking the chain (or removing a domino) of the series of events at any location. While it is generally understood that most accidents are preceded by a series of events, it is common practice to place most of the attention on the action that occurred immediately before the accident. A more pr oactive approach would be to focus on the "upstream" activities or events that se t the series of events in motion. Figure 2-7. Domino accident causation model suggested by Heinrich (1959) Many other theorists used Heinrich’s domi no theory as the starting point for their own work. For example, in 1971 Weaver modifi ed the original theory to propose that the last three dominos in the sequence were caused by management omissions (Cooper, 1998). In essence, Weaver’s model placed the immediate responsibility for accidents squarely on the shoulders of poor supervision and line management, instead of the injured workers. Personal failings or mistakes Unsafe behavior/physical hazards Accident Injury/Property damage/Near miss Social environment and heredity

PAGE 42

26 Today, as the zero injuries objective is accepted and sought by many owners and contractors, total safety culture (TSC) and be havior-based safety (BBS) have become the popular approaches being implemented by many proactive firms. The basic logic of these approaches is to prevent construction accidents by continually addressing the very front-end of the accident chain – the culture and habits. Safe ty management is based on the scientific approach that is focused on the psychology and the behavior of workers. Geller (2001) summarized the esse nce of TSC as the following: Promotes a work environment based on employee involvement, ownership, teamwork, education, training, and leadership. Builds self-esteem, empowerment, pride, enthusiasm, optimism, and encourages innovation. Reinforces the need for employees to actively care about their fellow coworkers. Promotes the philosophy that safety is not a priority that can be reordered, but is a value associated with every priority. Recognizes group and individual achievement. He further pointed out that total safety cu lture requires continual attention to three domains environmental factors (including equipm ent, tools, physical layout, procedures, standards, and temperature), pe rsonal factors (including people ’s attitudes, beliefs, and personalities), and behavior f actors (including safe and at-ris k work practices, as well as going beyond the call of duty to intervene on behalf of another person’s safety). In a total safety culture: Everyone feels responsible for safety and does something about it on a daily basis; People go beyond the call of duty to id entify unsafe conditions and at-risk behaviors, and they intervene to correct them; Safe work practices are supported interm ittently with reward ing feedback from both peers and managers;

PAGE 43

27 People “actively care” continuously for th e safety of themselves and others; Safety is not considered to be a priority, but it is a value. This change in philosophy is based on the premise that pr iorities can change but values cannot. Weinstein (1997) proposed another approach toward total safety culture through incorporating Total Quality Management with safety. He suggested a Total Quality Safety Management System which is achie vable through the implementation of Total Quality Management Principles, Process Safety Management Guidelines, ISO-9000 Quality Guidelines, and OSHA VPP (Volunt ary Protection Program) guidelines. The core of this approach relies on the commitm ent to and leadership of safety, and the involvement of each individual and each work team in safety. Cooper (1998) criticized the approach th at most accidents were blamed solely on employees rather than the work processes, poor management practices or a combination of all three. He analyzed di fferent theoretical models to ex plain the accident causation, and supported the argument that the main fo cus of accident prevention should be shifted away from the worker’s unsafe acts a nd more onto the organization’s overall management system, particularly in relation to the implementation of the organization’s strategic decisions. Cooper conclu ded that safety culture should be the product of multiple goal-directed interactions between people (psychological), jobs (behavioral) and the organization (situational) (Cooper, 1998, page 17). Viewed from this perspective, an organization’s prevailing safety culture is reflected in the dynamic interrelationships between members’ perceptions about, and attitudes towards, organizational safety goals; members’ day-to-day goal-direc ted safety behavior; and the presence and quality of organizational safety system s to support goal-directed behavior.

PAGE 44

28 In recent years, behavior theories are wi dely used in safety management. The behavior-based safety approach is being utilized more extens ively on construction projects. The objective of be havior-based safety is to eliminate unsafe behaviors and encourage safe behaviors through effective ma nagement interventions, based on behavior and psychological theories. The core of the be havior-based safety approach is to track both safe and unsafe behaviors of workers. According to Sutherland et al. (2000), the basic premise of the behavior-based safety approach can be stated very simply: Behavior is determined by its consequences That is, people tend to repeat those behaviors that produce 'positive' consequences, and not repeat those that result in either no positive or 'negative' consequences. Therefore, by giving workers with safe beha viors "nice" things or removing negative effects, safety behavi ors can be positively reinforced. On the contrary, by giving workers with unsafe behaviors negative f eedback or by taking away "nice" things, unsafe behaviors will be punished and eliminated. This is also known as negative reinforcement. Workers who wo rk safely could be rewarded (positive reinforcement), while workers who ar e unsafe could be punished (negative reinforcement). Management must decide which of these are most appropriate. Peterson (1988) summarized the most co mmon ways the behavioral scientist has contributed to safety management. The major approaches include the following: Survey methodology. Managers concerne d about characteristics of their organizations have often used employee att itude surveys to identify problem areas, to assess the effects of organizational ch ange or policies on job satisfaction, and as an index of management performance. Selection and assessment of personnel. The contribution of the behavioral sciences, particularly psychology, to improved selection, placement, and assessment of personnel is a most common use. Programmed Instruction. This concept includes such sensible elements as behavioral definition of training goals, the analysis of learning tasks into

PAGE 45

29 manageable units, and systematic applications of principles of reinforcement to the learning situation. Motivation theory. Intervening safety behavior through motivation techniques is the core of the theory. In the minds of some observers, the impact of the motivation theory and the application of th e practices they advocate constitute the major contribution of the behavioral sc ientists to the construction industry. Participative management. Workers would be more fully committed if they had a voice in setting the goals and conditi ons of their work environment. Organizational development. Organizat ional development is a process of systematic, planned change in an orga nization. It incor porates a definite philosophy that organizational change should be in the direction of mutual trust, openness in communication, readiness to deal with feelings as legitimate data, and understanding intraand intergroup relations. Today, behavior-based safety approach es are widely used by many proactive owners and contractors on their construction projects. The common approaches include safety observations, worker participation (throug h job safety analysis, safety survey etc.), and safety recognition programs. Summary From the literature review, it is apparent that although various recommendations exist, there is still no solid ev idence in the previous research to describe how owners can efficiently promote construction safety on their projects, and how to assess the relationship between safety performance and th e owner’s involvement in project safety. A detailed research study at the project level to evaluate the owner’s impact on construction safety has not been conducted. This conclusion gave impetus for conducting this research. Certain aspects to be inve stigated include the owner’s commitment to safety, the selection of contractors, c ontractual requirements, and the owner's participation in safety management during construction.

PAGE 46

30 CHAPTER 3 RESEARCH MOTHODOLOGY Since previous studies and pe rspectives expressed in th e literature suggested that the owner's impact on construction safety can be demonstrated in va rious ways, this study will focus on the influence of these factors on the safety performance of construction projects. This chapter will describe the rese arch methodology to disclo se the relationship of the owners’ actions on pr oject safety performance. Research Design Research Components The objective of the study: To compare the effectiveness of several different factors of the owner's involvement in construc tion safety on project safety performance. In order to identify the relationship betw een a specific factor and project safety performance, the median safety performan ces of projects on which owners implement specific safety approaches will be compared with the median safety performances of projects on which owners do not have such an approach. Universe: All construction projects in North Amer ican with U.S. and Canadian owners Experimental Unit: Any individual c onstruction project Population: Collection of safety performances of projects in the universe U-Sample: Collection of construction projects included in the study P-Sample: Collection of safety performa nces of projects in the study Inclusion criteria:

PAGE 47

31 The projects selected in the study shou ld have at least 100,000 worker hours accumulated to provide viable information on safety performance; 1. The projects should be on-going or have b een completed within the past two years; 2. The safety personnel of the projects must be willing to participate in the study. Response Variable: Project safety performance is the response variable in the study. It is measured by the total OSHA record able injury rate (TRIR) of the project, which is the number of OSHA recordable in juries occurring per 100 full-time workers annually (or injuries per 200,000 worker hour s of exposure). TRIR can be easily calculated with the number of OSHA recordable injuries and total worker hours expended on a project. Factors and levels: Each specific aspect of owne r involvement in construction safety can be identified as a factor. Theref ore, there are multiple factors included in this study. Based on the results of the literature re view and previous study results, the factors of interest can be classified in to four categories, as follows: Project context : the characteristics of the pr ojects, including the following: The size of the project as measured by the estimated worker hours expended on the project (quantit ative factor); The construction effort on the project: wh ether a new project, renovation project, maintenance project, or shutdown/turnaround/ outage project (qua litative factor); The type of the project: whether a manufact uring project, a petrochemical project, a civil work, a residential project, a uti lity project, or a commercial project (qualitative factor.); The labor relationship of the project: whet her the labor on site is union, open shop or merit shop (qualitative factor); Type of owner for the project: whether public or private ( qualitative factor); Basis for contract award, whether the c ontractor is selected through competitive bidding or not (qualitative factor);

PAGE 48

32 Type of contract arrangement used on the project: whether general contract, multiple primes, design-build, C.M. at risk, or C.M. agency contract (qualitative factor); Type of contract used on the project: wh ether lump sum contract or cost plus contract, similar to a time and material s (T&M) contract (qualitative factor); Firm providing the workers' compensation insurance on the project, whether the owner (OCIP), the project (PCIP), the ge neral contractor ( CCIP), or if each employer provides their own insuran ce coverage (qualitative factor); Number of work shifts per day: whether one, two, or three ( quantitative factor); Number of workdays per week: whether four, five, six, or seven (quantitative factor); Number of workers at the peak level of employment on the project (quantitative factor); Total estimated monetary value of the project (quantitative factor); Percent of workers on site that do not speak English (quantitative factor); Number of subcontractors awarded on the project (quantitative factor); The prime contractor's commitment to sa fety, as evaluated by the owner's site personnel (quantitative factor); Selection of contractor : the factors related to safety that owners consider when selecting contractor(s) to execu te their projects, including: Whether or not the owner uses a prefe rred list when selecting a contractor (qualitative factor); Whether or not safety performance is a consideration in contractor selection (qualitative factor); Measurements used to evaluate safety pe rformance of a contr actor including TRIR, EMR, safety personnel qualifications, proj ect team qualifications quality of overall safety program, the OSHA log, and OSHA in spection history (qualitative factor); Contractual safety arrangement : the safety requirements that are included in the contract between the owner and contractor. This category will incl ude different safety

PAGE 49

33 management techniques that contractually mu st be employed by the contractor. They are typically qualitative factors. Owner's involvement during project execution : the practices of the owner during the project execution to imp rove safety, including: Responsibilities of owner's site safety representative (qua litative factor); Implementation of a safety observation progr am on the project (qua litative factor); Specific items that must be included in th e contractor's safety program (qualitative factor); Means of addressing injury statistics on the project: whether by project, by contractor, or blended with owner's safe ty performance sta tistics (qualitative factor); Treatment of near misses: whether recorded and investigated, or recorded without further investigation (qualitative factor); Owner support of project safety, specifical ly whether extra funds are provided to promote safety (qualitative factor); Owner participation in an OSHA Voluntar y Protection Program (VPP) (qualitative factor); Whether or not a nurse or EMT is provided for the construction site by the owner (qualitative factor); Extent of owner involvement in the safe ty training program ( qualitative factor); Implementation of a safety recognition/rewa rd program on the project, and the level of the owner’s involvement (qualitative factor); Distribution of a safety newsletter on the project (quali tative factor); Use of safety lunches to promot e safety (qualitative factor); Incorporation of safety in the design of the project (qua litative factor); Most of the factors are qualitative factors, and a few are quantitative factors. For each qualitative factor, the parameters of interest are:

PAGE 50

34 Y = the true median TRIR of all projects where the owners employed the safety technique N = the true median TRIR of all proj ects where the owners did not employ the safety technique The statistics of interest are: YM = The observed median TRIR of all pr ojects where the owners employed the safety technique NM = The observed median TRIR of all pr ojects where the owners did not employ the safety technique Thus, the objective of this study could be achieved by testing the hypothesis (H0: Y = N vs. Ha: Y < N) for each qualitative factor. This is a one-tail test, since TRIR can be expected to be less when the owner implem ents a safety approach or when the project possesses a specific characteris tic identified in the literat ure review. The hypothesis testing will be further discussed in chapter 5. Method Design This study is a retrospective observationa l study, in which both response variable and factor levels are observed from the sample s selected. Personal interviews based on a questionnaire were used to collect most of the data necessary for the study on each project selected. This is one of the methods frequently us ed in studies on construction projects. Since construction projects generally cost millions of dollars, and they are often under the influence of many different factors, it is not practical to conduct a prospective experimental research study. The interview was therefore employed to obtain as much information as possible.

PAGE 51

35 Because little prior research had been conduc ted on this topic, an initial study was necessary to get some idea of the most co mmonly employed techniques of owners to promote construction safety, and whether they mi ght be related with safety performance. Also, a follow-up questionnaire su rvey with contractor safe ty personnel was conducted to help verify some of the findings and determin e how contractors would like owners to get involved in project safety management. As the study is an uncharted one, it is difficult to reasonably determine which factors might be extraneous factors and wh ich might be major factors that influence project safety. Therefore, all the factors s uggested in previous st udies and the industry pilot study were included and analyzed. Stat istical analysis and discussion will identify which factors make a difference on project sa fety performances. Finally, based on the analysis, a model was built to explain how owners can influence construction safety. Sample Size Determination Since the objective of this study was to compare Y to N, the type of inference is hypothesis testing. In order to determine the minimum acceptable sample size required for the study, the following steps to perf orm hypothesis testing were followed: 1. Determine the null hypothesis (Ho). The null hypothesis is Ho: Y = N. 2. Determine the alternative hypothesis (Ha). The alternativ e hypothesis is Ha: Y < N (one-tail test). When the literature suggest ed that projects wher e owners implemented specified practices would have better safety performance, and the median TRIRs also supported the statement, a one-ta il test was conducted, i.e. Ha: Y < N. 3. Determine a suitable significance level and power for the statistical test. The significance level is set as = 10%, instead of 5%. Since this study intends to identify those potential infl uences owners can have on pr oject safety, the significance level is set higher than usual. The power of the statistical test is set to be 1= 1-20% = 80% ( is set to be 20%). That is to sa y, the probability of a Type I error (the probability of erroneously rejecting Ho when it is actually true) is 10% and the

PAGE 52

36 probability of Type II error (the probability of erroneously accepting Ho when it is actually false) is 20%, respectively. 4. Choose and compute the test statistic. The test statistics ha ve been discussed. 5. Make correct conclusions. This will be discussed in Chapter 5. According to a previous study (Hinze, 2002) it was estimated that the practical difference between TRIR of two projects that makes sense in the industry is B = 1.0. This is determined to be the clinical significant difference to determine the sample size of this study. Since it was assumed that most of the projects that would participate in the interview study would have better safety pe rformances than the construction industry average (which is 7.8 for TRIR in 2001, BLS, 20 03), the response variable is estimated to be between 0 and 6. Therefore, based on the Empirical Rule, the estimate of the standard deviation of TRIR is = (6-0)/4 = 1.5. Then the sample size determination can be calculated with the following indexes: DELTA = B/ = 1/1.5 = 0.67 = 10%, = 20% By referring to the tables listed in (Rosne r, 1995), the sample size for each factor level (when owners do or do not implement a particular safety ma nagement approach) should be 28 projects. Therefore, the to tal sample size for this study is 28*2 = 56 projects. Computation of the sample size needed to provide valid research findings was found to be 56. This means that a sample size of 56 would be expected to yield reliable results.

PAGE 53

37 Sample Selection Most of the projects in th e project interview phase (the major phase of this study) were recommended by the Construction Industr y Institute Project Team #190 (with the title "the role of owners in construction sa fety"), based on the inclusion criteria. The contact persons of the projects were then aske d whether they could accept an interview. The person was then interviewed with the quest ionnaire after he or she agreed to accept the interview. Essentially, the project sampling was a voluntary sampling, since attitudes of owners towards safety varied significantly, and only those that woul d participate in the study and responded could be included in the samp le. Note that all projects identified in the research was contacted to request their participation. These tended to be projects known to the CII project team members. Ho wever, whether or not the data could be regarded as representative would then depend on the variability of the data collected and the size of the projects. That is, if the variability of the re sponse variable was similar to the large projects in the population (projects of the large owners), it could be regarded as representative of the population. If the sizes of the projects in the study summed up to a significant portion of the total size of proj ects in North American, the data could be reasonably regarded as represen tative. Also, in order to test whether the injury data (which might be sensitive to some owners and therefore biased) was true, the injury triangle by Heinrich (1959) was used to test the results. Data Collection Three studies were conducte d to collect data for this research effort: the pilot study, the project interviews, and a follow-up survey. The pilot study consisted of a mailed survey sent to large owner firms that were known to have significant construction budgets. This research was followed by a study in which representatives of owners were

PAGE 54

38 interviewed at participating c onstruction sites. The interv iew data consisted of more detailed information than was obtained in the mailed survey study, and therefore constituted the major data collection phase. Most of the findings of this research are generated based on the data collected thr ough the project interviews. The follow-up survey of contractors was a supplementary part of the study, w ith the purpose of comparing the findings of the project owner in terviews with the opinions of construction firms. Mailed Survey Study (Phase 1) The first phase consisted of surveys to study the practices of owners and their resultant impact on construction safety at the company level. The purpose of the survey was to establish and refine the hypotheses to be tested in the subse quent interview study. Also, the survey was to help identify new f actors to be tested. This included the acquisition of information on why safety is esse ntial to owners, means used to ensure that safe contractors are employed on their projects how safety is addr essed in construction contracts, how contractors ar e expected to address safety during construction, the means used by owners to monitor contractors duri ng construction, and the nature of the active role played by owners during project ex ecution. The questionnaire was developed by relying on the literature review and the input from constr uction industry safe ty personnel. The survey questionnaire asked questions rela ted to safety performances of the owner projects, safety coordination of the owne r, owner safety management practices on projects, safety in design phases, and other safety related topics (see Appendix a). The U.S. owners surveyed were those making la rge annual expenditures for construction, as listed in the Top 425 owners provided by EN R (Engineering News Report, 2001). Their construction budgets in year 2000 ranged fr om 23.8 million to 3.8 billion dollars.

PAGE 55

39 Project Interview (Phase 2): The second phase of the study was designed to acquire more detailed information, which would not be feasible or realistic in a mailed surve y. The purpose of this study was to test the hypotheses that had been deve loped. This study, with its focus at the project level, was conducted th rough in-depth interviews. In addition to obtaining information about the various practices of ow ners to ensure the safety of construction workers on their sites, information was also ob tained on measures of safety performance. Based on the results obtained from the literatu re review and the mailed survey, a detailed interview questionnaire was designed. The questionnaire went through a series of iterative improvements as suggestions were offered by members of CII Project Team #190. The questionnaire mainly focused on four aspects of the owner’s involvement in construction safety: Project description, Selection of safe contractors, Safety requirements in contracts, and Owner's involvement in safety during construction. Factors such as project characteristics, contractual arrangements, contractor’s commitment to safety, and so on were also a ddressed in an attempt to control external impacts on project safety performance. The questionnaire was modified through appr oximately 15 iterations before it was used to conduct interviews. After three telephone interv iews were conducted, a final version was devised. After the first three in terviews, the interviewees were asked about the clarity and integrity of the questionnaire. This feedback was used to further modify the questionnaire before it was finalized and us ed in the remaining interviews. The final version of the questionnaire is included in Appendix b. Approximately 100 construction

PAGE 56

40 projects were identified for inclusion in this research. The criteri a to select projects included: Under construction or newly comple ted (within the past two years); Worker exposure of at least 100,000 hours (some exceptions were made, as explained later); Variety in the types of projects: public or private; petrochemical, utility, manufacturing, commercial or residential projects Whenever the owner had several smalle r projects at the same location and implemented the same safety management policy, the multiple projects were regarded as a combined project. Persons interviewed included site repr esentatives of the owner, including construction managers, safety managers, and safety coordinators. The face-to-face project interviews generally took one and a half hours to two and a half hours to conduct. Whenever face-to-face interviews were econom ically infeasible, the interviewee was asked to fill out the questionnaire and return it to the researcher by Fax or email. For questionnaires returned this way, a follow-up telephone interview, la sting about half an hour, was conducted to clarify any questions. Most of the projects were introduced by members of the CII Project Team #190. Often either the owner or the contractor on the project, or both, were CII members; howev er, CII membership was not a criterion for inclusion in the study. Many owners of proj ects were not affiliated with the CII. At the conclusion of this study, 81 project s had been interviewed. Among them 59 projects provided the TRIR data and satisfied all the criteria to be included in the data analysis. The final sample of 59 projects included 49 U.S. projects, seven projects in Canada, and three international projects w ith U.S. owners and U.S. contractors.

PAGE 57

41 Follow-Up Survey To Contractors (Phase 3) A survey of contractors was conducted af ter the project owne r interviews were completed. This was not a major part of th e data collection. This study was to obtain information from contractors, because cont ractors may hold different views towards the same questions regarding owners. Therefor e, it was deemed reasonable to conduct a separate survey of contractors to identify possible bias of th e data obtained from owners. Also, some questions asked how the contract ors would like the owne rs to help promote project safety. The information could he lp communication and cooperation of the two parties on future projects. This survey was conducted with a short questionnaire developed on the basis of the analysis result s obtained from the project interviews (see Appendix c). The participating contractors ma y or may not have been involved in the owners’ projects in the primary study. Data Analysis The main objective of the project was to test whether the involvement of owners in safety management made a difference to th e safety performances of projects. Project safety performance, defined as the OSHA record able injury rate (also referred to as Total Recordable Injury Rate, TRIR), was the de pendable variable in the analysis. The different safety management techniques implemented by the owners were regarded as the independent variables. Other information obtained relating to project properties and contractor’s commitment to safety constitute d extraneous factors and were not the major variables to be considered in the research. They were also regard ed as the independent variables in the analysis. Analysis of the data obtained from th e mail-out questionnaire survey to owners (Phase 1) used non-parametric methods, univa riate methods and the logistic regression

PAGE 58

42 method based on the type of the response variab les. The analysis attempted to recognize the involvement of owners that might help project safety. Analysis of the data set obtained through the project interviews (Phase 2) was the major part of the data analysis It can be divided into thr ee stages. The first stage was focused on the descriptive statistics of the vari ables, that is, means, medians and standard deviation of the interval vari ables, and the proportions of nominal and ordinal variables were calculated. By this means, the frequenc y and popularity of differe nt safety practices and techniques were identified. The s econd stage aimed at testing reasonable associations among the variables, especially the a ssociations between safety performance and safety management approaches. Effort s in the third stage were conducted on the establishment of a model to define the rela tionship between project safety performance and various types of owner involvement. A scorecard was developed to evaluate the owner’s involvement in construction safety management. All the data, once obtained, were inpu t into a document developed in the Statistical Package for Social Sciences ( SPSS v10.0), according to the data structure of the questionnaire. In the first stage of data an alysis, the software was used to describe the properties of the projects interviewed, includi ng their location, size, safety performance, contractual arrangement, type s and labor arrangement. Fre quencies of different safety management techniques were also summari zed to provide a holistic picture of the information of the projects interviewed. As one aspect of the analysis, data from projects in Canada were examined separately from th e projects in the U.S to address any possible regional variations. It should be noted that no clear differences were found between the U.S. and Canadian projects.

PAGE 59

43 In the second stage, multi-variable st atistical analysis methods, especially nonparametric methods (which treat the respons e variable as an or dinal variable) were conducted to find the relationship between safety performance and the safety management techniques. Safety performan ces of projects with different safety management techniques were compared to identify those techniques that made a difference on safety performance. In this st age, the data were stratified to get a better idea of how some owners emphasized safety on large projects, and how owners would focus on safety on petrochemical projects. This was done because the pilot study showed that larger contractors and petrochemical projects appear ed to have better safety performances. When consideri ng the variety of the projects interviewed, the significance level was set at 0.10, to detect all the sa fety management techniques implemented by owners that could potentially impact safety performance. In the third stage of the data analysis answers to some open-ended questions, as well as some detailed cases, were considered to test how the owner’s involvement could make a difference to project safety performan ce. Based on the analysis in the first and second stages, this stage was focused on estab lishing a model to describe the cause-effect relationship between the owner’s involvement a nd project safety performances. Also, the most effective and important points for the fu ture reference by owners were summarized into a scorecard to evaluate the owne r’s involvement in safety management. Analysis of data collected in the surv ey of contractors (P hase 3) was mainly composed of descriptive statistical analysis. Those statistics were then referenced to support the findings in the inte rview questionnaire, and provide suggestions for owners to better cooperate with contractors in promoting construction safety.

PAGE 60

44 CHAPTER 4 QUESTIONNAIRE SURVEYS FOR LARGE OWNERS In this chapter, the analysis of the data obtained from the questionnaire survey is presented with different nonparametric statistical met hods, including the Mann-Whitney U test, Wilcoxon test, Kruskal-Wallis test, and Friedman two-way ANOVA. These nonparametric tests were selected because of the low response rate. By using the nonparametric statistical methods, the medians (or distributions) of the re sponse variable for different factor levels are compared. The Data Set The population surveyed was the ENR-list of the top (based on expenditures on construction) 425 owners in the U.S. The questionnaire was mailed out on February 21, 2002. Thirty-one responses were received by April 12, 2002, after which no additional replies were received. The response rate was about 7.3%, which was lower than the expected response rate of 10%. Possible r easons of the low response rate can be: The envelope of each survey was addresse d to “Construction Contract Manager”. Although this was the title suggested by some owner members of CII, the appropriate person to fill out the survey may have a different title in different companies. It is quite probable that th e questionnaires were not received by the appropriate individuals. The questionnaire may have been too le ngthy, requesting too much information. Some individuals may not be patient e nough to complete the surveys and return them. This was considered during the de sign of the survey, and it was determined that people familiar with safety could comp lete the survey within 20 minutes. Most of the information sought required a multiple-choice response. Only companies that emphasize safety and had better safety performances may have decided to participate in the study. It is possible that safety is still not the top

PAGE 61

45 priority in many owner comp anies. Some companies ma y be reluctant to reveal information about an unacceptable safety performance. As the consumer of construction products and services, most owner companies are expected to have their primary interests in the revenue generating sectors of their own businesses. It is suspected that many owners still regard c onstruction safety as resting totally with the contractors who c ontract their construction services. Thus, many owners are not familiar with the safety issues on their projects, and may not have felt capable of completing the survey. Because of the above reasons, the mail-out survey was regarded as a voluntary field experiment, with responses being received fr om those who had an interest in and had a strong focus on safety management. Therefor e, the results are probably biased and not representative of all owner companies. Th is further supports the use of non-parametric statistics in the analysis. However, as men tioned in the previous chapter, the aim of the questionnaire survey was to id entify the practices of the owners associated with better safety performances, and to develop hypotheses to be tested in the subsequent project interview phase of the resear ch. The survey data, although limited and probably biased, can still serve a valuable purpose and provide enlightenment for the follow-up research study. It should be noted that for non-parametric methods such as the Kruskal-Wallis tests, or Friedman two-way ANOVA, the objectiv e is to compare the medians of different groups, and the response variable is treated as an ordinal variable. The mean ranks that will show in the analysis result tables simply mean the average rank of the coded response variable, and a larger mean rank suggests a larger median of the category. Because only 19 of the respondents provi ded the TRIR of their construction projects in the past year, it was difficult to develop any strong conc lusions about the types of projects that have better safety pe rformance. However, among the 19 owners providing the TRIR data, ten ha d primary manufacturing proj ects (with the average TRIR

PAGE 62

46 of 2.92) and four had primary petro-chemical projects (with the aver age TRIR of 1.13). In order to discover the underlying reason for the difference, the rating of the importance of safety performance when selecting contract ors was considered as a safety index of each owner (ten was the most important and one was the least important). The ranking fo the importance of safety when selecting cont ractors was the primary measure used in the analysis since 31 respondents provided this in formation. If the TRIR had been used, the sample size was redeemed to 19, an unacceptabl e level. The use of the ranking measure was felt to be a viable appro ach as the ranking value was found to be relatively consistent with the TRIR. Analysis Of The Mailed Survey Data The Projects Context Of the responding owners, the annual cons truction budget, the pe rcentage of the construction budget on new projects, and the domin ant type of construc tion projects were the characteristics of partic ular interest. These factor s can influence the safety performance of the owner. The Kruskal-Wallis test was conducted to test whether the ranking of safety importance was dependant on the major type of projects. The results of the test are shown in Table 4-1. Even with a small numbe r of responses, it is apparent that petrochemical project owners place a higher importa nce on safety when selecting contractors, while owners of building projects have the lowest importance. The underlying reason may be associated with the at titudes of the owner towards construction safety in their own businesses. For example, the chemi cal industry and nuclear power generation industry are famous for their strict concern for safety. Therefore, owners in these industry sectors may similarly prioritize safety in their construction projects, and actively

PAGE 63

47 participate in safety management on the project s. The OSHA recordable injury rates of the four owners with petro-chemical projects (0.50, 1.10, 1.42, and 1.50, respectively) were also much lower than those of the othe r owners. This indicates that these owners have much better safety performa nces than the industry average. Table 4-1. Importance of safety during selecti on of contractors to owners of different types of projects (1 to 10 with 10 being the highest possible rating) Project Type Count Mean Median Mean Rank of Kruskal-Wallis Test Kruskal-Wallis Test Sign. (2-tail) Building 5 5.2 5 6.5 Others 2 7.5 7.5 13.25 Manufacturing 18 7.94 8 17.14 Utilities 2 8 8 16.5 Petro-Chemical 4 9.25 9.5 23.88 Total 31 7.65 8 0.05 Note: In this table, a higher Mean Rank m eans that the safety index of the group of owners tends to be higher, or the median of the safety index tends to be higher, and therefore, the owners are more concerned with safety. It is the same as in Table 4-2. The same statistical test was conducted to determine if there was a difference among owners with different annual constructio n budgets. The result is shown in Table 4-2. It is evident that owners with larger construction budgets are more concerned with safety than owners with small construction budgets. The Mann-Whitney test is conducted to compare the importance of sa fety to owners with $25-100 million annual construction budgets to those with >$500 millio n annual budgets, and the results showed that those with budgets >500 million placed a higher priority on safety (p<0.05). Table 4-2. Importance of safety when selec ting contractors to ow ners with differing construction budgets Annual Budget Count Mean Median Mean Rank of Kruskal-Wallis Test Kruskal-Wallis Test Sign. (2-tail) 25-100 million 12 7 7 12.46 100 500 million 13 7.62 8 15.96 >500 million 6 9 10 23.17 Total 31 7.65 8 0.05

PAGE 64

48 The percentage of new projects in the construction budget was examined, but this was not a significant factor influencing th e owners’ concern for safety. This is reasonable because owners with more ne w projects may not necessarily emphasize construction safety more than owners with fewer new projects, and they may not be good at construction safety performance, either. Commitment Of The Owner To Construction Safety Commitment of the owner to constructi on safety was measured by two questions: the earliest time the owner began to emphasi ze safety on the project (question 5) and the philosophy held by the owner concerning the res ponsibility for safety (question 13). In the questionnaire, both questions were coded w ith ordinal variables to show the degree of owner involvement in construction safety. The earliest stage to emphasize safety (e.g. during concept and feasibility phase) was c oded as a smaller number, and beginning to emphasize safety at a later period (e.g., during construction) was assigned a higher number. For the philosophy concerning safet y, a small number was assigned to more proactive attitudes (e.g. taking total control of safety). Owners with different types of major projec ts and owners with different sizes of annual construction budgets were compared. The answers to que stion 5 (earliest stage to emphasize construction safety) were significan tly different among owners with different annual budgets and owners of different types of projects (refer to Table 4-3 and Table 44). It is evident that larger owners and owners of petrochemical projects started to emphasize construction safety at a significantly earlier stage of the projects. As to the philosophy towards construction sa fety liability, owne rs with different size projects did not show significant difference s. It was noted that petrochemical owners tend to have

PAGE 65

49 more proactive attitudes towards safety (addres s safety in early project stages) and they would assume greater liability risks. Table 4-3. Earliest stage to emphasize safety vs. annual construction budget Annual Budget Count Mean Median Mean Rank of Kruskal-Wallis Test Kruskal-Wallis Test Sign. (2-tail) 25-100 million 12 3.33 3 20.96 100 500 million 13 1.92 1 13.54 >500 million 6 2 1 11.42 Total 31 2.48 3 0.03 Note: In this table, a smaller Mean Rank means that the earliest stage to concern for safety is earlier, and therefore, the owners ar e more concerned with sa fety. It is the same as in Table 4-4. Table 4-4. Earliest stage to emphasize safety vs. major type of projects Project Type Count Mean Median Mean Rank of Kruskal-Wallis Test Kruskal-Wallis Test Sign. (2-tail) Petro-Chemical 4 1 1 7.5 Utilities 2 2 2 13.5 Manufacturing 18 2.33 2 14.67 Building 5 4.2 3 22.8 Total 29 2.48 3 0.03 Note: Refer to footnote for Table 4-3. Selection Of Safe Contractors Most of the owners (80% of the res pondents) would establ ish minimum safety performance requirements in addition to the OSHA regulations when selecting contractors to construct their projects. When selecting cont ractors, safety performances of the contractors were considered by most owners (84%). After counting the number of different safety approaches taken into cons ideration when selecti ng contractors, it was found that the mean number of such re quirements was 2.4, ranging from a maximum count of 5 to a minimum count of zero. With the Kruskal-Wallis test, the number of the safety selection criteria was co mpared between owners of diffe rent sizes and owners with different project types. The results are shown in Table 4-5 and Table 4-6, respectively.

PAGE 66

50 From the results, it is reasonable to expect th at larger owners have more stringent safety criteria when selecting contra ctors. In order to analy ze the relationship between the importance of safety and the criteria used, the Mann-Whitney test was conducted and the results are shown in Table 4-7. It is ev ident that using each criterion or not is significantly related to the importance of safety in the selection of contractors. The loss ratio was the only exception, which was implemented by very few owners. Table 4-5. Number of safety crite ria vs. annual construction budget Annual Budget Count Mean Median Mean Rank of Kruskal-Wallis Test Kruskal-Wallis Test Sign. (2-tail) 25-100 million 13 1.58 1.5 11.85 100 500 million 12 3.15 3 17.21 >500 million 6 4.00 5 22.58 Total 31 2.71 3 0.04 Table 4-6. Number of safety criteria vs. major type of projects Project Type Count Mean Median Mean Rank of Kruskal-Wallis Test Kruskal-Wallis Test Sign. (2-tail) Building 5 0.8 0 6.7 Manufacturing 18 3 3 16.14 Utilities 2 3 3 15.25 Petro-Chemical 4 4 4 20.13 Total 29 2.71 3 0.08 Table 4-7. Results of Mann-Whitney test: sa fety selection criterion vs. importance of safety in the selection of contractors Contractual Safety Requirements Ans. Cnt. Mean Median m. r. s. r. Sign. N 14 6.86 7.5 13 182 Experience Modification Rating (EMR) of the contractor. Y 17 8.29 8 18.47 314 0.05 N 19 7.05 8 13.24 251.5 OSHA recordable injury rate of the contractor. Y 12 8.58 9 20.38 244.5 0.02 N 28 7.71 8 16.23 454.5 Loss Ratio of the contractor, should be less than Y 3 7 8 13.83 41.5 0.33 N 16 6.88 7.5 12.38 198 Site-specific safety program prepared by the contractor. Y 15 8.47 9 19.87 298 0.01 N 18 6.72 7 11.61 209 Qualifications of the safety staff of the contractor. Y 13 8.92 9 22.08 287 <0.01 N 9 6.11 6 9.17 82.5 Quality of the overall safety program of the contractor. Y 22 8.27 8 18.8 413.5 <0.01 Note: the title of the columns are: Ans = answers, Cnt = counts, mean = mean impor tance of safety (1-10, with 10 being the most important), median = median importance of safety (1-10, with 10 being the most important), m.r. = mean of rank, s.r. sum of rank, sign.= significance level (1-tail)

PAGE 67

51 Safety Management Practices Of The Owners 1) Safety requirements in the construction contract All the owners responded that they will include some safety requirements in their construction contracts. Two requirements, namely, “contractor must comply with the local, state and federal safety regulations”, and “contractor must report all OSHA recordable injuries to the owner” were requ ired by all responding owners. Less than 25% of the owners required contractors to place at least one full-time safe ty representative on the project, or that the contractor must subm it a safety policy signed by its CEO. As to other safety requirements liste d in the questionnair e, about 50% to 75% of the owners included them in the contract. When the counts of contractual safety requi rements of the owners were calculated, it was found that for owners with different t ypes of facilities cons tructed, the counts are considerably different (refer to Table 48). Petrochemical companies included the largest number of safety requirements in th eir contracts, followed by utility owners. Owners of buildings and other manufacturing plants have fewer safety requirements included in their contracts. Also, it was found that owners with larger construction budgets have more contractual safety requirements (refer to Table 4-9). Table 4-8. Number of contractual safety requirements vs. major type of projects Project Type Count Mean Median Mean Rank of Kruskal-Wallis Test Kruskal-Wallis Test Sign. (2-tail) Building 5 4.6 5 8.8 Manufacturing 18 6.7 7 14.75 Utilities 2 7.5 7.5 17.5 Petro-Chemical 4 9 9 22.63 Total 29 6.77 7 0.11

PAGE 68

52 Table 4-9. Number of contractual safety requirements vs. annual construction budget Annual Budget Count Mean Median Mean Rank of Kruskal-Wallis Test Kruskal-Wallis Test Sign. (2-tail) 25-100 million 12 5.33 5 10.96 100 500 million 13 7.85 8 19.65 >500 million 6 7.33 8 18.17 Total 31 6.77 7 0.05 2) Safety management involvement of owners taken during construction One question asked about safety management involvement of owners to promote construction safety on site. It was found that four of the seven listed safety management activities were frequently performed by owners, namely, owners’ personnel conduct periodic job site safety inspect ions or safety audits; owners ’ personnel participate in some contractor safety meetings; owners’ personnel pa rticipate in the investigation of all lost workday injury accidents; and owners monitor injury incidence rates on each project. More than 2/3 of the owners were involved in this manner. Other safety approaches were taken by less than 1/3 of the owners. A Kruskal-Wallis test was conducted to id entify the strength of the relationship between owner's involvement in safety mana gement and the owner' s annual construction budget/the major type of projects. The result s were quite consistent with the previous findings (refer to Table 4-10 and Table 4-11). Petrochemical owners and owners with larger construction budgets were more activel y involved in safety during construction. Table 4-10. Total number of safety practices implemented vs. major type of projects Project Type Count Mean Median Mean Rank of Kruskal-Wallis Test Kruskal-Wallis Test Sign. (2-tail) Building 5 1.6 2 6.5 Utilities 2 3.5 3.5 14.25 Manufacturing 18 3.89 4 15.39 Petro-Chemical 4 5.75 5.5 24.25 Total 29 3.74 4 0.06

PAGE 69

53 Table 4-11. Total number of safety practices implemented vs. annual construction budget Annual Budget Count Mean Median Mean Rank of Kruskal-Wallis Test Kruskal-Wallis Test Sign. (2-tail) 25-100 million 12 2.75 2 11.38 100 500 million 13 4.23 4 18.04 >500 million 6 4.67 5.5 20.83 Total 31 3.74 4 0.05 The Mann-Whitney U test was conducted to compare the rank of the importance of safety in the selection of contractors and whether or not owners were involved in construction safety in a particular way. Th e results are shown in Table 4-12. It is reasonable to suspect that owners who emphasi ze construction safety tend to be involved in more ways than owners who care less about safety. Table 4-12. Importance of safety vs. safety approaches taken Safety approaches Ans Cnt. Mean Median m. r. s. r. Sign. Y 7 8 8 9.93 69.5 Owner’s personnel conduct periodic job site safety inspections or safety audits. N 24 6.43 7 17.77 426.5 0.02 Y 21 8.3 9 14.07 295.5 Owner places a safety person on the project to support the contractor on project safety. N 10 7.33 8 20.05 200.5 0.04 Y 22 9.22 10 12.95 285 Owner provides a nurse or emergency medical technician (EMT) for the construction project. N 9 7 8 23.44 211 <0.01 Y 7 8.12 8 8.86 62 Owner’s personnel participate in some contractor safety meetings. N 24 6 7 18.08 434 0.01 Y 9 8.05 8 11.28 101.5 Owner’s personnel participate in the investigation of all OSHA lost workday injury accidents. N 22 6.67 6 17.93 394.5 0.03 Y 26 8.8 9 15.02 390.5 Owner implements a safety incentive that can be earned by the contractor for completing the project below a specified OSHA recordable injury rate. N 5 7.42 8 21.1 105.5 0.01 Y 9 8.18 8 8.78 79 Owner monitors injury incidence rates on each project. N 22 6.33 6 18.95 417 <0.01 Note: the title of the columns are: Ans = answers, Cnt = counts, mean = mean impor tance of safety (1-10, with 10 being the most important), median = median importance of safety (1-10, with 10 being the most important), m.r. = mean of rank, s.r. sum of rank, sign.= significance level (1-tail)

PAGE 70

54 If Spearman’s correlation (non-parametric correction) is calculated for the number of safety approaches, the number of safety re quirements included in the contract, and the rating of safety performance during the select ion of contractors (as shown in Table 4-13), it is evident that they are significantly pos itively correlated. Owners emphasizing safety tend to include more safety requirements in th eir contracts and utilize more approaches to promote safety on their projects. Table 4-13. Non-Parametric corr elation matrix between impor tance of safety, number of contractual requirements and sa fety involvement approaches Importance Contract Approaches Correlation Coefficient 1 0.734 0.651 Sig. (2-tailed) <0.01 <0.01 <0.01 Importance N 31 31 31 Correlation Coefficient 0.734 1 0.578 Sig. (2-tailed) <0.01 <0.01 <0.01 Contract N 31 31 31 Correlation Coefficient 0.651 0.578 1 Sig. (2-tailed) <0.01 <0.01 <0.01 Approaches N 31 31 31 3) Safety training and safety in design Safety training was one area of focus in th is survey, namely to explore how owners addressed safety training on their projects. The Mann-Whitney test was conducted with the rating of safety performance as the depe ndent variable. Four questions related to safety training were tested against the depe ndent variable, namel y, whether the owner’s safety representative attends safety orienta tion (q07), whether the owner makes monetary allocations for safety training (q10), whether standardized training is required for all the projects (q14), and whether basic safety tr aining is required for all workers (q17). The results are shown in Table 4-14. Owners who are more concerned about safety tend to emphasize the safety training and orientation program.

PAGE 71

55 Table 4-14. Mann-Whitney test: Importance of safety vs. safety training inputs Safety training inputs Ans. Cnt. Mean Median m. r. s. r. Sign. N 11 5.82 6 8.05 88.5 safety representative attends safety orientation (q071) Y 20 8.65 8.5 20.38 407.5 0.01 N 17 6.53 7 10.85 184.5 does the owner make allocations for safety training (q10) Y 14 9 9.5 22.25 311.5 0.03 N 7 6.29 6 8.86 62 standardized training is required for all the projects (q14) Y 24 8.04 8 18.08 434 0.02 N 6 5.67 5.5 6 36 basic safety training is required for all the workers (q17) Y 25 8.12 8 18.4 460 0.01 Note: the title of the columns are: Ans = answers, Cnt = counts, mean = mean impor tance of safety (1-10, with 10 being the most important), median = median importance of safety (1-10, with 10 being the most important), m.r. = mean of rank, s.r. sum of rank, sign.= significance level (1-tail) The total number of positive answers to the four safety training and orientation questions was calculated. The Kruskal-Wallis test was conducted to examine the owners’ involvement in training with the size of the construction budget and the major project type of the owners (refer to Tables 4.15 a nd 4.16). Not surprisingly, larger owners and petro-chemical owners tend to have more input into safety training and orientation. Table 4-15. Total amount of safety training input vs. project type Project Type Count Mean Median Mean Rank of Kruskal-Wallis Test Kruskal-Wallis Test Sign. (2-tail) Building 5 1 1 6.5 Manufacturing 18 2.72 3.5 15.64 Utilities 2 3.5 3.5 18.75 Petro-Chemical 4 3.75 4 20.88 Total 29 2.68 3 0.06 Table 4-16. Total amount of safety input vs. annual construction budget Annual Budget Count Mean Median Mean Rank of Kruskal-Wallis Test Kruskal-Wallis Test Sign. (2-tail) 25-100 million 12 1.83 1 12.08 100 500 million 13 3.15 3 17.83 >500 million 6 3.33 4 20.83 Total 31 2.68 3 0.08

PAGE 72

56 Safety can be addressed as early as the design phase of a proj ect. About half the owners stated they would address safety during the design stage. The Mann-Whitney U test was conducted with the importance of safety to the owners and whether or not designing for safety was a concern. The resu lts are shown in Tabl e 4-17. It can be concluded that owners addre ssing safety during the project design will also emphasize on the importance of safety performance dur ing the selection of contractors. Table 4-17. Relationship between the importance of safety and whether or not safety is addressed in the design phase Address safety in the design phase Count Mean Median Mean Rank of MannWhitney U Test Mann-Whitne y U Test Sign. (2-tail) No 12 6.67 7.5 12.08 Yes 19 8.26 8 18.47 Total 31 7.65 8 0.05 4) Ranking of different safety approaches Question 23 of the survey asked respondent s to rank the importance of different safety approaches. A Friedman two-way ANOVA (for repeated measures) was conducted to compare the means of ranking for different approaches (see Table 4-18). Table 4-18. Priority of different approaches pr eferred by owners (1 to 5 with 1 being the highest possible priority) Safety approaches Mean Median m. r.* Sign. Owner selects safe contractors to carry out the work. 1.45 1 1.83 Owner emphasizes safety and constructability in design. 2.1 2 2.5 Owner participates in and monitors safety during the entire life of a construction project. 2.03 2 2.54 Owner develops an effective safety recognition and reward program. 3.46 4 4 Owner dedicates funds to support th e contractor’s efforts in safety. 3.63 4 4.13 <0.01 mean ranking of Frie dman two-way ANOVA test. It is evident that significant differences exist between th e priorities owners placed on different safety approaches. The order of approaches, be ginning with the top priority, is as follows: 1. Owner selects safe contract ors to carry out the work.

PAGE 73

57 2. Owner emphasizes safety a nd constructability in design. 3. Owner participates in and monitors safe ty during the entire life of a construction project. 4. Owner develops an effective safe ty recognition and reward program. 5. Owner dedicates funds to support th e contractor’s efforts in safety. A Wilcoxon matched-pair sign ed-rank test (for repeat ed measures) was conducted to compare the mean ranking of selecting safe contractors (referred to as variable Q232) and participating in and mon itoring safety (Q233). It was found that the mean rank of selecting safe contractors is significantly higher than that of participating in and monitoring safety at the 0.04 le vel (refer to Table 4-19). This indicates that for many owners, the most efficient way for them to improve project safety performance is in selecting safe contractors, instead of activ ely participating in and monitoring safety. Table 4-19. Comparison of the ranks of selectin g safe contractor (Q232) v.s. participating in and monitoring safety (Q233) N Mean Rank Sum of Ranks Sign. Negative Ranks 7(a) 15.36 107.5 Positive Ranks 20(b) 13.52 270.5 0.04 Ties 4(c) Q233 Q232 Total 31 a Q233 < Q232; b Q233 > Q232; c Q232 = Q233 Note: the mean ranks and sum of ranks are the results of Wilcoxon matched-pair signed-rank test (for repeated measures). Summary Based on the statistical anal ysis of the survey responses, the following hypotheses were developed for further testing: The main type of projects of the owner, which may suggest the main business of the owner, has strong effects on the owner’s commitment to safety and practices to promote construction safety on their projec ts. Owners of petrochemical and utility projects tend to emphasize safety more th an owners of manufacturing facilities and buildings.

PAGE 74

58 The size of the annual construction budget s eems to be an important factor that impacts the safety practices of owners. Owners with larger construction budgets are more concerned about construction safety. Owners of projects with be tter safety performances te nd to emphasize safety during the earlier stages of project development. They will emphasize safety performance more during the selection of contractors, and they will also include more safety requirements in their contracts. Owners with safer projects tend to be heavily involved in project safety management, especially by implementing va rious approaches to promote safety management. Although these approaches may not be extensively used in the construction industry, they are widely im plemented by safe owners with large construction budgets. Instead of trying to contractually avoid liab ilities from injuries owners with safer projects tend to be more involved with contractors on project safety management, especially safety training. Owners who place greater emphasis on safety tend to address safety in the design of their projects. Among different approaches taken by owners to achieve success in project safety, owners proactive on safety tend to believe that selecting safe contractors to construct their projects and emphasizing safe ty in the design phase are important. The relationship between the overall safe ty performances of projects and the different safety management practices of th e owners could not be thoroughly examined, because only 19 owners provided their injury rate data. Despite this, the statistical analysis was able to identify some possible relationships between safety performance and the project type, selection of safe cont ractors, safety commitment, contractual requirements, and the safety involvement of owners.

PAGE 75

59 CHAPTER 5 RESULTS: OWNER'S ROLE IN PROJECT SAFETY In this chapter, the results obtained in th e major data collection stage are presented. The methods of interviews and selection of projects are intr oduced. The most frequent practices implemented by owners are then pr esented. Statistical analysis was conducted to test associations between project safety performance and owners' practices. Finally, a model to describe the owner's influence on pr oject safety and a sc orecard to evaluate owner involvement in safety were developed. The Interviews The second and major phase of this st udy was to collect information from construction projects through interviews to de termine the role and influence that owners have on construction project sa fety. As mentioned in the description of the research methodology, the focus of this study was to te st the null hypothesis that project safety performance was independent of the facility owner’s involve ment, against the alternative hypothesis that project safety performance was dependent on the owner’s involvement. Project safety performance was the dependent or response variable, and the responses to different questions were different factor levels or indepe ndent variables. A total of 81 personal interviews were conducted. Each interview took about one and a half hours. When the data were an alyzed, a constraint was imposed such that projects included in the an alysis must have had at least 100,000 hours of worker exposure. This was to ensure that the sa fety performance measures gave a reliable

PAGE 76

60 indication of the project safety performance. Since some projects were in their early stages of construction, they did not have the requisite num ber of hours to satisfy the criteria for inclusion in the data analysis By excluding the projects with fewer than 100,000 hours of worker exposure and those project s for which complete injury data were not provided, the final anal ysis included 59 projects. Safety performances of the projects were measured by using the total number of OSHA recordable injuries per 200,000 hours of worker exposure, commonly known as the Total OSHA recordable injury rate (TRI R). Whenever the term TRIR appears, it should be clear that the measur e of safety performance consists of all OSHA recordable injuries, including lost-time injuries and rest ricted work injuries. For the 59 projects included in the analysis, the average TRIR was 1.95, with six proj ects reporting zero OSHA recordable injuries. One of the pr ojects reporting a TRIR of zero had amassed nearly 500,000 hours of worker exposure. The safety performances of projects were compared with the different responses to the questions in the questionnaire. The hypothe ses were tested to determine if project safety performances were associated with each of the different responses. The major statistical methods employed were non-para metric methods, including the Mann-Whitney U test (for comparison between two factor levels) and the Kruskal-Wallis test (for comparison between more than two factor levels). The rationale for using nonparametric methods was the difficulty of sa tisfying the normality assumption and equal variance assumption in each factor level, whic h were required for ttest and analysis of variance (ANOVA) for comparing means of quantitative variables. By using nonparametric methods, the purpose was to compar e the distribution or medians of project

PAGE 77

61 safety performances for different factor le vels. In these non-parametric tests, the response variable, TRIR, was treated as an ordinal variable, instead of an interval variable. No assumptions were required fo r the non-parametric te sts. The level of significance indicated whether th e medians of the response vari able were different when responses differed. As a comparison, the means (or averages) of TRIR of different factor levels were also compared by using the univariate analys is method, one-way analysis of variance (ANOVA). The statistical assumptions for ANOVA were: (1) the response variable, TRIR, had a normal distribution for different factor levels; and (2) the variance of the response variable in different levels were the same. Si nce ANOVA is a robust statistical method, which means the results are still corre ct even if the assumptions are slightly violated, the significance leve l of each ANOVA test is also shown in the tables. When there were only two factor levels to compare, and the medians or means (or averages) of the TRIR were in the same direc tion as the literature s uggested (which factor level might have better safety performances), a one-tail test was conducted. If there were more than two factor levels included in the test, a two tail test was conducted since the comparison of the response variable in multiple factor levels (more than two) cannot be made, even if the null hypothesis is rejected. It was only when the differences of the TR IR medians were statistically significant at the 0.1 level (the significan ce level or P-value was less th an 0.1) that the results are presented, unless specifically noted otherwise. Statistical significance was assumed when the level of significance was 0.05 or smaller, meaning that there was less than a five percent probability that the finding was due to chance. Since the study was an uncharted

PAGE 78

62 and exploratory study, the results are also pr esented if the level of significance was between 0.05 to 0.1, suggesting a strong tendenc y toward statistical significance. Projects Interviewed In The Study Project size descriptors (a s measured in terms of worker hours and in total contracted project cost) and th e TRIRs of the 59 projects incl uded in the final analysis are shown in Table 5-1. Of these projects, se ven were in Canada, three projects were overseas with U.S. owners, and 49 projects were located in 18 states of the U.S, including one in the Virgin Islands. The numbers of interviews conducted in different states are shown in Figure 5-1. Although the safety perfor mances and sizes of the projects ranged widely, it is apparent that th e safety performances of most projects were much better than the construction industry average of about 7.8 for the year 2001 (BLS, 2003). Thus, the projects included in this resear ch generally enjoyed much better success in safety than the construction industry as a whole. Prior rese arch has shown that safer performances were noted among employers with larger numbers of workers. Thus, in general, larger projects would be expected to have better safety reco rds than smaller projects. It was in this context that the analysis was intended to id entify those practices of owners that had a particularly strong impact on the resultant TRIR. It may not be surprising to expect better safety performances on projects where CII members were the project owners, however, this research was not restricted to project s involving CII members. This may account for the relatively broad range in TRIRs. The average TRIR of CII members was reported as being about 1.03 in year 2000 (CII, 2001), significantly better th an the industry average. This research investigated those owner practi ces that had a direct impact on influencing the safety performances realized on large pr ojects. The interviews focused on obtaining information about the demographics of the projects, the manner in which contractors

PAGE 79

63 were selected, the types of safety-related prov isions included in the contracts, and the extent of owner involvement in safety management during project execution. Table 5-1. Safety performances and sizes of the projects include d in the research Total worker hours expended Total estimated cost of the project Total Recordable Injury Rate Mean 2,426,210 $ 379,440,000 1.95 Mode 200,000 $ 15,000,000 0 Std. Deviation 5,155,050 $ 861,740,000 1.94 Minimum 100,000 $ 3,500,000 0 Maximum 26,300,000 $ 5,000,000,000 9.25 Sum 143,146,400 $ 21,248,700,000 115.11 25 275,000 $ 15,250,000 0.58 50 627,000 $ 78,350,000 1.48 Percentiles 75 2,284,000 $ 306,250,000 2.73 Note: These data are based on 59 projects, however, only 56 projects provided the project cost information. Figure 5-1. Locations of the proj ects included in the research Although the 59 projects provi ded the information on OS HA recordable injuries (TRIRs), only 46 were able to provide informa tion on all types of inju ries, including losttime injuries, OSHA recordables, and first-ai d injuries. All the injuries occurring on projects were tallied by seve rity category, and the ratio be tween the different types of injuries was determined, as shown in the in jury pyramid in Figure 5-2. The ratio that 17 3 3 Overseas: 3 Canada: 7 Virgin Isl.: 1 2 1 1 2 1 1 1 2 2 4 2 1 2 3

PAGE 80

64 exists between injuries on the basis of sever ity has been discussed for many years. While these ratios may vary, the general trend is relatively consistent. Heinrich (1959) suggested that the ratio between major inju ries, minor injuries and no-injury accidents was 1:29:300. Note that Heinrich used differe nt definitions of accidents than those used in this research. The pyramid in Figure 5-2 could be simplified by reporting the ratio between lost-time, OSHA recordable, and fi rst-aid injuries as being roughly 1:10:300. Figure 5-2. Injury pyramid for the projects On the 59 projects, 42 different facility owners are represented, since several owners provided more than one project for in clusion in the study. Most of the owners (about 70 percent) were CII owner members with large annual construction budgets. Many owners were listed in the ENR (2001) top 425 owners. Although the data set cannot be referred to as represen tative of all the projects in North America, the projects were considered to be representati ve of the large-scale projects. The Most Common Safety Practices Before testing whether different practices of owners could make a difference in the safety performances achieved on their projec ts, the most popular and frequently used safety practices were listed. These practices included: 1 10.4 296.4 Lost-time In j uries OSHA Recordable In j uries First-aid In j uries

PAGE 81

65 Prime contractor reported injury statistic s to the owner (100%), and the types of injuries reported were: OSHA recordable IR (100%); lost-time injury (100%); near misses (88.1%); environmental issues (91.5%). The owner's site safety representative was generally an employee of the owner (88.1%) instead of a consultant (11.9%). This person was generally a member of the project management team (91.5%), and had authority to stop unsafe work (94.9%). The job responsibilities of the safety representative generally included: Monitoring safety management and perf ormance of the contractor on a daily basis (89.8%) Enforcing safety rules (89.8%) Conducting site safety inspect ions and audits (89.8%) Reviewing safety performance on site and submitting reports to the home office (88.1%) Reviewing contractors’ safety reports (88.1%) Coordinating safety efforts on site (81.4%) Owner's site safety representative reviewed the safety performance of the contractor on a regular basis (98.3%). They would check the project lost-workday injury rate (94.9%), check the project reco rdable injury rate (94.9%), and check the project first-aid in jury rate (88.1%). When selecting contractors, safety was ge nerally a consideration of most owners (94.9%). Among the criteria used by owners to evaluate the safety performances of contractors, the overall quality of the sa fety program was the most frequently mentioned subject (88.1%). The owner' s evaluation of contractor safety performance would make a difference between getting the contract or not (88.1%). If a contractor had some safe ty statistics of concern, the contractor could show that they had made major changes in the program and still be considered for contract award (94.9%). Owners had a variety of contractual safety requirements, the most frequently used provisions include the following: Contractor must comply with the local, state and federal safety regulations. (100%) Contractor must report all lost tim e injuries to the owner. (98.3%) Contractor must report all OSHA record able injuries to the owner. (96.6%) Contractor must report all in juries to the owner. (96.6%) Contractor is required to provide spec ified PPE (hard hats, safety glasses, gloves). (96.6%)

PAGE 82

66 Contractor must implement a substa nce abuse program. (96.2% of U.S. projects only, since drug testing is not allowed in Canada) Contractor must conduct weekly safety meetings for the workers. (93.2%) Contractor must comply with safety requirements beyond the OSHA regulations. (88.1%) Contractor must participate in site safety audits. (88.1%) Contractor must implement a permit system when performing hazardous activities (line breaks, lo ckout/tagout, excavations, proximity to power lines, confined space entry, hot work, etc.). (88.1%) Contractor must submit a site -specific safety plan. (84.7%) Owners impose the same safety requir ements on subcontractors and lower tier subcontractors. (91.5%) Most owners require specific items to be included in the contractor's safety program (98.3%), and included the subcontra ctors in the safety program as well (94.9%). The items that were re quired by most owners included: Regular safety meetings (94.9%) Incident reporting and investigations (93.2%) Conduct regular safety inspections (91.5%) Training on the hazards related to the tasks being performed (89.8%) Substance abuse program (89.8%) OSHA specific regulations (88.1%) Specific safety trai ning sessions (86.4%) Pre-project safety planning (86.4%) Other safety practices of owners included: During the design of this project, constr uction safety issues were specifically addressed. (98.3%) Owner required every worker on site to receive orientation training. (96.6%) Project Descriptions And Safety Performance The size of the project, labor arrangement s for the project, type of project, and other characteristics of the project may all be related to the project safety performance. These factors may be influenced by the owner to some extent and therefore were also analyzed.

PAGE 83

67 Shutdown Projects Although new projects did not show a signi ficantly better or worse safety performance when compared to renovation pr ojects, shutdown projects (operating plants that must stop operations to perform upgrad e and modification work) were found to have poorer safety performances. The median TR IR of the eight shutdown projects included in this research was higher than the median TRIR of the other projects (see Table 5-2). Shutdowns are characterized as having tight schedules (typically f our to eight weeks), significant amounts of overtime work, frequen tly working multiple shifts, and generally having a rapid buildup of the workforce. Wh en workers and managerial personnel work extended hours for one or two months, the possi bility of human erro rs increases, and so will the probability of injury causation. Table 5-2. Injury rates of shutdown projects and all other projects Type of projects Counts Mean Std. Dev.Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) All Others 51 1.80 1.76 1.30 Shutdown 8 2.91 2.76 2.20 Total 59 1.95 1.94 1.48 0.06 0.07 Explanation: In this table, for the Mann-Whitney test, the null hypothesis was: the median TRIR of shutdown projects was equal to the median TRIR of all other projects. The alternative hypothesis was: the median TRIR of shutdown projects was larger than the median TRIR of all other projects. The significance level was 0.07, which shows a strong tendency to rej ect the null hypothesis and to support the alternative hypothesis. For the ANOVA test, the results were simi lar to the Mann-Whitney test except it compared the mean (or average) TRIR of shutdown projects with the mean of other projects. The significance level of 0.06 suggested a tendency that the mean TRIR of shutdown projects was larger than the mean TRIR of all other projects. Other tables in this ch apter are presented in a similar manner. Public Or Private Project The comparison of the TRIRs of public pr ojects and private projects, excluding the shutdown projects (which were all private pr ojects), is illustrated in Table 5-3. It was suspected that private projects may have an advantage in achieving better safety performances than public projects. Since ma ny public projects must be awarded through

PAGE 84

68 the competitive bidding proce ss, public projects are freque ntly awarded to contractors without regard to their ability to deliver a safe project. Private owners may take into account factors other than simply awarding the contract to the lowest bidder. Nevertheless, some public agencies, especi ally federal agencies, may require the contractor to comply with their own safety requirements in addition to the OSHA 1926 regulations. However, the involvement of th e owners in the safety management of the public projects was generally viewed as being minimal, when compared to the extent of owner involvement on private projects. Table 5-3. Type of facility owner and safety performance (excludes shutdown projects) Counts Mean Std. Dev.Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) Private 45 1.62 1.63 1.20 Public 6 3.16 2.28 2.67 Total 51 1.80 1.76 1.30 0.02 0.02 Petrochemical Projects And Manufacturing Projects Petrochemical projects, which accounted for nearly half of the projects analyzed (30), reported better safety pe rformances than manufacturing projects (see Table 5-4). Note that the shutdown projects have been ex cluded in Table 5-5 in order to provide an accurate depiction of the diffe rences between the safety pe rformances of petrochemical and manufacturing projects. Petrochemical owners interviewed c onsistently reported having strong upper management commitment to construction safety and having successfully integrated safety into their company culture. They had a clear understanding of the zero injuries philosophy, and jobsite safety responsibilities were defined to strengthen the safety culture. A behavior-b ased safety approach was widely accepted and implemented on petrochemical projects. Thus, many accidents were avoided by

PAGE 85

69 addressing the front end of the accident causati on chain. This will be further discussed in Chapter 7. Safety performances on manufacturing proj ects were consistently not as good as on petrochemical projects. Note that resi dential and commercial projects were not included in this comparison, primarily because on ly a few such projects were in the entire sample. From the limited data, it appeared as if the residential and commercial projects were not as good as the manuf acturing projects in the area of safety. It should be mentioned that owners of many manufacturing pr ojects were aggressive in their efforts to improve project safety performance. Seve ral respondents stated that the enhanced emphasis on safety in the manufacturing sector was a relatively recent initiative. Since it takes time to be successful in making signifi cant changes in the safety culture of a company, it may be only a matter of time before additional improvements in safety performance are realized on ma nufacturing projects. Although their safety performances were not as good as the petroche mical projects, the manufactur ing projects in this study were already much better than the overal l construction industry TRIR average of 7.8 (BLS, 2003). Table 5-4. Type of project facility (inc luding shutdown projects) and injury rates Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) Petrochemical 30 1.23 1.04 1.03 Manufacturing 19 2.99 2.44 2.53 Total 49 1.91 1.91 1.33 <0.01 <0.01 Table 5-5. Type of project facility (exc luding shutdown projects) and injury rates Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) Petrochemical 25 1.11 1.00 0.84 Manufacturing 16 2.66 2.07 2.47 Total 41 1.71 1.67 1.20 <0.01 <0.01

PAGE 86

70 Union Projects Or Open Shop Projects There was a fairly even distribution of the number of open shop projects, union shop projects, and merit shop projects in the stud y. In general, owners may have little to say about employing union shop or open shop contractors, as the local labor conditions often dictate the type of firms that are avai lable to perform the work. Thus, in strong union areas, union shop contract ors would generally be employed. Merit shop projects are those in which the labor arrangements ar e not a consideration in the selection of contractors. It is on the “merit shop” projects that both union shop and open shop contractors may be employed at the same ti me. Although labor condi tions are not readily influenced by facility owners, it is interest ing to note that there was a difference between the safety performances reported on union shop versus open shop projects (see Table 56). Since Canadian projects are all union shop and since interna tional projects are in environments different from the United States the table presents information on only the U. S. projects. Table 5-6 shows that safe ty performances tended to be better on open shop projects than on union shop projects. When the union shop project and merit shop projects were combined as a category, the medi an TRIR of these proj ects is significantly larger than open shop projects. Note that th is same pattern of TRIR values was found to exist when only petrochemical projects (exc lusive of shutdown projects) were examined. Table 5-6. Type of labor and safe ty performance (US projects only) Counts Mean Std. Dev. Median ANOVA Sign. (2tail) Kruskal-Wallis Sign. (2-tail) Open shop 16 1.32 1.23 1.05 Merit shop 19 2.07 1.51 2.00 Union shop 14 2.71 2.64 2.03 Total 49 2.01 1.87 1.62 0.12 0.13

PAGE 87

71 On projects, especially in the southern portion of the U.S., site safety personnel commented that union workers were more likely to refuse to follow their safety instructions than were workers on open shop projects, possibly because they were in a more secure job position than were open shop workers. It is ofte n alleged that employing union contractors may cost more, but that the sk ill levels of the workers are consistently higher. On the other hand, some contend th at open shop workers are more mobile and are more likely to travel considerable dist ances from one project to the next. The open shop workers are reportedly more willing to foll ow management’s instructions to ensure that their jobs are not placed in jeopardy. Th erefore, the owners may simply have to alter their style of management when promoting th eir safety initiatives with union contractors and open shop contractors. O bviously, this can be a delicat e issue as any prejudice or discrimination in the implem entation of safety management policies and in the enforcement of safety regulations could eventually cause more problems. When implementing the project safety program, the owner must ensure that the program is implemented firmly and consistently. Type Of Contract Two aspects of the project contract were i nvestigated in the study: type of contract (how the owner would make payments to the contractor); and cont racting methods (the contractual relationship between the owner and contractor). One reason that the contract type should be considered when addressing project safety is that the contract establishes the basis on which the owner will make payments to contractor. Essentially, the payments can be made on the basis of unit prices (unit price cont racts), a schedule of values (lump-sum contracts), or a reimburse ment of actual incurred costs (cost plus contracts). As safety can be enhanced thr ough the concerted efforts of different parties

PAGE 88

72 (owner, contractor, and desi gner), the manner in which th e contract defines their relationships might read ily impact project safety. For ex ample, if the contractor has a close, long-term relationship with the owne r, the owner and the contractor may be inclined to use a cost reimbur sable contract (job order contra cting or cost plus). Under such contracts, the contractor’s investment in safety will be reimbursed by the owner. With greater support for the contractor’s e fforts in safety, ther e is less contractor reluctance to dedicate funds for safety. If the owner and contractor emphasize safe ty on the project, there is naturally a greater probability that the project will have good safety performance. During the interviews, one safety manager expressed his preference for lump-sum contracts, since he thought that the contractor coul d not commence safely with si te work until the design was finished. He argued that fewer design ch anges during project execution would reduce hazards on the project. However, anothe r safety manager preferred reimbursable contracts, since reasonable safety costs we re reimbursed and the contractor was never placed in a situation of having to decide be tween spending funds to enhance safety or increasing profits by cutting safety expend itures. This study did not identify any significant differences between the TRIRs of projects using lump sum contracts and cost reimbursable contracts. Different approaches will influence the safety efforts of all parties involved, including the owner, designer, contractor a nd subcontractor(s). One method of enhancing safety is to conduct a construc tability review as part of the design process (Hinze and Wiegand, 1992; Hinze and Gambatese, 1996; Jergeas and Van der Put, 2001). This review helps to coordinate th e safety efforts of designers and the work performed on site

PAGE 89

73 (Fischer and Tatum, 1997). Comparisons we re made of the safety performances of design-build projects with pr ojects constructed under othe r contracting arrangements. Design-build firms, including engineering, pr ocurement, and construction (EPC) firms, have a direct incentive to focus on construc tion safety during the design phase as it is their own employees that are impacted by their design efforts. In other arrangements, the design team is often considered to be separate from the c onstruction effort and does not address construction safety in the design. The owner could impose duties on the design firm to address construction worker safety, re gardless of the contract type, but this was not examined in the study. Results (see Table 5-7) show that designbuild (EPC) projects had significantly better safety performances than did projects with other forms of contracting arrangements. Table 5-7. Type of contract used on the project Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) Other 44 2.22 2.08 1.68 Design-build 15 1.15 1.14 0.69 Total 59 1.95 1.94 1.48 0.03 0.03 Size Of The Projects The size of the project may indicate the complexity of conducting site work and coordinating the related safety efforts. Size might be measured in terms of total constructed cost, the number of subcontractors on site, the nu mber of workers on site, or the number of worker hours expended. In this research it was felt that the total number of worker hours on site gave a more accurate po rtrayal of the difficu lty of implementing a safety program. On small projects, with fewer hours of worker exposure (less than 100,000 worker hours), simpler designs or engine ering plans, and a smaller workforce, safety efforts can often be more effective. On the other hand, on larger projects (with

PAGE 90

74 more than one million hours of worker expos ure) safety performances may also be expected to be good since a large prime contract or was selected to execute the project. Large contractors typically ha ve safety programs that are formalized, and they employ more advanced techniques to promote safet y. Despite the complexity involved, safety performances on the large projects were quite go od (see Table 5-8). This held true for all large projects, including shutdown projects, petrochemical projects, and manufacturing projects. Strong safety performances on la rge projects have been reported in other construction safety research. In general, th e very large and quite small contractors have better safety performances, while medium sized companies have poorer safety performances (Hinze, 1997). Table 5-8. Worker hours expended (in thousan ds of hours) and safety performance Hours Worked (1000s) Counts Mean Std. Dev. Median ANOVA Sign. (2-tail) Kruskal-Wallis Sign. (2-tail) 100~200 8 1.72 1.31 1.51 200~1000 28 2.42 2.27 2.03 1000 up 23 1.45 1.57 0.92 Total 59 1.95 1.94 1.48 0.193 0.14 Work Shift And Workdays The number of shifts worked and the number of workdays worked per week are often dictated by the owner’s schedule requir ements. Tight deadlin es often mean that shift work or overtime work will be necessita ted. On the projects involved in this research, it was found that projects with one shift had significantly better safety performances than those with more than one sh ift (refer to Table 5-9) Projects with fourday (primarily those working four-tens) or five-day workweeks had significantly better safety performances than those working more than five days per week (refer to Table 510). From these results, it is reasonable to suspect that fatigue can contribute to

PAGE 91

75 increasing the number of human errors, and th at days off for rest and recovery are necessary to ensure injury-free work. Table 5-9. Number of shifts worked and safety performance Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) 1 shift 37 1.44 1.45 1.17 2 or 3 shifts 22 2.80 2.35 2.20 Total 59 1.95 1.94 1.48 <0.01 <0.01 Table 5-10. Number of days worked per week and safety performance Counts Mean Std. Dev.Median ANOVA Sign. (1-tail) Mann-Whitne y Si g n. (1-tail) 4 or 5 34 1.54 1.50 1.18 6 or 7 25 2.51 2.32 2.00 Total 59 1.95 1.94 1.48 0.03 0.03 Owner’s Selection Of The Contractor Selecting a safe contractor for project execution is an important function for the owner to achieve better safety performance. In this research, it was found that most owners emphasize the importance of selecting sa fe contractors in the pursuit of the zeroinjury objective. Demonstrated safety pe rformance is a major prerequisite for many contractors to be awarded contra cts. While all owners seem to be aware of the need to select safe contractors, they differ in the approaches used to accomplish this objective. Most private owners will not consider awardi ng contracts to contract ors with bad safety performances. Some owners maintain their own database of the safety performance history of all parties with whom they have contracted, including contractors, subcontractors, and vendors. From this, th ey develop and maintain an approved bidder list and only these firms are given the opportu nity to submit bids on their projects. Preferred Contractors List Projects on which contracts were awar ded through competitive bidding or through a negotiated process did not show any statistically significant difference in their safety

PAGE 92

76 performances. Similarly, projects for which the contracts were awar ded to contractors on a preferred contractors list (T RIR median = 1.32) reported safety performances that were not significantly different from those for which the bidding was open to all interested contractors (average TRIR=1.92). Importance Of Safety During Selection Of Contractors Safety appears to be an important consid eration for most owners when contractors are selected. To get an impression of th e importance placed on safety, each owner respondent was asked about the extent to which safety played a role in the evaluation of contractors in the selection process. The fi ndings show that projects had better safety performances when the owners placed a higher priority on safety when evaluating contractors (see Table 5-11), i.e., projects ha d better safety performances when the contractors had already esta blished a proven safety record on past projects. Table 5-11. Emphasis placed on safety in the ove rall review of contractors (rating from 1 to 7 with 7 being the most important) Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) <=5 24 2.33 1.95 2.13 >=6 35 1.69 1.91 1.20 Total 59 1.95 1.94 1.48 0.11 0.05 Criteria Used To Evaluate Safe ty Performances Of Contractors Questions were asked about how owners eval uated the past safety performances of contractors. Possible options for the ques tion were developed based on the research results of Diaz and Cambrera (1997), Garza et al. (1998), Sawach a et al. (1999), and suggestions from the CII PT-190 members. Th e results show that owners used varying measures of safety performances. One such measure used was the experience modification rating (EMR) on the workers’ compensation insurance. This is a measure that has been widely used in the past few decades, but has lost favor with some

PAGE 93

77 companies as a viable measure of safety performance. The EMR is considered to be a lagging indicator in that it repres ents historical data rather than a current indication of a contractor’s safety commitment (Hinze, et al., 1995). The EMR is based on three years of loss history, but the value of the EMR is dramatically influenced by the number of workers employed by the firm and by the hourly wages paid to the employees. Thus, it is difficult to make valid comparisons between fi rms; especially if they differ in size (number of employees) or in the wages they pay. Most owners stipulat e that they will not employ a contractor whose EMR is greater th an 1.0. Some owners recognize there are shortcomings associated with making comparis ons between companies on the basis of the EMR. These owners no longer utilize EMRs because they are not sufficiently accurate for the owner to evaluate the probable safety performance of a firm for a project. Analysis shows that the average safety perf ormances of projects where the owners did not use EMRs were just as good as projects where EMRs were used (see Table 5-12). Table 5-12. Is the EMR used to evaluate safety performance of contractors? (No significant differences) Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 11 1.45 1.16 1.54 Yes 45 2.13 2.10 1.48 Total 56 2.00 1.96 1.51 0.16 0.23 Note: Three owner representatives di d not answer the entire question in the questionnaire. These three projects were not included in the analysis in Tables 5.12 through 5.17. The TRIR is also a lagging or after-the-fact indicator, but this is a measure that is widely utilized in the industry (Garza, et al ., 1998). The TRIR is a measure of how many failures have occurred, as each injury represen ts a failure on the project. The results of this research show that safety performances of projects were signi ficantly better when the owners used the TRIR as one of the measur ements for evaluating contractors (see Table 5-13). Those owners using the TRIR were asked if a threshold value of TRIR was

PAGE 94

78 established, namely a value above which safety performance was deemed to be unacceptable. The safety performances of projects with more stringent TRIR requirements (threshold values no greater than 2) tended to be significa ntly better than on projects using more lenient th reshold values (threshold values greater than 2), and these were better than where no TRIR limits were established (see Table 5-14). The findings indicated that setting a string ent objective resulted in better performance. Conversely, setting a weaker objective result s in weaker performance. Note that the category ‘none’ includes projects that did not use the TRIR as a requirement and also projects that used the TRIR, but that did not establis h a specific thres hold value. Table 5-13. Is the TRIR used to evaluate safety performance of contractors? Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 7 2.88 1.82 2.50 Yes 49 1.91 1.96 1.43 Total 56 2.04 1.95 1.58 0.11 0.04 Table 5-14. Threshold value of TRIR se t for contractor safety performance Counts Mean Std. Dev. Median ANOVA Sign. (2-tail) Kruskal-Wallis Sign. (2-tail) <2 12 1.06 0.92 0.99 >=2 21 1.66 1.23 1.92 None 23 2.89 2.51 2.18 Total 56 2.04 1.95 1.58 0.01 0.03 The most proactive owners turn to dynami c measurements of safety performance, which can better portray the safety perfor mance potential and safety management capabilities of contractors. Generally, owne rs will not focus on a single measure, but will try to assess the overall safety performa nce of contractors based on a number of measures. Viable measures include an assessm ent of the contractor’s safety program and the qualifications of the safety personnel. A thorough and careful investigation into the

PAGE 95

79 underlying safety commitment held by the cont ractor is conducted to ensure that the selected contractor will be able to achi eve an acceptable safety performance. Qualifications of the contra ctor’s safety personnel and qualifications of the project management team were used by some owners fo r the selection of cont ractors. The more proactive owners would review these qualifi cations by conducting pe rsonal interviews with them and also by making site visits to pr ojects where they were assigned at the time. The resultant TRIR was found to be lower on projects where the ow ner’s had a practice of considering the qualifications of the c ontractor’s safety pe rsonnel and also the qualifications of the project management team (see Table 5-15 and Table 5-16). Table 5-15. Are qualifications of safety sta ffs reviewed when evaluating contractors? Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 14 2.60 2.32 2.48 Yes 42 1.79 1.81 1.32 Total 56 2.00 1.96 1.51 0.09 0.06 Table 5-16. Are qualifications of the project t eam reviewed when ev aluating contractors? Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 18 3.04 2.54 2.48 Yes 38 1.50 1.40 1.20 Total 56 2.00 1.96 1.51 <0.01 <0.01 Many owners asked contractors to provide copies of the OSHA log for the past year and any records of OSHA inspections on past projects. The an alysis did not find any significant differences be tween the safety performances of projects when owners used or did not use these measures as crit eria for the evaluation of contractors. Further analysis was conducted on the me rits of using the following proactive criteria to measure safety performance. Qualifications of the safety staff of the contractor, Qualifications of the project manage ment team of the contractor, and

PAGE 96

80 Quality of the overall safety program of the contractor. The impact of using none or some of thes e proactive criteria a nd of using all three was examined. The analysis essentially determ ined the TRIR of those projects when the owners used all three measures in their asse ssment of contractors, when owners used less than three measures, and those that did not evaluate contractors on the basis of safety when making their selection of the contractor. It should be noted that most owners (33) used all three proactive criteria. The re lationship between the number of proactive criteria used and the resultant safety performances is shown in Table 5-17. From this, it is evident that the use of all three proactiv e criteria is associated with better safety performances than when none, one or two measures are used. Table 5-17. Number of proactiv e criteria utilized for eval uating contractors on safety Counts MeanStd. Dev. Median ANOVA Sign. (2-tail) Kruskal-Wallis Sign. (2-tail) Safety is not a factor 5 3.73 3.22 3.25 <=2 18 2.45 2.17 2.26 3 33 1.48 1.39 1.19 Total 56 2.00 1.96 1.51 0.02 0.04 By using more proactive criteria for safe ty evaluations, owners make it clear that safety is important. As shown in Table 517, when safety did not influence the contract award or when fewer proactive criteria were used, safety performances on the projects were not as good. Owner’s Contractual Arrangement The construction contract is the legal document that specifies the responsibilities of different parties involved in the project (Hinze, 2001). Many c onstruction contracts include provisions for safety that specify additional requirements and responsibilities concerning safety. These pr ovisions are often found in th e general conditions or the supplementary provisions of the contract. In this study, the contract s between the owners

PAGE 97

81 and contractors were investigated. The pr imary focus was on the safety requirements established by the owners. Findings show th at contractual arrangements influence the safety performances reali zed at the project level. The construction contract may impose a va riety of requirements on the contractor. Only the provisions clearly focused on safety were examined in this research. Many different requirements were noted to be includ ed in the contracts. While other contract requirements may have a favorable impact on sa fety, the nature of this influence could not be established when all respondents employe d that contract requirement. There were questions related to seventeen different types of contract requirements. The frequency use of each contract provision is show n in parenthesis (see also appendix b). Contractor must comply with the local, st ate and federal safety regulations. (100%) Contractor must comply with safety requirements beyond the OSHA regulations (88.1%) Contractor must place at least one fulltime safety representative on the project (83.1%) Contractor must submit the rsums of key safety personnel for the owner’s approval (71.2%) Contractor must provide specified minimum training for the workers (62.7%) Contractor must report all lost tim e injuries to the owner (98.3%) Contractor must report all OSHA record able injuries to the owner (96.6%) Contractor must report all in juries to the owner (96.6%) Contractor must include personnel from the owner in coordination meetings (67.8%) Contractor must submit subcontractor list to owner for approval (79.7%) Contractor must implement a substance abuse program. (93.2%) Contractor must participate in site safety audits (88.1%)

PAGE 98

82 Contractor must conduct weekly safety meetings for the workers (93.2%) Contractor must submit a site -specific safety plan (84.7%) Contractor must submit a safety policy signed by its CEO (52.5%) Contractor is required to provide specifie d PPE (hard hats, safety glasses, gloves) (96.6%) Contractor must implement a permit system when performing hazardous activities (line breaks, lockout/ta gout, excavations, proximity to power lines, confined space entry, hot work, etc.) (88.1%) Note that some contract provi sions appear in all or near ly all contracts used on the projects. For the purpose of th is research, only those requireme nts related to better safety performances are presented. Two particular pr ovisions were noted to be associated with better safety performances. One was that be tter safety performances were reported on projects where the contractor was required to assign at least one full-time safety representative to the construc tion site (see Table 5-18). Al so, better safety performances were reported on projects wher e the contractor was required to submit the rsums of the key safety personnel (to be assigned to the proj ect) for the owner’s approval (see Table 519). Impacts of other leading indicators used in the contracts are shown in Figure 5-3, although the differences between TRIRs were not statistically significan t. When the total number of safety requirements in the cont ract was calculated, it was found that those projects having contracts that had more re quirements had significantly better safety performances (see Table 5-20). Table 5-18. Contract requires the contractor to place at least one full-time safety representative on site? Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 10 2.96 2.64 1.87 Yes 49 1.75 1.72 1.30 Total 59 1.95 1.94 1.48 0.04 0.08

PAGE 99

83 Table 5-19. Contract requires the contractor to submit the rsums of key safety personnel for the owner’s approval? Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 17 3.01 2.34 2.81 Yes 42 1.52 1.58 1.20 Total 59 1.95 1.94 1.48 <0.01 <0.01 1.94 1.33 1.32 1.96 1.22 2.30 0.00 0.50 1.00 1.50 2.00 2.50 NoYesTRIR Contractor must provide specified minimum safety training to workers Contractor must submit a site-specific safety plan Contractor must submit a safety policy signed by CEO Figure 5-3. Other leading indica tors used in the project contract (not statistically significant) Table 5-20. Relationship between TRIR and total number of contractual safety requirements Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) <=15 38 2.38 2.24 2.13 16 and 17 21 1.17 0.74 1.2 Total 59 1.95 1.94 1.48 <0.01 0.02 Based on the statistical analys is, five requirements of c ontracts were identified as being “leading indicators”, as these can be used in a manner to predict safety performances. Leading indicators are practices that are associated with improved safety performances. These are listed as follows: Contractor must place at least one fulltime safety representative on the project

PAGE 100

84 Contractor must submit the rsums of key safety personnel for the owner’s approval Contractor must provide specified minimum training for the workers Contractor must submit a si te-specific safety plan Contractor must submit a safe ty policy signed by its CEO The data analysis showed that there was a relationship between the number of leading indicator safety require ments and the safety performa nces realized on the various projects (see Table 5-21). Projects on wh ich more leading indicator safety requirements were included in the contracts had better safe ty performances than projects with fewer leading indicator safety requirements. Table 5-21. Number of leading indicator sa fety requirements included in the contract Counts MeanStd. Dev. Median ANOVA Sign. (2-tail) Kruskal-Wallis Sign. (2-tail) 1 and 2 11 2.77 1.89 2.73 3 and 4 34 1.99 2.16 1.64 5 14 1.22 0.99 1.18 Total 59 1.95 1.94 1.48 0.14 0.07 Owner Involvement During Project Execution In addition to promoting project safe pe rformance through the car eful selection of contractors and the inclusion of carefully selected safety provisions in the contract, owners can be active participants in safe ty management during project execution. Several questions were asked a bout specific practices of owne rs that were expected to favorably influence safety performances of projects. These practices included owner participation in safety recognition programs, monitoring of safety performance, funding safety initiatives, accident reporting, acci dent investigations, safety training and orientation programs, and so on. Only those practices that showed a clear influence on safety performances are presented.

PAGE 101

85 Safety Program Of The Contractor Questions were asked about specific safety topics that must be included in the project safety programs of contractors. Fi fteen possible safety program elements were listed in the questionnaire, including: OSHA specific regulations Specific safety training sessions Contractor’s employees must have 10-hr OSHA cards Contractor’s supervisors must have CPR and First-Aid cards Training on the hazards related to the tasks Pre-project safety planning Task specific PPE analysis Conduct regular safety inspections Incident reporting and investigations Emergency plans (medical and hazardous materials) A substance abuse program must be implemented Regular safety meetings Safety responsibility defined for all levels Emergency response team maintained on the project Daily JSA (job safety analysis) conducted on the project site The inclusion of the following safety progr am elements was associated with better project safety performances (see Tables 5-22 through 5-24): Emergency plans (medical and hazardous materials) Daily JSA (job safety analysis) conducted on the project site A substance abuse program must be implemented Projects where these three elements were required in the safety program had significantly better safety performances (T RIR median=1.20, mean=1.75) than those that implemented one or two of them (TRIR median=2.46, mean=2.56) at the significance level of 0.01. Table 5-22. Owner requires the emergency plan (medical and hazardous materials) to be included in the contractor's safety program? Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 8 2.39 1.05 2.59 Yes 49 1.86 2.07 1.22 Total 57 1.93 1.96 1.43 0.24 0.04

PAGE 102

86 Table 5-23. Owner requires the daily JSA (job safety analysis) to be included in the contractor's safety program? Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 9 2.12 0.78 2.18 Yes 48 1.90 2.12 1.21 Total 57 1.93 1.96 1.43 0.38 0.06 Table 5-24. Owner requires a substance abuse program to be included in the contractor's safety program? Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 4 3.86 2.48 3.25 Yes 53 1.79 1.87 1.30 Total 57 1.93 1.96 1.43 0.02 0.02 Among these three elements in the safety program, drug testing is the most widely accepted on projects. Drug testing is the most important method to identify and screen substance abusers. Although dr ug testing has been controversia l in the past, a majority of large projects require the implementation of substance abuse programs. Today drug testing is restricted or pr ohibited only in the states of Minnesota, Montana, Maine, Connecticut, Rhode Island, and Vermont. However, according to a Substance Abuse and Mental Health Services Admi nistration (SAMHSA, 1999) repo rt, the popularity of drug testing in the construction indus try is still lower than the na tional average. For example, in 1997, 55.6% of all full-time construction wo rkers between the ages of 18-49 reported that their workplace had a wr itten policy concerning alcohol or other drug use (do not necessarily address testing) while the national average was 70.3%. Pre-employment screening tests were conducted for 25.8% of the construction workers and 38.6% of all other industries. Various studies on drug testing program s have shown the effects of the drug testing programs on construction safety. Duri ng the past two decades, drug testing in the

PAGE 103

87 workplace has gone from the point of virt ual nonexistence to widespread employer acceptance. SAMHSA (1999) reported that in 1983, less than 1% of the employees were subject to drug testing, and t oday, approximately 49% of the full-time workers are subject to some type of workplace drug testing. C oble’s study (1992) showed that drug testing was very much influenced by company size, i. e., most larger companies (those with over 100 employees) have drug testing programs, while few small companies would have a drug testing program. In a recent study by Gerber and Yacoubian (2001), it was found that companies with drug-testing programs e xperienced a 51% reduction in incident rates and a significant reduction in their EMR with in 2 years of implementation. Also, most company officials believed that drug-testing programs had a positive impact on all organizational indicators, includ ing safety, quality and costs. It was discovered that owners requiring more of the 15 safety program elements listed in the questionnaire had better project safety performa nces. While not statistically significant, the following inclusions in th e safety programs were associated with noticeable differences in the reported safety performances: Specific safety training program; Task specific PPE analysis; Safety responsibility defined for all levels; Emergency response team maintained on the project. Owner Monitors Near Misses On The Project It is almost an industry st andard for project owners to monitor contractor safety performance on the basis of the TRIR. Sin ce OSHA mandates that these records be kept, contractors can readily provide such inform ation. A more proactive approach is for owners to monitor near misses. By doing th is, the future occurrence of such accidents

PAGE 104

88 may be prevented, before an injury actually occurs. Safety performances achieved on projects where the owner repres entatives monitored near misse s were better than on those projects where this was not a practice (see Table 5-25). These results suggest that there is a beneficial impact of owners requiring near miss reporting and near miss investigations. Table 5-25. Does the owner’s representative monitor near misses on the project? Counts Mean Std. Dev.Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 11 2.48 1.90 2.18 Yes 47 1.82 1.96 1.22 Total 58 1.95 1.95 1.46 0.16 0.06 Accident/Incident Investigations Although all owners reported that they part icipated in the inve stigation of site accidents to a certain degree and generally main tained the accident records, the manner of using these accident records made a significant difference in project safety performances. Table 5-26 shows that owners tracking the individual safety performances of each contractor on site had significantly better sa fety performances. Evaluating the safety performance of each contractor can help in selecting safe contractors on future projects and this can also help the owner to iden tify any weaknesses in the current safety programs being implemented by each contractor Additionally, it wa s found that if the owners incorporated the safe ty statistics of the contra ctors into their own safety performance statistics, the projects achieved be tter safety performances (see Table 5-27). By including the safety records of the contractors in their own safety statistics, the owners essentially adopt the philosophy that any injuries on the project are a negative reflection on their own safety performances. Ideologically, the owner actually regards

PAGE 105

89 the contractor’s employees as its own employ ees, and recognizes th e value of protecting and caring for them. Table 5-26. Does the owner maintain injury statistic s by contractor? Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 26 2.50 2.26 2.13 Yes 33 1.52 1.53 1.19 Total 59 1.95 1.94 1.48 0.03 0.03 Table 5-27. Are the contractor’s safety perf ormance statistics incl uded in the owner’s safety performance statistics? Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitne y Si g n. (1-tail) No 22 2.59 2.42 1.88 Yes 37 1.57 1.49 1.19 Total 59 1.95 1.94 1.48 0.02 0.03 Safety Recognition Program Positive reinforcement is one mechanism by which individuals are encouraged to repeat certain types of behavior. One such a pproach in the area of safety is to implement a safety recognition program that recognizes and rewards workers who have exhibited good safety behavior. Owners held different opinions about safety recognition programs. Such a program essentially reflects the owner’s philosophy about safety and their ge neral concern for the well being of the workers. For example, some owners’ representatives stated that safety recognition and safety incentives were the same, and they would set aside certain monetary amounts to support safety bonuses an d to award the workers who met specified performance criteria. Some would also set up incentives for the contractors. They believed that when more funds were set aside, the safer the projects would be. However, other owners’ representatives he ld the opposite view with the belief that safety cannot be purchased through economic in centives. They would pref er spending money on better

PAGE 106

90 training of workers and recognizing workers with better safety performances. They carefully structured the safety recogniti on program so that the workers would be positively encouraged to work safely, instead of trying to cover up the accidents to get the monetary reward. For example, the owner might recognize the qualifying safe workers with a baseball cap, a gas coupon, a lunch c oupon, or some other small token to remind the workers about the priority being placed on safety. The impact of praising workers in front of their peers should never be underest imated. Some proactive owners would award workers based on both their safety performan ces and attendance performances, and some awarded workers based on their crew performances. These have proven to be useful and effective positive reinforcement approaches (Hinze, 1997). Analysis of the data shows that when the owner provided some funds above and beyond the contract amount to promote safety, projects were more likely to achieve better safety performances (see Table 528). If the owner participated in the safety recognition program, the safety performance would also be better (see Table 5-29). These results demonstrated the positive influence owners could have on project safety performances. Table 5-28. Are some funds provided to the contractor, above and beyond the contract amount, to promote project safety? Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 24 2.39 1.86 2.28 Yes 35 1.65 1.95 1.20 Total 59 1.95 1.94 1.48 0.08 0.01 Table 5-29. The owner participates in the safety recognition program Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitne y Si g n. (1-tail) No 12 3.29 2.12 2.92 Yes 44 1.61 1.81 1.15 Total 56 1.97 1.98 1.46 <0.01 <0.01

PAGE 107

91 Safety Education And Training New worker orientation and safety traini ng are considered by many to form the core of any project safety program (Hinze, 1997; Heinrich, 1959). Orientation of workers is essential to provide workers with necessa ry knowledge, so they can work safely. Orientation is the very foundation for the eff ective implementation of many other aspects of safety programs, i.e., training is requi red to successfully implement other safety initiatives. For example, without adequate safety training, pretask safety planning cannot be effectively implemented within a crew At the same time, safety training is an on-going program, which requires continuous and dynamic efforts to be expended by management. Some aspects of training may also be included in such safety initiatives as safety recognition programs, safety observati on programs, job safety analyses (JSA), safety committees, and other programs th at encourage worker participation. Opportunities to provide ongoing training should be rec ognized and understood. The primary objective is to help workers to always be aware of safety and to develop safe working habits. This is very important for the development of the project safety culture. Owners can help make new worker orientation more effective and they can also assist in the continuing safety training efforts in many ways. This is especially true for owners who are quite familiar with the hazards commonly encountered on sites. Owners can stipulate the minimum requirements of safety training, and the site safety representatives of owners can participate in safety orie ntation sessions. Owners can provide funds and personnel for the safety orie ntation of the workers, and they can also assess the results of safety training by requiri ng a test to be administered at the end of each training session.

PAGE 108

92 The results of data analysis support the fact that the ow ner’s involvement in safety training can make a difference in project safety performance. The methods employed to deliver safety training include showing videos, contractor presentations, owner presentations, consultant presentations, and pr oviding workers with reading materials. Of these, the contractor presentation along with owner presentations were identified as being particularly viable training approaches. Statistical analysis showed that these tw o training methods were most effective at impacting safety performance. Table 530 shows that when both the owner and the contractor were involved in the safety training, projects repor ted better safety performances. Also, more hours (more than th ree) of monthly refres her safety training for workers can improve project safety perfor mances (see Table 5-31). When the owner has a means to verify the comprehension of the safety orientation training received by workers, the safety performances are signifi cantly better (see Table 532). Generally, the owner will require a test or exam after th e safety orientation session to verify the comprehension of the training. Table 5-30. Safety training me thods used on the project Counts Mean Std. Dev.Median ANOVA Sign. (2-tail) Kruskal-Wallis Sign. (2-tail) Contractor only 11 3.11 2.74 1.69 Owner only 17 1.89 1.12 2.18 Both 29 1.58 1.91 0.84 Total 57 1.97 1.96 1.48 0.05 0.06 Table 5-31. Hours of monthly refresher sa fety training received by the workers Counts MeanStd. Dev. Median ANOVA Sign. (2-tail) Kruskal-Wallis Sign. (2-tail) None 11 2.77 1.89 2.73 1 to 3 hours 34 1.99 2.16 1.64 4 hours and above 14 1.22 0.99 1.18 Total 59 1.95 1.94 1.48 0.14 0.07

PAGE 109

93 Table 5-32. Is there any means of verifying the comprehension of safety orientation? Counts Mean Std. Dev.Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 11 3.20 2.13 2.53 Yes 48 1.66 1.79 1.20 Total 59 1.95 1.94 1.48 <0.01 <0.01 Responsibilities Of Owner’s Site Representative The owner’s site safety representative is the person who e xpresses the owner’s concerns for safety and helps to coordi nate the contractor’s efforts on safety management. The responsibilities and authority of the owner’s representative will impact the project safety performance. On all pr ojects included in this research, the owner would assign at least one manager as a fu ll-time project safety representative. The responsibilities of the safety representatives were explored on the different projects. One of the responsibilities consisted of active participation in safety meetings. Active participation in safety meetings and/or tool-box meetings resulted in better safety performances (see Table 5-33). Table 5-33. The owner participates in safety meetings and toolbox meetings Counts Mean Std. Dev.Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 10 3.33 2.80 2.26 Yes 46 1.67 1.66 1.19 Total 56 1.97 1.98 1.46 <0.01 0.01 When more responsibilities were assumed by the safety representative, the project tended to achieve much better sa fety performance than when fewer responsibilities were assumed. Based on the statistical analysis, fi ve key responsibilities assumed by the safety representatives were identified, including: Enforcing safety rules Reviewing safety performance on site and submitting reports to the home office Monitoring pre-task analysis programs Participating in safety recognition programs

PAGE 110

94 Participating in safety an d/or tool box meetings The number of key safety activities perfor med by the owner’s safety representative is related to project safety performance (see Table 5-34). Table 5-34. Total number of key activiti es performed by the owner’s safety representatives Counts MeanStd. Dev. Median ANOVA Sign. (2-tail) Kruskal-Wallis Sign. (2-tail) 1 and 2 6 3.30 2.20 2.92 3 and 4 18 2.26 2.50 1.26 5 32 1.55 1.49 1.23 Total 56 1.97 1.98 1.46 0.10 0.06 Another consideration relate s to how the owner’s safety representative monitors project safety performance. Nearly all the owner's representatives monitored project safety performance by monitoring the incide nt rates on the projects, including lost workday injury rate, TRIR, and first aid injury rate. It was noted that monitoring safety inspection records was associated with signifi cantly better safety perf ormances (see Table 5-35). Table 5-35. The owner’s represen tative monitors project sa fety inspection records on a regular basis Counts Mean Std. Dev.Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 11 3.14 2.71 1.69 Yes 47 1.67 1.65 1.20 Total 58 1.95 1.95 1.46 0.01 0.01 Also, when the owner’s safety representati ve monitored the project near miss rate on a regular basis, the project achieved better safety performance (see Table 5-25). The leading indicator measures were identifie d as (1) monitoring near misses and (2) monitoring project inspection reco rds. Statistical analysis shows that projects where the owner did not monitor either of these two lead ing indicators had rela tively weaker safety performances (see Table 5-36).

PAGE 111

95 Table 5-36. The owner’s represen tative monitors the project near miss rate and project inspection records on a regular basis Counts Mean Std. Dev.Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) Neither 6 3.01 2.35 2.26 Both 42 1.65 1.71 1.16 Total 48 1.82 1.83 1.27 0.04 0.03 Setting Zero-Injury Objectives Regarding the owner’s expectations about sa fety performance, th e results show that owners that established specific safety exp ectations reported better safety performances on their projects, especially t hose owners who set zero OSHA r ecordable injuries as their safety objective before project commencem ent (see Table 5-37). One owner commented “One can achieve the level of safety as he demonstrates to expect." Table 5-37. Is zero TRIR set as a safety objective by the owner before project commencement? Counts MeanStd. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) Yes 19 1.38 1.14 1.43 No 40 2.22 2.18 1.64 Total 59 1.95 1.94 1.48 0.06 0.10 Additional Findings In the research, there are some findings that are related to project safety performances, but not directly related to owne r involvement. These are also of interest and would be described. Safety Recognition Program Projects with safety recognition progra ms achieved significantly better safety performances than projects without re cognition programs (see Table 5-38). Table 5-38. Is there a safety re cognition program on the project? Counts Mean Std. Dev.Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 9 3.47 2.31 2.53 Yes 50 1.68 1.75 1.21 Total 59 1.95 1.94 1.48 <0.01 <0.01

PAGE 112

96 As to the extensiveness of the safety recognition programs, only subjective means are available to describe them. As discusse d in the previous sec tion of this chapter, safety recognition programs should be caref ully developed, in order to positively reinforce safe behavior and es tablish the project safety cultu re. For example, an owner might reward only workers without any inju ries and without a ny absences during a designated period of time (say, a quarter), with the belief that sa fety and attendance go hand in hand. Some owners awarded workers on a crew basis, in order to encourage workers to care for their colleagues. Thes e efforts generally le d to better safety performances through better recogn ition of the priority placed on safety and this resulted in higher morale. One way to recognize work ers with particularly good safety records is through safety dinners. Statisti cal results also show that pr ojects with safety dinners had better safety performances (refer to Table 5-39). Table 5-39. Is safety dinn er held on the project? Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 15 2.36 1.72 2.46 Yes 44 1.81 2.00 1.32 Total 59 1.95 1.94 1.48 0.17 0.05 Other owners implemented totally different approaches with the belief that safety performance should be proportionate to the rewa rds. However, this has resulted in new concerns that accidents might be cove red up. Hinze (1997) had examined the effectiveness of safety incentive progra ms and drew similar conclusions. Safety Observation Program Although different owners and contractors may have different definitions for a safety observation program, it is widely recogni zed as one of the most effective practices that can help promote safety performance. The focus of the safety observation program

PAGE 113

97 is on reducing unsafe behaviors and hazardous working conditions, instead of focusing on incidents only. The objective of the obser vation program is to prevent the occurrence of accidents by eliminating the front-end acts or conditions that might cause unexpected events. Although some owners may think that safety tours conducted by the safety supervisors or safety managers are essentia lly the same as safety observation programs, they are more accurately called safety inspec tions. A preferred definition of a safety observation program is a formal program in which some or all workers are trained and encouraged to recognize and report unsafe behavior or conditions. Management then tracks, records and analyzes th ese unsafe behaviors and condi tions on a regular basis, and implements interventions to eliminate them. Some programs have designated foremen as full-time safety observers around the site to make continuously observation. Other programs have designated and trained work ers making observations one to two hours each week. An observation program can encourag e the participation of all or a portion of the workers on site. Whether or not a sa fety observation program was conducted on the projects tended to make a difference in th e project safety performances (see Table 5-40). Table 5-40. Are safety observers used on the project? Counts Mean Std. Dev.Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 9 3.16 3.16 2.07 Yes 50 1.73 1.58 1.38 Total 59 1.95 1.94 1.48 0.02 0.10 Another important approach is the monito ring and investigation of near misses on projects. Generally, near misses were defi ned as “any unplanned event” or “any event that could have potentially caused personal in jury or property damage or environmental issues.” Although different owners and contra ctors may have varying definitions for near misses, the degree to which near misses are tr acked and investigated are quite different.

PAGE 114

98 One question asked if the contr actor reported near misses to th e owner. The results show that a strong difference exists in the sa fety performances of projects depending on whether or not near misses were reported by the contractors (see Table 5-41). When the number of near misses recorded on each proj ect was compared with the number of OSHA recordable injuries recorded on the same project, it was found that projects reporting more near misses than OSHA recordable in juries had significantly better safety performances (see Table 5-42). Table 5-41. Does the contractor re port near misses to the owner? Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) No 19 2.57 2.01 2.18 Yes 40 1.66 1.85 1.20 Total 59 1.95 1.94 1.48 0.05 0.07 Table 5-42. Comparison between number of OS HA recordable injuries recorded on the project with number of near misses recorded Counts Mean Std. Dev. Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) More recordables 26 2.75 2.20 2.43 More near misses 33 1.32 1.44 0.84 Total 59 1.95 1.94 1.48 <0.01 <0.01 Collaborative Efforts On Safety With Contractors And Designers Collaborative efforts on safety between th e owner, contractor and designer would appear to be essential to the achievement of better safety performance. As to the owner’s cooperation with contractors on safety issues many practices, such as the selection of safe contractors, the inclusion of safety requ irements in the contract and participation of the owner during construction, have been di scussed in the previous sections. The contractor’s commitment to safety an d communications between the owner and contractor can also impact pr oject safety performance. Th ese were evaluated through the subjective assessment of the owner’s safety representative.

PAGE 115

99 Statistical analysis shows that when the owner’s rating of the prime contractor’s commitment to safety was larger than or equal to 6 (out of 7, as the highest rating), safety performances of the projects were signifi cantly better (see Table 5-43). Table 5-44 shows that when the owner’s rating of the communication and cooperation efforts between the owner and contractor was 7 out of 7, the project had significantly better safety performance. Although the ratings were subjective, the findings supported the expectation that a stronger commitment to safety by the contractor, and better cooperation and communication between the ow ner and contractor will help promote project safety performance. Table 5-43. How does the owner rate the prim e contractor’s commitment to safety? (rating from 1 to 7, with 7 being the best) Counts Mean Std. Dev.Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) <=5 23 2.47 2.08 2.00 >=6 36 1.62 1.79 1.18 Total 59 1.95 1.94 1.48 0.05 0.03 Table 5-44. How would the owner rate the cooperation and communication between the owner and contractor? (rating from 1 to 7, with 7 being the best) Counts Mean Std. Dev.Median ANOVA Sign. (1-tail) Mann-Whitne y Si g n. (1-tail) <=6 42 2.30 2.11 1.81 7 17 1.09 1.03 0.69 Total 59 1.95 1.94 1.48 0.01 0.01 Although design for safety was beyond the fo cus of this research, a few open-ended questions were asked to explore how owners c ould impact safety in the design stage. It was found that constructability reviews, accessibility review s, operability reviews and maintainability reviews were addressed duri ng the design or engineering phase. These reviews are conducted by in-house teams that typically include personnel with varied types of expertise. These revi ews could also be performed by outside consultants. Most owners admitted that their efforts to empha size safety in design or in the engineering

PAGE 116

100 stage of their projects had occurred in th e past decade. The underlying reasons for addressing safety in design include the following: Owners have realized that the economic costs of construction worker accidents, which have risen drastically in recent years, are a si gnificant financial burden. Also, each injury accident negatively influences the owner's reputation and business profitability. Despite the contractual provisions, owne rs are not protected by indemnification provisions. Owners often share the liability for construction worker injuries because of defects or hazards resulting from the design or engineering of the project that are considered the responsibility of the owner. Many owners, especially i ndustrial project owners, ar e taking a more proactive approach towards safety. They emphasize safety as early as possible, and try to eliminate unsafe physical conditions (thr ough design for safety, and jobsite safety analysis) and unsafe human behavior (through safety training and safety observation programs). Owners have realized that many safety h azards not eliminated in the design phase will eventually pose a threat to thei r own employees in the operation and maintenance of the facilities. Many owners, especially indus trial owners, are experienced in identifying hazards in the design of their facilities and th ey feel obligated and responsible for participating in the design for safety efforts of their projects. As a result of these factors, owners are now taking a more active role in designing for safety. Owners of petrochemical projects ap pear to be most proactive at this time. The techniques they use to address safety dur ing the design or engineering phases of their facilities include the following: Clearly define their expectations for sa fety, health and the environment on their projects, and clearly communicate this to all parties. Establish a standard procedure for regular safety reviews of the project design by the personnel of the owner, contractor/CM, and designer. Some owners will combine safety, health and environmen tal reviews with constructability, operability, and maintainability review s in their total loss control program. In project design safety reviews, the owner will use a matrix to check different items included in the design and their related hazards. This includes a focus on accessibility and interference of different activities, fall protection, electrical

PAGE 117

101 equipment safety, confined space entry, and hazardous materials. Many owners implement portions of the more stringen t OSHA 1910 safety regulations (general industry) instead of OSHA 1926 safety regul ations (construction industry) in their safety requirements. Focus of safety reviews in the design diffe rs from project to project, depending on the characteristics of the facility to be constructed. When necessary, the design and schedule are reviewed and modified concurrently to avoid accidents. Computer graphical technologies are often applied in the safety reviews of project designs. Nearly half of the project ow ners reported reviewing the design through a 3-D model simulated by computers. Design modifications made solely becau se of safety reasons are common on projects. Although it might increase costs, owners regard the goal of zero injuries as outweighing the economic costs. Answers To Open-Ended Questions In the study, some open-ended questions we re asked. One question asked about the most important way for the owner to improve project safety performance. This question was not formally included in the questionnai re, but was asked as a concluding question. Although the answers varied considerab ly, the following points were made: Owner’s management commitment: Both site management and home office personnel of the owner should have a clea r understanding of the value of safety. Safety is no longer regarded as a priorit y, instead, it should be integrated into the owner’s values, always being placed fi rst. Management should have a common view that zero-injuries can and should be the safety objective. Safety observation program: The philosophy is to eliminate the front end of the accident chain and remind everyone on the pr oject about safety. Techniques can include training and encouraging everyone to report unsafe acts, hazards on site, and near misses. This is followed by tr acking the records and intervening when necessary to avoid unsafe acts. Personal accountability: Safety responsi bilities of each site person, whether employed by the owner or the contractor, should be clearly defined and closely related to the overall persona l performance evaluations. Integration of the safety plan with the sc hedule and cost plan: Safety will be built into the project schedule and cost plan ( by being directly included in the budget). Safety objectives, although set as the prior ity, are fulfilled along with time, quality and cost objectives.

PAGE 118

102 Safety communication: Owners should set th eir expectations on safety from the very beginning and strengthen their emphasi s on safety through continual efforts. Safety communication conveys safety expe rience and knowledge to the contractor. This can be done by discussing safety issu es at the beginning of each meeting and by providing firm and physical suppor t of the safety efforts on sites. Implementation of the safety program: The safety program should be carefully developed, evaluated and modified in respons e to the changes on the project. Once defined, the safety program should be im plemented firmly and consistently, and there should be no differences in the implementation for the employees of the owner, contractor or subcontractors. Evaluation of safety performance: Evaluati ons of workers should be based on their physical performance instead of incident st atistics only. Characteristics of the project being constructed a nd site safety observation r ecords should be considered when evaluating the safety performance of contractors. Overall quality of the safety program, as well as the project mana gement team, should also be considered. Physical walk-out and inspection: Owner personnel should not focus solely on the safety statistics reported by the contractors. They s hould “walk the talk”, and be visible on the site to monitor the contract or’s safety performance and show their support of the contractor’s safety efforts. Safety/constructability review of the desi gn: Proactive owners start their safety efforts as early as the design phase, and it is an essential part of the total loss control program of many owners. Safety culture: Caring for human life a nd health, caring for colleagues, and recognizing safe acts are the philosophy commo nly held on the project. The safety culture can be cultivated only throu gh close cooperation of the owner and contractor, based on the valu es they share on safety. Industrial Versus Commercial Construction (A Case Summary) One of the findings of this research show ed that petrochemical projects had better safety records than the other types of proj ects. Is there a logical explanation that industrial projects tend to have better safety records than do commerc ial projects? More specifically, is there any charac teristic of commercial projec ts that prevent them from achieving the level of safety performan ce achieved on industrial projects? The experience of one particular construction company provides a clear answer to these questions and that is a clear “No!” This particular company has two large divisions,

PAGE 119

103 namely an industrial divisi on and a commercial division. From 1992 to 2002, these two divisions worked nearly 100,000,000 hours, with the industrial division being somewhat larger than the comm ercial division. In 1992, the safety performances of these two divisions were quite different. The OSHA recordable injury rate of the industria l division was 3.26 while the injury rate of the commercial division was 9.74. At this point in time, the i ndustrial division had implemented an aggressive safety program while the commercial division was not as fully engaged in safety. Over the next y ears, the commercial divi sion began to adopt a more proactive stance on safety. This effort has continued to where the safety performance of the commercial division (0.60) has now slightly outpaced that of the industrial division (0.68). Wh ile the industrial division show ed a sustained improvement in the injury frequency rate, the commerci al division showed much more dramatic improvements. This is shown in Figure 5-4. 3.26 2.85 1.88 1.11 0.9 0.76 0.87 0.92 0.61 0.68 2.49 2.4 3.23 3.71 9.74 9.87 6.39 6.27 6.77 6.07 5.86 0.60 2 4 6 8 10 12 19921993199419951996199719981999200020012002Recordable Injury Rate Industrial Commercial Figure 5-4. Corporate histor y of safety performance

PAGE 120

104 From Figure 5-4 it is appare nt that the safety perfor mance on commercial projects can be as good as on industrial projects. On ce a strong safety culture is established, the type of construction project makes no differe nce in the safety performance that is realized. Thus, one can state th at the type of project should no t be a factor in the safety performance realized. The zero accident objec tive seems to be a realistic goal for both industrial and commercial projects. Presumab ly, similar safety performances could be achieved on civil projects, and resi dential projects. It is perhap s realistic to then conclude that the zero accident objective should be pursued on every type of project. Owner Safety Model Building In the previous sections of this chapter, the relationship between different safety practices and the safety perfor mances of projects were disc ussed. Those practices that make significant differences in project safety performances were pr esented with a brief explanation of the possible reasons. However, it is still not clear which factor has the most influential cause and eff ect relationship with project sa fety performance, i.e., what practices cause the greatest diffe rences in the safety perfor mances in the presence of other practices. According to Agresti and Fi nley (1997), a relationshi p must satisfy three criteria to be a causal one, these criteria are: Association between the variables; An appropriate time order; The elimination of alternative explanations. At this stage, the first criterion is already satisfied, with the significant relationships shown between the safety performance and differe nt practices discussed in this chapter. The second criterion should be reasonably satisfied, since safety performance (or accidents) occurred after projects commenced and different safety management practices were implemented. Therefore, the major task to conclude a causal relationship between

PAGE 121

105 the safety performance and different aspect s of the projects, especially the owner’s involvement in safety management, is to discuss the relationship between the safety performance and each practice, after controlling for other factors. If statistical tests disclose that the relationship between safety performance and the selected practices is still significant, then conclu sions can be made that a cau sal relationship between the factor and safety performance exists. Otherwis e, the relationship could be superficial. It should be noted that since nearly all pr ojects conducted the common practices, the analysis was conducted with the ig norance of the common practices. In this section, several indi ces are generated to quantif y different aspects of the practices that are significantly related to safety performance, based on the significant relationships presented in the previous sections. With a multi-linear regression model, their separate relations to project safety performance are evaluate d, after controlling for other alternative explanations. Thus, a tenta tive cause effect relati onship between project safety performance and different practic es employed by owners are presented. Quantification Of The Index The response or dependent variable is the TRIR of each pr oject, which is an effective measure of project safety perfor mance. The explanatory variables are the indices generated based on th e significant findings in this chapter, including: Project index: noted as P, which quantifie s different significant factors that are helpful for the explanation of the injury ra te of different projects. For the purpose of this study, these factors should only be those which are infl uenced by the owner; Selection of contractor inde x: noted as S, which quan tifies different significant factors in the selection of contractors that are associated with differences in safety performances; Index of contractual safety requirements : noted as C, which quantifies different significant factors in the cont ractual safety requirements that help promote project safety performance;

PAGE 122

106 Owner’s involvement index: noted as O, which quantifies different significant factors in the owner’s in volvement in project safety management that are significantly related to proj ect safety performance. All the significant factors or characteristics that are rela ted to safety performance and can be influenced by the owners, can be included in one of the four groups above. The quantification of the variables in each index is based on the “all or nothing” principle. If the project implements a practice or possesse s a characteristic which is significantly associated with better project sa fety performance, the factor value of this characteristic will be 1, otherwis e, it will be 0. The values of all factors in each group are added up to form the index value of this group. For the project index (P) with values poten tially ranging from 0 to 2, the significant factors and their quantifica tion include the following: Does the project work one shift? (q09 : yes=1, no=0, q09 means the ninth question in the questionnaire); Does the project work five days a week or less? (q09: yes=1, no=0). For the contractor selection index (S) with values potentially ranging from 0 to 4, the significant components and their quantification include the following: Is the TRIR requirement for the contr actors less than 2? (q27: yes=1, no=0); Are the qualifications of the project team reviewed? (q27: yes=1, no=0); Are the qualifications of the safety staff reviewed? (q27: yes=1, no=0); Can the evaluation of the contractor’s safe ty performance make a difference in the award of the contract ? (q27a: yes=1, no=0). For the contract index (C) with valu es potentially ranging from 0 to 4, the significant components and their quan tification include the following: Does the project use a design-build or EPC type contract? (q03: yes=1, no=0);

PAGE 123

107 Does the contract require the contractor to place at least one full-time safety representative on site ? (q29: yes=1, no=0); Does the contract require the contractor to submit all safety personnel rsums for the owner’s approval? (q29: yes=1, no=0); Of the listed contractual safety requirements, is the total number of contractual safety requirements 16 or greater? (q29: yes=1, no=0). For the owner’s involvement index (O) w ith values potentially ranging from 0 to 10, the significant components and their quantification includ e the following: Number of critical indicators require d by the owner to be included in the contractor's safety program (q33, if the number is equal to 3 the value is 1, otherwise it is 0). The critical indicators are: Emergency plan (medical and hazardous materials) Daily JSA (job safety analysis) conducted on the project site Substance abuse program Does the owner's safety representative revi ew the near miss rate on a regular basis? (q25: yes=1; no=0); Does the owner's safety representative re view safety inspection records of the contractor on a regularly basis? (q25: yes=1: no=0); Are injury statistics on the owner’s proj ects maintained separately for each contractor? (q37: yes=1, no=0); Are the contractor’s safety performance statistics included in the owner’s safety performance statistics? (q41: yes=1, no=0); Does the owner's safety representative participate in the site safety recognition program? (q24: yes=1; no=0) Does the owner make a presentation in th e worker's safety orientation session? (q44b: yes=1, no=0) Does the owner have any means to verify the comprehension of safety training? (q44f: yes=1, no=0); Number of key responsibili ties assumed by the owner's safety representative (q24, if the number is 4 or 5, the value is 1, otherwise th e value is 0): Enforcing safety rules Reviewing site safety performance and submitting reports to the home office Monitoring pre-task analysis programs

PAGE 124

108Participating in safety recognition programs Participating in safety an d/or tool box meetings Does the owner set zero injuries as its sa fety expectation before the commencement of site work? (q16: yes=1, no=0). As noted, the total values of the differe nt indices are differe nt. If the owner’s involvement in safety management is intensiv e, the value of owner’s index (O) can be as high as 10. For the selection index and the c ontract index, the highest value is 4, and for the project index, the highest value is 2. The descriptive statistics of the different indices are listed in Table 5-45. Table 5-45. Distributi on of the indices Project index Selection index Contract index Owner index N Valid 59 59 59 56 Missing 0 0 0 3 Mean 1.20 2.49 2.15 6.93 Median 1 3 2 7.5 Mode 2 3 3 8 Std. Deviation 0.87 1.19 1.11 1.86 Minimum 0 0 0 2 Maximum 2 4 4 10 25 0 2 1 6 50 1 3 2 7.5 Percentiles 75 2 3 3 8 In order to investigate th e relationship between each index and the TRIR, an analysis was conducted for each category inde x. The results are shown in Table 5-46 through Table 5-49. Table 5-46. Relationship between TR IR and the project index (P) P N Mean Std. Deviation Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) 0 and 1 30 2.50 2.35 1.85 2 29 1.39 1.19 1.17 0.03 Total 59 1.95 1.94 1.48 0.01 Table 5-47. Relationship between TR IR and the selection index (S) S N Mean Std. Deviation Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) <=2 25 2.52 2.40 2.07 3 and 4 34 1.54 1.40 1.20 0.05 Total 59 1.95 1.94 1.48 0.03

PAGE 125

109 Table 5-48. Relationship between TR IR and the contract index (C) C N Mean Std. Deviation Median ANOVA Si g n. (1-tail) Mann-Whitney Sign. (1-tail) <=2 31 2.72 2.30 2.50 3 and 4 28 1.09 0.83 1.15 <0.01 Total 59 1.95 1.94 1.48 <0.01 Table 5-49. Relationship between TR IR and the owner’s index (O) O N Mean Std. Deviation Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) <=6 25 3.03 2.28 2.50 >=7 and <=10 34 1.15 1.12 0.70 <0.01 Total 59 1.95 1.94 1.48 <0.01 The means and medians of project safety performances of the different indices show that the safety performances of the projects were significantly different when the project index, the selection i ndex, the contract index and th e owner index were different, when ignoring the effects of the other i ndices (see Tables from 5.46 to 5.49). The groupings in the tables were established so that the selected values roughly split the samples into two categories w ith relatively equal counts. If all the indices are included in the ANO VA model, the mean TRIRs of projects with different index values w ith the presence of other i ndices can be compared. For example, the mean TRIRs of projects with ow ner index (O) <= 6 can be compared with the mean TRIRs of projects with owner inde x (O) >= 7, after contro lling for other indices (P, S, and C). Table 5-50 shows the ANOVA model with main effects only. From the table, it is evident that the main effects of the project index (P), contract index (C), and owner index (O) were significant. That is, th e mean TRIRs of project s with project index (P)<=1 are statistically significantly worse (h igher) when compared with the mean TRIRs of projects with a project inde x (P) = 2 at the 0.05 level, after controlling for the other indices (S, C and O). The mean TRIRs of projects with contr act index (C) <= 2 was

PAGE 126

110 significantly worse when compared with the me an TRIRs of projects with contract index (C) >= 3 at 0.01 level, after controlling for ot her indices (P, S, and O). The mean TRIRs of projects with owner index (O) <= 6 was significantly worse when compared with the mean TRIRs of projects with ow ner index (O) >= 6 at the 0.0 1 level, after controlling for other indices (P, S, and C). However, after controlling for P, C, and O, the TRIR with different selection index values (S) were not significantly different. This may suggest that selecting safe contractors is widely accep ted by owners to guarantee project safety. Owners that focus on other aspects of the proj ects (P, C and O) are also careful in the selection of safe contractors. If the owner does not care a bout the other aspects (P, C and O), selection of safe contractor may also not be a major concern. Table 5-50. ANOVA table for between-subject effects Source Type III Sum of Squares Df Mean Square F Sig. Corrected Model 78.979 4 19.745 7.711 <0.01 Intercept 224.409 1 224.409 87.636 <0.01 P 10.262 1 10.262 4.007 0.05 S 0.224 1 0.224 0.088 0.77 C 17.471 1 17.471 6.823 0.01 O 18.03 1 18.03 7.041 0.01 Error 138.277 54 2.561 Total 441.82 59 Corrected Total 217.256 58 Dependent Variable: Total Recordable Injury Rate Multi-Linear Regression Be tween TRIR And The Indices Spearman’s correlation (a non-parametric correlation) between TRIR and the different indices and the sum of the indices are listed in Table 5-51. It indicates that TRIR has a negative correlati on with the four indices a nd the sum of the indices. Meanwhile, the contract index (C), the selection index (S), and the owner index (O) are significantly positively correlated with each other. The negative relationship between TRIR and different indices is reasonable, sin ce each index consists of different factors

PAGE 127

111 that are strongly associated with project sa fety performance. The positive correlation between the different indices (C and S, C a nd O, S and O) indicates that owners who carefully select safe contractors also emphasi ze safety in contracts, and employ different practices and techniques to promote safety performance. It also suggests that safety cannot be achieved through a single effort. The relationships between the TRIR and the indices can be linear. A multi-linear re gression model might disclose the more significant causal relationships between the sa fety performances and different aspects of the projects. The model may also help estimate the safety performance that can be realized on a project with the invo lvement of the owner in safety. Table 5-51. Spearman’s Correlation between TRIR and the index TRIR P S C O SUM TRIR 1 -0.328(*) -0.318(*) -0.452(**) -0.484(**) -0.636(**) (Sign.) 0.011 0.014 <0.01 <0.01 <0.01 P -.328(*) 1 0.228 -0.022 0.116 0.380(**) (Sign.) 0.011 0.082 0.866 0.395 0.003 S -.318(*) 0.228 1 0.434(**) 0.277(*) 0.579(**) (Sign.) 0.014 0.082 0.001 0.039 <0.01 C -0.452(**) -0.022 0.434(**) 1 0.362(**) 0.664(**) (Sign.) <0.01 0.866 0.001 0.006 <0.01 O -0.484(**) 0.116 0.277(*) 0.362(**) 1 0.752(**) (Sign.) <0.01 0.395 0.039 0.006 <0.01 SUM -0.636(**) 0.380(**) 0.579(**) 0.664(**) 0.752(**) 1 (Sign.) <0.01 0.003 <0.01 <0.01 <0.01 Note: means significant at 0.05 level and ** means signific ant at 0.01 level. Multi-variable regression was conducted, w ith TRIR as the response variable and the indices as the explanatory variables. A stepwise auto-selecti on of the explanatory variables procedure was applied, in order to firs t select the most significant variable. The final results of the regres sion are shown in Table 5-52, Table 5-53 and Table 5-54.

PAGE 128

112 Table 5-52. Model Summary of the multi-variable regression Model Predictors R R Square Std. Error of the Estimate 1 O 0.481 0.231 1.7561 2 O, P 0.580 0.336 1.6471 3 O, P, C 0.652 0.426 1.5466 The response variable was TRIR in all the models. Table 5-53. ANOVA of the multi-variable regression Model Sum of Squares df Mean Square F Sig. 1 Regression 50.05 1 50.05 16.229 <0.01 Predictors: Residual 166.534 54 3.084 O Total 216.584 55 2 Regression 72.806 2 36.403 13.419 <0.01 Predictors: Residual 143.779 53 2.713 O, P Total 216.584 55 3 Regression 92.203 3 30.734 12.849 <0.01 Predictors: Residual 124.381 52 2.392 O, P, C Total 216.584 55 Dependent Variable: Total Recordable Injury Rate Table 5-54. Coefficients of the multi-variable regression Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta (Constant) 6.606 0.85 7.773 <0.01 Owner index -0.36 0.123 -0.337 -2.939 <0.01 Project index -0.752 0.237 -0.334 -3.178 <0.01 Contract index -0.576 0.202 -0.326 -2.848 <0.01 In order to validate the linear regression analysis, several assumptions should be satisfied, including: the data are randomly se lected, the response variable (TRIR) should be roughly normally distributed, the data s hould fit the linear model well, and the response variable should have constant variance at different levels of the explanatory variables. The selection of th e data was described previously in this chapter. Although the data were not purely randomly selected, th e sample projects were independent of each other. The normality assumption was tested by the histogram of regression standardized residuals (Figure 5-5) and the P-P plot (Figure 5-6). Figure 5-5 shows the histogram of the standardized residual of the TRIR, wh ich represents the distribution of the

PAGE 129

113 standardized value of the observed TRIR less the predicted TRIR (predicted by the calculated linear model). Figure 5-6 shows the level of conformity of the expected cumulative probability of the standardized resi dual of the TRIR when compared to its expected distribution (45 de gree line). Both figures i ndicate that the normality assumption is roughly satisfied: the histogram roughly fits the norma l distribution curve, and the P-P plot roughly fits the 45 degree line. The linearity and constant variance assumption was tested by the predicted value vs standardized residual plot (Figure 5-7, if the equal variation as sumption is satisfied, the scatters s hould be randomly scattered around the y=0 line), which shows that the assump tion is roughly satisf ied. Therefore, it can be concluded that the results of linear regression are valid, and the findings are reliable. Figure 5-5. Histogram of the re gression standardized residual

PAGE 130

114 Figure 5-6. P-P plot of the regression Figure 5-7. The predicted value vs standardized residual plot

PAGE 131

115 Explanation And Discussion Of The Regression Results The results shown in Tables 5-52 through 5-54 demonstrate how the indices were selected in the multi-variable linear regression model. Tables 5-52 and 5-53 show that the first index selected was the owner index (O), which means that the owner index was the independent variable that can explain the largest proportion of the variability of the TRIR. The next independent variable selected was P, followed by C. With the inclusion of more explanatory variable s in the model, the sum of squares explained by the model and R square increased, and the error sum of squares decreased (see Tables 5.52 and 5.53). In the final model selected in the mu lti-variable linear regression, the R square was 0.426, which means that 42.6% of the variat ion of TRIR could be explained by the three indices: project index (P ), contract index (C) and ow ner index (O) in the linear model. The F value and signifi cance level of the final model in Table 5-53 shows that the null hypothesis (that each of the co efficients of the indices is equal to zero) is rejected. The results in Table 5-54 indicate that all th e three coefficients are not equal to 0 at the significance level of 0.01. Th e relationship between TRIR and the indices can be expressed as Equation (1). TRIR = 6.606 -0.752P 0.576C 0.36 O (Equa. 1) In which, P means the project index ranging from 0 to 2, C means the contract index ranging from 0 to 4, O means the owner index ranging from 0 to 10. If P = C = O = 0 (i.e., the owner does not take any practices in cluded in P, C, and O), then the project TRIR might be as high as 6.61. If P = 2, C = 3, and O = 9 (i.e., the

PAGE 132

116 owner takes most of the practices included in P, C, and O), then the estimated TRIR can be as low as 0.13. The selection index is not included in the equation, because of the multicollinearity effects that o ccur when the selection inde x and the other indices are significantly correlated. That is, the power of the selec tion index in explaining the variability of TRIR was trivial in the presen ce of the other explan atory variables (P, O, C). This does not mean that the selection of th e contractor is not rela ted to project safety performance, as the results in Table 5-51 illu strate that they are significantly related, although not as high as the othe r indices. Meanwhile, it also suggests that unlike twenty years ago, the selection of safe contractors is no longer the only and most effective way to improve project safety performance. Most owners consider safety performance when selecting contractors to execu te their projects. More im portantly, owners positively influence project safety performances by taking an active role in project safety management and by including stringent safe ty requirements in their construction contracts. Referring to the standardized coefficients in Table 5-54, the owner’s involvement index (O) is more influential than the other coefficients in reducing TRIR. This means that the owner’s actual practices can be at l east as important as the safety requirements in the contract and the project context. As a ll three indices in the regression model can be directly influenced by the owne r, it can be concluded that owners do play a very strong role in influencing the safety performance realized on projects. The R2 of the regression (0.426) is not quite cl ose to 1 (which w ould mean that all injuries are impacted by the owner). This finding is reasonable since the contractor-

PAGE 133

117 related factors have not been considered in this study. Although the safer owners tend to select safer contractors, and their safety pe rformances tend to be better, there are also circumstances in which the contractor could ta ke the leadership role in project safety management and achieve better safety perfor mance without very much involvement of the owner. Also, with the active involvement of the designer (A/E firms), safety might be improved as well. Therefore, some of the va riation of the project safety performance is left to be explained by the involvement of the contractors and desi gners. However, as explained in Chapter 1, because of the positi on held by owners on ea ch project, owners are an important party that can influence project safety pe rformance to a large extent. Safety Performance Model: How To Achieve Better Safety Performance To illustrate how owners can (help) achi eve better safety performances on their projects, a safety performance model was deve loped, based on the data analysis in this chapter and the previous chapter. The thre e indices in Equation (1) were included this model. When comparing means and medians, it can be concluded that the project index, contract index, and owner index all make a diff erence in project safety performance. The model includes the indices of the project (P), the contract (C) and the owner (O) as shown in Figure 5-8. The TRIR means and median s shown on the right branch (with larger index values) at different leve ls are always less than the left branch (with smaller index values). At the lowest level, when control ling for the project index (P) and the contract index (C), it is evident that the owner’s dir ect involvement can make quite a significant difference in the project safety performance. The leg on the very right, with high values for all three indices, has the lowest TR IR mean (TRIR=0.86) and median (TRIR=0.70) values, while the very left leg, with low valu es for all three indices has the highest TRIR mean (TRIR=4.60) and median (TRIR=3.41) values.

PAGE 134

118 Therefore, an increase of each index contri butes to reducing the TRIR. To test this conclusion, the Spearman’s correlation between TRIR and the total of the three indices (P, C and O) is -0.636 with a two-tailed signi ficance level of less than 0.001 (see Table 551). This provides strong evidence that th e owner’s involvement, together with the project characteristics and the contractual sa fety requirements will significantly influence project safety performance. Based on this model, a project scorecard was developed to evaluate the safety performance of a project. Any owner can use th e scorecard to evaluate the possibility of its project achieving good safety performance by referring to the scores obtained in the card. Also, the score obtained in the different sections of the scorecard can help the owner identify which areas can be improved to achieve better safety performance. The scorecard is shown in Figure 5-9. Note that th e scorecard is valid only after the following practices are already implemented on the projects: the owner requires the primary contractor reports OSHA recordable injuries; the owner assigns at least one site safety representative with high aut hority and responsibilit ies clearly defined; the owner’s site safety representative monitors contractor ’s safety performance regularly; the owner prioritize safety when selecting contractors, and maintains an approve d contractor list for awarding the contract; the owner specifies ba sic safety requirements in the contract document; the owner requires basic components included in contractor’s safety program; owner address safety in the design phase.

PAGE 135

119 Legend: Figure 5-8. Project safety model: owner’s impact All Projects 59 1.95 1.48 Project Index<=1 30 2.50 1.85 Project Index=2 29 1.39 1.17 Contract Index>=3 15 1.28 1.22 Contract Index<=2 16 1.80 1.88 Contract Index<=2 15 3.71 2.81 Contract Index>=3 13 .88 .72 Own Index<=6 10 4.60 3.41 Own Index>=7 8 1.19 .54 Own index>=7 5 1.94 2.41 Own Index<=6 6 1.57 1.26 Own Index>=7 9 1.09 1.13 Own Index<=6 8 2.41 2.45 Own Index<=6 1 1.17 1.17 Own Index>=7 12 .86 .70 Category Count Mean Median .013/.030 .002/.001 .062/.034 .036/.065 .162/.221 .033/.033 Si g n. Of ANOVA ( 1-tails ) / Si g n. Of Mann-Whitne y test ( 1-tail )

PAGE 136

120Owner’s Influence on Construction Safety Scorecard Project Context, Contractor Selection, Contractual Safety Requirements and Owner Involvement in Project Safety Answer Project context: (1) Does the project work one shift? Y N (2) Does the project work five days a week or less? Y N Selection of contractor (3) Is the TRIR requirement for the contractor selection less than 2.0? Y N (4) Are the qualifications of the project team reviewed? Y N (5) Are the qualifications of the safety staff reviewed? Y N (6) Does the evaluation of each contractor’s safety performance make a difference in awarding the contract? Y N Contractual safety requirements (7) Does the project use a design-build contract? Y N (8) Does the contract require the contract or to place at least one full-time safety representative on the project site? Y N (9) Does the contract require the contractor to submit all safety personnel rsums for the owner’s approval? Y N (10) Does the contract require the contract or to prepare a site-specific safety plan? Y N (11) Does the contract require the contractor to submit a safety policy signed by its CEO Y N (12) Does the contract require the contractor to provide a minimum specified amount of training to the construction workers? Y N Contractor Safety Program Requirements Which of the following are required to be included in the contractor’s safety program? (13) Contractor must prepare a plan for site emergencies Y N (14) Contractor must conduct pre-task safety planning on the project site Y N (15) Contractor must implement a substance abuse testing program Y N Owner’s involvement in pr oject safety management (16) Does the owner's safety repr esentative investigate near misses? Y N (17) Are injury statistics on the projects maintained separately on each contractor? Y N (18) Are all project injuries in cluded in the owner’s overall measure of safety performance? Y N (19) The owner actively participates (gives pres entations) during worker safety orientation? Y N (20) Comprehension of safety training is evaluated through testing? Y N Which of the following activities are performed by the owner’s site safety representative? (21) Enforcing safety rules and regulations Y N (22) Monitoring of the implementation of pre-task planning Y N (23) Participating in safety recognition programs Y N (24) Participating in safety and/or tool box meetings Y N (25) Does the owner set zero injuries as its safety expectation before the commencement of site work? Y N Total Count of Yes Responses Multiply the Yes Count by 4 (X 4) % What the score means: 88% or better is strong owner involvement 52% or less indicates weak owner involvement Figure 5-9. Owner's influence on construction safety scorecard

PAGE 137

121 The highest possible score of the scorecard is 100 percent. The scores of the projects in the study ranged from 40% to 92% If the projects w ith at least 22 points (scoring over 88%) are compared with those w ith less than or equa l to 52% scores, the TRIRs are significantly different (refer to Ta ble 5-55). The Spearman’s correlation of the TRIR and the score generated on the scorecard for the 59 proj ects is – 0.60, significant at the 0.001 level. This shows that the results of the scorecard are negativ ely correlated to TRIR, and the score can be used to evaluate the owner’s impact on safety. Table 5-55. The safety performances of pr ojects with highest and lowest scores Score N Mean Std. Deviation Median ANOVA Sign. (1-tail) Mann-Whitney Sign. (1-tail) <=13 (52%) 8 4.64 2.90 2.92 >=22 (88%) 7 0.53 0.34 0.50 <0.01 Total 15 2.72 2.96 1.62 <0.01 Follow-up Survey: Contractors' Responses As a supplementary effort of this study, a second survey (shown in Appendix c) was conducted to collect opinions from contractors on the owner’s impact on construction safety. Owner involvement in safety management was the focus of the survey. Questions were asked about the owne r's responsibilities in construction safety, which type of project tends to be safer, and preferences of contractors for owner's participation in safety management. Sixty que stionnaires were sent to contractors with large construction projects in North America n. Forty-one responses were received. It was a brief survey, but the results of the su rvey support several findings of the project interview study. Also the contractors' assessment of how safety is emphasized by owners, as well as the contractor’s preferen ces for owners who participate in safety management were discussed. Since this was not the focus of this study, the results of the contractor responses ar e shown in Appendix d.

PAGE 138

122 Summary This research was focused on the owner’s involvement in safety management as demonstrated through the project context, the se lection of safe contr actors, inclusion of safety requirements in the contract, and the owner’s active participa tion in safety during project execution. Through anal ysis of the projec t interview data, it can be concluded that owners can favorably influence projec t safety performances. Better safety performances are related to the following pr oject characteristics and practices of the owners: Project context: Project characteristics rela ted to better safety performance include petrochemical projects, priv ate projects, open-shop projec ts, projects with designbuild contracts, projects of fairly larg e or small size (not medium size), projects working one shift, and projects worki ng five or less workdays per week. Careful selection of safe c ontractors: Proactiv e criteria are used to evaluate and select contractors by owners aggressive in safety. These criter ia include the TRIR on past projects, qualifications of the contra ctor’s safety staff, qualifications of the contractor’s project management team, and the assessed quality of the contractor’s overall safety program. Contractual safety requirements: Contract ual safety requirements should clearly convey the owner’s emphasis on safety a nd the owner’s expectation of a safe project. Five contractual requirements we re identified as bein g leading indicators, and they are listed as follows: Contractor must place at least one fulltime safety representative on the project Contractor must submit the rsums of key safety personnel for the owner’s approval Contractor must provide specified minimum training for the workers Contractor must submit a si te-specific safety plan Contractor must submit a safe ty policy signed by its CEO Owner’s proactive involvement in the safety practices of projects: In this study, the key measurements employed by the owners with better safety performances are identified as: Owners set their expectations on safety fr om the very beginning of a project, especially the zero-injury objective;

PAGE 139

123 Owners impose requirements on the safety programs developed by contractors and emphasize specific items, including: emergency plan (medical and hazardous materials), daily JSA (job safe ty analysis) conducted on the project site, and substance abuse program; Owners monitor near miss rates and the safety inspection records on the projects, besides other types of injury st atistics (TRIR, lost-time injury rate, etc); Owners maintain accident statistics by c ontractor on their projects, and also include the contractor’s injury stat istics in their own accident records; Owners establish a safety recognition program and contribute funds to the program; Owners actively participate in safety tr aining and orientati on and verify the comprehension of the training (such as by testing); Owners assign a full-time safety representative on site with various responsibilities including: en forcing safety rules; revi ewing safety performance on site and submitting reports to the home office; monitoring pre-task analysis programs; participating in safety reco gnition programs; and participating in safety and/or tool box meetings. Based on the significant factors listed, four indices, namely the project index, the selection index, the contract index, and the owner index, can be generated to quantitatively measure the vari ous types of involvement of the project owner in safety management. Statistical analysis of the data demonstrates all indices, except the selection index, have a strong relationship with project safety performa nce after controlling for the other indices. Using the multiple linear regr ession method, the impact of each index was evaluated. A model to describe the impact of the important indices was established. A scorecard was developed to evaluate the owner's overall impact on project safety.

PAGE 140

124 CHAPTER 6 HOW CAN OWNER ACHIEVE PROJECT SAFETY In this chapter, the safety management pr actices of two owners on several projects are examined. The scorecard developed in Chapter 5 is used to evaluate each owner’s involvement in construction safety manageme nt and this score is compared with the safety performances achieved. The reasons why the projects may or may not achieve good safety performances are diagnosed, with recommendations for future improvement. Case 1: Owner R As part of the study, intervie ws were conducted on five pr ojects of the same owner, hereinafter referred to as Owner R. These pr ojects are coded as projects A through E. Arranged by Owner R, the individuals intervie wed on these projects were all in project management positions, and they were all familiar with safety management on the projects. The information gathered can help develop an overall pi cture of how Owner R addressed safety issues on its construction pr ojects. While keeping the similar pattern of safety management on the different projects, sa fety performances of these five projects varied. Analysis of safety management prac tices on these five proj ects can help explain the differences in the safety performances realized. The scores for the projects were evaluated. The incident rates were calculated, together with the indicators to specify the owner’s involvement in safety management and the cooperation between the owner and cont ractor. The data are listed in Table 6-1.

PAGE 141

125 Table 6-1. Comparison of injury rate and scorecard evaluation Project A B C D E Scorecard evaluation score 76% 84% 84% 92%* 92% Near miss injury rate 109.60 8.70 31.38 75.00 16.45 TRIR 3.20 1.92 2.00 0.00 0.18 Lost-time injury rate 0.00 0.00 0.00 0.00 0.00 First-aid injury rate 58.40 34.78 12.31 0.00 6.89 Worker Compensation case injury rate 3.20 0.00 0.00 0.00 0.00 Contractor’s Commitment 6.5 4.0 3.5 5.0 6.0 Importance of safety in selecting contractors 6.5 7.0 7.0 6.5 7.0 TRIR threshold for contractors 4.0 3.5 2.0 1.0 1.0 Rating of safety communication with contractors 4.5 5.0 3.0 6.0 7.0 *:Evaluation score of project D was calculated to estimate the owner’s involvement although it did not satisfy the 100,000 worker hours criterion. Among the five projects, three (Project A, C, and E) were completed or essentially completed, project D has just been started, and pr oject B was at the point of half completion. While the project phase of construction may impact safety performance, this could not be assessed in this limited study. While the five projec ts (average lost-time injury rate of 0) were well below the averag e lost-time incident ra te of both CII members (0.26) and the construction industry (3.55 in 2000), project A, B and C had higher recordable incident rates (TRIR = 3.20, 1.92, and 2.00) than the average of CII owners (0.92) and CII overall (1.03). With limite d worker hours expended, the overall safety performances of project D and project B were still not clear, although the trend of safety management on both projects was promising. Project E achieved very good safety performance through sustained efforts of both the owner and contractor. The evaluation scores of the scorecard follow the same order as the safety performance, with project E being the best and project A be ing the worst (see Table 6-2). Table 6-2. Comparison between TRIR and the scores of the scorecard Project TRIR Score of scorecard (%) A 3.20 76 B 1.92 84 C 2.00 84 D 0.00 92 E 0.18 92

PAGE 142

126 By reviewing the information in Table 6-1, project E had the most stringent threshold for TRIR in the selection of the contractor (less than 1.0) and safety was regarded as the most important factor in the overall review of the contractor’s performance (7 out of 7). With a good safety commitment of the contractor (6 out of 7) and good communication between the owner and th e contractor (7 out of 7), the safety performance realized was a testimony of the su ccess of the owner’s involvement in safety management (with the evaluation score of 23 out of 25, or 92%). In the interview, it was found that project E employed other techni ques to improve safety performance, including: Strict selection of the contractor: Be sides a high threshold for past safety performance, the owner reviewed all the in cident records of the contractor. When evaluating the contractor’s safety pers onnel, the owner not only requested the rsums of the key safety personnel of the contractor, but also inspected the current projects for which they were responsible. During the construction of the project, one safety person of the contractor was la id off or dismissed by the owner because of incompetence in safety management, wh ich sent a strong message of the owner’s concern for safety. Management commitment: The constructi on manager took the responsibility for safety on the project. About 25% of th e construction manager’s responsibilities were related to safety management. Safety professionals and safety representatives were regarded as good resour ces for construction safety. Safety observation program: The project kept close track of unsafe activities. Both unsafe acts observed by individual work ers and near misses were carefully investigated by the project management team and reported to the construction manager. If any trend of unsafe acts was identified, the safety plan was modified, and notice would be made to everyone on site. Effective safety recognition program: The safety reward program recognized an individual’s safety performa nce by contributing 50 cents per hour per worker for a safety bonus for workers who worked for a month without any recordable incident. Also, safety lunches and random giveaway s were used to recognize good safety performances of workers. Focus on communication between the owner an d contractors: Safety bulletin that was "refreshed" every week provided nece ssary safety knowledge to workers based on the results of safety observations. Th e project manager specified the weekly

PAGE 143

127 safety meeting topics. Safety was always the first issue discussed at the beginning of each meeting. Project A and project C had weaker safety statistics than proj ect E. Based on the scorecard, some contributing reasons were: Loose safety thresholds during selection of the contractors: The TRIR thresholds of the two projects were 4 and 2, respectively. Reviewing the current job site of the contractor and safety perf ormance of the key safety persons was not carefully addressed, which might send a message to th e contractors of the owner’s lack of concern for safety. The contractor on project C was selected after completing another project for the owner without furt her evaluating the safety performance of the contractor and without reemphasizing th e safety expectations of the owner. This might contribute to the low commitme nt of the contractor to safety on the project (3.5 out of 7). Lack of communication betw een the owner and contract or and between different levels of the project: The ratings of communication between the owner and contractor on the two projects were 4.5 and 3 out of 7, respectively. The high number of subcontracts (56) awarded on pr oject C might be a c ontributing factor. Communication with the contractor’s work ers was regarded as the main problem on project C, and it was stated that the union workers were not as cooperative as open shop workers. Ineffective safety observation program: On both projects, the safety observations were essentially conducted by the owner’s staff. Workers were not adequately trained to make the observations and to re port to management. Unsafe trends were not easy to discover. The ratios between the total number of workers at the peak level and the number of owner staff on site were relatively large (more than 65 for project C), i.e., safety supervision may ha ve been inadequate. Near misses were not thoroughly investigated, which limited the capability of management to identify potential hazards on site and prevent accidents. Lack of effective metrics for safety performance on site: Without major efforts focused on near misses and unsafe behaviors, project management used EMR (for project A) and the amount of training, do llars spent on safety issues, management commitment (in project C) as the main me trics to ensure good safety performance on site. As discussed in Chapter 5, the EMR is a lagging indicator of the contractor, instead of being project specific. It was unwise for the safety manager on project A to depend on the EMR in monito ring the project safety performance. For project C, dollars spent on safety may suggest that safety could be purchased, and the philosophy was not good to safety. Job complexity: Project A was a shut down project, with 250,000 worker hours exposure within 2 months, and worked two sh ifts at six ten-hour days per week. Since project C involved considerable electrical equipment and specialty

PAGE 144

128 subcontractors, these factors made safety management more difficult than on the other projects. With reference to the experiences on projec ts A, C, and E, the strong points of projects B and D are listed below: Management commitment: Construction ma nagers and safety managers on the projects understood the importance of safe ty and set their ex pectations of zero injuries for everyone on the pr oject from the very beginning. Emphasis on design for safety: Project B wa s very cautious about rigging activities and implemented a lifting permit system, when considering that most of the elements of the project were prefabricate d. During the safety review of the design plan for project D with virtual reality technology, more than 100 changes were made to eliminate safety hazards. This could help considerable to reduce hazards and improve safety performance. Thorough implementation of safety obs ervation program: Project B had 10 supervisors and one group of workers that did weekly safety observations of the 150 workers on site. Project D had observations made by every worker and each submitted an observation card daily. All the unsafe trends would be tracked and analyzed. Near misses would also be caref ully recorded and investigated to help modify the safety plan. This could help update the safety program regularly and reduce the unsafe behaviors on site. Safety communication: Comm unication on safety information between different levels on the project was enhanced by frequent safety meetings and other dissemination systems. Steering meetings on project D and a formalized injury prevention information system implemente d on project D should effectively help communication on the projects. Significant safety incentive system for i ndividuals and contract ors: owners on both projects made significant contributions to recognizing good safety performance of the contractors and their workers. These rewards reinforce safety expectations of the owner. The weak point of these two projects was the selection of safe contractors. The threshold value for the TRIR for project B was 3.5, and the contractor did not need the owner’s approval when selecting subcontractors. Project D had not reviewed the rsums of the contractor’s key safety persons or their performances on prior projects.

PAGE 145

129 Projects B and D had established a good sa fety culture from the very beginning. Effective techniques were employed to promot e construction safety on site. Good safety performance could be expected on both project s if the weak points were recognized and addressed. In general, the five projects of Owner R had good safety performances. Performing work that consumed 3,857,000 worker hours with 19 recordable incidents, the average TRIR of 0.985 was slightly higher than the CII owne r average 0.92 in year 2000. However, it was significantly less than the industry average of 7.8 in 2001. Success of safety management came from the commit ment to construction safety from top management to the owner staff on each proj ect. The owner’s involvement in safety management on the different projects was inte nsive, although with different degrees. Approaches that Owner R employed to prom ote safety can be summarized as follows: Have high expectation on construction safety (zero-injury objectives); Implement safety recognition programs; Set safety requirements in the construction contract; Assign full-time safety representative on site; Monitor and investigate near mi sses and first aid cases, and Emphasize safety training and orientation programs. It should be mentioned that certain char acteristics of the projects made safety management more difficult. A good example wa s that most workers of contractors were far from their head offices, which generall y causes difficulty in communications between the project and the contractor (Hinze, 1997). In turn, co mmunications and cooperation between the owner and contractors suffe red, resulting in some problems between management and the workers on the projects. While the projects were working well w ith safety management, communications between owner staff on different projects a ppeared inadequate, i. e., there was little

PAGE 146

130 sharing of lessons learned between projects. If management techniques and information concerning safety could be shared regul arly among different projects, between management of both owner staffs and cont ractor staffs, safety performance should improve significantly in the future. Case 2: Owner S Project Description A manufacturing plant, referred to as Own er S, was included in the study. It executed a shutdown project for one furnace, coded as project F, in year 2001. The project accumulated 627,931 field worker hours. There were no lost time injuries on this project. There were 29 OSHA recordable in juries for a TRIR of 9.24. There were two reported near misses or incidents without inju ry. While Owner S expended considerable efforts to achieve better safe ty performance on the project, the TRIR on the project was even worse than that of the industry average. At the time of the interview, the owner was eager to know the root causes of poor safety performance on the project. The scorecard developed in Chapter 5 was used to dia gnose the owner’s involvement in safety management on project F. Project F employed one prime contractor and two large specialty contractors to execute the construction work. These th ree contractors and Owner S formed a construction team, and they developed and impl emented a project specific safety program as one cohesive unit. Owner S and the union business agents participated in the review process after the initial plan was devel oped. Owner S and the three contractors developed and signed the project safety charte r. The charter’s 24point safety program was designed specifically for project F.

PAGE 147

131 The safety program included a crew in centive program and drug and alcohol testing. The program included: issuing work gloves with a project safety logo at the safety orientation class; pr oviding luncheons for the crafts for good safety performances and awareness; and spot rewards for good safety practices. Gift certif icates were issued to crews with no recordable injuries, viola tions, or observed unsafe behaviors for each two-week period. The CII Zero Injury T echniques, including the “Five High-Impact Techniques” were utilized for the project’s 24-point program. Task planners (a task plan sheet) were de signed specifically for this project. Each day the foreman completed the planner and di scussed the job with the crew. When the assignments were changed or the crafts we re re-assigned to another activity, new task planners were completed and discussed with the crew. All employees of contractors and subcontractors went through a four-hour safety orientation and the foremen were required to have the 10-hour OSHA card. This safety or ientation included proj ect specific safety policies and procedures. The executive comm ittee was required to investigate all lost time accidents within 24 hours, and to invest igate recordable injuries and non-injury incidents within 48 hours. There was a da ily update of worker hours and recordable injuries at the daily constr uction review meeting. A "tar get zero audit" was conducted about half way through the outage phase of the project. The audit de termined that 87.5% of the CII techniques were being utilized. Th e overall total score for the audit was 77%, indicating good communications of safety practices. Analysis And Diagnosis The safety manager of Owner S showed a deep concern for construction safety during the interview. Howeve r, the poor safety record on project F was difficult to explain. The safety manager complained that there were two weak points on the projects

PAGE 148

132 that might be contributing reasons: (1) During investigations of the recordable or losttime injuries, management seldom asked th e victims to participate, so valuable information concerning accident causation could not be discovered; (2) Injury management was inconsistent. The owner safety manager complained that some employees of contractors sustained first aid in juries but they were reported as recordable injuries. The scorecard developed in Chapter 5 was us ed to diagnose the level of the owner’s involvement in construction safety manage ment on project F. The results of the assessment of owner involvement for project F are shown in Figure 6-1. The score of project F was only 48%, which demonstrates weak involvement of the owner in safety management. From the result s in the scorecard and some other findings presented in Chapter 5, the project had the following weak points unfavorable for safety: Project context The project was a shutdown project, and a medium size project. It had a tight schedule, and therefore, used two work sh ifts, six workdays per week, and, at the least, 10 hours were worked per shift. The project employed union workers. The owner had great difficulties when trying to regulate safety among union workers. Even the monetary incentive program did not help to remind the workers of safet y. During the outage of the furnace, only 58% of the employees were able to earn the incentive bonus. Selection of contractors The owner did not conduct a stringent c ontractor evaluation procedure. The selection of the prime co ntractor was bound by a previ ous contract. The prime contractor was actually predetermined, because of its success (not measured in terms of safety) on a similar past project. There was no specified TRIR threshold va lue for contractors to satisfy to be selected. Qualifications of the project management team and safety staff of the contractors were not evaluated when selecti ng contractors, either Safety was not a consideration in cont ractor selection.

PAGE 149

133 The project employed low-caliber workers to perform work on the project, in order to save about 5% of the labor cost. Th e foremen had essentially no more working experience than the workers. Owner’s Influence on Construction Safety Scorecard Project Context, Contractor Selection, Contractual Safety Requirements and Owner Involvement in Project Safety Answer Project context: (1) Does the project work one shift? Y N (2) Does the project work five days a week or less? Y N Selection of contractor (3) Is the TRIR requirement for the contractor selection less than 2.0? Y N (4) Are the qualifications of the project team reviewed? Y N (5) Are the qualifications of the safety staff reviewed? Y N (6) Does the evaluation of each contractor’s safety performance make a difference in awarding the contract? Y N Contractual safety requirements (7) Does the project use a design-build contract? Y N (8) Does the contract require the contract or to place at least one full-time safety representative on the project site? Y N (9) Does the contract require the contractor to submit all safety personnel rsums for the owner’s approval? Y N (10) Does the contract require the contract or to prepare a site-specific safety plan? Y N (11) Does the contract require the contractor to submit a safety policy signed by its CEO Y N (12) Does the contract require the contractor to provide a minimum specified amount of training to the construction workers? Y N Contractor Safety Program Requirements Which of the following are required to be included in the contractor’s safety program? (13) Contractor must prepare a plan for site emergencies Y N (14) Contractor must conduct pre-task safety planning on the project site Y N (15) Contractor must implement a substance abuse testing program Y N Owner’s involvement in pr oject safety management (16) Does the owner's safety repr esentative investigate near misses? Y N (17) Are injury statistics on the projects maintained separately on each contractor? Y N (18) Are all project injuries in cluded in the owner’s overall measure of safety performance? Y N (19) The owner actively participates (gives pres entations) during worker safety orientation? Y N (20) Comprehension of safety training is evaluated through testing? Y N Which of the following activities are performed by the owner’s site safety representative? (21) Enforcing safety rules and regulations Y N (22) Monitoring of the implementation of pre-task planning Y N (23) Participating in safety recognition programs Y (24) Participating in safety and/or tool box meetings Y N (25) Does the owner set zero injuries as its safety expectation before the commencement of site work? Y N Total Count of Yes Responses 12 Multiply the Yes Count by 4 (X 4) 48% What the score means: 88% or better is strong owner involvement 52% or less indicates weak owner involvement Figure 6-1. Owner's influence on construction safety scorecard

PAGE 150

134 Contractual safety requirements The owner made detailed safety requirem ents in the contract. However, the number of safety representatives for every one hundred workers was one, which was probably too small for the t ype of work being performed. The contract did not specify minimum safety training requirements for the workers. Owner’s involvement during construction Safety observation program. The project ha d a safety observation program, but the workers were not among the observers. Al so, near misses and first aid cases were not investigated on the projec t. Modifications to the safety program were based on the investigations of th e lost-time accidents and OSHA recordable accidents. Accident investigation program. Only lost-time accidents and OSHA recordable accidents were investigated. The owner did not maintain the safety performances of different contractors, nor did it include the contractor’s injury statistics with its own injury data. Safety recognition program. The owner pa rticipated in the recognition program, but the program was mainly based on the injury records. Since there were 29 injuries during the two mont hs, only 58% of the workers earned the bonus. The effect of the recognition program was modest. Safety training program. The owner did not participate in safety training and orientation. Responsibilities of safety re presentatives. Owner’s safe ty representatives did not participate in toolbox meetings, and they di d not participate in job hazard analysis. Safety objective. The safety objective of the TRIR set by the owner was 2 per 200,000 workhours. This was a relative weak expectation. If the model established in Chapter 5 was used to evaluate the owner’s involvement in safety management, the project would be lo cated on the left most leg of the model. Each index, P, S, C, and O was less than the median of the index. Thus, poor safety performance could be expected on the project. In general, the safety program of Owner S was focused on the elimination of injuries through incentives and punishment. Compliance and enforcement of the safety regulations were the major characteristics of safety management. Near misses and safety observations were not emphasi zed. The safety management of Owner S was not as

PAGE 151

135 proactive as Owner R in case 1. If Owner R was typical of owners with petrochemical projects and if Owner S was typical of owners with othe r manufacturing pr ojects, this may be indicative of why petrochemical projec ts tend to have better safety performances. Summary In this chapter, two case st udies were discussed and eval uated in terms of the owner involvement model and scorecard developed in Chapter 5. From the results obtained in both cases, it is evident that an owner’s influence on project safety is important. The scorecard proved to be an accurate predictor of the level of safety performance actually realized. The model and scorecard can be used to effectively evaluate the owner’s involvement in construction safety.

PAGE 152

136 CHAPTER 7 CONCLUSIONS, DISCUSSIONS AND RECOMMENDATIONS Conclusions In the past decade, safety performance in the construction industry has been dramatically improved. This is due, in pa rt, to the concerted efforts of owners, contractors, subcontractors and designers. In this study, the owner's role in construction safety was investigated. The relationship between project safety performance and the owner's influence was examined, with particular focus on the project c ontext, selection of safe contractors, contractual safety requirements, and the owner's proactive involvement in safety management. It can be concluded that the owner's involvement can significantly influence project safety performances. The findings show that the owner sets the tone of the safety culture for projects. The owner can impact safety management and the safety commitment of the designer, the contractor, and the subcontractors in various means. Figure 7-1 illustrates how owners can influence project safety through communica ting safety, selecting safe contractors and participating in safety management. At the core of the owner's involvement is the zero injuries objective. Owners should consider using various evaluating measures of safety performance when selecting contractors, when setting performance objectives, when monitoring safety performance, and when partic ipating in project safety programs. These measures include risk involved in the cons truction process, past safety performance (TRIR), overall quality of safety pr ogram, management safety commitment,

PAGE 153

137 qualifications of project management team and safety personnel, worker participation, safety observation results, near misses, and so on. Although there are some aspects of the project that owners cannot impact significantl y, such as the local labor conditions, owners can help promote safety performance by implementing a carefully designed, dynamic safety program. These programs can then serv e as a viable model fo r the contractor to emulate. By being directly involved in project safety management, many proactive owners have achieved excellent safety performances. It should be noted that the involvement of owners in safety management should be a continuous and integral effort. Simply im plementing some proactive safety practices cannot guarantee a safe project. Owner's con cern and participation in safety should start from the very beginning of the project desi gn until the completion and even the operation and maintenance of the project. Safety management should be a dynamic and flexible effort with a firm safety philosophy of zero injuries. Safety should not consist primarily of a safety manual left on the table or in the statistics that are gene rated. The safety program should be changed with reference to the site conditions and the dynamic evaluation of safety performance (e.g. safety observations, near misses, safety inspections). While the specific practices implemente d on the project can vary, the safety program should be implemented consistently and firmly once developed. There should be no gray area in safety management. A be havior or condition can be either safe, or unsafe. No unsafe behaviors or unsafe physical conditions should be ignored or tolerated.

PAGE 154

138 Referring to the results in this study, th e zero injuries object ive is attainable, particularly with the owner’s proactive invol vement in safety. The study also found that petro-chemical owners are among the most pro active owners in construction safety. This may be due to the traditional concern for safety in the chemical industry. It was accepted by many petro-chemical owners that the safe ty attitudes in their major business also impact their philosophy in construction safety. Safety is necessitated by the considerable hazards existing in the petrochemical industry. It was said that one petrochemical owner who had been proactive in safety started to implement even more stringent safety requirements after they started to produce e xplosive materials. These reasons may help explain why the safety performances of petr o-chemical projects are better than other types of projects. Discussions As to the safety management method, three levels of management can be identified. The first level of management is the accident-and-prevention method. All the efforts on construction safety are led by lessons learned from past accidents. The second level is the regulation-and-enforcement level. Un safe behaviors are the target of the enforcement. The third and most advanced le vel is behavior-based safety. The focus is to establish a safety culture by discouragi ng unsafe behavior and recognizing/rewarding safe behavior, based on observations. The focu s of the different levels is on different elements in the accident causation chain (refer to Figure 7-2). A comparison of the three levels of safety management is listed in Table 7-1. Many petrochemical owners in terviewed in the study were at the third level of the safety management method. In comparison, many manufacturing owners were either at the second level of the method, or they are m oving from the second level toward the third

PAGE 155

139 level. The effectiveness of the different safety management methods can be observed in the differences of the project safety performances. As to the motivation of safety management for the owners, costs and legal entanglements are still the major reasons. Because of the high risks in project construction, many owners find that safety is an essential part of their loss control program. At the same time, caring for peopl e and business reputation are also reasons why owners have become more inclined to pa rticipate in safety management in the past decade. As to the challenges owners confronted in construction safety, some owners commented that labor conditions and the decrease of labor skill levels are the major problems. Although labor conditions largel y depend on the location of the project, statistical analysis showed that open-shop projects had significantly better safety performances than merit shop and union projec ts. This result was also supported by the survey conducted with contractors. Howeve r, it should be noted that no matter what the labor conditions, the zero-injury objectives have been achieved by proactive owners through appropriate safety management met hods. For example, among the six projects without any OSHA recordable injuries, one union project work ed 380,000 injury-free hours, and a merit shop project worked 266,000 injury-free hours. The means by which the owner can help improve construction safety was found to be similar to methods found in other studies to be used by contract ors and subcontractors who had positive impacts on project safety (Hinze, 2002; Hinze and Figone, 1988; Hinze and Talley, 1988). From various studies, th e practices related to better safety performance have been summarized in Ta ble 7-2. Although the study on subcontractor

PAGE 156

140 safety was conducted approximately 15 years before the contractor study and the owner study, the comparison still shows that the prac tices to achieve bette r safety performance are relatively consistent for these three parties. In effect, the cont ractor is acting as a bridge between the owner and subcontractor. The owner’s role in construction safety can then be defined as the party to oversee and f acilitate safety management on the project. Owners may not always take th e leadership role for projec t safety management, but the owner’s attitudes towards safety and their phys ical involvement in safety will favorably impact the safety performance of ge neral contractors and subcontractors. Recommendations All owners, regardless of the type and size of their proj ects, should recognize that they have a responsibility for construction safe ty. Safety should be integrated into the overall project objective of the owner. This study focused on the owner’s individual impact on project safety performance. The influence of owners on designers and contractors was not thoroughly explored. Further research is suggested to develop a fuller understanding of how owners can impact the way designers address safety in their desi gns. Research is also needed to address constructability issues and to establish how owners can facilitate contractors in their efforts to implement safety programs. For example, more information is needed on how owners can develop an effective safety and constructability review procedure, and how owners can work in concert wi th contractors to combine th eir own safety programs with the safety programs of contractors. Th ese studies could be conducted at both the company level and the project level. Thus, a fuller picture could be generated on how to achieve better safety performance.

PAGE 157

141 While it is felt that findings of this research apply to all types of projects, research should be conducted to confirm this. The data were heavily influenced by industrial and manufacturing projects and further study c ould determine how well the findings can be applied to projects of all types. A larger and broader based research study could be focused on the role of owners on smaller proj ects. This would determine if the owner’s role changes with project size. This study did not conduct a cost and benef it analysis when owners get involved in project safety management. However, the be nefit of a safe project should be apparent, for example, the increased morale, the in creased productivity and so on. Also, by avoiding one singe lost-time injury (with th e average cost of 25,000 dollars, according to Hinze and Appelgate (1991)), the owner can ne arly employ a full-time safety supervisor on site. Future study can investigate how owne rs regard the benefit and cost of owner’s involving in construction safety management. Since separate studies have been conducted on the safe ty roles played by owners, contractors, subcontractors a nd designers, future research should be conducted on the impact on project safety when a ll parties are considered to be members of a project team. Prospective and experimental methods can be implemented in the research, and the most effective practices of each part y with the presence of the ot her parties can be evaluated. The interactive impact on project safety perfor mances of the different parties should be evaluated. Thus, an overall model can be developed with consideration of both the separate influence of each party and their interactive impacts. This would provide information by which to develop a ho listic approach to project safety.

PAGE 158

142 Figure 7-1. Owner's involvement in Safety management Zero injuries Select Contract type Contract arrangement Project design Project schedule Construction methods Designers Contractors Subcontractors Outsourcin g and vendors Participate Safety program Safety observations Safety inspections Safety orientation Accident investigations Safety recognition Safety enforcement Communicate Set project expectations for every party involved through the contract and other communications.

PAGE 159

143 Figure 7-2. Focus of different levels of safety management methods Table 7-1. Comparison of the three levels of safety management methods Driving event Aim Means Result Accident-andprevention Accident Avoid repetition of the accident by not making the same mistake Accident investigation, safety training, thorough safety program. Good Regulation-andenforcement Unsafe acts Prevent accidents through the elimination of unsafe acts Enforcement and punis hment, safety inspection, training, thorough safety regulations Better Behavior-based safety Unsafe and safe behavior Zero-injuries through a total safety culture Safety training, observation, near miss and first-aid investigation, safety recognition program, design for safety Best Personal failings or mistakes Unsafe behavior/physical hazards Accident Injury/Property damage/Near miss Social environment and heredity Behaviorb ased safety Regulation-andenforcement Accident-andprevention

PAGE 160

144Table 7-2. Comparison of findings in different studies Owner study Contractor study Subcontractor study Management commitment No solid data support Company president reviews safety reports Home office inspects safety frequently Top management involvement in every recordable accident investigation Home office manager comes to visit the job frequently Staffing for safety Owner assigns full-time safety rep. on site Safety rep. assumes various responsibilities Contractor employ full-time safety representative Safety representative reports to main office staff A designated person for safety on the project. Safety planning Owner requires certain items to be included in the safety program (inc. JSA) Safety rep. monitors safety planning program Require site-specific safety program Hold pre-task meetings TSA conducted in each phase General contractor reviews safety program of sub Safety training Owner involved in safety training; Owner has means to verify comprehension of training Owner’s safety rep. participate in tool-box meeting Formal safety training plan in print Safety training is a line item in the budget Every worker receives safety orientation Formal safety training is conducted Intense safety training conducted monthly for workers and project management Tool box meetings held daily Subcontractors are required to hold safety meetings Separate safety meetings for supervisors Worker involvement Conduct safety observation program with participation of workers; Formal worker safety observation program Management and supervisory person receive behavior overview training Safety perception surveys implemented Subs participate in project meetings with owner Safety incentive and recognition Owner funds safety recognition program Owner safety rep. participates in safety recognition program Safety dinners held on the project Safety award given to workers frequently Safety incentive based on zero-injury objective Family members attend safety dinner Field supervisors evaluated on safety No solid data support Subcontractor management Subcontractors are required to follow same safety regulations as contractor Sub is required to submit site-specific safety plan All sub workers attend formal safety training Sub holds safety meetings daily Sanction subs for non-compliance with safety stds. Subcontractor submits safety report to general contractor. General inspects sub safety Accident Report and investigation Incident stats. maintained by contractor Include safety stats. of contractor in owner’s Safety rep. reviews accident reports More near misses recorded on the project Top management investigates every accident General contractor investigates accidents of subs Drug test Substance abuse testing required in safety program Implement substance abuse test No solid data support

PAGE 161

145 APPENDIX A OWNER SURVEY QUESTIONNAIRE Project Owner’s Role in Co nstruction Worker Safety Please provide answers for the following questi ons. For most questions, there is only one answer to check, however, in some cases a short written re sponse is requested. A. General information 1 What is the typical annual capital budget of your firm for construction projects in North America? About $ Million per year. (a) How was the budget distributed among constr uction projects in the last fiscal year? % for new projects; ____% for alteration, revamp or rehabilitation projects. (b) What percent of the annual cap ital budget is spent on design services ? %. (c) What percent of the annu al capital budget is performe d by construction contractors (not done in-house)? %. 2 As the owner, does your firm track sa fety performance of contractors on the construction project(s)? Yes No (a) If yes, how long has your firm kept track of the safety performance of contractors? For the past years. (b) If yes, what was the overall OSHA recordable injury rate of contractors working on your capital projects in 2001? About OSHA recordable injuries per 200,000 work hours. (c) What other safety performance measures does your firm monitor? (d) Does your firm establish/ negotiate specific safety goals for contractors performing capital work? Yes No 3 What is the dominant type of constr uction project in your firm’s construction budget? (Please check one and show the percentage) Manufacturing facilities _____% Petro-chemical plants _____% Buildings _____% Utilities _____% Land development _____% Civil works _____% Other (please specify): B. Safety coordination of the owner 4 Does your firm include safety as part of the project performance review? Yes No 5 When, at the earliest, does your firm begin to emphasize safety on projects? ( Check one ) During the concept and feasibility phase. During the design phase. Before the start of bidding. During the bidding phase. After bidding, before starting site work. After the start of site work. Other (please specify):

PAGE 162

146 6 Does your firm assign full-time owner safe ty representative(s) to the construction projects being built? Yes No If yes, what criterion is used to place safety representatives on projects? 7 Which of the following does the owner’s site representative/manager attend? (Please check all that apply.) Safety orientations. Site safety meetings. Site safety audits. None of them. 8 In the past fiscal year, what percent of your projects were undertaken under an owner controlled insurance progr am (OCIP)? _____% If applicable, what is the smallest size (c ost) of project that would be constructed under an OCIP? $______ million 9 Do you have minimum safety performance requirements for contractors, other than complying with the OSHA regulations? Yes (please specify): No. 10 As the owner, does your firm make any a llocation (personnel or financial) for the training of the contractor’s employees? Yes No If yes, please explain: C. Safety management practice in projects. 11 How are contractors initially selected? By competitive bid By negotiated contracts Through references and by reviewing past performance Other (please specify): What is the breakdown of the contracts? Lump-sum contracts: % Cost reimbursable contracts: % 12 During the selection of a contractor, is safety performance a factor in the prequalification process? Yes No (a) If yes, what measures are used to co mpare the safety performances of different contractors? (Please check all that apply, and provide the related data requested) Experience Modification Rating (EMR) of the contractor, should be less than OSHA recordable injury rate of the contractor, should be less than per 200,000 work hours. Loss Ratio of the contractor, should be less than Site-specific safety program prepared by the contractor. Qualifications of the safety staff of the contractor. Quality of the overall safety program of the contractor. Other (please specify): (b) When selecting a contractor, how would you rate the importance of safety in your overall performance assessment of the contractor? The rating of the importance of safety is _____. (assume 1 is very low and 10 is very high). 13 As the owner, which philosophy does the firm hold concerning safety? We assume as much liability for safety as possible, and we keep safety totally under our control.

PAGE 163

147 We equitably share the liability for safety with the contractor, and we emphasize cooperation with the contractor on safety matters. We try to avoid safety responsibility as much as possible, and expect the contractor to assume responsibility for project safety. Other (please specify): 14 As the owner, does your firm require a ll employees of contractors doing capital projects to attend standardized site orientation training sessions? Yes No 15 As the owner, does your firm contribute funds for the implementation of a safety incentive program by the construction contractor? Yes No If yes, please specify: 16 What safety requirements are specifically included in the co nstruction contract? (Please check all that apply) Contractor must comply with the local state and federal safety regulations. Contractor must place at least one fulltime safety representative on the project. Contractor must submit the rsums of key safety personnel for the owner’s approval. Contractor must submit a safety plan for the owner’s approval. Contractor must report all lost time injuries to the owner. Contractor must report all OSHA recordable injuries to the owner. Contractor must include personnel from the owner in coordination meetings. Contractor must implement a drug-testing program. Contractor must conduct weekly safety meetings for the workers. Contractor must submit a s ite-specific safety program. Contractor must implement a permit sy stem when performing hazardous activities. Contractor must submit a safety policy signed by its CEO. Other (please specify): . 17 As the owner, does your firm require basic safety orientation for all workers before they can work on your projects? Yes No If yes, how is safety orientation conducted on your project(s)? We totally control the safety training by conducting the training sessions. Safety training is essentially a joint or shared effort between the contractor and us (the owner). Contractor conducts safety traini ng for everyone who enters the site; we only monitor the records and results of safety training. Contractor is totally responsible for safety training, and we don’t get involved. Other (please specify): 18 Contractually, what sanctions can your firm impose on a contractor for noncompliance with safety requirements? (Please check all that apply) Certain amount of money could be deducted from the payment earned by the contractor. Site work could be suspended until the contractor complies with safety requirements specified in the contract. Contract could be terminated. Other (please specify): What has been the most extreme sanction actually imposed on a contractor?

PAGE 164

148 19 Please select among the safety practices that are implemented by your firm, as the owner, to support safety during construction. (Please check all that apply) Owner’s personnel conduct periodic job site safety inspections or safety audits. Owner places a safety person on the projec t to support the contractor on project safety. Owner provides a nurse or emergency medical technician (EMT) for the construction project. Owner’s personnel participate in some contractor safety meetings. Owner’s personnel participate in the investigati on of all OSHA lost workday injury accidents. Owner implements a safety incen tive that can be earned by the contractor for completing the project below a specified OSHA recordable injury rate. Owner monitors injury in cidence rates on each project. Other (please specify): 20 Do any managers in your firm have their bonuses tied directly to the level of safety performance of contractors working on your projects being constructed? Yes No 21 Does your firm consider anyone in your firm accountable for injuries sustained by employees of the construction contractor? Yes No, this is the contractor’s sole responsibility. D. Safety in design phase and others. 22 Is construction safety actively addres sed in the design pha se, whether through designer decisions or engineering input during the design phase? Yes No If yes, please specify how your firm evalua tes designers (prior to obtaining their services) on their ability to address construction safety in their designs. . 23 Please rank the following measures in term s of their priorities in achieving success in project safety? (Please rank these, beginning w ith 1 as the most important.) ( ) Owner emphasizes safety and constructability in design. ( ) Owner selects safe contractors to carry out the work. ( ) Owner participates in and monitors safety during the whole life of a construction project. ( ) Owner develops an effective safety recognition and reward program. ( ) Owner dedicates funds to support th e contractor’s efforts in safety. ( ) Other (please specify): 24 Provide any additional co mments or suggestions you wish to share regarding construction safety? . .

PAGE 165

149 APPENDIX B OWNER INTERVIEW QUESTIONNAIRE Owner Study Interview Questionnaire Information specifically focuses on one current project or a recently completed project. Code: Date: ______________, 2002 Project: Project location: Owner Co.: General contractor: Interviewee: Title of Interviewee: Telephone: Email: A. The project 1 What type of project is this? (a) public 53 (89.8%) private 6 (10.2%) (b) new 30 (50.8%) renovation 11 (18.6%) maintenance 3 (5.1%) shutdown 8 (13.6%) other: ___________________combined 6 (10.2%) (c) Residential 0 Commercial 1 (1.7%) Manufacturing 19 (32.2%) Petrochemical 30 (50.8%) Utilities 5 (8.5%) Civil works 1 (1.7%) Others _________________ (d) union 19 (32.2%) open shop 19 (32.2%) mixed 21 (35.6%) 2 Was this project awarded through competitive bidding? Yes 43 (72.9%) No 16 (27,1%) How was the contractor selected? negotiated with one firm 16 (27.1%) short bidder's list 34 (57.6%) open to all 7 (11.9%) Not known 2 (3.4%) 3 What type of contracting arra ngement is used on this project? General Contract 21 (35.6%) Multiple Primes 11 (18/6%) Design-Build 15 (25.4%) C.M. at risk 6 (10.2%) C.M. (agency) 6 (10.2%) Other: _______ 4 What type of contract is used on this project? lump sum 23 (39.0%) cost plus 19 (32.2%) time and materials (T&M) 17 (28.8%) other 5 What is the total estimated cost of this construction project: $ M. 6 Who provides the workers’ compensation in surance on this construction project? Each firm provides their own 45 (76.3%) OCIP 13 (22.0%) CCIP PCIP NA 1 (1.7%) 7 What is the expected project durati on (construction phase only): _______ months. (a) What is the current percent of completion for this construction project? _______% of the duration (b) Anticipated completion date (m,y): ___________, ______ 8 How many total worker hours have been expended on this project? _______________thousand hrs Since January 1,2002 In the past 12 months ________________ Does this include only the hours of the prime contractor or does it include subcontracted work? prime subs both 9 How many shifts are worked on this project? _____ How many hours per shift? ______ (hrs) and how many days per week? ______.days 10 Total number of workers at peak level for the project is: _________

PAGE 166

150(a) How many field workers are on the project at this time? _______ which is about % the peak. This number includes: prime only all workers (b) What percent of the workers do not speak English? __________% 11 What percent of the construction work is performed by the prime contractor? _______% 12 How many subcontracts have b een awarded on this project: ________ 13 How would you rate the prime contractor's commitment to safety? (1-7, 7=best) ______. 14 Does the prime contractor re port accident statistics to you? Yes 59 (100%) No If yes, what kind of data? OSHA recordable IR 59 (100%) Lost-time injury 59 (100%) Near misses 52 (88.1%) Environmental issues 54 (91.5%) Others, please specify _____________________________ 15 How many incidents have been reco rded on this construction project? _______ near misses _______ first aid cases _______ OSHA recordable _______ w. comp cases _______ lost workday Do these incidents include subcontractor statistics? Yes No 16 Does the owner establish/ne gotiate specific safety goals for contractors performing this project? Yes 47 (79.3%) No 12 (20.7%) If yes please specify The OSHA RIR should be < ____ Lost workdays should be < _____ Other __________ B. Safety representative of the owner 17 Who has field or site responsibility fo r the project from the owner’s staff? _______ Total number of owner staff members w ho are full-time on the project? _______ 18 Title of owner's rep. having the most safety responsibilities? __________________ Is this an owner’s employee or a consultant? Employee 52 (88.1%) Consultant 7 (11.9%) 19 What percent of the owner's rep's responsibilities relate to safety? ________% 20 Who does this person report to? Owner company 24 (40.7%) Project Manager; 21 (35.6%) Other__________14 (23.7%) 21 Is this person a member of the project management team? Yes 54 (91.5%) No 5 (8.5%) Does the owner's safety rep have the authority to stop unsafe work? Yes 56 (94.9%) No 3 (5.1%) 22 Is this person copied on information related to th e project progress (costs, scheduling, quality, etc.)? Yes 49 (83.1%) No 10 (16.9%) 23 Does this person have any responsibilities on any other projects? Yes 28 (47.5%) No 31 (52.5%) 24 What are the job responsibilities and dutie s of owner’s jobsite safety rep.? Monitoring safety management and performance of the contractor on a daily basis 53 (89.8%) Reviewing safety performance on site and submits reports to the home office 52 (88.1%) Coordinating safety efforts on site 48 (81.4%) Participating in site safety training and orientation sessions 42 (71.2%) If applicable to what extent are the owner’s personnel involved in orientation training? participate in every session 15 (25.4%) participate in some sessions 24 (40.7%) Participating in safety meeting and tool box meetings 46 (78%) Frequency : d 12 (20.3%) w 27 (45.8%) m 6 (10.2%) Conducting site safety inspection and auditing 53 (89.8%) Frequency : d 21 (35.6%) w 25 (42.4%) m 5 (8.5%) Assisting in job hazard analysis 46 (78%) Frequency : d 19 (32.2%) w 16 (27.2%) m 9 (15.3%) Issuing permits 27 (45.8%) -------------Monitoring pre-task analysis program 42 (71.2%) Frequency : d 12 (20.3%) w 18 (30.5%) m 6 (10.2%)

PAGE 167

151 Participating in safety recognition program 44 (74.6%) -------------Investigating near miss 50 (84.7%) -------------Investigating first aid accident 46 (78%) -------------Investigating recordable accident 45 (76.4%) -------------Reviewing contractors’ safety reports. 52 (88.1%) Frequency : d 15 (25.4%) w 27 (45.8%) m 6 (10.2%) Enforcing safety rules 53 (89.8%) -------------Other, please specify. ___________________________________________________________ 25 Does the owner’s rep. review the safety performance of the contractor on a regular basis? Yes 58 (98.3%) No If yes, how? By checking the project lo st workday injury rate. 56 (94.9%) By checking the project recordable injury rate. 56 (94.9%) By checking the project first-aid injury rate. 52 (88.1%) By checking the project n ear miss rate. 47 (79.7%) By checking the project sa fety inspection records. 47 (79.7%) By checking worker training records. 41 (69.5%) Other, please specify_____________________________________________________________ C. Selection of the contractor 26 Does the owner have a preferred list of contr actors to receive contract awards? Yes 48 (81.4%) No 11 (18.6%) (a) If yes, is safety performance a consideration? Yes 56 (94.9%) No (b) If yes, how is safety rated in the overall review of the contractor? (1 to 7, 7= most important) (c) If yes, who performs this evaluation? contract’s/purchasing 31 (52.5%) safety person 34 (57.6%) ___________ 27 During the selection of contractors, what measures are used to evaluate the safety performances of different contractors? Experience Modification Rating (EMR) of the contractor, should be less than 45 (76.3%) OSHA recordable injury rate of the contractor, should be less than per 200,000 work hours. 49 (83.1%) Loss Ratio of the contractor, should be less than 7 (11.9%) Qualifications of the safety staff of the contractor. 42 (71.2%) Qualifications of the project management team of the contractor. 38 (64.4%) Quality of the overall corporate safety program of the contractor. 52 (88.1%) Copy of OSHA log for the past year. 40 (67.8%) OSHA inspection history on past projects. 27 (45.8%) Other (please specify): (a) How are these criteria used? In other words, can this information make the difference between getting the contract or not? Yes 52 (88.1%) No 7 (11.9%) If yes, has it ever? Yes 50 (84.7%) No (b) If a contractor has some bad statistics, can the contractor show that they have made major changes in the program and still be considered for contract award? Yes 56 (94.9%) No (c) What leading metrics or indicators do you use to predict and improve safety performance or to deliver a safe construction project? _______________________________________________ 28 During selection of contractors, how do you ensure that safety is given at least equal weight to quality, cost and schedule? ________________________________ D. Construction contract

PAGE 168

152 29 What safety requirements are specifically included in the co nstruction contract? Contractor must comply with the local, state and federal safety regulations. 59 (100%) Contractor must comply with safety requir ements beyond the OSHA regulations. 52 (88.1%) Contractor must place at least one full-time safe ty representative on th e project. 49 (83.1%) Number of employee per safe ty representatives? _______ Sub requirement? yes 28 (47.5%) no 5 (8.5%) Contractor must submit the rsums of key safety personnel for the owner’s approval. 42 (71.2%) Contractor must provide specified minimum training for the workers 37 (62.7%) If applied, please specify _______ hours per worker per month Contractor must report all lost time injuries to the owner. 58 (98.3%) Contractor must report all OSHA recordable injuries to the owner. 57 (96.6%) Contractor must report all injuries to the owner 57 (96.6%) Contractor must include personnel from the owner in coordination meetings. 40 (67.8%) Contractor must submit subcontractor list to owner for approval 47 (79.7%) Contractor must implement a substance abuse program. 55 (93.2%) Contractor must participate in site safety audits. 52 (88.1%) Contractor must conduct weekly safety meetings for the workers. 55 (93.2%) Contractor must submit a site-s pecific safety plan. 50 (84.7%) Contractor must submit a safety policy signed by its CEO. 31 (52.5%) Contractor is required to provide specified PPE (hard hats, safety glasses, gloves) 57 (96.6%) Contractor must implement a permit system when performing hazardous activities (line breaks, lockout/tagout, excavations, proximity to power lines confined space entry, hot work, etc.). 52 (88.1%) Other (please specify): . 30 Are the same safety requirements im posed on subcontractors and lower-tier subcontractors? Yes 54 (91.5%) No 3 (5.1%) NA 2 (3.4%) If yes, has any subcontractors of you been refused by the owner for safety reasons? Yes 40 (67.8%) No 11 (18.6%) Do you know why? ________________________________________________________________ E. Owner’s involvement in safety 31 As the owner, which philosophy does the firm hold concerning safety? We assume as much liability for safety as possible, and we keep safety totally under our control. 7 (11.9%) We equitably share the liability for safety with the contractor, and we emphasize cooperation with the contractor on safety matters. 43 (72.9%) We try to avoid safety responsibility as much as possible, and expect the contractor to assume responsibility for project safety. 4 (6.8%) Other (please specify): 5 (8.5%) 32 Are safety observers used on this project?` Yes 50 (84.7%) No 9 (15.3%) If yes, how? 33 As the owner, do you require specific items to be included in the site specific safety program prepared by the contractor? Yes 58 (98.3%) No 1 (1.7%) If yes, what are they? OSHA specific regulations 52 (88.1%) Specific safety training session 51 (86.4%) Prime contractor’s employees to have 10-hr OSHA cards 13 (22.0%) Prime contractor’s supervisor’s have CPR and First-Aid cards 18 (30.5%) Training on the hazards related to the tasks 53 (89.8%) Pre-project safety planning 51 (86.4%)

PAGE 169

153 Task specific PPE analysis 50 (84.7%) Conduct regular safety inspections 54 (91.5%) Incident reporting and investigation 55 (93.2%) Emergency plan (medical and hazard ous materials) 48 (81.4%) Substance abuse program 53 (89.8%) Regular safety meetings 56 (94.9%) Safety responsibility defined for all levels 41 (69.5%) Emergency response team maintained on the project 27 (45.8%) Daily JSA (Job Safety Analysis) are conducted on the project site 48 (81.4%) Other, please specify:____________________________________________________________ 34 Are subcontractors specifically in cluded in the safety program? Yes 56 (94.9%) No 2 (3.4%) 35 As the owner, do you have safety program requirements that the contractors must follow? Yes 56 (94.9%) No 2 (3.4%) 36 When a prime contractor’s employee sust ains an OSHA recordable injury, what must be reported to the owner? Supplementary (101) ______________________ 37 How are injury statistics maintained? by project 45 (76.3%) by contractor 33 (55.9%) blended with owner 27 (45.8%) 38 Does owner monitor near misses? Yes 55 (93.2%) No 3 (5.1%) If yes, how are they defined? near death incidents any incidents that can hurt any unplanned event others (specify)________________________________________________________________ 39 How are near misses handled? Carefully recorded and investigated; 36 (61.0%) Recorded without further investigation; Nothing is done with near misses, we only deal with accidents; Other_________________ 40 Are some funds provided to the contract or, above and beyond the contract amount, to promote project safety? Yes 35 (59.3%) No 24 (40.7%) If yes, explain: ____________________________________________________________________ 41 Is the contractor’s safety performance in cluded in the owner’s safety performance statistics? Yes 37 (62.7%) No 22 (37.3%) 42 Does the owner participate in OSHA’s Volunt ary Protection Program (VPP)? Yes 17 (28.8%) No 34 (57.6%) 43 Is a nurse or EMT provide d on the construction site? Yes 50 (84.7%) No 9 (15.3%) If yes, who is responsible for providing these services? Owner 25 (42.4%) Prime Contractor 20 (33.9%) BOTH 4 (6.8%) 44 The following questions are concerni ng safety training and orientation. (a) As the owner, does your firm make any a llocation (personnel or financial) for the safety training of the contractor’s employees? Yes 44 (74.6%) No 15 (25.4%) If yes, please explain: (b) What is involved in orientation training? videos 50 (84.7%) contractor presentations 39 (66.1%) owner presentations 45 (76.3%) consultant presentations 10 (16.9%) reading materials 37 (62.7%) (c) Does every worker on site receive orientation training? Yes 57 (96.6%) No 1 (1.7%) (d) Do all workers receive the same orie ntation training? Yes 50 (84.7%) No 8 (13.6%) (e) What is the typical duration of orientation training for a worker? ______ hours (f) Does the owner have a means of verifying the comprehension of the orientation training? Yes 48 (81.4%) No 11 (18.6%) (g) Is there any special safety training for non-English-speaking employees? Yes 23 (39%) No 34 (57.6%)

PAGE 170

154 If yes, please specify: orientation tool box talks pre-task planning ________ (h) How many hours of training do workers a nd supervisors receive monthly, not counting toolbox meetings? _____hrs for workers; _____hrs for supervisors 45 Is there a safety recognition/reward program on this project? Yes 50 (84.7%) No 9 (15.3%) If yes, please explain how it works: (distinguish between incentives for workers and for the firm) Workers: _____________________________________________________________________ Supervisors: __________________________________________________________________ Contractor: ___________________________________________________________________ (a) Are worker incentives based on safe behavior or on injury occurrences? behavior 15 (25.4%) injuries 18 (30.5%) BOTH 20 (33.9%) (b) Is the owner involved in any way in the safety recognitio n/reward program? Yes 46 (78%) No 7 (11.9%) Describe? _______________________________________________________________________ 46 Are safety dinners held for pe rsonnel at the construction project? Yes 44 (74.6%) No 15 (25.4%) If yes, are family members of workers asked to attend? Yes 5 (8.5%) No 42 (71.2%) 47 Is there a newsletter or a safety newsletter? newsletter 21 (35.6%) safety newsletter 14 (23.7%) none 24 (40.7%) (a) If yes, what percent of the newsletter is devoted to safety? ______% (b) If yes, is the newsletter also published in Spanish? Yes 1 (1.7%) No 29 (49.2%) 48 Other than toolbox meetings, how do you ensure the communication of safety information to all levels of the project? _________________________________ ______________________________________________________________________________ How do you rate the communication and coopera tion efforts between the prime contractor and the owner on this project concerning safety? (1 to 7 points, with 7 to be the best.) _________ 49 How do you ensure that the safety plan is kept active with frequent review and that it is modified as needed? . 50 During the design of this project, were construction safety issues specifically addressed? Yes 58 (98.3%) No 1 (1.7%) If yes, please describe: ______________________________________________________________ Can you think of any examples of how the design was changed to address safety?

PAGE 171

155 APPENDIX C CONTRACTOR SURVEY QUESTIONNAIRE The Owner’s Role In Construction Safety The following questions are to identify the owner’s role in construction safety based on your general experience. Pl ease check the appropriate optio n(s) and give brief answers. The questionnaire can be finished within five mi nutes. Thank you fo r your participation. 1 How long have you been working in the construc tion industry? _____ years. 2 What percentage of the safety burden is generally assumed by the owner? ____% (a) What is the lowest level of owner i nvolvement in safety you have observed? ____% (b) What is the highest leve l? ____ %; Was it successful? Yes No How did the owner show this involvement? _____________________________________________3 In the past decade, owner's involvement in construction safety on your projects has: increased remained the same decreased Reason of change (if any):____________________________________________ 4 In the past fiscal year, what is the firm's total construction volume? $_______M What is the firm's overall OSHA Recordable Injury Rate? ____ per 200,000 work hours. 5 In your experience, what type of proj ect has the best safety performance? (a) Residential Commercial Manufacturing Petrochemical Utilities Civil works No difference Reason: (optional)__________________________________________________________________ (b) union open shop no difference Reason: (optional)__________________________________________________________________ (c) lump sum cost plus no difference Reason: (optional)__________________________________________________________________ 6 When awarding contracts, what level of importance does the owner place on the following? (use 1 for top prior, 2 for the second, and so on) ___ economic ___ safety ___ schedule ___ quality 7 In general, do you think the owner’s staff provides sufficient support on safety on your projects? Yes No Depends, please explain ___________________________________________ 8 In what area do you think th e owner can do the most to promote safety? (check one) Safety training Safety inspection Safety incentive Accident investigation Job hazard analysis Other ____________________ 9 What is the most efficient means for the owner to address thei r concern for safety? ________________________________________________________________________________ 10 To what extent do you expect the owner to be involved in safety management? Show their concern, but let the contractor manage safety in their methods Set their expectation, support and guide the contractor in safety management Totally take control of the safety management on the project, taking the responsibility Other, please explain _____________________________________________________________ 11 What do you think is the key to the succes s of safety manageme nt on the owner side? ________________________________________________________________________________

PAGE 172

156 APPENDIX D ANALYSIS OF THE CONTRACTOR SURVEY To supplement and verify th e results obtained through the analysis of the interview data, a survey was conducted among contractors. Also, the purpose of the survey was to collect opinions of contractors on how they w ould like owners to facilitate construction safety management, and to share these opini ons with owners. The questionnaire shown in Appendix c was used to obtain the in formation about contractor opinions. The Contractors Interviewed The survey questionnaires were sent to safety managers of contractors at the corporate or project level, who were familiar with safety management on large projects. A total of 41 responses were received. Among them, 10 responses were from Canada and the other 31 were from the U.S. Table A1 shows the years of participation in the construction industry of the respondents, tota l construction volume of the respondents’ firms in the last year, and the TRIR information for the last year. Table A-1. Characteristics of the respondents Years of participation in the industry The firm's total construction volume The firm's overall TRIR Number of replies 41 36 40* Mean response 23.63 $854,000,000 1.56 Median response 24 $160,000,000 1.22 Minimum 3 $4,000,000 0.00 Maximum 50 $9,000,000,000 6.80 25 20 $26,000,000 0.59 50 24 $160,000,000 1.22 Percentiles 75 30 $622,000,000 2.23 Note:* number of replies was 41 with 40 providing the TRIR of the firm.

PAGE 173

157 It is evident that most respondents ha ve a long history of working in the construction industry and are qui te experienced. The construction firms they are working with have much better safety performances (median TRIR of 1.22) than the construction industry average (average TRIR of 7.8 in 2001, BLS, 2003). Note that these respondents had a cumulative total volume of 30.75 billion dollars. The sample size is small, when compared with the large number of constr uction firms in North America. The data collected shed light on the owners' involveme nt in project safety management, and how these contractors would like the owners to get involved. Project Content And Safety Performance In answering the question "In your experience, what type of project has had the best safety performance", contractors gave differe nt answers. Summarization of the answers is shown in Figures A.1 to A.3. NA 4.9% cost reimbursable 48.8% lump sum 4.9% no difference 41.5% Figure A-1. What type of proj ect has the best safety perf ormance? (type of contract)

PAGE 174

158 NA 4.9% no difference 36.6% open shop 46.3% union 12.2% Figure A-2. What type of project has the best safety perfor mance? (union or open shop) manufacturing 9.8% NA 7.3% no difference 7.3% petrochemical 63.4% commercial 12.2% Figure A-3. What type of proj ect has the best safety perf ormance? (type of project) The results are quite similar to the results obtained through the project interviews. Petrochemical projects, and open shop proj ects reportedly have better safety performances than other projects. Also, cost reimbursable contracts are preferred by more contractors to promote safety, alt hough this was not supported by the interview data.

PAGE 175

159 The Owner's Involvement In Project Safety Management The survey asked questions about the invol vement of owners in project safety management from various aspects. These que stions were concerned with responsibilities owners assumed, level of importance owners pl aced on safety, and whether or not owners provided adequate support fo r safety on their projects. What Percentage Of The Safety Responsibility Is Assumed By The Owner The distribution of the contractors' answers is shown in Table A-2. It is reasonable to conclude that the general involvement of owners in project safety management is between the range of 20-30 per cent. Although some owners may take 0 or 100% of the safety responsibilities on their projects, the overall involvement is generally modest. Table A-2. What percent of safety re sponsibility is assumed by the owner Percentage of the safety responsibility generally assumed by the owner Lowest level of owner's involvement (Percentage) Highest level of owner's involvement (Percentage) Mean 36.63 8.43 72.63 Median 25 0 80 Std. Deviation 32.45 15.26 26.98 Minimum 0 0 5 Maximum 100 70 100 25 10 0 50 50 25 0 80 Percentiles 75 63.75 10 90 Despite the owner’s overall modest invol vement in safety, most contractors mentioned that the owner's involvement in proj ect safety had increased in the past decade (see Figure A-4).

PAGE 176

160 decrease 7.3% increase 70.7% remained 22.0% Figure A-4. Change of owner involvement in project construction sa fety in the past decade Respondents gave various examples of increased involvement, including: Safety representatives of owner are generally assigned to each project; Participation in safety incentive programs; More limitations on bid lists ba sed on safety performance; Participation in weekly safety audits; Detailed and constructive reviews of c ontractor health and safety programs; Require contractors to have a JSA on all pr oject and tasks to be done, in addition to contractor pre-task checklists; More involvement in day-to-day activities Do not view contracted work as contracti ng away responsibility, but rather that the contractor is an ex tension of their internal work force who possesses a unique skill set or has access to labor providers; More stringent reporting re quirements, more emphasis on training and orientation; Higher standards are generally enforced, little or no tolerance for substandard practices; Assumption of responsibilities by management and supervision;

PAGE 177

161 Most plant owners at least show an interest in the statistical out come of the project and support our programs and initiatives from a distance, a small number seek to actively participate; More emphasis on behavior based safety. Level Of Importance Owners Place On Safety When Awarding Contracts A question asked the levels of importance owners placed of different factors when awarding contracts (with 1 being the most impor tant and 4 being the l east important). In comparison with cost, schedule and qualit y, safety is not the most important consideration for many owne rs. The Friedman two-wa y non-parametric ANOVA shows there were significant differences between the ranks of importance owners placed on each of the four factors (refer to Table A-3). Cost is the mo st important concern for most owners, with schedule and safety being statis tically tied for second, followed by quality. Table A-3. Level of importance of the factors when owners award contracts FACTOR Mean Std. Error Median Mean Rank Sign. (two-tail) Cost 1.27 0.55 1 1.442 Safety 2.73 1.1 3 3.078 Schedule 2.51 0.75 2 2.748 Quality 3.39 0.8 4 3.644 <0.01 Does The Owner Provide Sufficient Support For Project Safety Most contractors think that owners do pr ovide sufficient support for safety on their projects (see Figure A-5). For those who mentioned "depends", the following explanations were offered: Depends on how concerned they are about human welfare. Depends on industry: petrochemical is best, civil and utilities are worst. Most contractors are able to provide the proper level of safety support for their employees if adequate funds are available to support it. The only complaint with reimbur sable contracts is that c ontractors tend to abuse or overstate their need for safety expenditures.

PAGE 178

162 Depends, more interested in a good record than how it was achieved, hiding of injuries acceptable in some cases, or spending large sums to keep workers on modified work (working wounded) to maintain LTI free status. It varies from owner to owner. Money is key with some and safety is key with others. There is a wide range in the level of support coming from the owners. Many owners view safety performance solely through statistics. You cannot improve safety by managing statistics better. Each owner has a different approach, so me support well and others not enough. depends 19.5% no 24.4% yes 56.1% Figure A-5. Does the owner provide suffi cient support for safety on the projects How Can Owners Facilitate Contract ors In Project Safety Management In the questionnaire, questions were al so asked how owners generally help contractors with safety management and how contractors would like owners to facilitate their safety efforts. In What Area Can The Owner Do The Most To Promote Safety The question listed six options, including sa fety training, safety inspections, safety incentives, accident investigations, job hazar d analysis, and "others". The responses of contractors varied considerably, and many pr ovided expanded answers or several answers to this question. Among the listed options 13 responses mentioned safety training, 6

PAGE 179

163 responses mentioned safety incentives, 5 responses mentioned safety inspections, 5 responses mentioned job hazard analysis, and none mentioned accident investigations. Other than the five specific opti ons, respondents also mentioned: Behavior based safety incentives. Show leadership in safety. Reduce double standard, same rules for contractor employees and owner employees. Contractor review, comparison. Overall support, being vi sible and participating. Contract terms and conditions co mbined with full compliance. Complete engineering and develop a schedul e that is reasonable and able to be accomplished safely. Leadership: setting the st age for success and holding contractors accountable for compliance to the established management processes and taking action on noncompliance. Continuing the evolution of OHS&E in a management system, integrated into every aspect of business. What Is The Best Way That Owners Ca n Address Their Concern For Safety This was an open-ended question. The an swers provided by contractors varied. These answers can be grouped into different categories: (1) Intensive involvement: 100% proactive participation in a project safety program, one that is not owned by the owner or contractor but by the project Be involved in the process by performing a udits and assisting in training programs Support and be involved in the process By showing their willingness to stop a project in order to make it safe for construction employees

PAGE 180

164 Continuous involvement: can assist the c ontractor with regular assessments and audits VISIBLY lead by example get involved at a personal level They need to help promote and drive the overall site safety program (2) Partnering with contractor: Discussing with the contractor on the proj ect and addressing c oncerns in a team approach Work with the contractor to identify and re solve potential hazards, forget cost and schedule issues By supporting the efforts of the contractor s and requiring the same actions of their employees, one program for all Make safety a priority and work with the contractor to perform to a high standard Demonstrate their commitment to the pr oject team and the workers regularly by being part of the team. The owner's re presentative should be visible. Owners could support the constructor by understandi ng the effort being made to be injury free and not get focuse d on statistics only Show leadership and integrate thei r team with the contractor team (3) Carefully selecting contractors and se t high safety expectations in contracts Bid list open only to companies with good safety levels Set expectations in the contract and hold contractors accountable Draft contracts that focus on safety activ ities and actions, NOT numbers. Measure positive performance and compliance, allow time and monies for training and manage proactively Work with the contractor to address c oncerns and be willi ng to award work on parameters other than just price Select the correct GC or CM and have a gr eat contractor selection process for every contractor who performs work (4) Other safety practices Walk the talk. No double standards.

PAGE 181

165 Have regular safety meetings, with all primes and subs involved Support the cost of training and safe ty professionals with resources Training seminars focus on job hazard analysis Do not start field construction until the engineering is 80% complete and do not allow contractors to start work until the complete construction package, including material, is available; also complete each phase of the project before starting the next phase, Continue to place a priority on safety, insi st on trained persons from contractors, do not place schedule above safety, always tr y and take a practical approach, do not have double standards, i.e., owner forces, direct hire forces, nonunion forces and other forces should not be treated differently Safety meetings (communications) Hold their employees to the same level as contractors Many contractors pointed out the importance of treating the employees of owners and contractors equally on safety issues. Once safety regulations are developed, they should be implemented firmly a nd consistently. This is a ve ry important point which was not included in the interview questionnaires fo r owners. However, some results in the interview data analysis also support this argum ent. For example, owners who include the project injury statistics in their own injury statistics reported significantly better safety performances. To What Extent Do You Prefer Owners To Be Involved In Project Safety Twenty-five respondents prefer owners "setting their expectations, support and guide the contractor in safety management". Seven respondents would like the owners to "show their concern, but let the contractor manage safety". Only two respondents prefer owners to totally take contro l of safety on the projects. Other responses include the following:

PAGE 182

166 100% participation in a jointly owned project safety program Attract experienced safe worker s to projects if necessary Set expectations, let the contractor ma nage safety and hold the contractor accountable Set their expectations, dem onstrate leadership through re gular participation at the project management level but let the contractor manage safety The research results support the importance of owners setting expectations before project commencement. This is also prefe rred by the contractors: setting high safety expectations at project star t, followed by support and guidance in safety management throughout project execution. Summary From the analysis of the surv ey to contractors, it is evident that owner involvement in construction safety is increasing in r ecent years. Many findings obtained from the project interviews are supported by the contr actors. Also, contra ctors expressed their opinions on how they would like owners to be involved in safety management, that is, carefully selecting safe contractors to execu te their projects, setting safety objectives, facilitating contractors in sa fety during construction, and never setting double safety standards on the project. The suggestions offered by contractors can be helpful for owners in the future.

PAGE 183

167 LIST OF REFERENCES Agresti, A., and Finlay, B. (1997). Statistical Methods For The Social Sciences. Prentice Hall, Upper Saddle River, New Jersey. Applebaum, H. (1998). Construction Workers, U.S.A. Greenwood Press, Connecticut. American Society of Civil Engineers. (1988). Report Of The Task Committee On Construction Site Safety. ASCE, Reston, VA. Baggs, R.E. and Cunningham, J. (1988). “Safet y and Efficiency in Steel Construction – The Broadgate Experience.” Civil Engineering, May 1988. 35. Bureau of Labor Statistics (B LS). (2003). “Injuries, illnesses, and fatalities.” U.S. Department of Labor, Bureau of Labor St atistics. http://www .bls.gov/iif/home.htm. February, 2003. Business Roundtable. (1982). “Improving Constr uction Safety Performance: The User’s Role.” A Companion Publication to Construc tion Industry Cost Effectiveness Project Report A-3, New York. Census of Fatal Occupational Injuries (CFO I). (2003). Fatalities by detailed industry (all sectors of the construction industry). U.S. Department of Labor, Bureau of Labor Statistics. http://dat a.bls.gov/cgi-bin/surveymost. February, 2003. Construction Industry Institute. (1991). Managing Subcontractor Safety. The University of Texas at Austin. Publi cation 13-1, Austin, Texas. Construction Industry Institute. (2001). Construction Industry Institute Benchmarking And Metrics, 2001 Safety Report. The University of Texas at Austin, Austin, Texas. Coble, R. (1992). An Empirical Investigation Of Factors Related To The Drug-Free Workplace. Ph.D. Dissertation, Univ ersity of Florida, Ga inesville, Florida. Coble, R. J., and Blatter, R. L. Jr. (1999). “C oncerns with safety in design/build process.” Journal Of Architectural Engineering. 5(2), June, 1999. 44-48. Coble, R. J., and Hinze, J. (2000). Analysis Of The Magnitude Of Underpayment Of 1997 Construction Industry Workers' Compensa tion Premiums In The State Of Florida. Dec. 2000. (unpublished internal research report). University of Florida, Gainesville, Florida.

PAGE 184

168 Cooper, D. (1998). Improving Safety Culture. John Wiley & Sons, New York. Dagostino, F.R. (2002). Estimating In Building Construction, Sixth Edition. Prentice Hall, Upper Saddle River, New Jersey. Diaz, R. I. and Cambrera, D. D. (1997). “S afety climate and attitude as evaluation measures of organizational safety.” Accident Analysis And Preview. 29(5), 643650. Duff, A. R. (2000). “Behavior measurement for continuous improvement in construction safety and quality.” The Management Of Constr uction Safety And Health. Coble, R., Haupt, T., and Hinze J. (edit). Balkema, Rotterdam. 1-18. Engineering News Report, ( 2001). “The Top Owners.” ENR, Nov. 12, 2001. 34-48. Everett J. G., and Frank, P. B. (1996). “Costs of accidents and injuries to the construction industry.” Journal Of Construction Engineering And Management, 122(2), June 1996, 158-164. Fischer, M., and Tatum, C. B. (1997). “Char acteristics of design-relevant constructability knowledge.” Journal Of Construction Engineering And Management, 123(3), May/June 1997, 253–260. Gambatese, J. (1996). Addressing Construction Worker Safety In The Project Design. Ph.D. Dissertation, 1996, University of Washington, Seattle, WA. Gambatese, J. (1998). “Liability in desi gning for construction worker safety.” Journal Of Architectural Engineering, 4(3), Sep. 1998, 107-112. Gambatese, J. (2000a). “Saf ety in a designer’s hands.” Civil Engineering. 70(6), June, 2000, 56-59. Gambatese, J. (2000b). “Owner involveme nt in construction site safety.” Proceedings Of Construction Congress VI. February 20-22, 2000. Orlando, Florida. 661-669. Garza J. M., Hancher, D. E., and Decker, L. (1998). “Analysis of safety indicators in construction.” Journal Of Construction Engineering And Management. 124(4), July/August, 1998, 312-314. Geller, E.S. (2001). The Psychology Of Safety Handbook. Lewis Publishers, New York. Gerber, J.K., and Yacoubian Jr. G.S. ( 2001). “Evaluation of drug testing in the workplace: study of the construction industry.” Journal Of Construction Engineering And Management, 127(6), Nov./Dec. 2001, 438-444. Godfrey, K. A. (1988). “A fter L’Ambiance plaza.” Civil Engineering, 3(1), 32-41.

PAGE 185

169 Heinrich, H.W. (1959). Industrial Accident Prevention. McGraw-Hill Book Company, New York. Helander, M. (1981). Human Factors/Ergonomics For Building And Construction. John Wiley & Sons, New York. Hinze, J. (1997). Construction Safety. Prentice Hall, Upper Saddle River, New Jersey. Hinze, J. (2001). Construction Contracts (2nd edition). McGraw Hill, Boston, MA. Hinze, J. (2002). Safety Plus: Making Zero Accidents A Reality. CII Research Report 160-11. March 2002, University of Texas at Austin, Austin, Texas. Hinze, J. and Figone, L. (1988). Subcontractor Safety As Influenced By General Contractors On Small A nd Medium Sized Projects. CII report No. 2. 1988. University of Texas at Austin, Austin, Texas. Hinze, J., and Talley, D. (1988). Subcontractor Safety As Influenced By General Contractors On Large Projects. CII Research Report No. 1. August, 1988, University of Texas at Austin, Austin, Texas. Hinze, J., and Appelgate, L. (1991). “Costs of construction injuries.” Journal Of Construction Engineering And Management, 117(3), Sep. 1991, 537-550. Hinze, J., and Wiegand, F. (1992). “Role of designers in construction worker safety.” Journal Of Construction Engineering And Management, 118(4), Dec. 1992, 677684. Hinze, J., Bren, D., and Piepho, N. (1995). “E xperience modification ra ting as measure of safety performance.” Journal Of Construction En gineering And Management. 121(4), Nov./Dec. 1995, 455-458. Hinze, J. and Gambatese, J. (1996). Addressing Construction Worker Safety In The Project Design. CII Research Report 101-11, June 1996. University of Texas at Austin, Austin, Texas. Hinze, J and R. Godfrey. (2003). "An eval uation of safety performance measures for construction projects," Special Issu e On Construction Health & Safety, Journal Of Construction Research, 4(1), March, 2003. Hislop, R. D. (1999). Construction Site Safety: A Guide For Managing Contractors. Lewis Publishers, Boca Raton. Holt, A. J. (2001). Principles Of Construction Safety. Blackwell Science, Malden, MA. Jergeas, G., and Van der Put, J. (2001). “B enefits of constructa bility on construction projects.” Journal Of Construction Engineering And Management, 127(4), Jul./Aug. 2001, 280-290.

PAGE 186

170 Levitt, R. E., and Samelson, N. M. (1993). Construction Safety Management, 2nd edition. John Wiley & Sons, New York. Levitt, R. E., Samelson, N. M., and Mummy, G. (1981). “Improving construction safety performance: the user’s role.” Department Of Civil Engi neering Technical Report No. 260, Stanford University, Stanford, CA. Liska, R. W., Goodloe, D., and Sen, R. (1993). Zero Accident Techniques. Construction Industry Institute. Source Document 86, Ja n. 1993. University of Texas at Austin, Austin, Texas. MacCollum, D. V. (1995). Construction Safety Planning. Van Nostrand Reinhold, New York. McKee, B. (1994). “Architect s as construction mangers.” Architecture, 84(12), Dec. 1994, 111-115 Peterson, D. (1988). Safety Management, A Human Approach. Aloray, New York. Peyton, T.X., and Rubio, T.C. (1991). Construction Safety Practices And Principles. Van Nostrand Reinhold, New York. Rosner, B. (1995). Fundamentals Of Biostatistics. Duxbury Press, Belmont, CA. Samelson, N. M., and Levitt, R. E. (1982). “Owner’s guidelines for selecting safe constructors.” Journal of The Construction Division. 108(4), December, 1982. 617623. Sawacha, E., Naoum, S. and Fong, D. (1999). “Factors affecting safety performance on construction sites.” International Journal Of Project Management. 17(5), 309-315. Sikes, R. W., Qu, T., and Coble, R. J. (2000) “An owner looks at safety.” In Coble R., Hinze J., and Haupt T. (eds.), Construction Safety And Health Management, 2000. Prentice Hall, Upper Saddle River, New Jersey. 193-205. Spangerberg, S., Mikkelsen, K. L., Kines, P., Dyreborg, and J., Baarts, C. (2002). “The construction of the resund Link between Denmark and Sweden; the effect of a multi-faceted safety campaign”. Safety Science, 40 (2002), 457-465. Substance Abuse and Mental Health Se rvices Administration (SAMHSA). (1999). Worker Drug Use And Workplace Policie s And Programs: Results From The 1994 And 1997 National Household Survey On Drug Abuse, United States Department of Health and Human Se rvices, Rockville, MD. Sutherland, V., Makin, P., and Cox, C. (2000). The Management Of Safety: The Behavioral Approach To Changing Organizations. SAGE Publications, London.

PAGE 187

171 Suraji, A., Duff R.A., Peckitt, S.J. (2001). “D evelopment of causal model of construction accident causation.” Journal Of Construction Engineering And Management, 127(4), Jul/Aug. 2001, 337-344. Toole, T. M. (2002). “Construction site safety roles.” Journal Of Construction Engineering And Management, Vol. 128, No. 3, May/June 2002, 203-210. Weistein, M.B. (1997). Total Quality Safety Management And Auditing. Lewis Publishers, New York.

PAGE 188

172 BIOGRAPHICAL SKETCH Xinyu Huang was born on October 21, 1975, in Kunming, China. In September 1993, Xinyu was admitted to the Department of Civil Engineering at Tsinghua University in China, where he achieved both his bachel or’s (in 1998) and mast er’s (in 2001) degrees majoring in construction management. In the fall of 2001, Xinyu was admitted to the University of Florida to pursue his Ph.D. unde r the guidance of Professor Jimmie Hinze. He was admitted to doctoral candidacy in September, 2002, and has since been working on his dissertation as well as other avenues of research. Xinyu’s research interests in clude construction safety a nd project management. He has published several safety-related articles and conference papers. He has co-authored a textbook in China with his supervisors for master’s study and Ph.D. study. Xinyu Huang married Jie Bai in 2002.


Permanent Link: http://ufdc.ufl.edu/UFE0001376/00001

Material Information

Title: The Owner's Role in Construction Safety
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0001376:00001

Permanent Link: http://ufdc.ufl.edu/UFE0001376/00001

Material Information

Title: The Owner's Role in Construction Safety
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0001376:00001


This item has the following downloads:


Full Text












THE OWNER'S ROLE IN CONSTRUCTION SAFETY


By

XINYU HUANG














A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2003

































Copyright 2003

by

Xinyu Huang















This dissertation is dedicated to my wife, Jie Bai; my parents, Xiaoyi Huang and Guizhen
Zhang; my family and everyone engaged daily in the battle against the poor safety and
health performance of the construction industry.















ACKNOWLEDGMENTS

First I wish to thank my wife and my parents for their enduring support, trust, and

encouragement for my study and research. Their unquestioning belief in my ability to

complete my study was often the only inspiration and motivation I needed to keep me

from succumbing to frequent feelings of inadequacy and ineptitude. I am indebted to

them all for not demanding too much.

I wish to acknowledge my supervisor for PhD study, Professor Jimmie Hinze, for

his patience, generosity, encouragement, and guidance during my study at the University

of Florida. The invaluable assistance and guidance of him in the course of completing

this project were essential for me to complete the study. The academic and profession

spirit he demonstrates will inspire me in my future life and career.

I am also indebted to all the safety personnel I interviewed or surveyed. The study

could not have been completed without their enthusiasm for construction safety and their

experience in safety management.

I want to thank Professor Jo Hasell, who guided and helped me so warmly ever

since the first day I arrived at the University of Florida. I want to thank Drs. Dennis

Fukai, Charles Glagola, Leon Wetherington, and Kwaku Tenah, for their help with my

research. I want to thank Dr. Richard Coble for encouraging and recommending me to

study for my PhD at the University of Florida.









The statistical courses I took with Drs. Alan Agristi, Bhramar Mukherjee, and Ron

Marks at the Department of Statistics helped me to conduct research design and data

analysis. Their guidance is appreciated.

This research was supported primarily by funds provided by the National Institute

of Occupational Safety and Health (NIOSH). Support was also provided through the

cooperative efforts of the Construction Industry Institute (CII). This research could not

have been successfully undertaken without the assistance of both organizations. Their

contributions of funds and effort are gratefully acknowledged.

Finally, I owe an unquantifiable debt to many of my friends and teachers who have

encouraged and helped me to achieve my degree.
















TABLE OF CONTENTS

Page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TA BLE S .............................. ........... ..... .. ...... ....... ....... ix

LIST OF FIGURES ............................................................. ......... xiii

A B S T R A C T .......................................... ..................................................x v

1 INTRODUCTION ............... ................. ........... ................. ... ..... 1

2 LITER A TU RE REV IEW .................................................. ............................... 5

A c c id e n t F a cts ........................ ... ................................................ 5
Ow ners' M ore A ctive Involvem ent In Safety ........................................ ....................7
How Owners Take Their Role In Safety .................................. .................12
Establish And Communicate Attitudes Towards Safety ...................................14
Consider Safety In Contractor Selection ..........................................................14
Contractual Safety Arrangements.............................. ...... .................. 16
Address Safety During Design And Constructability Review ............................18
Participate In Safety During Construction .................................. ............... 21
Total Safety Culture And Behavior-Based Safety.............................................24
Sum m ary ...................................... ................. ................. .......... 29

3 RESEARCH M OTHODOLOGY ..................................... ........................... ......... 30

R research D esig n ..................................................... ................ 3 0
R research C om ponents .............................................................. .....................30
M ethod D design .....................................................................34
Sam ple Size D eterm nation ...................................................................... 35
S am p le S ele ctio n ........................................................................................... 3 7
D ata Collection ............................. ...................37
M ailed Survey Study (Phase 1) ...................................... ........................ 38
Project Interview (Phase 2): ..................................................................... ...39
Follow-Up Survey To Contractors (Phase 3) ............................................... 41
D ata A n aly sis ............................. ...................................................... ............... 4 1









4 QUESTIONNAIRE SURVEYS FOR LARGE OWNERS ........................................44

T he D ata Set .................... .......................44
Analysis Of The M ailed Survey Data.................................... ......................... 46
The Projects Context ................ ............... ............ .. .... .. ................. 46
Commitment Of The Owner To Construction Safety ......................................48
Selection O f Safe C contractors ........................................ ......... ............... 49
Safety Management Practices Of The Owners............................... ...............51
S u m m a ry ................................................... ................... ................ 5 7

5 RESULTS: OWNER'S ROLE IN PROJECT SAFETY...........................................59

T h e In terv iew s ......................................................... ................ 5 9
Projects Interview ed In The Study ........................................ ........................ 62
The M ost Com m on Safety Practices ........................................ ....... ............... 64
Project Descriptions And Safety Performance .........................................................66
Shutdow n Projects ............................................... .. ...... .. ............ 67
Public O r Private Project ..................... ........ .............. ............................... 67
Petrochemical Projects And Manufacturing Projects............... ...................68
Union Projects Or Open Shop Projects ...........................................................70
Type Of Contract .................. ............................ ........ ................. 71
Size O f The Projects ............................................ .. .. .... .. ........ .... 73
W ork Shift A nd W orkdays........................................................ ............... 74
Ow ner's Selection Of The Contractor .................................... ................................. 75
Preferred C contractors L ist ................................................. ............... ....75
Importance Of Safety During Selection Of Contractors ...................................76
Criteria Used To Evaluate Safety Performances Of Contractors ......................76
Owner's Contractual Arrangement.................................................... ................80
Owner Involvement During Project Execution .....................................................84
Safety Program Of The Contractor............................................. ...............85
Owner Monitors Near Misses On The Project.................................................87
Accident/Incident Investigations ...................................................................... 88
Safety R recognition Program .......................................... ........... ............... 89
Safety Education And Training .. ........................... ......................... ............. 91
Responsibilities Of Owner's Site Representative..................... ..............93
Setting Zero-Injury Objectives ...................................................................... 95
A additional Findings .......................................... ... .... ........ ......... 95
Safety R recognition Program .......................................... ........... ............... 95
Safety Observation Program.................... .......... ............... 96
Collaborative Efforts On Safety With Contractors And Designers...................98
Answers To Open-Ended Questions ...................................... ............... 101
Industrial Versus Commercial Construction (A Case Summary).............................102
O w ner Safety M odel B building ....................................................... ..................... 104
Quantification Of The Index ........................................... ...............105
Multi-Linear Regression Between TRIR And The Indices ..............................110
Explanation And Discussion Of The Regression Results .............................115
Safety Performance Model: How To Achieve Better Safety Performance....... 117









Follow-up Survey: Contractors' Responses .............................. ......... ...............121
Su m m ary ...................................... ....................................................122

6 HOW CAN OWNER ACHIEVE PROJECT SAFETY .................................. 124

C ase 1: O w n er R ................................................................................. 124
C ase 2 : O w n er S ...............................................................130
Project D description ............................................................ .......... 130
Analysis And Diagnosis ...... .... ...................... ..... ........... .... 131
Summary ............. ...................... ............. 135

7 CONCLUSIONS, DISCUSSIONS AND RECOMMENDATIONS .....................136

C o n c lu sio n s......................................................................................................... 1 3 6
Discussions ............ .......... ................... 138
R ecom m endations....... ........ .............................. .. .... ........ ....... 140

APPENDIX

A OWNER SURVEY QUESTIONNAIRE ...................................... ............... 145

B OWNER INTERVIEW QUESTIONNAIRE ................................ .....................149

C CONTRACTOR SURVEY QUESTIONNAIRE.....................................................155

D ANALYSIS OF THE CONTRACTOR SURVEY........................................156

The Contractors Interview ed ............................................................................. 156
Project Content And Safety Performance........................... ....................157
The Owner's Involvement In Project Safety Management.................................... 159
What Percentage Of The Safety Responsibility Is Assumed By The Owner.... 159
Level Of Importance Owners Place On Safety When Awarding Contracts .....161
Does The Owner Provide Sufficient Support For Project Safety ......................161
How Can Owners Facilitate Contractors In Project Safety Management ..............162
In What Area Can The Owner Do The Most To Promote Safety ...................162
What Is The Best Way That Owners Can Address Their Concern For Safety .163
To What Extent Do You Prefer Owners To Be Involved In Project Safety......165
Sum m ary .................................... ........................... ........ .......... 166

L IST O F R E FE R E N C E S ........................................................................ ................... 167

BIOGRAPHICAL SKETCH ............................................................. ............... 172
















LIST OF TABLES


Table page

2-1. Average cost of construction site injuries........................................ .....................10

4-1. Importance of safety during selection of contractors to owners of different types of
projects (1 to 10 with 10 being the highest possible rating)................. ........ 47

4-2. Importance of safety when selecting contractors to owners with differing
construction budgets............. ............................................................. ... .... ....... 47

4-3. Earliest stage to emphasize safety vs. annual construction budget ............................49

4-4. Earliest stage to emphasize safety vs. major type of projects .................................49

4-5. Number of safety criteria vs. annual construction budget.......................................50

4-6. Number of safety criteria vs. major type of projects ............................................50

4-7. Results of Mann-Whitney test: safety selection criterion vs. importance of safety in
the selection of contractors......................................................... .............. 50

4-8. Number of contractual safety requirements vs. major type of projects................... 51

4-9. Number of contractual safety requirements vs. annual construction budget..............52

4-10. Total number of safety practices implemented vs. maj or type of projects............... 52

4-11. Total number of safety practices implemented vs. annual construction budget.......53

4-12. Importance of safety vs. safety approaches taken ......................................... 53

4-13. Non-Parametric correlation matrix between importance of safety, number of
contractual requirements and safety involvement approaches..............................54

4-14. Mann-Whitney test: Importance of safety vs. safety training inputs......................55

4-15. Total amount of safety training input vs. project type............................................55

4-16. Total amount of safety input vs. annual construction budget................................55









4-17. Relationship between the importance of safety and whether or not safety is
addressed in the design phase........................................................ ............... 56

4-18. Priority of different approaches preferred by owners (1 to 5 with 1 being the highest
possible priority) ......................................................................56

4-19. Comparison of the ranks of selecting safe contractor (Q232) v.s. participating in
and m monitoring safety (Q233) ............................................................................ 57

5-1. Safety performances and sizes of the projects included in the research...................63

5-2. Injury rates of shutdown projects and all other projects...........................................67

5-3. Type of facility owner and safety performance (excludes shutdown projects)..........68

5-4. Type of project facility (including shutdown projects) and injury rates ..................69

5-5. Type of project facility (excluding shutdown projects) and injury rates ..................69

5-6. Type of labor and safety performance (US projects only) ......................................70

5-7. Type of contract used on the project ...... .......................... ...................73

5-8. Worker hours expended (in thousands of hours) and safety performance ...............74

5-9. Number of shifts worked and safety performance............... ............ ..............75

5-10. Number of days worked per week and safety performance ...................................75

5-11. Emphasis placed on safety in the overall review of contractors (rating from 1 to 7
w ith 7 being the m ost im portant) ........................................ ......................... 76

5-12. Is the EMR used to evaluate safety performance of contractors? (No significant
differences) ............................................................... ... .... ......... 77

5-13. Is the TRIR used to evaluate safety performance of contractors?.........................78

5-14. Threshold value of TRIR set for contractor safety performance..................................78

5-15. Are qualifications of safety staffs reviewed when evaluating contractors? ............79

5-16. Are qualifications of the project team reviewed when evaluating contractors? .......79

5-17. Number of proactive criteria utilized for evaluating contractors on safety .............80

5-18. Contract requires the contractor to place at least one full-time safety representative
o n site ? ............................................................................8 2









5-19. Contract requires the contractor to submit the resumes of key safety personnel for the
owner's approval? ............... ...... .............. .............. 83

5-20. Relationship between TRIR and total number of contractual safety requirements..83

5-21. Number of leading indicator safety requirements included in the contract..............84

5-22. Owner requires the emergency plan (medical and hazardous materials) to be included
in the contractor's safety program ? ........................................ .......................... 85

5-23. Owner requires the daily JSA (job safety analysis) to be included in the contractor's
safety program ? ........................................................................86

5-24. Owner requires a substance abuse program to be included in the contractor's safety
program ? ............................................................... ..... ..... ......... 86

5-25. Does the owner's representative monitor near misses on the project? .....................88

5-26. Does the owner maintain injury statistics by contractor? ........................................89

5-27. Are the contractor's safety performance statistics included in the owner's safety
perform ance statistics? ........................................ ................. ......... 89

5-28. Are some funds provided to the contractor, above and beyond the contract amount,
to promote project safety? .......................... ..........................90

5-29. The owner participates in the safety recognition program ....................................90

5-30. Safety training methods used on the project................................. ............... 92

5-31. Hours of monthly refresher safety training received by the workers ..................92

5-32. Is there any means of verifying the comprehension of safety orientation? ..............93

5-33. The owner participates in safety meetings and toolbox meetings............................93

5-34. Total number of key activities performed by the owner's safety representatives .... 94

5-35. The owner's representative monitors project safety inspection records on a regular
b a sis ............................................................................ .9 4

5-36. The owner's representative monitors the project near miss rate and project
inspection records on a regular basis ............................ ........ ............. .................. 95

5-37. Is zero TRIR set as a safety objective by the owner before project commencement?95

5-38. Is there a safety recognition program on the project? ................... ........ ........95

5-39. Is safety dinner held on the project? ................. ........................ ..................... 96









5-40. Are safety observers used on the project? ...........................................................97

5-41. Does the contractor report near misses to the owner? ...........................................98

5-42. Comparison between number of OSHA recordable injuries recorded on the project
with number of near misses recorded.................................................................... 98

5-43. How does the owner rate the prime contractor's commitment to safety? (rating from
1 to 7, w ith 7 being the best) .............................................................................. 99

5-44. How would the owner rate the cooperation and communication between the owner
and contractor? (rating from 1 to 7, with 7 being the best)................................99

5-45. D distribution of the indices .......................................................................... ..... 108

5-46. Relationship between TRIR and the project index (P) .................................. 108

5-47. Relationship between TRIR and the selection index (S) ............. ... ..................108

5-48. Relationship between TRIR and the contract index (C) .............. ... ...............109

5-49. Relationship between TRIR and the owner's index (0)............... ... .................109

5-50. ANOVA for between-subject effects ...........................................................110

5-51. Spearman's Correlation between TRIR and the index................. ...............111

5-52. Model Summary of the multi-variable regression................................................. 112

5-53. ANOVA of the multi-variable regression .........................................................112

5-54. Coefficients of the multi-variable regression ........................................................ 112

5-55. The safety performances of projects with highest and lowest scores..................... 121

6-1. Comparison of injury rate and scorecard evaluation................................................ 125

6-2. Comparison between TRIR and the scores of the scorecard ................. ........... 125

7-1. Comparison of the three levels of safety management methods ............................143

7-2. Comparison of findings in different studies .................................. ....... ..........144

A -1. Characteristics of the respondents ...... ........ ............................... ..................156

A-2. What percent of safety responsibility is assumed by the owner.............................159

A-3. Level of importance of the factors when owners award contracts ..........................161
















LIST OF FIGURES


Figure pge

1-1. The organizational chart of a typical general contract agreement.............................2

2-1. Injury rate of construction and all private industry (data source: www.bls.gov) .........5

2-2. Fatalities in construction industry and all industries .........................................6

2-3. Causes of construction fall accidents investigated by OSHA (01/90-10/01) ..............6

2-4. Breakdown of OSHA investigated accidents in construction (01/90-10/01) ............7

2-5. Construction Accident Causation Model: Constraints and Responses.....................9

2-6. Relationship between Design Effort and Project Cost.....................................19

2-7. Domino accident causation model suggested by Heinrich (1959) ...........................25

5-1. Locations of the projects included in the research .............. ...................... 63

5-2. Injury pyram id for the projects............................................................ ............... 64

5-3. Other leading indicators used in the project contract (not statistically significant) ...83

5-4. Corporate history of safety performance.................................................103

5-5. Histogram of the regression standardized residual.............................. ...............13

5-6. P-P plot of the regression............................ ................................................... 114

5-7. The predicted value vs. standardized residual plot.............................. ...............114

5-8. Project safety model: owner's impact ....................... ........ ...............119

5-9. Owner's influence on construction safety scorecard .............................................120

6-1. Owner's influence on construction safety scorecard .............................................133

7-1. Owner's involvement in Safety m management ...................................... .................142

7-2. Focus of different levels of safety management methods ....................................... 143









A-1. What type of project has the best safety performance? (type of contract) ............157

A-2. What type of project has the best safety performance? (union or open shop).........158

A-3. What type of project has the best safety performance? (type of project) ..............158

A-4. Change of owner involvement in project construction safety in the past decade....160

A-5. Does the owner provide sufficient support for safety on the projects.....................162















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE OWNER'S ROLE IN CONSTRUCTION SAFETY


By

Xinyu Huang

December 2003

Chair: Jimmie Hinze
Major Department: Building Construction

The construction industry has long been regarded as one of the most dangerous

industries. The construction industry has a history of poor safety performance. This is

despite the fact that dramatic improvements have been made in the safety performance of

the construction industry in the past decade. The improvements are due, in part, to the

concerted efforts of owners, contractors, subcontractors, and designers. While past

studies have investigated the safety roles of contractors, subcontractors, and designers,

the owner's impact on construction safety has not been thoroughly studied.

The owners of projects are the primary consumers of construction services, the

sources of project finances, and, in many cases, the end users of the facilities. They are

often at the pivotal position of the projects. Traditionally, owners have not been directly

involved in construction safety, often in order to avoid economic losses and legal

entanglements resulting from injuries. With the increase of accident costs and legal cases









involving owners as the third-party defendants, owners have come to realize the

importance of safety.

In this study, the owner's role in construction safety was investigated. The

relationship between project safety performance and the owner's influence was examined,

with particular focus on the project context, selection of safe contractors, contractual

safety requirements, and the owner's proactive involvement in safety management. A

questionnaire survey sent to owners with large construction budgets and interviews

conducted on projects with large worker hour exposures were the primary means of data

collection for the study. Statistical analysis shows that there is a strong relationship

between project safety performance and owner involvement. Practices of owners

significantly associated with project safety performances were identified. A model to

demonstrate how owners can help improve safety was established. A scorecard that can

be used to assess the owner's involvement in safety was developed and tested.

It was concluded that the owner's involvement can significantly influence project

safety performance. Owners can achieve better project safety performances by setting

safety objectives, selecting safe contractors, and participating in safety management

during construction.














CHAPTER 1
INTRODUCTION

With peak levels of employment reaching eight million workers, the United States

construction industry has the dubious distinction of perpetually being one of the

industries with the worst injury and fatality records. The mining industry and the

agriculture industry are the only industries that have worse records than the construction

industry. While the actual incidence of construction worker injuries has declined over the

past three decades, the number of injuries and fatalities is still at an unacceptable level

(Hinze, 1997). According to the Census of Fatal Occupational Injuries (CFOI, 2003), an

average of 1,115 workers were killed annually on construction sites from 1995 to 2000,

which accounted for nearly 20% of all industrial fatalities in the United States. This

number is a disproportionate distribution of construction worker fatalities since

construction workers account for only 7% of the industrial workforce. While many

strides have been made to reduce the incidence of injuries, there is considerable room for

improvement. The industry is trying to find new ways to improve safety performance.

In the construction industry, most construction projects involve the participation of

owners, designers, and contractors. The owners (also called facility clients, or project

buyers) of projects are the primary consumers of construction services, the sources of

project finances, and, in some cases, the end users of the facilities. Owners include both

public and private entities. Some owners obtain the services of a construction firm to

build a facility with virtually no interaction between the owner and the construction firm

during the construction process. Other owners play varying roles during the construction









phase, with some being closely involved with the construction effort during every stage

of project execution.

Generally, an owner of a project will require various objectives to be satisfied by

the designer and contractor (Hinze, 2001). The owner defines the scope of a project to

become a reality, which requires the services of several parties. First, the designer will

develop the guidance document for building the project. The contractor will then follow

by executing the construction of the project.

Typical objectives required by the owners include, but are not limited to the

following: when a particular project should be finished; what quality requirements the

project must satisfy; the owner's cost of the project; and possibly the safety standards that

must be met during project construction. The roles of some of the parties are shown in

the organizational chart for a general contract agreement in Figure 1-1. In all types of

contract agreements, the owner is at the top position of the project organization.

Essentially, the owner has the overall authority on the project.


Figure 1-1. The organizational chart of a typical general contract agreement









Traditionally, safety responsibilities have rested solely on the shoulders of

contractors. Owners and designers have held the view that they should not get involved

in construction worker safety for fear of incurring increased liability exposure. However,

in the past decades more and more parties to the construction process have come to

realize that zero accidents is an attainable objective, but only through the concerted

efforts of all parties involved in the construction process. Owners, A/E firms (also called

designers), and contractors/subcontractors have different roles in preventing accidents to

achieve an injury-free worksite (Hinze, 1997; Gambatese, 1996; Toole, 2002). The

contractor is undoubtedly the pivotal party to control jobsite safety. Strategies and

approaches taken by contractors to improve project safety have been thoroughly

investigated in past research studies (Levitt and Samelson, 1993; Hinze, 1997; Hinze,

2002). Subcontractor safety influenced by the general contractor in various sizes of

projects was investigated by Hinze and Figone (1988, for small and medium projects),

and Hinze and Talley (1988, for large projects). Designers can reduce safety hazards in

the working environment by considering worker safety issues in their design decisions.

Hinze and Gambatese (1996) gathered various "best practices" for designers to address

safety issues in their designs and developed a safety design tool to help designers

eliminate hazards when making decisions. Today, more and more owners have come to

realize that the costs of construction accidents are ultimately their own financial burden,

and that they cannot with certainty shield themselves from the legal liabilities associated

with worker injuries. As a result, many owners are taking more active roles in

construction safety. This begins with the selection of safer construction firms and can

also be observed in the safety provisions included in the construction contract. Owners









will continue to influence safety performance by the nature of their involvement during

the construction process.

Despite all the research that has been conducted with the various parties to the

construction process, little prior work has examined the specific role played by project

owners in promoting safety performance. How owners regard their role in safety and

how they attempt to exert their influence in the construction process have not been

previously examined in detail. This study is focused on the investigation of the impact of

owners on construction safety.


















CHAPTER 2
LITERATURE REVIEW




Accident Facts

Accident data prepared by the Bureau of Labor Statistics (2003) show that the


construction industry has performed much worse than the average of all industries (see


Figure 2-1). Although the statistics have improved dramatically in the 1990's, accident


rates in the construction industry are still 50% higher than that of all industries, lagging


all industries by about 10 years. An alarming fact is that the number of fatalities in the


construction industry has increased in the past decade (see Figure 2-2).




-*-Lost Workday Injury Rate of construction industry -*-Lost Workday Injury Rate of all private industry
-A-OSHA Recordable Injury Rate of construction industry ----OSHA Recordable Injury Rate of all private industry
18

15816
16 14 9152 147154 15 154 1


E 1 \128129
S14 2812 9

) 12
120 115
00 4
10 6 9 9 92 92 9-A 9-6
8 8
85 83 82 83 83 84
8 8 79 79 77 8


59 58
5 63 65 63 6 62 Do 6 -6
S 54 5454
44 42 39,44 44 9 9
3* 7 8 39 3 3 36 3 36 3 4 4 41 4 39
2 1 31 3434 8 28 2b

0 1
1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001
Year

Figure 2-1. Injury rate of construction and all private industry












SFatalities in all industries -- Fatalities in construction industry

7000

7 6000 6632
S- 6217 6331 6275 6202 6238 6055 6054
5000
4000
4-



z 1000 -

0 Q3 Q71 1077 1083 1086 1123 1198 1210 1159
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Year


Figure 2-2. Fatalities in construction industry and all industries

Figure 2-3 shows the major types of construction accidents that have occurred in


the past decade, based on the analysis of all serious injuries and fatalities investigated by


OSHA. Falls, struck-bys, electric shocks, caught-in-or-betweens, and other are the major


causes of injuries, although the relative proportion of each type has changed somewhat in


recent years (see Figure 2-4).



Others
5.7%
Shock
13.4% Caught in or /Cardio-
Betw een Vascular/Respirator
11.6% y System Failure
4.2%
Struck By Struck Against
24.3% 2.3%
Inhalation
Other 1.6%
145% Rubbed/Abraded
0.2%
Absorption
0.2%

Bite/Sting/Scratch
0.1%
Falls Repeated
36.3% NMotion/Ressure
Ingestion 0.1%
0.1%


Figure 2-3. Causes of construction fall accidents investigated by OSHA (01/90-10/01)










0 Falls U Struck By E Shock O Caught in or Between 0 Other

100%


80% -


60% -


40% -


20%


0% -
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

Figure 2-4. Breakdown of OSHA investigated accidents in construction (01/90-10/01)

Owners' More Active Involvement In Safety

With a higher proportion of injuries, the construction industry has long been

regarded as a dangerous industry. With an average employment of approximately 7% of

the industrial workforce, the construction industry has generally accounted for nearly

20% of all industrial worker fatalities (CFOI, 2003). The research conducted by Everett

and Frank (1996) concluded that the total costs of construction accidents accounted for

7.9-15.0% of the total costs of new, non-residential projects. A more recent but

unpublished research by Coble and Hinze (2000) showed that the average worker's

compensation insurance costs could be conservatively estimated as constituting 3.5% of

the total project cost.

Safety performance of the construction industry did improve during the 1990's (see

Figure 2-1). As suggested by Hislop (1999), perhaps the most significant factor

attributed to this improvement is the increased management commitment to safety.

Management has been motivated to make a greater commitment to safety, based on its









increased awareness of the impact of the high costs of worker's compensation payments,

the higher dollar value settlements in lawsuits, the increased amounts of OSHA fines for

safety violations, and the adverse impacts of poor safety performance on the corporate

image.

In order to reduce and eventually eliminate construction accidents, researchers have

explored techniques to fulfill the "zero injury objective" through the concerted efforts of

all parties involved. The involvement of owners has been regarded as an essential

requirement for this objective. For example, in the research conducted by Liska et al.

(1993), it was found that an important prerequisite attributed to excellent safety

performance was the involvement of the owner not only in pre-project planning including

financially supporting the contractor's safety program, but also in the day-to-day project

safety activities. In the construction accident causation model developed by Suraji et al.

(2001, see Figure 2-5), construction accidents are caused by different responses to certain

constraints and the environment on construction projects. In the model, owner (client)

responses are the actions (or inactions) of the owner in response to constraints during the

development of a project brief (Duff, 2000). These include, for example, reducing the

project budget, adding new project criteria, changing project objectives, and accelerating

the design or construction efforts of the project. All these elements will play an essential

role in accident occurrences.










L!tL% L L riL I -I nr ij c


I I *n ,I1 A RL 'R
4 .


Figure 2-5. Construction Accident Causation Model: Constraints and Responses

In the past, there was a reluctance of owners to become involved in matters related

to construction safety issues for the fear of added liability exposure (Sikes et al., 2000).

However, since the 1980's more and more owners, especially owners with larger

construction budgets, have voluntarily expanded their role in ensuring worker safety. A

series of studies conducted at the University of Washington in the early 1990's









demonstrated that owners' concern for construction safety was increasing (Hinze, 1997).

The major reasons include the following:

S The rising costs of health care and workers' compensation are not being ignored by
owners (see Table 2-1).

Table 2-1. Average cost of construction site injuries
S. Job Costs Estimated Total Cost to
Type of iury Direct Indirect Liability Costs Employer
Medical Only $520 $440 $240 $1,200
Lost Work Day $6,900 $1,600 $16,500 $25,000
(Hinze and Appelgate, 1991)

* Owners realize that the costs of injuries are ultimately reflected in the cost of
construction (Gambatese, 2000b).

* Litigation involving third parties has escalated in the past three decades. For
example, in the case of Phillips v. United Engineers & Constructors, Inc., and
Plasteel Products Corp., 500 N.E. 2d 1265 (1986), owner was sued but was not held
responsible for a worker's fall from catwalk during steel erection. In another case
Rigatti v. Reddy, 723 A.2d 1283 (1999), the owner was similarly cleared of being
responsible for a roofer's fall. In the case of Stark v. Rotterdam Square, 603
N.Y.S.2d 347 (1993), the owner of a mall was held liable for injuries suffered by a
roofer when he fell through a hole cut into the roof.

Because of these types of lawsuits, many owners have come to realize that reducing

the frequency and severity of construction injuries is the only sure way of reducing their

potential liability for worker injuries (Levitt and Samelson, 1993).

Since the early 1980s, several efforts have been undertaken to formally require

owners to participate in construction site safety. A major effort to expand safety

legislation resulted after 28 workers died in the tragic collapse of the L'Ambiance Plaza

Building in Bridgeport, Connecticut (Godfrey, 1988). This accident became the driving

force behind U.S. Senate Bill 2581 to amend the Occupational Safety and Health Act to

"require all construction projects to be supervised by ... a professional engineer-architect

designated by the owner and registered in the state where the construction is to be









performed" (ASCE, 1988). Opposition to this bill from a large segment of the

construction industry ultimately resulted in its failure.

Although these legislation efforts failed in the U.S., the liability of owners in the

United Kingdom was specified by the Construction (Design and Management)

Regulations 1994 (CDM), in which the owner's main duties are contained in Regulations

6, 10, 11 and 12 (Holt, 2001). In CDM, a owner is defined as "any person for whom a

project is carried out, whether carried out by another person or in-house." CDM imposed

criminal liability onto the owner who ignores construction safety and where this results in

an accident. Each project owner has a responsibility for safety on the project. In cases

where there are multiple owners, the owners can appoint an agent or another owner to

carry out the owner's duties, and then have to make a declaration to the enforcing

authority (the Health and Safety Executive) that the transfer of duties has been made.

Under CDM, the owner must do the following:

* Appoint a Planning Supervisor and a Principal Contractor for each project, being
satisfied that these "duty holders" are competent and have the resources to perform
their duties adequately.

* Not permit the construction work to start unless a health and safety plan, which
complies with the safety regulations, is in place for that project.

* Provide the planning supervisor with information about the state or condition of the
premises where the work is to be carried out. This is information which is relevant,
and which the owner either has or could get after making reasonable inquiries.

* Verify that any designer or contractor that is appointed directly is competent for the
task and has allocated sufficient resources to it.

* Make the health and safety file available for inspection by anyone who may need
information to comply with legal requirements. The owner will sell or pass on the
file to a future owner or a person acquiring the interest in the property of the
structure to which it refers.









The trend is clear that owners are getting more concerned about construction safety,

but many owners are confronted with the problem of how to effectively influence project

safety. For example, a safety campaign supported by the owner was conducted during

the construction of the Oresund Link between Denmark and Sweden. Primarily, the

safety campaign consisted of a comprehensive information campaign aimed at promoting

positive attitudes towards safety among the employees, and a specific campaign aimed at

the behavioral aspects to increase the level of safety when performing routine work at the

construction site (Spangerberg, 2002). In the end, the results of the campaign did not

satisfy the owner's expectations of low injury rate, which led people to question what

techniques by the owner can effectively improve construction safety.

Therefore, it would be essential to investigate the current techniques employed by

owners to achieve good safety performances on construction sites, and if possible

establish the effect the techniques on the resultant safety performance.

How Owners Take Their Role In Safety

Owners can actively impact construction safety by selecting safe contractors,

addressing safety issues in design, and participating in safety management during

construction (Hinze, 1997). To the extent possible, the owners, through their project

representatives, should participate with the contractors in all project safety activities,

including but not limited to new employee orientation, safety meetings, audits and

accident investigations, training, incentive programs and other safety related programs

(Gambatese, 2000b).

One of the earlier studies on the owner's role in construction safety by Levitt et al.

(1981) reached the conclusion that construction owners who selected or prequalified

contractors based on safety performance, and/or who got involved in construction safety









management, had fewer accidents on their projects. The owner's involvement in

construction safety management not only reduces the number and severity of accidents

but probably also reduces the owner's total liability exposure.

The American Society of Civil Engineers (ASCE) moved to the forefront in the

trend to involve owners in safety following the issuance of ASCE's Policy Statement 350

on construction site safety in 1998. The statement outlines ASCE's belief that

"improving construction site safety requires attention and commitment from all parties

involved." The policy states that safety should be addressed "for each project on a

project specific basis," and that owners should "take an active role in project safety."

Various ways owners can actively address safety were given in the policy, including the

following:

* Assigning overall project safety responsibility and authority to a specific
organization or individual (or specifically retaining that responsibility) that is
qualified in construction safety principles, rules, and practices appropriate for the
particular project.

* Including prior safety performance as a criterion for contractor selection.

* Designating an individual or organization to monitor safety performance during
construction.

* Designating in the contract documents those parties responsible for the final
approval of shop drawings and details.

Gambatese (2000b) summarized various ways in which owners can actively

address safety, including the following:

* Establish a clear position on safety.
* Ensure that safety is addressed in project planning and design.
* Consider safety performance when selecting a contractor.
* Address safety in the construction contract.
* Assign safety responsibility during construction.
* Participate in project safety during construction.









Establish And Communicate Attitudes Towards Safety

All owners have a legal and moral responsibility to insist on the safe performance

of their construction contractors and to use reasonable care to prevent contractors from

injuring others on the site. Owners should understand that their involvement in

construction safety management not only reduces the number and severity of injuries, but

is also likely to reduce the owners' total liability. In addition, the owner should realize

that safe performance is generally related to lower costs, better quality work, improved

productivity, adherence to schedule, reduced exposure to bad publicity, and minimal

disruption of the work being performed. As Hinze (1997) stated, owners whose

objectives are to avoid injuries are likely to have a proactive and more direct involvement

in construction operations. Therefore, the first step of the owner's involvement in safety

is commitment, especially management commitment, to create an injury-free project.

Each owner in the survey conducted by Levitt and Samelson (1993) stated the belief that

active involvement in construction safety, done properly, served to reduce rather than

increase the firm's potential liability for construction accidents.

The owner's position on safety should be clearly communicated to the project team

at the beginning of the project and to all team members joining the project during the

construction phase (Gambatese, 2000b). During project development, owners should

convey their commitment to safety to the contractors through various means, including

devising incentives for safety and by implementing sanctions for failing to support the

owners' safety initiatives.

Consider Safety In Contractor Selection

Traditionally, selecting safe contractors was recognized by many owners as the

most effective way to guarantee safety performance on their projects. Criteria for









selecting safe contractors have been extensively investigated in the past. Levitt and

Samelson (1993) stated that screening contractors in terms of their expected safety

performance is an easy and effective way for construction owners to enhance

construction safety. Levitt and Samelson (1981) suggested that the criteria must be (1)

predictive of safe project performance; (2) equally applicable to different construction

firms; and (3) objective. Possible selection measures were EMRs, OSHA recordable

injury rates, and management safety accountability. Findings from the research that are

indicative of safety accountability include the following:

1. Who in the organization receives and reviews accident reports, and what is the
frequency of distribution of these reports? (Frequent reports, detailed by operating
units, and reviewed by the president of the firm would be an indication of high
company accountability for safety.)

2. Frequency of safety meetings for field supervisors.

3. Compilation of accident records by foremen and superintendents and the frequency
of reporting. (Contractors who categorize their accident statistics by superintendent
and foremen, rather than by larger units, have a more detailed accountability
system.)

4. Frequency of project safety inspections and the degree to which they involve
project managers and field superintendents.

5. Use of an accident cost system allocating or assigning responsibility to individual
foremen and superintendents, as well as project managers.

Certain safety criteria during the selection of contractors suggested by Hinze

(1997), and Hinze and Godfrey (2003) include the following:

* Injury incident rates (including lost workday injury rates, OSHA recordable injury
rates, first aid injury rates, and etc.)
* Job site safety inspections
* Behavior based worker observations
* Experience modification rates (EMR)
* Loss ratios of workers' compensation
* Records of OSHA citations and fines
* Litigation related to injuries
* Performance records of key personnel









* Project safety plans
* Contractor qualification safety surveys
* Worker's conception surveys

Contractual Safety Arrangements

Owners must also make sure that contractors recognize their contractual

responsibility to perform safely (Levitt and Samelson, 1993). Contractually, most owners

have the contractor indemnify them from any losses or liabilities resulting from injuries,

but it is also essential to include specific and thorough safety requirements in the

construction contract. The requirements might include the following(Gambatese, 2000b):

* The requirement that the contractor abide by all applicable safety laws and
regulations;

* A delineation of the responsibility for safety on the jobsite;

* The submission of a written contractor safety program before work begins;

* A requirement for implementing a substance abuse program; and

* The submission of an emergency plan and accident reporting procedure.

Questionnaires sent to owners and contractors (Business Roundtable, 1982) were

used by Stanford University to identify safety requirements owners placed on

construction contractors and specific practices of owners to emphasize safety with

contractors. The responses are arranged in decreasing order of use by the respondents.

Require use of a system of permits for potentially hazardous activities.

1. Require the contractor to designate a responsible supervisor for safety coordination
on the job site.

2. Provide the contractor with safety guidelines that must be followed.

3. Discuss safety at owner-contractor meetings.

4. Periodically discuss safety audits of the contractor operations.

5. Require immediate reporting of contractor accidents.









6. Stress safety as part of the contract during pre-bid walk-arounds.

7. Investigate contractor accidents.

8. Maintain statistics of contractor accidents.

9. Conduct periodic safety inspections.

10. Set goals for construction safety.

11. Consider safety in prequalifying contractors for bidding on projects.

12. Set up a construction safety department to monitor contractor safety.

13. Set safety guidelines in the body of the contract.

14. Be involved in orientation sessions alerting workers to safety hazards on the job.

An owner, to be actively involved in construction safety, might consider several

contractual issues. Many of the issues relate to the safety obligations placed on the

contractor. Hinze (1997) suggested that contract provisions may include the following

requirements:

* Submittal of a project-specific safety plan
* Job hazard analysis
* Regular safety meetings with supervisory personnel
* A designated project safety coordinator
* Mandatory reporting of accidents, safety inspections, and safety meetings
* Inclusion of subcontractors in the safety program
* Compliance with the owner's safety guidelines
* Establishment of an effective worker orientation program
As the major binding document between the owner and contractor, the contractual

arrangement (lump-sum or cost reimbursable) will also impact how safety will be

addressed on the project. One reason that the contract type should be considered when

addressing project safety is that the contract essentially determines how the owner will

make payment to contractors to compensate them for their services on the project, and

therefore, can impact the relationship between owner and contractor (Dagostino, 2002).

As safety can be obtained only through the concerted efforts of different parties (owner,









contractor, and designer), the nature of their relationships can be expected to impact

safety. For example, if the contractor has a close, long-term relationship with the owner,

they will generally use cost reimbursable contracts (cost plus). Thus, safety investments

will be reimbursed by the owner to support the contractor's efforts. This arrangement

encourages the contractor to spend the necessary funds on safety.

Address Safety During Design And Constructability Review

Design decisions can dictate the structure, elements, materials and even the

construction methods to be used on the project. For example, the selection of forming

systems, cycle times, sequences, equipment, and design of temporary structures are

significantly impacted by the design of the permanent construction. Therefore, designers

are very influential, perhaps unknowingly, in impacting the way safety issues will be

addressed during construction. Consideration of construction worker safety by designers

(hereinafter also called A/E) can lead to a reduction in injuries and associated costs

during the construction phase (Hinze and Wiegand, 1992; Gambatese, 1998; Gambatese

2000a). Toole (2002) stressed that the A/Es are in the best position to implement specific

safety design recommendations, thereby preventing less safe conditions on the site. Also,

A/Es may be best able to identify questionable structural situations such as temporary

loadings on the permanent structure or temporary work platforms, provided they are

explicitly requested to do so and possess all of the data necessary to perform the analysis.

One reason why the Corps of Engineers has had particularly low injury rates on their

construction projects is that it coordinates the design with erection procedures and

coordinates their efforts on safety (MacCollum, 1995). Although the involvement of

design professionals in construction site safety has been minimal to nonexistent on many

projects, when they are involved their influence can be significant, and can help reduce










accident occurrence. This influence has favorable cost implications for the projects

(Gambatese, 2000a. see Figure 2-6).



High Additional costs: Additional costs:
Delays, design changes, Plan review, drafting, plan
CM efforts, etc. coordination, etc.


Project
costs

Optimal _____
costs

Low

Low Optimal High
Extent of Design
Efforts
Figure 2-6. Relationship between Design Effort and Project Cost

In the past, design professionals typically would distance themselves from the

responsibility for the safety of the construction workers. The reasons include their lack

of safety education and training, the lack of safety design tools, their restricted role on the

project team, and an attempt to limit their liability exposure (Gambatese, 1996).

However, with the unacceptable injury and fatality rates, the construction industry has

concluded, "The need for changes in attitude does not stop at educating erectors to work

more safely. It has to go back to the architect and engineer who should not only ask

themselves if it can be built, but can it be built safely." (Baggs, and Cunningham, 1988)

In the UK, the Construction (Design and Management) Regulations 1994 place duties on

designers and owners to make assessments of the impact of the proposed design on the

life cycle safety of the facility, including the construction workers, maintenance

personnel, and the facility occupants (Holt, 2001). Although a similar U.S. bill to expand









the designer and owner responsibilities on construction safety did not pass, the OSHA

standards are certainly not foreign to designers, which is a potential source of designer

liability. This occurs not solely by assuming the contractor's responsibilities for the

means or methods of the construction process, but through the understanding of their duty

as designers to help provide a safe workplace which is basic in their review of the

contractor's work product (Coble and Blatter, 1999). In response to the needs of the

industry, some architects started to work as construction managers on small projects and

thus coordinated the construction and design efforts to reduce accidents and costs

(McKee, 1994). MacCollum (1995) ranked the methods used for reducing safety risks

according to their priority and effectiveness, including the following:

* Design to eliminate or minimize the hazards;
* Guard the hazards;
* Give warnings;
* Provide special procedures and training; and
* Provide personal protective equipment.
He further suggested that the design team should include at least one safety

professional, and should develop and implement a hazard prevention plan for each

project.

Owners can impact designers and contractors through their proactive participation

in construction safety issues. In the construction accident causation model developed by

Suraji et al. (2001, see Figure 2-5), designer responses to project design constraints

(which may arise from the owner) can in turn place constraints on the construction

process. This may result in a chain reaction effect on accident causation. For example,

the constraints imposed by the owner may include the following:

* Accelerated design program
* Inadequate design budget
* Conflict of objectives or demands of other proj ects









Designer response is the action or inaction by designers to respond or react to the

constraints existing during the project design stage. These include the following:

* Increased design complexity
* Sublet part of design process
* Reduction in design resources
* Reduction in quality of components
* Ignore legal liabilities
Therefore, owners can take an active role in design for safety through:

* Setting their expectations in the design phase that construction safety is one of their
major concerns and is to be built into the project design;

* Addressing safety issues as early as the feasibility study and conceptual design
phases and integrate safety into the objectives of the project;

* Actively participating and coordinating the efforts of the designer and the
contractor through regular safety/constructability reviews of the project design; and

* If possible, awarding the contract to an engineering and construction company to
help promote safety performance.

Participate In Safety During Construction

In the research conducted by Levitt and Samelson (1993), it was found that the

owners with the safest construction projects tended to use many of the following

strategies with their contractors:

* Stress safety as part of the contract during the pre-job walk-around. This gives
contractors the opportunity to include all safety-related items in their bids and lets
contractors know early that the owner is serious about safety.

* Require short-term permits, rather than ongoing permits, for hazardous activities.
This means that contractors must check daily or more frequently to ensure that any
planned hazardous activities are coordinated with other contractors and with the
owner's own plant work force.

* Conduct safety audits of the contractor during construction. The owner's or
construction manager's safety staff conducts these audits to ensure compliance with
the owner's safety requirements and with all state and federal safety regulations.
They are aimed at systems and procedures rather than at specific hazards.


* Conduct periodic safety inspections.









* Require safety training of all project employees.

* Maintain statistics on the contractor's safety performance.

* Set goals for construction safety. Project-wide safety goals should be set, along
with specific goals for contractors who need special attention because of past poor
performance or particularly hazardous work operations.

* Include general safety guidelines in the body of the contract.

* Set up a construction safety department to monitor contractor safety.

* Require immediate reporting of all worker accidents. Immediate rather than
periodic reporting gives the owner more time to intervene and ensure that the
contractor has corrected any identified hazards before others can be injured by
them. Such notification can also serve to initiate consultations with the contractor's
senior management, if needed.

* Investigate the contractors' accidents. The owner's involvement in investigating
the contractors' accidents gives the owner valuable insights about generic safety
hazards on the project, as well as additional insights about the contractors'
organization and capabilities in the area of safety.

* Always include safety on the agenda at owner-contractor meetings.

* Provide contractors with special safety guidelines they must follow.

* Require the contractor to assign safety coordination responsibility to someone on
site.

* Reimburse the contractor's safety costs in full.

The following is a sample of best practices suggested by Gambatese (2000b) that

an owner can implement:

* Adjust the scheduling of different activities or construction phases that would
otherwise occur at the same location and be performed simultaneously.

* Provide a list and location of toxic substances and other hazardous materials that
are located on the site.

* Do not allow schedules that contain sustained overtime or night work.

* Impose a ceiling on the number of workers on site or in a particular area.









* Confirm that the contractor knows of the potential hazards of all construction
materials and their proper storage and disposal.

* On renovation or retrofit projects, provide the contractor with complete and
updated as-built drawings of the existing structure.

* Conduct a pre-construction meeting with the contractor (including all
subcontractors) to discuss safety issues.

* Consider involving OSHA in planning for project safety prior to starting
construction.

Gambatese (2000b) also suggested that owners should:

* Establish a clear position on safety. The owner's position on safety should be
clearly communicated to the project team at the beginning of the project and to all
team members joining the project as part of the construction phase. The position
can be written in the project documents and contracts, and verbally communicated
in project team meetings during design and construction. The actions of all
members of the owner's organization during the course of the project must reflect
and reinforce the established position.

* Ensure that safety is addressed in project planning and design. Thus, owners must
provide the initial impetus, by requesting or even requiring, by contract terms, that
designers consider construction site safety in their designs.

* Consider safety performance when selecting a contractor.

* Address safety in the construction contract. There are numerous clauses that could
be included in a contract to promote safety.

* Assign safety responsibility during construction. The responsibility for overseeing
safety on a construction project should be held by a competent organization or
individual.

* Participate in project safety during construction. Perhaps the most effective way an
owner can influence safety is through jobsite participation. Another common
means by which owners can be involved in safety is by requiring the submittal of
regular safety reports. Reports can be requested that provide information on the
results ofjobsite inspections, a listing of all injuries, safety meeting minutes, and
investigations of major accidents.

Hinze (1997) summarized some unique owner practices as the following:

* Placing company representatives on every construction project
* Conducting safety meetings with the contractors
* Requiring the contractors to adhere to owner-developed safety practices









* Providing specialized safety training for the contractor personnel
* Requiring all workers to go through safety orientation
* Reviewing each contractor's safety program
* Conducting regular audits of contractor safety performance
* Implementing safety incentive programs on all construction projects
The participation of owners in safety during construction is an integral part of the

efforts of many owners, but it is far from the sole participation of owners in safety.

Safety involvement of owners often starts at the very beginning of the project, and lasts

throughout the life cycle of the facility. Owner involvement in safety includes selecting

safe contractors, addressing safety in design, including safety requirements in the

construction contract, and being actively involved in project safety management.

Total Safety Culture And Behavior-Based Safety

When participating in construction safety management, proactive owners generally

implement initiatives that promote safety. The concept of total safety culture and

behavior-based safety management are among the latest safety management initiatives

that will be introduced.

Originally, people believed that accidents were a result of pure chance, and could

happen to anyone at any time, and some people were more likely to suffer an accident

than others. This was referred to as the accident proneness model (Cooper, 1998). In the

past, the prevailing influence of this approach meant that most accidents were blamed

solely on employees rather than on the work processes, poor management practices or a

combination of these. Recent studies have found no solid evidence to show that certain

people were more likely to be involved in accidents, and thus concluded that accident

proneness was essentially related the individual's propensity to take risks (Hinze, 1997).

The most famous accident causation model is the Domino model suggested by

Heinrich (1959, see Figure 2-7). The model shows that an accident is the result of a









series of events related to the social environment and heredity, personal failings or

mistakes, physical hazards and unsafe behavior. These events, if permitted to occur in

continuity, result in an accident, which may result in injuries and damages. Any accident

is prevented by breaking the chain (or removing a domino) of the series of events at any

location. While it is generally understood that most accidents are preceded by a series of

events, it is common practice to place most of the attention on the action that occurred

immediately before the accident. A more proactive approach would be to focus on the

"upstream" activities or events that set the series of events in motion.



















1998). In essence, Weaver's model placed the immediate responsibility for accidents











squarely on the shoulders of poor supervision and line management, instead of the injured

workers.
workers.









Today, as the zero injuries objective is accepted and sought by many owners and

contractors, total safety culture (TSC) and behavior-based safety (BBS) have become the

popular approaches being implemented by many proactive firms. The basic logic of

these approaches is to prevent construction accidents by continually addressing the very

front-end of the accident chain the culture and habits. Safety management is based on

the scientific approach that is focused on the psychology and the behavior of workers.

Geller (2001) summarized the essence of TSC as the following:

* Promotes a work environment based on employee involvement, ownership,
teamwork, education, training, and leadership.

* Builds self-esteem, empowerment, pride, enthusiasm, optimism, and encourages
innovation.

* Reinforces the need for employees to actively care about their fellow coworkers.

* Promotes the philosophy that safety is not a priority that can be reordered, but is a
value associated with every priority.

* Recognizes group and individual achievement.

He further pointed out that total safety culture requires continual attention to three

domains environmental factors (including equipment, tools, physical layout, procedures,

standards, and temperature), personal factors (including people's attitudes, beliefs, and

personalities), and behavior factors (including safe and at-risk work practices, as well as

going beyond the call of duty to intervene on behalf of another person's safety). In a total

safety culture:

* Everyone feels responsible for safety and does something about it on a daily basis;

* People go beyond the call of duty to identify unsafe conditions and at-risk
behaviors, and they intervene to correct them;

* Safe work practices are supported intermittently with rewarding feedback from
both peers and managers;









* People "actively care" continuously for the safety of themselves and others;

* Safety is not considered to be a priority, but it is a value. This change in
philosophy is based on the premise that priorities can change but values cannot.

Weinstein (1997) proposed another approach toward total safety culture through

incorporating Total Quality Management with safety. He suggested a Total Quality

Safety Management System which is achievable through the implementation of Total

Quality Management Principles, Process Safety Management Guidelines, ISO-9000

Quality Guidelines, and OSHA VPP (Voluntary Protection Program) guidelines. The

core of this approach relies on the commitment to and leadership of safety, and the

involvement of each individual and each work team in safety.

Cooper (1998) criticized the approach that most accidents were blamed solely on

employees rather than the work processes, poor management practices or a combination

of all three. He analyzed different theoretical models to explain the accident causation,

and supported the argument that the main focus of accident prevention should be shifted

away from the worker's unsafe acts and more onto the organization's overall

management system, particularly in relation to the implementation of the organization's

strategic decisions. Cooper concluded that safety culture should be the product of

multiple goal-directed interactions between people (psychological), jobs (behavioral)

and the organization (situational) (Cooper, 1998, page 17). Viewed from this

perspective, an organization's prevailing safety culture is reflected in the dynamic inter-

relationships between members' perceptions about, and attitudes towards, organizational

safety goals; members' day-to-day goal-directed safety behavior; and the presence and

quality of organizational safety systems to support goal-directed behavior.









In recent years, behavior theories are widely used in safety management. The

behavior-based safety approach is being utilized more extensively on construction

projects. The objective of behavior-based safety is to eliminate unsafe behaviors and

encourage safe behaviors through effective management interventions, based on behavior

and psychological theories. The core of the behavior-based safety approach is to track

both safe and unsafe behaviors of workers. According to Sutherland et al. (2000), the

basic premise of the behavior-based safety approach can be stated very simply: Behavior

is determined by its consequences. That is, people tend to repeat those behaviors that

produce 'positive' consequences, and not repeat those that result in either no positive or

'negative' consequences. Therefore, by giving workers with safe behaviors "nice" things

or removing negative effects, safety behaviors can be positively reinforced. On the

contrary, by giving workers with unsafe behaviors negative feedback or by taking away

"nice" things, unsafe behaviors will be punished and eliminated. This is also known as

negative reinforcement. Workers who work safely could be rewarded (positive

reinforcement), while workers who are unsafe could be punished (negative

reinforcement). Management must decide which of these are most appropriate.

Peterson (1988) summarized the most common ways the behavioral scientist has

contributed to safety management. The major approaches include the following:

* Survey methodology. Managers concerned about characteristics of their
organizations have often used employee attitude surveys to identify problem areas,
to assess the effects of organizational change or policies on job satisfaction, and as
an index of management performance.

* Selection and assessment of personnel. The contribution of the behavioral
sciences, particularly psychology, to improved selection, placement, and
assessment of personnel is a most common use.

* Programmed Instruction. This concept includes such sensible elements as
behavioral definition of training goals, the analysis of learning tasks into









manageable units, and systematic applications of principles of reinforcement to the
learning situation.

* Motivation theory. Intervening safety behavior through motivation techniques is
the core of the theory. In the minds of some observers, the impact of the
motivation theory and the application of the practices they advocate constitute the
major contribution of the behavioral scientists to the construction industry.

* Participative management. Workers would be more fully committed if they had a
voice in setting the goals and conditions of their work environment.

* Organizational development. Organizational development is a process of
systematic, planned change in an organization. It incorporates a definite
philosophy that organizational change should be in the direction of mutual trust,
openness in communication, readiness to deal with feelings as legitimate data, and
understanding intra- and inter- group relations.

Today, behavior-based safety approaches are widely used by many proactive

owners and contractors on their construction projects. The common approaches include

safety observations, worker participation (through job safety analysis, safety survey etc.),

and safety recognition programs.

Summary

From the literature review, it is apparent that although various recommendations

exist, there is still no solid evidence in the previous research to describe how owners can

efficiently promote construction safety on their projects, and how to assess the

relationship between safety performance and the owner's involvement in project safety.

A detailed research study at the project level to evaluate the owner's impact on

construction safety has not been conducted. This conclusion gave impetus for conducting

this research. Certain aspects to be investigated include the owner's commitment to

safety, the selection of contractors, contractual requirements, and the owner's

participation in safety management during construction.














CHAPTER 3
RESEARCH MOTHODOLOGY

Since previous studies and perspectives expressed in the literature suggested that

the owner's impact on construction safety can be demonstrated in various ways, this study

will focus on the influence of these factors on the safety performance of construction

projects. This chapter will describe the research methodology to disclose the relationship

of the owners' actions on project safety performance.

Research Design

Research Components

The objective of the study: To compare the effectiveness of several different

factors of the owner's involvement in construction safety on project safety performance.

In order to identify the relationship between a specific factor and project safety

performance, the median safety performances of projects on which owners implement

specific safety approaches will be compared with the median safety performances of

projects on which owners do not have such an approach.

Universe: All construction projects in North American with U.S. and Canadian owners

Experimental Unit: Any individual construction project

Population: Collection of safety performances of projects in the universe

U-Sample: Collection of construction projects included in the study

P-Sample: Collection of safety performances of projects in the study

Inclusion criteria:









The projects selected in the study should have at least 100,000 worker hours

accumulated to provide viable information on safety performance;

1. The projects should be on-going or have been completed within the past two years;
2. The safety personnel of the projects must be willing to participate in the study.


Response Variable: Project safety performance is the response variable in the

study. It is measured by the total OSHA recordable injury rate (TRIR) of the project,

which is the number of OSHA recordable injuries occurring per 100 full-time workers

annually (or injuries per 200,000 worker hours of exposure). TRIR can be easily

calculated with the number of OSHA recordable injuries and total worker hours expended

on a project.

Factors and levels: Each specific aspect of owner involvement in construction

safety can be identified as a factor. Therefore, there are multiple factors included in this

study. Based on the results of the literature review and previous study results, the factors

of interest can be classified into four categories, as follows:

Project context: the characteristics of the projects, including the following:

* The size of the project as measured by the estimated worker hours expended on the
project (quantitative factor);

* The construction effort on the project: whether a new project, renovation project,
maintenance project, or shutdown/turnaround/outage project (qualitative factor);

* The type of the project: whether a manufacturing project, a petrochemical project, a
civil work, a residential project, a utility project, or a commercial project
(qualitative factor.);

* The labor relationship of the project: whether the labor on site is union, open shop
or merit shop (qualitative factor);

* Type of owner for the project: whether public or private (qualitative factor);

* Basis for contract award, whether the contractor is selected through competitive
bidding or not (qualitative factor);









* Type of contract arrangement used on the project: whether general contract,
multiple primes, design-build, C.M. at risk, or C.M. agency contract (qualitative
factor);

* Type of contract used on the project: whether lump sum contract or cost plus
contract, similar to a time and materials (T&M) contract (qualitative factor);

* Firm providing the workers' compensation insurance on the project, whether the
owner (OCIP), the project (PCIP), the general contractor (CCIP), or if each
employer provides their own insurance coverage (qualitative factor);

* Number of work shifts per day: whether one, two, or three (quantitative factor);

* Number of workdays per week: whether four, five, six, or seven (quantitative
factor);

* Number of workers at the peak level of employment on the project (quantitative
factor);

* Total estimated monetary value of the project (quantitative factor);

* Percent of workers on site that do not speak English (quantitative factor);

* Number of subcontractors awarded on the project (quantitative factor);

* The prime contractor's commitment to safety, as evaluated by the owner's site
personnel (quantitative factor);

Selection of contractor: the factors related to safety that owners consider when

selecting contractors) to execute their projects, including:

* Whether or not the owner uses a preferred list when selecting a contractor
(qualitative factor);

* Whether or not safety performance is a consideration in contractor selection
(qualitative factor);

* Measurements used to evaluate safety performance of a contractor including TRIR,
EMR, safety personnel qualifications, project team qualifications, quality of overall
safety program, the OSHA log, and OSHA inspection history (qualitative factor);

Contractual safety arrangement: the safety requirements that are included in the

contract between the owner and contractor. This category will include different safety









management techniques that contractually must be employed by the contractor. They are

typically qualitative factors.

Owner's involvement during project execution: the practices of the owner during

the project execution to improve safety, including:

* Responsibilities of owner's site safety representative (qualitative factor);

* Implementation of a safety observation program on the project (qualitative factor);

* Specific items that must be included in the contractor's safety program (qualitative
factor);

* Means of addressing injury statistics on the project: whether by project, by
contractor, or blended with owner's safety performance statistics (qualitative
factor);

* Treatment of near misses: whether recorded and investigated, or recorded without
further investigation (qualitative factor);

* Owner support of project safety, specifically whether extra funds are provided to
promote safety (qualitative factor);

* Owner participation in an OSHA Voluntary Protection Program (VPP) (qualitative
factor);

* Whether or not a nurse or EMT is provided for the construction site by the owner
(qualitative factor);

* Extent of owner involvement in the safety training program (qualitative factor);

* Implementation of a safety recognition/reward program on the project, and the level
of the owner's involvement (qualitative factor);

* Distribution of a safety newsletter on the project (qualitative factor);

* Use of safety lunches to promote safety (qualitative factor);

* Incorporation of safety in the design of the project (qualitative factor);

Most of the factors are qualitative factors, and a few are quantitative factors. For

each qualitative factor, the parameters of interest are:









,y = the true median TRIR of all projects where the owners employed the safety

technique

,N = the true median TRIR of all projects where the owners did not employ the

safety technique

The statistics of interest are:

Mr = The observed median TRIR of all projects where the owners employed the

safety technique

MN = The observed median TRIR of all projects where the owners did not employ

the safety technique

Thus, the objective of this study could be achieved by testing the hypothesis (Ho:

SY = N vs. Ha: &y < N) for each qualitative factor. This is a one-tail test, since TRIR can

be expected to be less when the owner implements a safety approach or when the project

possesses a specific characteristic identified in the literature review. The hypothesis

testing will be further discussed in chapter 5.

Method Design

This study is a retrospective observational study, in which both response variable

and factor levels are observed from the samples selected. Personal interviews based on a

questionnaire were used to collect most of the data necessary for the study on each

project selected. This is one of the methods frequently used in studies on construction

projects. Since construction projects generally cost millions of dollars, and they are often

under the influence of many different factors, it is not practical to conduct a prospective

experimental research study. The interview was therefore employed to obtain as much

information as possible.









Because little prior research had been conducted on this topic, an initial study was

necessary to get some idea of the most commonly employed techniques of owners to

promote construction safety, and whether they might be related with safety performance.

Also, a follow-up questionnaire survey with contractor safety personnel was conducted to

help verify some of the findings and determine how contractors would like owners to get

involved in project safety management.

As the study is an uncharted one, it is difficult to reasonably determine which

factors might be extraneous factors and which might be major factors that influence

project safety. Therefore, all the factors suggested in previous studies and the industry

pilot study were included and analyzed. Statistical analysis and discussion will identify

which factors make a difference on project safety performances. Finally, based on the

analysis, a model was built to explain how owners can influence construction safety.

Sample Size Determination

Since the objective of this study was to compare y to ,N, the type of inference is

hypothesis testing. In order to determine the minimum acceptable sample size required

for the study, the following steps to perform hypothesis testing were followed:

1. Determine the null hypothesis (Ho). The null hypothesis is Ho: y = NN.

2. Determine the alternative hypothesis (Ha). The alternative hypothesis is Ha: Y < KN
(one-tail test). When the literature suggested that projects where owners implemented
specified practices would have better safety performance, and the median TRIRs also
supported the statement, a one-tail test was conducted, i.e. Ha: y < KN.

3. Determine a suitable significance level and power for the statistical test. The
significance level is set as c = 10%, instead of 5%. Since this study intends to
identify those potential influences owners can have on project safety, the significance
level is set higher than usual. The power of the statistical test is set to be 1-3 = 1-20%
= 80% (P is set to be 20%). That is to say, the probability of a Type I error (the
probability of erroneously rejecting Ho when it is actually true) is 10% and the









probability of Type II error (the probability of erroneously accepting Ho when it is
actually false) is 20%, respectively.

4. Choose and compute the test statistic. The test statistics have been discussed.

5. Make correct conclusions. This will be discussed in Chapter 5.

According to a previous study (Hinze, 2002), it was estimated that the practical

difference between TRIR of two projects that makes sense in the industry is B = 1.0.

This is determined to be the clinical significant difference to determine the sample size of

this study. Since it was assumed that most of the projects that would participate in the

interview study would have better safety performances than the construction industry

average (which is 7.8 for TRIR in 2001, BLS, 2003), the response variable is estimated to

be between 0 and 6. Therefore, based on the Empirical Rule, the estimate of the standard

deviation of TRIR is o = (6-0)/4 = 1.5. Then the sample size determination can be

calculated with the following indexes:

DELTA = B/o = 1/1.5 = 0.67

a = 10%, P = 20%

By referring to the tables listed in (Rosner, 1995), the sample size for each factor

level (when owners do or do not implement a particular safety management approach)

should be 28 projects. Therefore, the total sample size for this study is 28*2 = 56

projects.

Computation of the sample size needed to provide valid research findings was

found to be 56. This means that a sample size of 56 would be expected to yield reliable

results.









Sample Selection

Most of the projects in the project interview phase (the major phase of this study)

were recommended by the Construction Industry Institute Project Team #190 (with the

title "the role of owners in construction safety"), based on the inclusion criteria. The

contact persons of the projects were then asked whether they could accept an interview.

The person was then interviewed with the questionnaire after he or she agreed to accept

the interview. Essentially, the project sampling was a voluntary sampling, since attitudes

of owners towards safety varied significantly, and only those that would participate in the

study and responded could be included in the sample. Note that all projects identified in

the research was contacted to request their participation. These tended to be projects

known to the CII project team members. However, whether or not the data could be

regarded as representative would then depend on the variability of the data collected and

the size of the projects. That is, if the variability of the response variable was similar to

the large projects in the population (projects of the large owners), it could be regarded as

representative of the population. If the sizes of the projects in the study summed up to a

significant portion of the total size of projects in North American, the data could be

reasonably regarded as representative. Also, in order to test whether the injury data

(which might be sensitive to some owners and therefore biased) was true, the injury

triangle by Heinrich (1959) was used to test the results.

Data Collection

Three studies were conducted to collect data for this research effort: the pilot

study, the project interviews, and a follow-up survey. The pilot study consisted of a

mailed survey sent to large owner firms that were known to have significant construction

budgets. This research was followed by a study in which representatives of owners were









interviewed at participating construction sites. The interview data consisted of more

detailed information than was obtained in the mailed survey study, and therefore

constituted the major data collection phase. Most of the findings of this research are

generated based on the data collected through the project interviews. The follow-up

survey of contractors was a supplementary part of the study, with the purpose of

comparing the findings of the project owner interviews with the opinions of construction

firms.

Mailed Survey Study (Phase 1)

The first phase consisted of surveys to study the practices of owners and their

resultant impact on construction safety at the company level. The purpose of the survey

was to establish and refine the hypotheses to be tested in the subsequent interview study.

Also, the survey was to help identify new factors to be tested. This included the

acquisition of information on why safety is essential to owners, means used to ensure that

safe contractors are employed on their projects, how safety is addressed in construction

contracts, how contractors are expected to address safety during construction, the means

used by owners to monitor contractors during construction, and the nature of the active

role played by owners during project execution. The questionnaire was developed by

relying on the literature review and the input from construction industry safety personnel.

The survey questionnaire asked questions related to safety performances of the owner

projects, safety coordination of the owner, owner safety management practices on

projects, safety in design phases, and other safety related topics (see Appendix a). The

U.S. owners surveyed were those making large annual expenditures for construction, as

listed in the Top 425 owners provided by ENR (Engineering News Report, 2001). Their

construction budgets in year 2000 ranged from 23.8 million to 3.8 billion dollars.









Project Interview (Phase 2):

The second phase of the study was designed to acquire more detailed information,

which would not be feasible or realistic in a mailed survey. The purpose of this study

was to test the hypotheses that had been developed. This study, with its focus at the

project level, was conducted through in-depth interviews. In addition to obtaining

information about the various practices of owners to ensure the safety of construction

workers on their sites, information was also obtained on measures of safety performance.

Based on the results obtained from the literature review and the mailed survey, a detailed

interview questionnaire was designed. The questionnaire went through a series of

iterative improvements as suggestions were offered by members of CII Project Team

#190. The questionnaire mainly focused on four aspects of the owner's involvement in

construction safety:

* Project description,
* Selection of safe contractors,
* Safety requirements in contracts, and
* Owner's involvement in safety during construction.
Factors such as project characteristics, contractual arrangements, contractor's

commitment to safety, and so on were also addressed in an attempt to control external

impacts on project safety performance.

The questionnaire was modified through approximately 15 iterations before it was

used to conduct interviews. After three telephone interviews were conducted, a final

version was devised. After the first three interviews, the interviewees were asked about

the clarity and integrity of the questionnaire. This feedback was used to further modify

the questionnaire before it was finalized and used in the remaining interviews. The final

version of the questionnaire is included in Appendix b. Approximately 100 construction









projects were identified for inclusion in this research. The criteria to select projects

included:

* Under construction or newly completed (within the past two years);

* Worker exposure of at least 100,000 hours (some exceptions were made, as
explained later);

* Variety in the types of projects: public or private; petrochemical, utility,
manufacturing, commercial or residential projects

* Whenever the owner had several smaller projects at the same location and
implemented the same safety management policy, the multiple projects were
regarded as a combined project.

Persons interviewed included site representatives of the owner, including

construction managers, safety managers, and safety coordinators. The face-to-face

project interviews generally took one and a half hours to two and a half hours to conduct.

Whenever face-to-face interviews were economically infeasible, the interviewee was

asked to fill out the questionnaire and return it to the researcher by Fax or email. For

questionnaires returned this way, a follow-up telephone interview, lasting about half an

hour, was conducted to clarify any questions. Most of the projects were introduced by

members of the CII Project Team #190. Often either the owner or the contractor on the

project, or both, were CII members; however, CII membership was not a criterion for

inclusion in the study. Many owners of projects were not affiliated with the CII.

At the conclusion of this study, 81 projects had been interviewed. Among them 59

projects provided the TRIR data and satisfied all the criteria to be included in the data

analysis. The final sample of 59 projects included 49 U.S. projects, seven projects in

Canada, and three international projects with U.S. owners and U.S. contractors.









Follow-Up Survey To Contractors (Phase 3)

A survey of contractors was conducted after the project owner interviews were

completed. This was not a major part of the data collection. This study was to obtain

information from contractors, because contractors may hold different views towards the

same questions regarding owners. Therefore, it was deemed reasonable to conduct a

separate survey of contractors to identify possible bias of the data obtained from owners.

Also, some questions asked how the contractors would like the owners to help promote

project safety. The information could help communication and cooperation of the two

parties on future projects. This survey was conducted with a short questionnaire

developed on the basis of the analysis results obtained from the project interviews (see

Appendix c). The participating contractors may or may not have been involved in the

owners' projects in the primary study.

Data Analysis

The main objective of the project was to test whether the involvement of owners

in safety management made a difference to the safety performances of projects. Project

safety performance, defined as the OSHA recordable injury rate (also referred to as Total

Recordable Injury Rate, TRIR), was the dependable variable in the analysis. The

different safety management techniques implemented by the owners were regarded as the

independent variables. Other information obtained relating to project properties and

contractor's commitment to safety constituted extraneous factors and were not the major

variables to be considered in the research. They were also regarded as the independent

variables in the analysis.

Analysis of the data obtained from the mail-out questionnaire survey to owners

(Phase 1) used non-parametric methods, univariate methods and the logistic regression









method based on the type of the response variables. The analysis attempted to recognize

the involvement of owners that might help project safety.

Analysis of the data set obtained through the project interviews (Phase 2) was the

major part of the data analysis. It can be divided into three stages. The first stage was

focused on the descriptive statistics of the variables, that is, means, medians and standard

deviation of the interval variables, and the proportions of nominal and ordinal variables

were calculated. By this means, the frequency and popularity of different safety practices

and techniques were identified. The second stage aimed at testing reasonable

associations among the variables, especially the associations between safety performance

and safety management approaches. Efforts in the third stage were conducted on the

establishment of a model to define the relationship between project safety performance

and various types of owner involvement. A scorecard was developed to evaluate the

owner's involvement in construction safety management.

All the data, once obtained, were input into a document developed in the

Statistical Package for Social Sciences (SPSS v10.0), according to the data structure of

the questionnaire. In the first stage of data analysis, the software was used to describe the

properties of the projects interviewed, including their location, size, safety performance,

contractual arrangement, types and labor arrangement. Frequencies of different safety

management techniques were also summarized to provide a holistic picture of the

information of the projects interviewed. As one aspect of the analysis, data from projects

in Canada were examined separately from the projects in the U.S to address any possible

regional variations. It should be noted that no clear differences were found between the

U.S. and Canadian projects.









In the second stage, multi-variable statistical analysis methods, especially non-

parametric methods (which treat the response variable as an ordinal variable) were

conducted to find the relationship between safety performance and the safety

management techniques. Safety performances of projects with different safety

management techniques were compared to identify those techniques that made a

difference on safety performance. In this stage, the data were stratified to get a better

idea of how some owners emphasized safety on large projects, and how owners would

focus on safety on petrochemical projects. This was done because the pilot study showed

that larger contractors and petrochemical projects appeared to have better safety

performances. When considering the variety of the projects interviewed, the significance

level was set at 0.10, to detect all the safety management techniques implemented by

owners that could potentially impact safety performance.

In the third stage of the data analysis, answers to some open-ended questions, as

well as some detailed cases, were considered to test how the owner's involvement could

make a difference to project safety performance. Based on the analysis in the first and

second stages, this stage was focused on establishing a model to describe the cause-effect

relationship between the owner's involvement and project safety performances. Also, the

most effective and important points for the future reference by owners were summarized

into a scorecard to evaluate the owner's involvement in safety management.

Analysis of data collected in the survey of contractors (Phase 3) was mainly

composed of descriptive statistical analysis. Those statistics were then referenced to

support the findings in the interview questionnaire, and provide suggestions for owners to

better cooperate with contractors in promoting construction safety.














CHAPTER 4
QUESTIONNAIRE SURVEYS FOR LARGE OWNERS


In this chapter, the analysis of the data obtained from the questionnaire survey is

presented with different non-parametric statistical methods, including the Mann-Whitney

U test, Wilcoxon test, Kruskal-Wallis test, and Friedman two-way ANOVA. These non-

parametric tests were selected because of the low response rate. By using the non-

parametric statistical methods, the medians (or distributions) of the response variable for

different factor levels are compared.

The Data Set

The population surveyed was the ENR-list of the top (based on expenditures on

construction) 425 owners in the U.S. The questionnaire was mailed out on February 21,

2002. Thirty-one responses were received by April 12, 2002, after which no additional

replies were received. The response rate was about 7.3%, which was lower than the

expected response rate of 10%. Possible reasons of the low response rate can be:

* The envelope of each survey was addressed to "Construction Contract Manager".
Although this was the title suggested by some owner members of CII, the
appropriate person to fill out the survey may have a different title in different
companies. It is quite probable that the questionnaires were not received by the
appropriate individuals.

* The questionnaire may have been too lengthy, requesting too much information.
Some individuals may not be patient enough to complete the surveys and return
them. This was considered during the design of the survey, and it was determined
that people familiar with safety could complete the survey within 20 minutes. Most
of the information sought required a multiple-choice response.

* Only companies that emphasize safety and had better safety performances may
have decided to participate in the study. It is possible that safety is still not the top









priority in many owner companies. Some companies may be reluctant to reveal
information about an unacceptable safety performance.

* As the consumer of construction products and services, most owner companies are
expected to have their primary interests in the revenue generating sectors of their
own businesses. It is suspected that many owners still regard construction safety as
resting totally with the contractors who contract their construction services. Thus,
many owners are not familiar with the safety issues on their projects, and may not
have felt capable of completing the survey.

Because of the above reasons, the mail-out survey was regarded as a voluntary field

experiment, with responses being received from those who had an interest in and had a

strong focus on safety management. Therefore, the results are probably biased and not

representative of all owner companies. This further supports the use of non-parametric

statistics in the analysis. However, as mentioned in the previous chapter, the aim of the

questionnaire survey was to identify the practices of the owners associated with better

safety performances, and to develop hypotheses to be tested in the subsequent project

interview phase of the research. The survey data, although limited and probably biased,

can still serve a valuable purpose and provide enlightenment for the follow-up research

study.

It should be noted that for non-parametric methods such as the Kruskal-Wallis

tests, or Friedman two-way ANOVA, the objective is to compare the medians of different

groups, and the response variable is treated as an ordinal variable. The mean ranks that

will show in the analysis result tables simply mean the average rank of the coded

response variable, and a larger mean rank suggests a larger median of the category.

Because only 19 of the respondents provided the TRIR of their construction

projects in the past year, it was difficult to develop any strong conclusions about the types

of projects that have better safety performance. However, among the 19 owners

providing the TRIR data, ten had primary manufacturing projects (with the average TRIR









of 2.92) and four had primary petro-chemical projects (with the average TRIR of 1.13).

In order to discover the underlying reason for the difference, the rating of the importance

of safety performance when selecting contractors was considered as a safety index of

each owner (ten was the most important and one was the least important). The ranking fo

the importance of safety when selecting contractors was the primary measure used in the

analysis since 31 respondents provided this information. If the TRIR had been used, the

sample size was redeemed to 19, an unacceptable level. The use of the ranking measure

was felt to be a viable approach as the ranking value was found to be relatively consistent

with the TRIR.

Analysis Of The Mailed Survey Data

The Projects Context

Of the responding owners, the annual construction budget, the percentage of the

construction budget on new projects, and the dominant type of construction projects were

the characteristics of particular interest. These factors can influence the safety

performance of the owner.

The Kruskal-Wallis test was conducted to test whether the ranking of safety

importance was dependant on the major type of projects. The results of the test are

shown in Table 4-1. Even with a small number of responses, it is apparent that petro-

chemical project owners place a higher importance on safety when selecting contractors,

while owners of building projects have the lowest importance. The underlying reason

may be associated with the attitudes of the owner towards construction safety in their

own businesses. For example, the chemical industry and nuclear power generation

industry are famous for their strict concern for safety. Therefore, owners in these

industry sectors may similarly prioritize safety in their construction projects, and actively










participate in safety management on the projects. The OSHA recordable injury rates of

the four owners with petro-chemical projects (0.50, 1.10, 1.42, and 1.50, respectively)

were also much lower than those of the other owners. This indicates that these owners

have much better safety performances than the industry average.

Table 4-1. Importance of safety during selection of contractors to owners of different
types of projects (1 to 10 with 10 being the highest possible rating)
Mean Rank of Kruskal-Wallis
Project Type Count Mean Median
Kruskal-Wallis Test Test Sign. (2-tail)
Building 5 5.2 5 6.5
Others 2 7.5 7.5 13.25
Manufacturing 18 7.94 8 17.14 0
0.05
Utilities 2 8 8 16.5
Petro-Chemical 4 9.25 9.5 23.88
Total 31 7.65 8
Note: In this table, a higher Mean Rank means that the safety index of the group of

owners tends to be higher, or the median of the safety index tends to be higher, and

therefore, the owners are more concerned with safety. It is the same as in Table 4-2.


The same statistical test was conducted to determine if there was a difference

among owners with different annual construction budgets. The result is shown in Table

4-2. It is evident that owners with larger construction budgets are more concerned with

safety than owners with small construction budgets. The Mann-Whitney test is

conducted to compare the importance of safety to owners with $25-100 million annual

construction budgets to those with >$500 million annual budgets, and the results showed

that those with budgets >500 million placed a higher priority on safety (p<0.05).

Table 4-2. Importance of safety when selecting contractors to owners with differing
construction budgets
Mean Rank of Kruskal-Wallis
Annual Budget Count Mean Median
Kruskal-Wallis Test Test Sign. (2-tail)
25-100 million 12 7 7 12.46
100 500 million 13 7.62 8 15.96 0
0.05
>500 million 6 9 10 23.17
Total 31 7.65 8___









The percentage of new projects in the construction budget was examined, but this

was not a significant factor influencing the owners' concern for safety. This is

reasonable because owners with more new projects may not necessarily emphasize

construction safety more than owners with fewer new projects, and they may not be good

at construction safety performance, either.

Commitment Of The Owner To Construction Safety

Commitment of the owner to construction safety was measured by two questions:

the earliest time the owner began to emphasize safety on the project (question 5) and the

philosophy held by the owner concerning the responsibility for safety (question 13). In

the questionnaire, both questions were coded with ordinal variables to show the degree of

owner involvement in construction safety. The earliest stage to emphasize safety (e.g.

during concept and feasibility phase) was coded as a smaller number, and beginning to

emphasize safety at a later period (e.g., during construction) was assigned a higher

number. For the philosophy concerning safety, a small number was assigned to more

proactive attitudes (e.g. taking total control of safety).

Owners with different types of major projects and owners with different sizes of

annual construction budgets were compared. The answers to question 5 (earliest stage to

emphasize construction safety) were significantly different among owners with different

annual budgets and owners of different types of projects (refer to Table 4-3 and Table 4-

4). It is evident that larger owners and owners of petro-chemical projects started to

emphasize construction safety at a significantly earlier stage of the projects. As to the

philosophy towards construction safety liability, owners with different size projects did

not show significant differences. It was noted that petro-chemical owners tend to have










more proactive attitudes towards safety (address safety in early project stages) and they

would assume greater liability risks.

Table 4-3. Earliest stage to emphasize safety vs. annual construction budget


Note: In this table, a smaller Mean Rank means that the earliest stage to concern for

safety is earlier, and therefore, the owners are more concerned with safety. It is the same

as in Table 4-4.


Table 4-4 Earliest stage to emphasize safety vs major type of projects


Note: Refer to footnote for Table 4-3.


Selection Of Safe Contractors

Most of the owners (80% of the respondents) would establish minimum safety

performance requirements in addition to the OSHA regulations when selecting

contractors to construct their projects. When selecting contractors, safety performances

of the contractors were considered by most owners (84%). After counting the number of

different safety approaches taken into consideration when selecting contractors, it was

found that the mean number of such requirements was 2.4, ranging from a maximum

count of 5 to a minimum count of zero. With the Kruskal-Wallis test, the number of the

safety selection criteria was compared between owners of different sizes and owners with

different project types. The results are shown in Table 4-5 and Table 4-6, respectively.


Mean Rank of Kruskal-Wallis
Annual Budget Count Mean Median
Annual Budget Count MeaKruskal-Wallis Test Test Sign. (2-tail)
25-100 million 12 3.33 3 20.96
100 500 million 13 1.92 1 13.54 0
0.03
>500 million 6 2 1 11.42
Total 31 2.48 3


Mean Rank of Kruskal-Wallis
Project Type Count Mean Median
Project Type Cot M n M n Kruskal-Wallis Test Test Sign. (2-tail)
Petro-Chemical 4 1 1 7.5
Utilities 2 2 2 13.5
Manufacturing 18 2.33 2 14.67 0.03
Building 5 4.2 3 22.8
Total 29 2.48 3










From the results, it is reasonable to expect that larger owners have more stringent safety

criteria when selecting contractors. In order to analyze the relationship between the

importance of safety and the criteria used, the Mann-Whitney test was conducted and the

results are shown in Table 4-7. It is evident that using each criterion or not is

significantly related to the importance of safety in the selection of contractors. The loss

ratio was the only exception, which was implemented by very few owners.

Table 4-5. Number of safety criteria vs. annual construction budget
Mean Rank of Kruskal-Wallis
Annual Budget Count Mean Median Sg. ( -
Kruskal-Wallis Test Test Sign. (2-tail)
25-100 million 13 1.58 1.5 11.85
100 500 million 12 3.15 3 17.21 0
0.04
>500 million 6 4.00 5 22.58
Total 31 2.71 3___


Table 4-6. Number of safety criteria vs. major type of projects
Mean Rank of Kruskal-Wallis
Project Type Count Mean Median
Kruskal-Wallis Test Test Sign. (2-tail)
Building 5 0.8 0 6.7
Manufacturing 18 3 3 16.14
Utilities 2 3 3 15.25 0.08
Petro-Chemical 4 4 4 20.13
Total 29 2.71 3


Table 4-7. Results of Mann-Whitney test: safety selection criterion vs. importance of
safety in the selection of contractors
Contractual Safety
Contractual Safety Ans. Cnt. Mean Median m. r. s. r. Sign.
Requirements
Experience Modification Rating N 14 6.86 7.5 13 182 05
(EMR) of the contractor. Y 17 8.29 8 18.47 314
OSHA recordable injury rate of N 19 7.05 8 13.24 251.5
0.02
the contractor. Y 12 8.58 9 20.38 244.5
Loss Ratio of the contractor, N 28 7.71 8 16.23 454.5
0.33
should be less than Y 3 7 8 13.83 41.5
Site-specific safety program N 16 6.88 7.5 12.38 198 01
prepared by the contractor. Y 15 8.47 9 19.87 298
Qualifications of the safety staff N 18 6.72 7 11.61 209
<0.01
of the contractor. Y 13 8.92 9 22.08 287
Quality of the overall safety N 9 6.11 6 9.17 82.5
program of the contractor. Y 22 8.27 8 18.8 413.5


Note: the title of the columns are: Ans


answers, Cnt


counts, mean


mean importance of safety (1-10,


with 10 being the most important), median = median importance of safety (1-10, with 10 being the most
important), m.r. = mean of rank, s.r. sum of rank, sign.= significance level (1-tail)










Safety Management Practices Of The Owners

1) Safety requirements in the construction contract

All the owners responded that they will include some safety requirements in their

construction contracts. Two requirements, namely, "contractor must comply with the

local, state and federal safety regulations", and "contractor must report all OSHA

recordable injuries to the owner" were required by all responding owners. Less than 25%

of the owners required contractors to place at least one full-time safety representative on

the project, or that the contractor must submit a safety policy signed by its CEO. As to

other safety requirements listed in the questionnaire, about 50% to 75% of the owners

included them in the contract.

When the counts of contractual safety requirements of the owners were calculated,

it was found that for owners with different types of facilities constructed, the counts are

considerably different (refer to Table 4-8). Petrochemical companies included the

largest number of safety requirements in their contracts, followed by utility owners.

Owners of buildings and other manufacturing plants have fewer safety requirements

included in their contracts. Also, it was found that owners with larger construction

budgets have more contractual safety requirements (refer to Table 4-9).



Table 4-8. Number of contractual safety requirements vs. major type of projects
Mean Rank of Kruskal-Wallis
Project Type Count Mean Median
Kruskal-Wallis Test Test Sign. (2-tail)
Building 5 4.6 5 8.8
Manufacturing 18 6.7 7 14.75
Utilities 2 7.5 7.5 17.5 0.11
Petro-Chemical 4 9 9 22.63
Total 29 6.77 7










Table 4-9. Number of contractual safety requirements vs. annual construction budget
Mean Rank of Kruskal-Wallis
Annual Budget Count Mean Median
Kruskal-Wallis Test Test Sign. (2-tail)
25-100 million 12 5.33 5 10.96
100 500 million 13 7.85 8 19.65 0
0.05
>500 million 6 7.33 8 18.17
Total 31 6.77 7___
2) Safety management involvement of owners taken during construction

One question asked about safety management involvement of owners to promote

construction safety on site. It was found that four of the seven listed safety management

activities were frequently performed by owners, namely, owners' personnel conduct

periodic job site safety inspections or safety audits; owners' personnel participate in some

contractor safety meetings; owners' personnel participate in the investigation of all lost

workday injury accidents; and owners monitor injury incidence rates on each project.

More than 2/3 of the owners were involved in this manner. Other safety approaches were

taken by less than 1/3 of the owners.

A Kruskal-Wallis test was conducted to identify the strength of the relationship

between owner's involvement in safety management and the owner's annual construction

budget/the major type of projects. The results were quite consistent with the previous

findings (refer to Table 4-10 and Table 4-11). Petrochemical owners and owners with

larger construction budgets were more actively involved in safety during construction.




Table 4-10. Total number of safety practices implemented vs. major type of projects
Mean Rank of Kruskal-Wallis
Project Type Count Mean Median
Kruskal-Wallis Test Test Sign. (2-tail)
Building 5 1.6 2 6.5
Utilities 2 3.5 3.5 14.25
Manufacturing 18 3.89 4 15.39 0.06
Petro-Chemical 4 5.75 5.5 24.25
Total 29 3.74 4__










Table 4-11. Total number of safety practices implemented vs. annual construction budget
Mean Rank of Kruskal-Wallis
Annual Budget Count Mean Median
Kruskal-Wallis Test Test Sign. (2-tail)
25-100 million 12 2.75 2 11.38
100 500 million 13 4.23 4 18.04 0
0.05
>500 million 6 4.67 5.5 20.83
Total 31 3.74 4___

The Mann-Whitney U test was conducted to compare the rank of the importance of

safety in the selection of contractors and whether or not owners were involved in

construction safety in a particular way. The results are shown in Table 4-12. It is

reasonable to suspect that owners who emphasize construction safety tend to be involved

in more ways than owners who care less about safety.


Tible 4-12 Importance o n


Note: the title of n


mean importance of sa 0,


with 10 being the most important), median = median importance of safety (1-10, with 10 being the most
important), m.r. = mean of rank, s.r. sum of rank, sign.= significance level (1-tail)


... I- ... .. . .. . ... .
Safety approaches Ans Cnt. Mean Median m. r. s. r. Sign.

Owner's personnel conduct periodic job Y 7 8 8 9.93 69.5
0.02
site safety inspections or safety audits. N 24 6.43 7 17.77 426.5

Owner places a safety person on the Y 21 8.3 9 14.07 295.5
project to support the contractor on 0.04
project safety. N 10 7.33 8 20.05 200.5
Owner provides a nurse or emergency Y 22 9.22 10 12.95 285
medical technician (EMT) for the <0.01
construction project. N 9 7 8 23.44 211

Owner's personnel participate in some Y 7 8.12 8 8.86 62 0.01
contractor safety meetings. N 24 6 7 18.08 434

Owner's personnel participate in the Y 9 8.05 8 11.28 101.5
investigation of all OSHA lost workday 0.03
injury accidents. N 22 6.67 6 17.93 394.5
Owner implements a safety incentive Y 26 8.8 9 15.02 390.5
that can be earned by the contractor for 0.01
completing the project below a N 5 7.42 8 21.1 105.5
specified OSHA recordable injury rate.

Owner monitors injury incidence rates Y 9 8.18 8 8.78 79 <
<0.01
on each project. N 22 6.33 6 18.95 417
-c --_


hT










If Spearman's correlation (non-parametric correction) is calculated for the number

of safety approaches, the number of safety requirements included in the contract, and the

rating of safety performance during the selection of contractors (as shown in Table 4-13),

it is evident that they are significantly positively correlated. Owners emphasizing safety

tend to include more safety requirements in their contracts and utilize more approaches to

promote safety on their projects.

Table 4-13. Non-Parametric correlation matrix between importance of safety, number of
contractual requirements and safety involvement approaches
Importance Contract Approaches
Correlation Coefficient 1 0.734 0.651
Importance Sig. (2-tailed) <0.01 <0.01 <0.01
N 31 31 31
Correlation Coefficient 0.734 1 0.578
Contract Sig. (2-tailed) <0.01 <0.01 <0.01
N 31 31 31
Correlation Coefficient 0.651 0.578 1
Approaches Sig. (2-tailed) <0.01 <0.01 <0.01
N 31 31 31


3) Safety training and safety in design

Safety training was one area of focus in this survey, namely to explore how owners

addressed safety training on their projects. The Mann-Whitney test was conducted with

the rating of safety performance as the dependent variable. Four questions related to

safety training were tested against the dependent variable, namely, whether the owner's

safety representative attends safety orientation (q07), whether the owner makes monetary

allocations for safety training (qlO), whether standardized training is required for all the

projects (q14), and whether basic safety training is required for all workers (q17). The

results are shown in Table 4-14. Owners who are more concerned about safety tend to

emphasize the safety training and orientation program.










Table 4-14. Mann-Whitney test: Importance of safety vs. safety training inputs
Safety training inputs Ans. Cnt. Mean Median m. r. s. r. Sign.

safety representative attends safety N 11 5.82 6 8.05 88.5
0.01
orientation (q071) Y 20 8.65 8.5 20.38 407.5

does the owner make allocations for safety N 17 6.53 7 10.85 184.5
0.03
training (q10) Y 14 9 9.5 22.25 311.5

standardized training is required for all the N 7 6.29 6 8.86 62 0.02
projects (q14) Y 24 8.04 8 18.08 434

basic safety training is required for all the N 6 5.67 5.5 6 36 0.01
workers (q17) Y 25 8.12 8 18.4 460

Note: the title of the columns are: Ans = answers, Cnt = counts, mean = mean importance of safety (1-10,
with 10 being the most important), median = median importance of safety (1-10, with 10 being the most
important), m.r. = mean of rank, s.r. sum of rank, sign.= significance level (1-tail)

The total number of positive answers to the four safety training and orientation

questions was calculated. The Kruskal-Wallis test was conducted to examine the owners'

involvement in training with the size of the construction budget and the major project

type of the owners (refer to Tables 4.15 and 4.16). Not surprisingly, larger owners and

petro-chemical owners tend to have more input into safety training and orientation.


Table 4-15. Total amount of safety training input vs. project type
Mean Rank of Kruskal-Wallis
Project Type Count Mean Median
Kruskal-Wallis Test Test Sign. (2-tail)
Building 5 1 1 6.5
Manufacturing 18 2.72 3.5 15.64
Utilities 2 3.5 3.5 18.75 0.06
Petro-Chemical 4 3.75 4 20.88
Total 29 2.68 3___



Table 4-16. Total amount of safety input vs. annual construction budget
Mean Rank of Kruskal-Wallis
Annual Budget Count Mean Median
Kruskal-Wallis Test Test Sign. (2-tail)
25-100 million 12 1.83 1 12.08
100 500 million 13 3.15 3 17.83
0.08
>500 million 6 3.33 4 20.83
Total 31 2.68 3___










Safety can be addressed as early as the design phase of a project. About half the

owners stated they would address safety during the design stage. The Mann-Whitney U

test was conducted with the importance of safety to the owners and whether or not

designing for safety was a concern. The results are shown in Table 4-17. It can be

concluded that owners addressing safety during the project design will also emphasize on

the importance of safety performance during the selection of contractors.

Table 4-17. Relationship between the importance of safety and whether or not safety is
addressed in the design phase
Address safety in Count Mean Median Mean Rank of Mann- Mann-Whitney U
Count Mean Median
the design phase Whitney U Test Test Sign. (2-tail)
No 12 6.67 7.5 12.08
Yes 19 8.26 8 18.47 0.05
Total 31 7.65 8

4) Ranking of different safety approaches

Question 23 of the survey asked respondents to rank the importance of different

safety approaches. A Friedman two-way ANOVA (for repeated measures) was

conducted to compare the means of ranking for different approaches (see Table 4-18).

Table 4-18. Priority of different approaches preferred by owners (1 to 5 with 1 being the
highest possible priority)
Safety approaches Mean Median m. r.* Sign.
Owner selects safe contractors to carry out the work. 1.45 1 1.83
Owner emphasizes safety and constructability in design. 2.1 2 2.5
Owner participates in and monitors safety during the entire life of a 2.03 2 2.54 <0.01
construction project.
Owner develops an effective safety recognition and reward program. 3.46 4 4
Owner dedicates funds to support the contractor's efforts in safety. 3.63 4 4.13
mean ranking of Friedman two-way ANOVA test.


It is evident that significant differences exist between the priorities owners placed

on different safety approaches. The order of approaches, beginning with the top priority,

is as follows:

1. Owner selects safe contractors to carry out the work.









2. Owner emphasizes safety and constructability in design.

3. Owner participates in and monitors safety during the entire life of a construction
project.

4. Owner develops an effective safety recognition and reward program.

5. Owner dedicates funds to support the contractor's efforts in safety.

A Wilcoxon matched-pair signed-rank test (for repeated measures) was conducted

to compare the mean ranking of selecting safe contractors (referred to as variable Q232)

and participating in and monitoring safety (Q233). It was found that the mean rank of

selecting safe contractors is significantly higher than that of participating in and

monitoring safety at the 0.04 level (refer to Table 4-19). This indicates that for many

owners, the most efficient way for them to improve project safety performance is in

selecting safe contractors, instead of actively participating in and monitoring safety.

Table 4-19. Comparison of the ranks of selecting safe contractor (Q232) v.s. participating
in and monitoring safety (Q233)
N Mean Rank Sum of Ranks Sign.
Negative Ranks 7(a) 15.36 107.5
0.04
Positive Ranks 20(b) 13.52 270.5
Q233 Q232
Ties 4(c)
Total 31
a Q233 < Q232; b Q233 > Q232; c Q232 = Q233
Note: the mean ranks and sum of ranks are the results of Wilcoxon matched-pair signed-rank test (for
repeated measures).

Summary

Based on the statistical analysis of the survey responses, the following hypotheses

were developed for further testing:

S The main type of projects of the owner, which may suggest the main business of
the owner, has strong effects on the owner's commitment to safety and practices to
promote construction safety on their projects. Owners of petrochemical and utility
projects tend to emphasize safety more than owners of manufacturing facilities and
buildings.









* The size of the annual construction budget seems to be an important factor that
impacts the safety practices of owners. Owners with larger construction budgets
are more concerned about construction safety.

* Owners of projects with better safety performances tend to emphasize safety during
the earlier stages of project development. They will emphasize safety performance
more during the selection of contractors, and they will also include more safety
requirements in their contracts.

* Owners with safer projects tend to be heavily involved in project safety
management, especially by implementing various approaches to promote safety
management. Although these approaches may not be extensively used in the
construction industry, they are widely implemented by safe owners with large
construction budgets.

* Instead of trying to contractually avoid liabilities from injuries, owners with safer
projects tend to be more involved with contractors on project safety management,
especially safety training.

* Owners who place greater emphasis on safety tend to address safety in the design
of their projects.

* Among different approaches taken by owners to achieve success in project safety,
owners proactive on safety tend to believe that selecting safe contractors to
construct their projects and emphasizing safety in the design phase are important.

The relationship between the overall safety performances of projects and the

different safety management practices of the owners could not be thoroughly examined,

because only 19 owners provided their injury rate data. Despite this, the statistical

analysis was able to identify some possible relationships between safety performance and

the project type, selection of safe contractors, safety commitment, contractual


requirements, and the safety involvement of owners.














CHAPTER 5
RESULTS: OWNER'S ROLE IN PROJECT SAFETY



In this chapter, the results obtained in the major data collection stage are presented.

The methods of interviews and selection of projects are introduced. The most frequent

practices implemented by owners are then presented. Statistical analysis was conducted

to test associations between project safety performance and owners' practices. Finally, a

model to describe the owner's influence on project safety and a scorecard to evaluate

owner involvement in safety were developed.

The Interviews

The second and major phase of this study was to collect information from

construction projects through interviews to determine the role and influence that owners

have on construction project safety. As mentioned in the description of the research

methodology, the focus of this study was to test the null hypothesis that project safety

performance was independent of the facility owner's involvement, against the alternative

hypothesis that project safety performance was dependent on the owner's involvement.

Project safety performance was the dependent or response variable, and the responses to

different questions were different factor levels or independent variables.

A total of 81 personal interviews were conducted. Each interview took about one

and a half hours. When the data were analyzed, a constraint was imposed such that

projects included in the analysis must have had at least 100,000 hours of worker

exposure. This was to ensure that the safety performance measures gave a reliable









indication of the project safety performance. Since some projects were in their early

stages of construction, they did not have the requisite number of hours to satisfy the

criteria for inclusion in the data analysis. By excluding the projects with fewer than

100,000 hours of worker exposure and those projects for which complete injury data were

not provided, the final analysis included 59 projects.

Safety performances of the projects were measured by using the total number of

OSHA recordable injuries per 200,000 hours of worker exposure, commonly known as

the Total OSHA recordable injury rate (TRIR). Whenever the term TRIR appears, it

should be clear that the measure of safety performance consists of all OSHA recordable

injuries, including lost-time injuries and restricted work injuries. For the 59 projects

included in the analysis, the average TRIR was 1.95, with six projects reporting zero

OSHA recordable injuries. One of the projects reporting a TRIR of zero had amassed

nearly 500,000 hours of worker exposure.

The safety performances of projects were compared with the different responses to

the questions in the questionnaire. The hypotheses were tested to determine if project

safety performances were associated with each of the different responses. The major

statistical methods employed were non-parametric methods, including the Mann-Whitney

U test (for comparison between two factor levels) and the Kruskal-Wallis test (for

comparison between more than two factor levels). The rationale for using non-

parametric methods was the difficulty of satisfying the normality assumption and equal

variance assumption in each factor level, which were required for t-test and analysis of

variance (ANOVA) for comparing means of quantitative variables. By using non-

parametric methods, the purpose was to compare the distribution or medians of project









safety performances for different factor levels. In these non-parametric tests, the

response variable, TRIR, was treated as an ordinal variable, instead of an interval

variable. No assumptions were required for the non-parametric tests. The level of

significance indicated whether the medians of the response variable were different when

responses differed.

As a comparison, the means (or averages) of TRIR of different factor levels were

also compared by using the univariate analysis method, one-way analysis of variance

(ANOVA). The statistical assumptions for ANOVA were: (1) the response variable,

TRIR, had a normal distribution for different factor levels; and (2) the variance of the

response variable in different levels were the same. Since ANOVA is a robust statistical

method, which means the results are still correct even if the assumptions are slightly

violated, the significance level of each ANOVA test is also shown in the tables.

When there were only two factor levels to compare, and the medians or means (or

averages) of the TRIR were in the same direction as the literature suggested (which factor

level might have better safety performances), a one-tail test was conducted. If there were

more than two factor levels included in the test, a two tail test was conducted since the

comparison of the response variable in multiple factor levels (more than two) cannot be

made, even if the null hypothesis is rejected.

It was only when the differences of the TRIR medians were statistically significant

at the 0.1 level (the significance level or P-value was less than 0.1) that the results are

presented, unless specifically noted otherwise. Statistical significance was assumed when

the level of significance was 0.05 or smaller, meaning that there was less than a five

percent probability that the finding was due to chance. Since the study was an uncharted









and exploratory study, the results are also presented if the level of significance was

between 0.05 to 0.1, suggesting a strong tendency toward statistical significance.

Projects Interviewed In The Study

Project size descriptors (as measured in terms of worker hours and in total

contracted project cost) and the TRIRs of the 59 projects included in the final analysis are

shown in Table 5-1. Of these projects, seven were in Canada, three projects were

overseas with U.S. owners, and 49 projects were located in 18 states of the U.S, including

one in the Virgin Islands. The numbers of interviews conducted in different states are

shown in Figure 5-1. Although the safety performances and sizes of the projects ranged

widely, it is apparent that the safety performances of most projects were much better than

the construction industry average of about 7.8 for the year 2001 (BLS, 2003). Thus, the

projects included in this research generally enjoyed much better success in safety than the

construction industry as a whole. Prior research has shown that safer performances were

noted among employers with larger numbers of workers. Thus, in general, larger projects

would be expected to have better safety records than smaller projects. It was in this

context that the analysis was intended to identify those practices of owners that had a

particularly strong impact on the resultant TRIR. It may not be surprising to expect better

safety performances on projects where CII members were the project owners, however,

this research was not restricted to projects involving CII members. This may account for

the relatively broad range in TRIRs. The average TRIR of CII members was reported as

being about 1.03 in year 2000 (CII, 2001), significantly better than the industry average.

This research investigated those owner practices that had a direct impact on influencing

the safety performances realized on large projects. The interviews focused on obtaining

information about the demographics of the projects, the manner in which contractors










were selected, the types of safety-related provisions included in the contracts, and the

extent of owner involvement in safety management during project execution.

Table 5-1. Safety performances and sizes of the projects included in the research
Total worker hours Total estimated cost of Total Recordable Injury
expended the project Rate
Mean 2,426,210 $379,440,000 1.95
Mode 200,000 $15,000,000 0
Std. Deviation 5,155,050 $861,740,000 1.94
Minimum 100,000 $3,500,000 0
Maximum 26,300,000 $5,000,000,000 9.25
Sum 143,146,400 $21,248,700,000 115.11
25 275,000 $15,250,000 0.58
Percentiles 50 627,000 $78,350,000 1.48
75 2,284,000 $306,250,000 2.73
Note: These data are based on 59 projects, however, only 56 projects provided the project cost information.


Figure 5-1. Locations of the projects included in the research

Although the 59 projects provided the information on OSHA recordable injuries

(TRIRs), only 46 were able to provide information on all types of injuries, including lost-

time injuries, OSHA recordables, and first-aid injuries. All the injuries occurring on

projects were tallied by severity category, and the ratio between the different types of

injuries was determined, as shown in the injury pyramid in Figure 5-2. The ratio that








exists between injuries on the basis of severity has been discussed for many years. While

these ratios may vary, the general trend is relatively consistent. Heinrich (1959)

suggested that the ratio between major injuries, minor injuries and no-injury accidents

was 1:29:300. Note that Heinrich used different definitions of accidents than those used

in this research. The pyramid in Figure 5-2 could be simplified by reporting the ratio

between lost-time, OSHA recordable, and first-aid injuries as being roughly 1:10:300.


A Lost-time Injuries


OSHA Recordable Injuries


PF~~I 29.4


First-aid Injuries


Figure 5-2. Injury pyramid for the projects
On the 59 projects, 42 different facility owners are represented, since several

owners provided more than one project for inclusion in the study. Most of the owners

(about 70 percent) were CII owner members with large annual construction budgets.

Many owners were listed in the ENR (2001) top 425 owners. Although the data set

cannot be referred to as representative of all the projects in North America, the projects

were considered to be representative of the large-scale projects.

The Most Common Safety Practices
Before testing whether different practices of owners could make a difference in the

safety performances achieved on their projects, the most popular and frequently used

safety practices were listed. These practices included:









* Prime contractor reported injury statistics to the owner (100%), and the types of
injuries reported were: OSHA recordable IR (100%); lost-time injury (100%); near
misses (88.1%); environmental issues (91.5%).

* The owner's site safety representative was generally an employee of the owner
(88.1%) instead of a consultant (11.9%). This person was generally a member of
the project management team (91.5%), and had authority to stop unsafe work
(94.9%). The job responsibilities of the safety representative generally included:

SMonitoring safety management and performance of the contractor on a daily
basis (89.8%)
SEnforcing safety rules (89.8%)
SConducting site safety inspections and audits (89.8%)
SReviewing safety performance on site and submitting reports to the home
office (88.1%)
SReviewing contractors' safety reports (88.1%)
SCoordinating safety efforts on site (81.4%)
* Owner's site safety representative reviewed the safety performance of the
contractor on a regular basis (98.3%). They would check the project lost-workday
injury rate (94.9%), check the project recordable injury rate (94.9%), and check the
project first-aid injury rate (88.1%).

* When selecting contractors, safety was generally a consideration of most owners
(94.9%). Among the criteria used by owners to evaluate the safety performances of
contractors, the overall quality of the safety program was the most frequently
mentioned subject (88.1%). The owner's evaluation of contractor safety
performance would make a difference between getting the contract or not (88.1%).
If a contractor had some safety statistics of concern, the contractor could show that
they had made major changes in the program and still be considered for contract
award (94.9%).

* Owners had a variety of contractual safety requirements, the most frequently used
provisions include the following:

SContractor must comply with the local, state and federal safety regulations.
(100%)
SContractor must report all lost time injuries to the owner. (98.3%)
SContractor must report all OSHA recordable injuries to the owner. (96.6%)
SContractor must report all injuries to the owner. (96.6%)
SContractor is required to provide specified PPE (hard hats, safety glasses,
gloves). (96.6%)









Contractor must implement a substance abuse program. (96.2% of U.S.
projects only, since drug testing is not allowed in Canada)
SContractor must conduct weekly safety meetings for the workers. (93.2%)
SContractor must comply with safety requirements beyond the OSHA
regulations. (88.1%)
SContractor must participate in site safety audits. (88.1%)
SContractor must implement a permit system when performing hazardous
activities (line breaks, lockout/tagout, excavations, proximity to power lines,
confined space entry, hot work, etc.). (88.1%)
SContractor must submit a site-specific safety plan. (84.7%)
* Owners impose the same safety requirements on subcontractors and lower tier
subcontractors. (91.5%)

* Most owners require specific items to be included in the contractor's safety
program (98.3%), and included the subcontractors in the safety program as well
(94.9%). The items that were required by most owners included:

SRegular safety meetings (94.9%)
SIncident reporting and investigations (93.2%)
SConduct regular safety inspections (91.5%)
STraining on the hazards related to the tasks being performed (89.8%)
SSubstance abuse program (89.8%)
SOSHA specific regulations (88.1%)
SSpecific safety training sessions (86.4%)
SPre-project safety planning (86.4%)
* Other safety practices of owners included:

SDuring the design of this project, construction safety issues were specifically
addressed. (98.3%)
SOwner required every worker on site to receive orientation training. (96.6%)
Project Descriptions And Safety Performance

The size of the project, labor arrangements for the project, type of project, and

other characteristics of the project may all be related to the project safety performance.

These factors may be influenced by the owner to some extent and therefore were also


analyzed.










Shutdown Projects

Although new projects did not show a significantly better or worse safety

performance when compared to renovation projects, shutdown projects (operating plants

that must stop operations to perform upgrade and modification work) were found to have

poorer safety performances. The median TRIR of the eight shutdown projects included

in this research was higher than the median TRIR of the other projects (see Table 5-2).

Shutdowns are characterized as having tight schedules (typically four to eight weeks),

significant amounts of overtime work, frequently working multiple shifts, and generally

having a rapid buildup of the workforce. When workers and managerial personnel work

extended hours for one or two months, the possibility of human errors increases, and so

will the probability of injury causation.


Table 5-2. Injury rates of shutdown projects and all other projects
Type of s M n M ia ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median
projects (1-tail) Sign. (1-tail)
All Others 51 1.80 1.76 1.30
Shutdown 8 2.91 2.76 2.20 0.06 0.07
Total 59 1.95 1.94 1.48

Explanation: In this table, for the Mann-Whitney test, the null hypothesis was: the median TRIR of
shutdown projects was equal to the median TRIR of all other projects. The alternative hypothesis was: the
median TRIR of shutdown projects was larger than the median TRIR of all other projects. The significance
level was 0.07, which shows a strong tendency to reject the null hypothesis and to support the alternative
hypothesis. For the ANOVA test, the results were similar to the Mann-Whitney test except it compared the
mean (or average) TRIR of shutdown projects with the mean of other projects. The significance level of
0.06 suggested a tendency that the mean TRIR of shutdown projects was larger than the mean TRIR of all
other projects. Other tables in this chapter are presented in a similar manner.

Public Or Private Project

The comparison of the TRIRs of public projects and private projects, excluding

the shutdown projects (which were all private projects), is illustrated in Table 5-3. It was

suspected that private projects may have an advantage in achieving better safety

performances than public projects. Since many public projects must be awarded through









the competitive bidding process, public projects are frequently awarded to contractors

without regard to their ability to deliver a safe project. Private owners may take into

account factors other than simply awarding the contract to the lowest bidder.

Nevertheless, some public agencies, especially federal agencies, may require the

contractor to comply with their own safety requirements in addition to the OSHA 1926

regulations. However, the involvement of the owners in the safety management of the

public projects was generally viewed as being minimal, when compared to the extent of

owner involvement on private projects.

Table 5-3. Type of facility owner and safety performance (excludes shutdown projects)
ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median ANOVA Si. Mann-Whit
(1-tail) Sign. (1-tail)
Private 45 1.62 1.63 1.20
Public 6 3.16 2.28 2.67 0.02 0.02
Total 51 1.80 1.76 1.30

Petrochemical Projects And Manufacturing Projects

Petrochemical projects, which accounted for nearly half of the projects analyzed

(30), reported better safety performances than manufacturing projects (see Table 5-4).

Note that the shutdown projects have been excluded in Table 5-5 in order to provide an

accurate depiction of the differences between the safety performances of petrochemical

and manufacturing projects. Petrochemical owners interviewed consistently reported

having strong upper management commitment to construction safety and having

successfully integrated safety into their company culture. They had a clear understanding

of the zero injuries philosophy, and jobsite safety responsibilities were defined to

strengthen the safety culture. A behavior-based safety approach was widely accepted and

implemented on petrochemical projects. Thus, many accidents were avoided by









addressing the front end of the accident causation chain. This will be further discussed in

Chapter 7.

Safety performances on manufacturing projects were consistently not as good as

on petrochemical projects. Note that residential and commercial projects were not

included in this comparison, primarily because only a few such projects were in the entire

sample. From the limited data, it appeared as if the residential and commercial projects

were not as good as the manufacturing projects in the area of safety. It should be

mentioned that owners of many manufacturing projects were aggressive in their efforts to

improve project safety performance. Several respondents stated that the enhanced

emphasis on safety in the manufacturing sector was a relatively recent initiative. Since it

takes time to be successful in making significant changes in the safety culture of a

company, it may be only a matter of time before additional improvements in safety

performance are realized on manufacturing projects. Although their safety performances

were not as good as the petrochemical projects, the manufacturing projects in this study

were already much better than the overall construction industry TRIR average of 7.8

(BLS, 2003).


Table 5-4. Type of project facility (including shutdown projects) and injury rates
ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median a Sg. (-ta
(1-tail) Sign. (1-tail)
Petrochemical 30 1.23 1.04 1.03
Manufacturing 19 2.99 2.44 2.53 <0.01 <0.01
Total 49 1.91 1.91 1.33


Table 5-5. Type of project facility (excluding shutdown projects) and injury rates
ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median ANOV Sign. Mann-Whitn
(1-tail) Sign. (1-tail)
Petrochemical 25 1.11 1.00 0.84
Manufacturing 16 2.66 2.07 2.47 <0.01 <0.01
Total 41 1.71 1.67 1.20









Union Projects Or Open Shop Projects

There was a fairly even distribution of the number of open shop projects, union

shop projects, and merit shop projects in the study. In general, owners may have little to

say about employing union shop or open shop contractors, as the local labor conditions

often dictate the type of firms that are available to perform the work. Thus, in strong

union areas, union shop contractors would generally be employed. Merit shop projects

are those in which the labor arrangements are not a consideration in the selection of

contractors. It is on the "merit shop" projects that both union shop and open shop

contractors may be employed at the same time. Although labor conditions are not readily

influenced by facility owners, it is interesting to note that there was a difference between

the safety performances reported on union shop versus open shop projects (see Table 5-

6). Since Canadian projects are all union shop and since international projects are in

environments different from the United States, the table presents information on only the

U. S. projects. Table 5-6 shows that safety performances tended to be better on open

shop projects than on union shop projects. When the union shop project and merit shop

projects were combined as a category, the median TRIR of these projects is significantly

larger than open shop projects. Note that this same pattern of TRIR values was found to

exist when only petrochemical projects (exclusive of shutdown projects) were examined.


Table 5-6. Ty e of labor and safety performance (US projects only)
s M S. D. M ANOVA Sign. (2- Kruskal-Wallis
Counts Mean Std. Dev. Median Sg. (-a
tail) Sign. (2-tail)
Open shop 16 1.32 1.23 1.05
Merit shop 19 2.07 1.51 2.00
0.12 0.13
Union shop 14 2.71 2.64 2.03
Total 49 2.01 1.87 1.62









On projects, especially in the southern portion of the U.S., site safety personnel

commented that union workers were more likely to refuse to follow their safety

instructions than were workers on open shop projects, possibly because they were in a

more secure job position than were open shop workers. It is often alleged that employing

union contractors may cost more, but that the skill levels of the workers are consistently

higher. On the other hand, some contend that open shop workers are more mobile and

are more likely to travel considerable distances from one project to the next. The open

shop workers are reportedly more willing to follow management's instructions to ensure

that their jobs are not placed in jeopardy. Therefore, the owners may simply have to alter

their style of management when promoting their safety initiatives with union contractors

and open shop contractors. Obviously, this can be a delicate issue as any prejudice or

discrimination in the implementation of safety management policies and in the

enforcement of safety regulations could eventually cause more problems. When

implementing the project safety program, the owner must ensure that the program is

implemented firmly and consistently.

Type Of Contract

Two aspects of the project contract were investigated in the study: type of contract

(how the owner would make payments to the contractor); and contracting methods (the

contractual relationship between the owner and contractor). One reason that the contract

type should be considered when addressing project safety is that the contract establishes

the basis on which the owner will make payments to contractor. Essentially, the

payments can be made on the basis of unit prices (unit price contracts), a schedule of

values (lump-sum contracts), or a reimbursement of actual incurred costs (cost plus

contracts). As safety can be enhanced through the concerted efforts of different parties









(owner, contractor, and designer), the manner in which the contract defines their

relationships might readily impact project safety. For example, if the contractor has a

close, long-term relationship with the owner, the owner and the contractor may be

inclined to use a cost reimbursable contract (job order contracting or cost plus). Under

such contracts, the contractor's investment in safety will be reimbursed by the owner.

With greater support for the contractor's efforts in safety, there is less contractor

reluctance to dedicate funds for safety.

If the owner and contractor emphasize safety on the project, there is naturally a

greater probability that the project will have good safety performance. During the

interviews, one safety manager expressed his preference for lump-sum contracts, since he

thought that the contractor could not commence safely with site work until the design was

finished. He argued that fewer design changes during project execution would reduce

hazards on the project. However, another safety manager preferred reimbursable

contracts, since reasonable safety costs were reimbursed and the contractor was never

placed in a situation of having to decide between spending funds to enhance safety or

increasing profits by cutting safety expenditures. This study did not identify any

significant differences between the TRIRs of projects using lump sum contracts and cost

reimbursable contracts.

Different approaches will influence the safety efforts of all parties involved,

including the owner, designer, contractor and subcontractorss. One method of enhancing

safety is to conduct a constructability review as part of the design process (Hinze and

Wiegand, 1992; Hinze and Gambatese, 1996; Jergeas and Van der Put, 2001). This

review helps to coordinate the safety efforts of designers and the work performed on site









(Fischer and Tatum, 1997). Comparisons were made of the safety performances of

design-build projects with projects constructed under other contracting arrangements.

Design-build firms, including engineering, procurement, and construction (EPC) firms,

have a direct incentive to focus on construction safety during the design phase as it is

their own employees that are impacted by their design efforts. In other arrangements, the

design team is often considered to be separate from the construction effort and does not

address construction safety in the design. The owner could impose duties on the design

firm to address construction worker safety, regardless of the contract type, but this was

not examined in the study. Results (see Table 5-7) show that design-build (EPC) projects

had significantly better safety performances than did projects with other forms of

contracting arrangements.

Table 5-7. Ty e of contract used on the project
ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median ANOV Sign. Mann-Whitn
(1-tail) Sign. (1-tail)
Other 44 2.22 2.08 1.68
Design-build 15 1.15 1.14 0.69 0.03 0.03
Total 59 1.95 1.94 1.48

Size Of The Projects

The size of the project may indicate the complexity of conducting site work and

coordinating the related safety efforts. Size might be measured in terms of total

constructed cost, the number of subcontractors on site, the number of workers on site, or

the number of worker hours expended. In this research it was felt that the total number of

worker hours on site gave a more accurate portrayal of the difficulty of implementing a

safety program. On small projects, with fewer hours of worker exposure (less than

100,000 worker hours), simpler designs or engineering plans, and a smaller workforce,

safety efforts can often be more effective. On the other hand, on larger projects (with









more than one million hours of worker exposure) safety performances may also be

expected to be good since a large prime contractor was selected to execute the project.

Large contractors typically have safety programs that are formalized, and they employ

more advanced techniques to promote safety. Despite the complexity involved, safety

performances on the large projects were quite good (see Table 5-8). This held true for all

large projects, including shutdown projects, petrochemical projects, and manufacturing

projects. Strong safety performances on large projects have been reported in other

construction safety research. In general, the very large and quite small contractors have

better safety performances, while medium sized companies have poorer safety

performances (Hinze, 1997).

Table 5-8. Worker hours expended (in thousands of hours) and safety performance
S ANOVA Sign. Kruskal-Wallis
Hours Worked (1000s) Counts Mean Std. Dev. Median ANOVA S.
(2-tail) Sign. (2-tail)
100-200 8 1.72 1.31 1.51
200-1000 28 2.42 2.27 2.03 .
0.193 0.14
1000 up 23 1.45 1.57 0.92
Total 59 1.95 1.94 1.48

Work Shift And Workdays

The number of shifts worked and the number of workdays worked per week are

often dictated by the owner's schedule requirements. Tight deadlines often mean that

shift work or overtime work will be necessitated. On the projects involved in this

research, it was found that projects with one shift had significantly better safety

performances than those with more than one shift (refer to Table 5-9). Projects with four-

day (primarily those working four-tens) or five-day workweeks had significantly better

safety performances than those working more than five days per week (refer to Table 5-

10). From these results, it is reasonable to suspect that fatigue can contribute to










increasing the number of human errors, and that days off for rest and recovery are

necessary to ensure injury-free work.

Table 5-9. Number of shifts worked and safety performance
ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median ANOVA Si. Mann-Whit
(1-tail) Sign. (1-tail)
1 shift 37 1.44 1.45 1.17
2 or 3 shifts 22 2.80 2.35 2.20 <0.01 <0.01
Total 59 1.95 1.94 1.48

Table 5-10. Number of days worked per week and safety performance
Couns M n Dv. ANOVA Sign. Mann-Whitney Sign.
Counts Mean Std. Dev. Median ( (-al
(1-tail) (1-tail)
4 or 5 34 1.54 1.50 1.18
6 or 7 25 2.51 2.32 2.00 0.03 0.03
Total 59 1.95 1.94 1.48

Owner's Selection Of The Contractor

Selecting a safe contractor for project execution is an important function for the

owner to achieve better safety performance. In this research, it was found that most

owners emphasize the importance of selecting safe contractors in the pursuit of the zero-

injury objective. Demonstrated safety performance is a major prerequisite for many

contractors to be awarded contracts. While all owners seem to be aware of the need to

select safe contractors, they differ in the approaches used to accomplish this objective.

Most private owners will not consider awarding contracts to contractors with bad safety

performances. Some owners maintain their own database of the safety performance

history of all parties with whom they have contracted, including contractors,

subcontractors, and vendors. From this, they develop and maintain an approved bidder

list and only these firms are given the opportunity to submit bids on their projects.

Preferred Contractors List

Projects on which contracts were awarded through competitive bidding or through

a negotiated process did not show any statistically significant difference in their safety









performances. Similarly, projects for which the contracts were awarded to contractors on

a preferred contractors list (TRIR median = 1.32) reported safety performances that were

not significantly different from those for which the bidding was open to all interested

contractors (average TRIR=1.92).

Importance Of Safety During Selection Of Contractors

Safety appears to be an important consideration for most owners when contractors

are selected. To get an impression of the importance placed on safety, each owner

respondent was asked about the extent to which safety played a role in the evaluation of

contractors in the selection process. The findings show that projects had better safety

performances when the owners placed a higher priority on safety when evaluating

contractors (see Table 5-11), i.e., projects had better safety performances when the

contractors had already established a proven safety record on past projects.

Table 5-11. Emphasis placed on safety in the overall review of contractors (rating from 1
to 7 with 7 being the most important)
ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median a i. (-t
(1-tail) Sign. (1-tail)
<=5 24 2.33 1.95 2.13
>=6 35 1.69 1.91 1.20 0.11 0.05
Total 59 1.95 1.94 1.48

Criteria Used To Evaluate Safety Performances Of Contractors

Questions were asked about how owners evaluated the past safety performances of

contractors. Possible options for the question were developed based on the research

results of Diaz and Cambrera (1997), Garza et al. (1998), Sawacha et al. (1999), and

suggestions from the CII PT-190 members. The results show that owners used varying

measures of safety performances. One such measure used was the experience

modification rating (EMR) on the workers' compensation insurance. This is a measure

that has been widely used in the past few decades, but has lost favor with some









companies as a viable measure of safety performance. The EMR is considered to be a

lagging indicator in that it represents historical data rather than a current indication of a

contractor's safety commitment (Hinze, et al., 1995). The EMR is based on three years

of loss history, but the value of the EMR is dramatically influenced by the number of

workers employed by the firm and by the hourly wages paid to the employees. Thus, it is

difficult to make valid comparisons between firms; especially if they differ in size

(number of employees) or in the wages they pay. Most owners stipulate that they will not

employ a contractor whose EMR is greater than 1.0. Some owners recognize there are

shortcomings associated with making comparisons between companies on the basis of the

EMR. These owners no longer utilize EMRs because they are not sufficiently accurate

for the owner to evaluate the probable safety performance of a firm for a project.

Analysis shows that the average safety performances of projects where the owners did

not use EMRs were just as good as projects where EMRs were used (see Table 5-12).

Table 5-12. Is the EMR used to evaluate safety performance of contractors? (No
significant differences)
ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median a Sg. (-t
(1-tail) Sign. (1-tail)
No 11 1.45 1.16 1.54
Yes 45 2.13 2.10 1.48 0.16 0.23
Total 56 2.00 1.96 1.51 _
Note: Three owner representatives did not answer the entire question in the questionnaire. These three
projects were not included in the analysis in Tables 5.12 through 5.17.

The TRIR is also a lagging or after-the-fact indicator, but this is a measure that is

widely utilized in the industry (Garza, et al., 1998). The TRIR is a measure of how many

failures have occurred, as each injury represents a failure on the project. The results of

this research show that safety performances of projects were significantly better when the

owners used the TRIR as one of the measurements for evaluating contractors (see Table

5-13). Those owners using the TRIR were asked if a threshold value of TRIR was










established, namely a value above which safety performance was deemed to be

unacceptable. The safety performances of projects with more stringent TRIR

requirements (threshold values no greater than 2) tended to be significantly better than on

projects using more lenient threshold values (threshold values greater than 2), and these

were better than where no TRIR limits were established (see Table 5-14). The findings

indicated that setting a stringent objective resulted in better performance. Conversely,

setting a weaker objective results in weaker performance. Note that the category 'none'

includes projects that did not use the TRIR as a requirement and also projects that used

the TRIR, but that did not establish a specific threshold value.

Table 5-13. Is the TRIR used to evaluate safety performance of contractors?
ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median a Sg. (-ti
(1-tail) Sign. (1-tail)
No 7 2.88 1.82 2.50
Yes 49 1.91 1.96 1.43 0.11 0.04
Total 56 2.04 1.95 1.58

Table 5-14. Threshold value of TRIR set for contractor safety performance
ANOVA Sign. Kruskal-Wallis
Counts Mean Std. Dev. Median a S. (-a
(2-tail) Sign. (2-tail)
<2 12 1.06 0.92 0.99
>=2 21 1.66 1.23 1.92
0.01 0.03
None 23 2.89 2.51 2.18
Total 56 2.04 1.95 1.58

The most proactive owners turn to dynamic measurements of safety performance,

which can better portray the safety performance potential and safety management

capabilities of contractors. Generally, owners will not focus on a single measure, but will

try to assess the overall safety performance of contractors based on a number of

measures. Viable measures include an assessment of the contractor's safety program and

the qualifications of the safety personnel. A thorough and careful investigation into the










underlying safety commitment held by the contractor is conducted to ensure that the

selected contractor will be able to achieve an acceptable safety performance.

Qualifications of the contractor's safety personnel and qualifications of the project

management team were used by some owners for the selection of contractors. The more

proactive owners would review these qualifications by conducting personal interviews

with them and also by making site visits to projects where they were assigned at the time.

The resultant TRIR was found to be lower on projects where the owner's had a practice

of considering the qualifications of the contractor's safety personnel and also the

qualifications of the project management team (see Table 5-15 and Table 5-16).

Table 5-15. Are qualifications of safety staffs reviewed when evaluating contractors?
ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median a Sg. (-ti
(1-tail) Sign. (1-tail)
No 14 2.60 2.32 2.48
Yes 42 1.79 1.81 1.32 0.09 0.06
Total 56 2.00 1.96 1.51


Table 5-16. Are qualifications of the project team reviewed when evaluating contractors?
ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median a S. (-t
(1-tail) Sign. (1-tail)
No 18 3.04 2.54 2.48
Yes 38 1.50 1.40 1.20 <0.01 <0.01
Total 56 2.00 1.96 1.51


Many owners asked contractors to provide copies of the OSHA log for the past

year and any records of OSHA inspections on past projects. The analysis did not find

any significant differences between the safety performances of projects when owners

used or did not use these measures as criteria for the evaluation of contractors.

Further analysis was conducted on the merits of using the following proactive

criteria to measure safety performance.

* Qualifications of the safety staff of the contractor,
* Qualifications of the project management team of the contractor, and










* Quality of the overall safety program of the contractor.
The impact of using none or some of these proactive criteria and of using all three

was examined. The analysis essentially determined the TRIR of those projects when the

owners used all three measures in their assessment of contractors, when owners used less

than three measures, and those that did not evaluate contractors on the basis of safety

when making their selection of the contractor. It should be noted that most owners (33)

used all three proactive criteria. The relationship between the number of proactive

criteria used and the resultant safety performances is shown in Table 5-17. From this, it

is evident that the use of all three proactive criteria is associated with better safety

performances than when none, one or two measures are used.

Table 5-17. Number of proactive criteria utilized for evaluating contractors on safety
ANOVA Sign. Kruskal-Wallis
Counts Mean Std. Dev. Median A A Sign K kal-Walli
(2-tail) Sign. (2-tail)
Safety is not a factor 5 3.73 3.22 3.25
<=2 18 2.45 2.17 2.2602 0
0.02 0.04
3 33 1.48 1.39 1.19
Total 56 2.00 1.96 1.51

By using more proactive criteria for safety evaluations, owners make it clear that

safety is important. As shown in Table 5-17, when safety did not influence the contract

award or when fewer proactive criteria were used, safety performances on the projects

were not as good.

Owner's Contractual Arrangement

The construction contract is the legal document that specifies the responsibilities of

different parties involved in the project (Hinze, 2001). Many construction contracts

include provisions for safety that specify additional requirements and responsibilities

concerning safety. These provisions are often found in the general conditions or the

supplementary provisions of the contract. In this study, the contracts between the owners









and contractors were investigated. The primary focus was on the safety requirements

established by the owners. Findings show that contractual arrangements influence the

safety performances realized at the project level.

The construction contract may impose a variety of requirements on the contractor.

Only the provisions clearly focused on safety were examined in this research. Many

different requirements were noted to be included in the contracts. While other contract

requirements may have a favorable impact on safety, the nature of this influence could

not be established when all respondents employed that contract requirement. There were

questions related to seventeen different types of contract requirements. The frequency

use of each contract provision is shown in parenthesis (see also appendix b).

* Contractor must comply with the local, state and federal safety regulations. (100%)

* Contractor must comply with safety requirements beyond the OSHA regulations
(88.1%)

* Contractor must place at least one full-time safety representative on the project
(83.1%)

* Contractor must submit the resumes of key safety personnel for the owner's
approval (71.2%)

* Contractor must provide specified minimum training for the workers (62.7%)

* Contractor must report all lost time injuries to the owner (98.3%)

* Contractor must report all OSHA recordable injuries to the owner (96.6%)

* Contractor must report all injuries to the owner (96.6%)

* Contractor must include personnel from the owner in coordination meetings
(67.8%)

* Contractor must submit subcontractor list to owner for approval (79.7%)

* Contractor must implement a substance abuse program. (93.2%)

* Contractor must participate in site safety audits (88.1%)









* Contractor must conduct weekly safety meetings for the workers (93.2%)

* Contractor must submit a site-specific safety plan (84.7%)

* Contractor must submit a safety policy signed by its CEO (52.5%)

* Contractor is required to provide specified PPE (hard hats, safety glasses, gloves)
(96.6%)

* Contractor must implement a permit system when performing hazardous activities
(line breaks, lockout/tagout, excavations, proximity to power lines, confined space
entry, hot work, etc.) (88.1%)

Note that some contract provisions appear in all or nearly all contracts used on the

projects. For the purpose of this research, only those requirements related to better safety

performances are presented. Two particular provisions were noted to be associated with

better safety performances. One was that better safety performances were reported on

projects where the contractor was required to assign at least one full-time safety

representative to the construction site (see Table 5-18). Also, better safety performances

were reported on projects where the contractor was required to submit the resumes of the

key safety personnel (to be assigned to the project) for the owner's approval (see Table 5-

19). Impacts of other leading indicators used in the contracts are shown in Figure 5-3,

although the differences between TRIRs were not statistically significant. When the total

number of safety requirements in the contract was calculated, it was found that those

projects having contracts that had more requirements had significantly better safety

performances (see Table 5-20).

Table 5-18. Contract requires the contractor to place at least one full-time safety
representative on site?
ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median a Sg. (-ti)
(1-tail) Sign. (1-tail)
No 10 2.96 2.64 1.87
Yes 49 1.75 1.72 1.30 0.04 0.08
Total 59 1.95 1.94 1.48










Table 5-19. Contract requires the contractor to submit the resumes of key safety personnel
for the owner's approval?
ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median a ig. (-ti
(1-tail) Sign. (1-tail)
No 17 3.01 2.34 2.81
Yes 42 1.52 1.58 1.20 <0.01 <0.01
Total 59 1.95 1.94 1.48


iContractor must provide specified minimum safety training to workers
*Contractor must submit a site-specific safety plan
OContractor must submit a safety policy signed by CEO


2.50


2.00


1.50


1.00


0.50


2.30


1.96


i.00 .I.
1.22


0.00 -
No Yes

Figure 5-3. Other leading indicators used in the project contract (not statistically
significant)

Table 5-20. Relationship between TRIR and total number of contractual safety
requirements
ANOVA Sign. Mann-Whitney
Counts Mean Std. Dev. Median i (-ta
(1-tail) Sign. (1-tail)
<=15 38 2.38 2.24 2.13
16 and 17 21 1.17 0.74 1.2 <0.01 0.02
Total 59 1.95 1.94 1.48

Based on the statistical analysis, five requirements of contracts were identified as

being "leading indicators", as these can be used in a manner to predict safety

performances. Leading indicators are practices that are associated with improved safety

performances. These are listed as follows:

* Contractor must place at least one full-time safety representative on the project









* Contractor must submit the resumes of key safety personnel for the owner's
approval

* Contractor must provide specified minimum training for the workers

* Contractor must submit a site-specific safety plan

* Contractor must submit a safety policy signed by its CEO

The data analysis showed that there was a relationship between the number of

leading indicator safety requirements and the safety performances realized on the various

projects (see Table 5-21). Projects on which more leading indicator safety requirements

were included in the contracts had better safety performances than projects with fewer

leading indicator safety requirements.

Table 5-21. Number of leading indicator safety requirements included in the contract
ANOVA Sign. Kruskal-Wallis
Counts Mean Std. Dev. Median a Sg. -ai
(2-tail) Sign. (2-tail)
1 and 2 11 2.77 1.89 2.73
3 and 4 34 1.99 2.16 1.64
0.14 0.07
5 14 1.22 0.99 1.18
Total 59 1.95 1.94 1.48
Owner Involvement During Project Execution

In addition to promoting project safe performance through the careful selection of

contractors and the inclusion of carefully selected safety provisions in the contract,

owners can be active participants in safety management during project execution.

Several questions were asked about specific practices of owners that were expected to

favorably influence safety performances of projects. These practices included owner

participation in safety recognition programs, monitoring of safety performance, funding

safety initiatives, accident reporting, accident investigations, safety training and

orientation programs, and so on. Only those practices that showed a clear influence on

safety performances are presented.