www.seipub.org/ce
Construction Engineering Volume 3, 2015 doi: 10.14355/ce.2015.03.006
Critical Success Factors of Construction Site Safety Management in Taiwan Wei Tong Chen*1, Jia-Kai You2, Hong Long Chen3 Construction Dept. of Construction Engineering, National Yunlin University of Sci. & Tech., No. 1, Section 3, University Rd., Touliu, Yunlin County, Taiwan 1, 2
3
Dept. of Business and Management, National University of Tainan, No. 33, Sec. 2, Shu-Lin St., Tainan 700, Taiwan chenwt@yuntech.edu.tw; 2M10216302 @yuntech.edu.tw; 3along314@mail.nutn.edu.tw
*1
Abstract This study explores the success factors (SFs) of safety management in Taiwan’s construction sites using Exploratory Factor Analysis (EFA). Research results show that four critical successful factors (management level attitude, work environment safety, worker safety, and correction & prevention measures) have a significant influence on the success of safety management in Taiwan’s construction sites. It is also found that ensuring a safe working environment is the primary focus of safety management in Taiwan’s construction sites while safety awareness and worker training are generally inadequate. Keywords Construction Site Safety Management, Critical Success Factors, Exploring Factor Analysis
Introduction The construction industry is a complex configuration of multi-level subcontractors and personnel. Usually, a construction project entails multiple subcontractors and diverse construction trades. Proper management of personnel, equipment, and environment are of particular importance. However, contractors often respond to time and cost constraints by reducing investment in safety management, which can lead to increased frequency and severity of construction site accidents. Construction sites feature a range of environmental characteristics, and construction workers need to adjust to different environments and operating procedures whenever they move from one site to another, thus significantly increasing the potential for accidents. According to Chi and Han (2013), construction accidents resulted from factors including unsafe working conditions and worker unsafe behavior. Unsafe working conditions can arise from problems involving safety equipment, warning signs and systems, and site conditions. Unsafe worker behavior can be categorized as being due to cognitive deficiencies, errors and bad habits. According to Feng (2013), safety investment can be classified into (1) basic safety investment, and (2) voluntary safety measures investment. Most contractors in Taiwan give priority to basic safety investment (that is, providing safety protection facilities in accordance with government laws and regulations). Generally, contractors only invest around 0.3% of the total project cost for safety management, which is insufficient to ensure proper safety. Despite the range of conditions found at different construction sites, site safety management can be enhanced by identifying key factors affecting the success of site safety management, clarifying the relationship between factors and the overall environment, and developing effective countermeasures. This study surveyed some experienced construction safety management professionals, superintendents and senior supervisors, using the EFA to extract the critical factors which impact site safety management. Safety Management in Construction Sites Most construction accidents occur on construction sites. Although the construction industry has focused more attention to safety management than that in the past, such concerns are still given lower priority than project duration and costs. Recent studies on safety management can be categorized as focusing on security risk management, safety management systems and safety performance. Using factor analysis and analytic hierarchy
30
Construction Engineering Volume 3, 2015
www.seipub.org/ce
process (AHP), Haadir et al. (2011) identified critical success factors for construction safety management. They investigated safety management improvements from three different perspectives, including: labour and management, safety prevention and control mechanisms, and safety equipment and training. Using various methods, Han et al. (2014) and Wu et al. (2015) identified related measures and factors useful for enhancing construction safety management performance. They also investigated the relationship between accident cause and effect. Tรถrner et al. (2009) described construction industry safety standards in terms of worker and site manager behavior. They also investigated the impact of management team values and personal attitudes on safety management performance, thus establishing an overall safety management system. Using hierarchical analysis and conceptualization techniques, Shin et al. (2014) established a safety management system model and identified impact factors including personal cognition, communication, personal protection equipment, and personal attitudes. An effective risk assessment model was developed by Fung et al. (2010) to enhance safety knowledge and risk perception among construction personnel. Puerto et al. (2014) explored perceptions of safety culture among construction personnel and found a significant safety awareness gap between workers in residential, commercial, and large public construction projects. Chi et al. (2013), construction hazards are generally divided into unsafe working conditions and unsafe acts. Unsafe working conditions vary in accordance with construction project conditions and environmental conditions. Working conditions and accident incidence are highly correlated. Research Methodology This work examines documents and theories regarding construction project safety management, and distributes a questionnaire to construction safety professionals to further analyze the factors involved in successful construction site safety management. The questionnaire compares the roles of different professionals in construction projects, and the effects of their various project attributes provide a reference for site safety management. This work also ranks critical success factors (CSF) of construction safety site management (CSSM). Questionnaire Development The questionnaire used a 5-point Likert scale, running from 1 (extremely unimportant) to 5 (extremely important). The second evaluation was conducted to ensure instrument validity and effectiveness. The original questionnaire design included 45 questions regarding construction site safety management success factors (SFs). In this work, validity was used to ensure accurate measurement of the characteristics and factors. The measurement results and forecasting characteristics are used to represent the degree of validity. Aspects of questionnaires from previous studies (Black et al., 1998; Chan, 2004) were incorporated to assess important factors related to safety management, performance evaluation, and SFs. Prior to distribution, the questionnaire was reviewed by five experienced construction safety experts who were asked to comment on readability, comprehensiveness, and accuracy. The corrected instrument contained 33 structural survey questions representing 33 SFs. Questionnaire Distribution The survey was distributed by website, e-mail, mail, fax, and personal delivery to construction safety professionals and experts in Taiwan. A total of 557 questionnaires were distributed during March 2015, and 284 valid responses were retrieved (a return rate of 55.25%). The majority of respondents were college graduates, between 31-40 years old, with 5-10 years working experience, and holding safety management certification issued by the Taiwanese government. SPSS V22 was used to perform further statistical analysis. Extracting CSFs of CSSM Factor Analysis Factor analysis was used to explore and detect the underlying relationships among the CSFs. This statistical
31
www.seipub.org/ce
Construction Engineering Volume 3, 2015
technique identifies a relatively small number of factors that can be used to represent relationships among sets of multiple interrelated variables. The appropriateness of the factor analysis for the factor extraction needs to be tested in various ways. Factor analysis can be used either in hypothesis testing or in searching for constructs within a group of variables. Factor analysis is a series of methods for identifying clusters of related variables and hence is an ideal technique for reducing numerous items into a more easily understood framework (Norusis, 2000). Factor analysis focuses on a data matrix produced from collecting numerous individual cases or respondents. When applying factor analysis, the total number of survey samples should be at least 5 times the number of questionnaire variables (Steven, 2002). This work follows Steven’s suggestion to perform EFA using SPSS statistical software V22 to explore the underlying constructs of the identified CSFs for site safety management. Thirty three identified SFs were subjected to factor analysis using principal component analysis and varimax rotation. Principal component analysis is commonly used in factor analysis, and involves generating linear combinations of variables through factor analysis to explain as much of the variance present in the collected data as possible. Such analysis summarizes the variability of the observed data via a series of linear combinations of ‘‘factors’’. Each factor can thus be viewed as a ‘‘supervariable’’ comprising a specific combination of the actual variables examined in the survey. The advantage of principle component analysis over other factor analytic approaches is that the mathematical representation of the derived linear combinations makes it unnecessary to use questionable causal models (Johnson, 1998). Results of the FA To investigate the relationship among the 18 identified CSFs, the correction analysis is performed to clarify the correlation of CSFs. Table 1 lists the matrix of the correlation coefficients among the 18 CSFs. The matrix is automatically generated along with the factor analysis. The correlation coefficients demonstrate that the SFs share common factors. The Bartlett test of sphericity is 2391.074, and the associated significance level is 0.000, indicating that the population correlation matrix is not an identity matrix. Moreover, the value of the Kaiser–Meyer–Olkin (KMO) measure of sampling accuracy is 0.915, significantly exceeding 0.5 and thus is considered highly acceptable. Figure 1 confirms that a four-factor model should be sufficient for the research model. To avoid confusion among the extracted factors and CSFs, the extracted factor was renamed as a ‘‘cluster’’. Four clusters with eigenvalues greater than 1 were extracted. Table 2 lists the cluster matrix following varimax rotation. Each of the CSFs weighs heavily on only one of the clusters, with the loading exceeding 0.5 (to round off). Generally, the loadings and the interpretation of the factors extracted were reasonably consistent. Table 2 also lists the final statistics of the principal component analysis, and the clusters extracted comprise 63.477% of the variance. TABLE 1 CORRELATION MATRIX OF CSF FOR SITE SAFET MANAGEMENT
Factor CSF18 CSF17 CSF16 CSF15 CSF14 CSF13 CSF12 CSF11 CSF10 CSF9 CSF8 CSF7 CSF6 CSF5 CSF4 CSF3 CSF2 CSF1 CSF18 1 CSF17 0.403 1 CSF16 0.360 0.393 1 CSF15 0.460 0.501 0.448 1 CSF14 0.197 0.215 0.192 0.245 1 CSF13 0.207 0.226 0.202 0.257 0.490 1 CSF12 0.228 0.248 0.222 0.283 0.540 0.567 1 CSF11 0.215 0.234 0.209 0.267 0.509 0.535 0.588 1 CSF10 0.264 0.287 0.257 0.328 0.299 0.314 0.346 0.326 1 CSF9 0.298 0.324 0.290 0.370 0.337 0.355 0.390 0.368 0.493 1 CSF8 0.301 0.328 0.294 0.374 0.342 0.359 0.395 0.373 0.499 0.563 1 CSF7 0.256 0.279 0.250 0.318 0.291 0.305 0.336 0.317 0.424 0.479 0.485 1 CSF6 0.318 0.346 0.310 0.395 0.360 0.379 0.417 0.393 0.526 0.594 0.601 0.511 1 CSF5 0.289 0.315 0.282 0.359 0.306 0.322 0.355 0.334 0.363 0.409 0.414 0.352 0.437 1 CSF4 0.303 0.330 0.295 0.376 0.321 0.337 0.371 0.350 0.379 0.428 0.434 0.368 0.457 0.526 1 CSF3 0.320 0.349 0.312 0.398 0.340 0.357 0.393 0.370 0.402 0.453 0.459 0.390 0.484 0.557 0.583 1 CSF2 0.306 0.333 0.298 0.380 0.324 0.340 0.375 0.353 0.383 0.432 0.438 0.372 0.461 0.531 0.556 0.589 1 CSF1 0.307 0.334 0.299 0.381 0.325 0.342 0.376 0.354 0.385 0.434 0.440 0.374 0.463 0.534 0.558 0.591 0.564 1 Notes: KMO measure of sampling adequacy = 0.915; Bartlett test of sphericity = 2391.074; degree of freedom = 153; significance = 0.000
32
Construction Engineering Volume 3, 2015
www.seipub.org/ce
FIG. 1 TOTAL VARIANCE ASSOCIATED WITH EACH FACTOR TABLE 2 CLUSTE OF MATRIX AFTER VARIMAX ROTATION AND PRINCIPAL COMPONENT ANALYSIS
Critical Success Factors Site managers’ safety decisions are fully supported and contractors are responsible for those decisions (CSF1) Contractor and labor establish mutual trust by implementing safety dialogue (CSF2) Contractor actively emphasizes attention of safety (CSF3) Various contractor divisions regularly coordinate on safety issues (CSF4) Site managers pay attention to worker behavior and mentality (CSF5) Qualified safety equipment and measures in place on site (CSF6) Qualified safety personnel on site (CSF7) Danger warning signs/systems in place (CSF8) Safety equipment regularly maintained and deficiencies repaired immediately (CSF9) Safety personnel are familiar with the location and use of safety equipment (CSF10) Workers are equipped with good work habits/experience and health conditions (CSF11) Workers adopt safe working practices (CSF12) Workers are capable of identifying hazards (CSF13) Workers will remind and help each other when jointly performing risky tasks (CSF14) Management staff make inspection tours of working environment, and actively improve safety problems (CSF15) Regular worker health examinations (CSF16) Active daily communication between management and workers and correction of unsafe behavior (CSF17) Workers follow managers’ competent advice and command (CSF18) Eigenvalues Percent of variance Cumulative percentage of variance
Cluster 2 Cluster 1 Management Safe working level manner environment
Cluster 3 Cluster 4 Worker Correction & Cronbach’s α safety prevention beheavior measures
.781 .699
0.863
.698 .693 .687 .709 .687 .684 0.841
.669 .647 .790 .763 .749
0.820
.685 .761 .698
0.741
.648 7.514 41.745 41.745
1.557 8.649 50.394
1.217 6.761 57.155
.645 1.133 6.292 63.447
Among the four aspects, Safe working environment is identified as the most important aspect followed by Management level attitude and Correction & prevention measures. Worker safety beheavior is ranked the least important factor. In terms of Management level attitude, CSF3 is the most important measure indicating that the perspectives of contractor will influence the actual performance of site management staff. Worker safety beheavior is ranked last
33
www.seipub.org/ce
Construction Engineering Volume 3, 2015
mainly because the respondents believe that the Management level attitude, Safe working environment and Correction & prevention measures can all be improved by improving company monitoring and remedial action. On the other hand, although Worker safety beheavior could be improved via the implementation of Tool Box Meeting and Labor education/training, contractors have difficulty influencing worker safety congition. CSF14 is the most important factor underneath this dimension. This is mainly because workers are close when performing construction jobs, and will impart personal safety experience to colleagues and partners engaged in dangerous tasks. Conclusions Although Taiwan has improved monitoring and labor inspection effectiveness, construction accidents still abound and a comprehensive investigation for preventing construction site accidents is urgently needed. Based on the recent site safety management literature, this study used expert interviews to compile 33 site safety management factors. Using exploratory factor analysis, 18 critical site safety management factors were extracted and classified into four clusters namely: management level attitude, work environment safety, worker safety, and correction & prevention measures. Correlation analysis shows moderate correlation between the factors of overall dimensions in a secure operating environment, with CSF11 and CSF13 being the most relevant (0.601). These findings suggest that contractors should focus on improving their safety oversight strategy and the safety atmosphere. Management attitudes, work environment safety, worker safety and behavior and improved preventive measures should be the top priorities for enhancing overall construction site safety performance. ACKNOWLEDGMENT
The authors would like to thank Taiwan Ministry of Science and Technology for financially supporting this research under Contract No. 103-2221-E-224 -069. REFERENCES
[1] Black, Carolynn, Akintoye, Akintola, Fitegerald, Eamon., “An Analysis of Success Factors and Benefits of Partnering in Construction.” International Journal of Project Management 18( 1999): 423–34. [2] Chan, Albert .P.C., Chan, DanielW.M., Chiang, Y.H., Tang, B.S., Chan, EdwinH.W., Ho, Kathy S.K., “Exploring Critical Success Factors for Partnering in Construction Projects.” Journal of Construction Engineering and Management 130(2004): 188–98. [3] Chi, Seokho, Han, Sangwon, and Kim, Dae Young, “Relationship between Unsafe Working Conditions and Workers’ Behavior and Impact of Working Conditions on Injury Severity in U.S. Construction Industry.” Journal of Construction Engineering and Management 139(2013): 826–38. [4] Chi, Seokho, Han, Sangwon. Analyses of systems theory for construction accident prevention with specific reference to OSHA accident reports. International Journal of Project Management 31(2013), 1027-41. [5] Feng, Yingbin, “Effect of Safety Investments on Safety Performance of Building Projects,” Safety Science 59(2013): 28–45. [6] Fung, Ivan W. H, Tam, Vivian W.Y., Lo, Tommy Y., and Lu, Lori L.H., “Developing a Risk Assessment Model for Construction Safety.” International Journal of Project Management 28(2010): 593–600. [7] Haadir-Al, Saeed, and Panuwatwanich, Kriengsak, “Critical Success Factors for Safety Program Implementation among Construction Companies in Saudi Arabia.” Procedia Engineering14(2011): 148–55. [8] Han, SangUk, Saba, Farzaneh, Lee, SangHyun, Mohamed, Yasser, and Peña-Mora, Feniosky, “Toward an Understanding of the Impact of Production Pressure on Safety Performance in Construction Operations.” Accident Analysis and Prevention 68(2014): 106–16. [9] Johnson, Dallas E., Applied Multivariate Methods for Data Analysts. New York: Duxbury, 1998. [10] Norušis, Marija J., SPSS 10.0 Guide to Data Analysis. New York: Prentice Hall, 2000.
34
Construction Engineering Volume 3, 2015
www.seipub.org/ce
[11] Puerto, Carla Lopez del, Clevenger, Caroline M., Boremann, Kane, and Gilkey, David P., “Exploratory Study to Identify Perceptions of Safety and Risk among Residential Latino Construction Workers as Distinct from Commercial and Heavy Civil Construction Workers.” Journal of Construction Engineering and Management 140(2013): 04013048. [12] Shin, Mingyu, Lee, Hyun-Soo, Park, Moonseo, Moon, Myunggi, and Han, Sangwon, “A System Dynamics Approach for Modeling Construction Workers’ Safety Attitudes and Behaviours.” Accident Analysis and Prevention 68(2014): 95–105. [13] Stevens, James Paul. Applied Multivariate Statistics for the Social Science, 4th edition, Mahwah, New Jersey: Lawrence Erlbaum Associates, 2002. [14] Törner, Marianne, and Pousette, Anders, “Safety in Construction– A Comprehensive Description of the Characteristics of High Safety Standards in Construction Work, from the Combined Perspective of Supervisors and Experienced WSrkers.” Journal of Safety Research 40(2009): 399–409. [15] Wu, Xianguo, Liu, Qian, Zhang, Limao, Skibniewski, Miroslaw J., and Wang, Yanhong, “Prospective Safety Performance Evaluation on Construction Sites.” Accident Analysis and Prevention 75(2015): 58–72.
35