Countries Around World Go For Development

Published Date: 02 Nov 2017

1.1 Introduction

This chapter discuss on the background of the study, the key problems of the study, the research questions as well as the research objectives. Moreover, the research framework and also the scope of the research will be discussed in this chapter. At the end of the chapter, the significant of the research was also being discussed.

1.2 Background of the study

In this globalization era, almost all of the countries around the world go for development (Tharaldsen et al., 2010). The globalization had resulted in the attraction of awareness of many parties which put safety issues a major issue around the world (Li et al., 2009). One point to be highlighted, however, is the accidents statistics reported as many as 591000 cases of non-fatal injuries by years 2011/2012 in United Kingdom (Health and Safety Executive, 2012a). Expectedly, United States recorded 760000 cases for years 2011 (U.S. Bureau of Labor Statistics, 2012). Health and Safety Executive (2012b) provided statistics which shows the trend of workplace accidents all over the years. Based on Health and Safety Executive (2012b), there is a decreasing trend of workplace accidents in United Kingdom. However, the total reported cases of workplace accidents did not reassuring. In light of these statistical numbers, the issues concerning safety had become the central issue for the researchers (Choudhry and Lingard, 2009).

Safety performance is the subsystem that include in the organizational performance (Wu et al., 2008; Tharaldsen et al., 2010). Wameedh et al. (2011), proposed that there are many factors that can influence safety performance, for instance, environmental, organizational, and psychological factors. Recent evidences (Chen and Jin, 2012; Choudhry et al., 2009; Yang et al., 2010) suggest that safety performance should be measured by using proactive and reactive indicators. Reactive indicators are incidents rates and compensation costs. Whereas, proactive indicators are safety behaviour and hazard identification (Chen and Jin, 2012; Wu et al., 2008; Powell, 2009).

Today, the knowledge to perform or to fulfil the jobs had never been enough (Hassanein and Hanna, 2008; Ministry of Finance Malaysia, 2012). The safety training practices can actually help in reducing the occupational accidents when the trainings are related to safety (Wrede, 2009). It can be clearly seen by the evidence provided by Cheng et al., (2012), which the likelihood of accidents can be reduced by the application of safety training practices. Therefore, the employers and also the government had invested tons of moneys for safety training. Encouragingly, the importance of safety training was addressed by Ministry of Finance Malaysia (2012), with the decision of allocating up to 21 percents from the Budget 2013 in generating human capital excellence, through safety training.

With regard to this, the government will set up some training scheme under the budget provided for the year 2013, for instance, Dual National Training Scheme and 1 Malaysia Training Scheme (SL1M) that covers all of the training available, including safety training. Due to this situation, the employees will need to upgrade or develop themselves from time to time. Refer to the evidences provided by Yang et al. (2010), training is a vital part which the employers are able to identify the strengths of their employees and make use of it to reach a better performance. Taken together, a total of RM 200 million was provided by Ministry of Finance Malaysia (2012) for the first time in contribute to the Social Security Organization (SOCSO) fund under PERKESO for the claimants of workplace accidents. However, in order to deal with workplace accidents, safety training alone is far more than enough.

Turning to the safety leadership, it seems that there is a significant influence of the quality of safety leadership towards safety performance (Wu et al., 2008; Hsu et al., 2012; Wameedh et al., 2011). In a recent study by Wameedh et al. (2011), workplace accidents are preventable and safety leadership is one of the approaches in achieving it. Krause and Weekley (2005), provide an in-depth explanation on the influence of safety leadership towards safety performance. Krause and Weekley (2005) argue that there is a great and positive effect of the safety leaders, who comprise themselves with excellence leadership and commitment towards safety issues. Yet, when it comes to the safety performance, the belief of employees is the focal (Krause, 2005).

The perceived and belief on the safety over productivity is greatly affecting the performance of employees (Colley et al., 2013; Shang and Lu, 2009). Safety climate had been defined as the reflection and manifestation of the safety culture of the organization and the perceived safety values of the organizations. (Mearns et al., 2003; Zohar, 1980). Recently, Lingard et al., (2010a), reveals an arguement on safety climate in construction in which it came out that the levels of safety performance are linked with the levels of safety climate consistently. According to Brondino et al., (2012), safety performance can be categorized into two components, which are safety compliance and safety participation. Safety compliance refers to the performing of tasks according to the safety rules and regulations (Lingard et al., 2010a; Brondino et al., 2012; Tharaldsen et al., 2010). On the other hand, safety participation is the engagement of the employees and the initiative of the employees to perform the tasks safely (Zohar, 2008; Lingard et al., 2010b).

1.3 Problem statement

The key problem which triggering this research is the workplace accidents that are critical. To date, workplace accidents in Malaysia experiencing a trend of decreasing (2000= 98281 cases; 2003= 81003 cases; 2006= 68008 cases; 2008= 56095 cases; 2011= 24290) from year to year (Department of Safety and Health, 2012; Social Security Organization, 2011). However, there is a controversy trend on the accidents percentages of manufacturing sectors. Statistical evidences shown that the percentages of total accidents in manufacturing sectors suffered a raise despite the dropped of total accidents from 2008 to 2011 (2008= 33.94%; 2011=67.89%) after the decrease trend from 2000 to 2008 (2000= 43.67%; 2003= 41.85%; 2006= 39.80%; 2008= 33.94%). The evidences indicates that there is a needs to identify the problems of safety issues in manufacturing sector (Social Security Organization, 2011).

While Malaysia developed constantly to climb on the position as a strong economic participant in the world, there are challenges of safety issues that we are bound to face. (Ministry of Human Resources Malaysia, 2009). The numbers of industrial accidents have always been high, especially when it involves in manufacturing industries. As at August 2012, there are 102 cases of permanent disablement accidents and 1112 cases of non-permanent disablement accidents reported for manufacturing sector (Department of Occupational Safety and Health (DOSH), 2012). As highlighted by the Ministry and DOSH, safety performance can be affected by safety leadership, safety training practices and safety climate and it should be further study (Mohamed, 1999; Nahrgang et al., 2007; Lingard et al., 2010a; Yang et al., 2010).

Another problem that needs to be addressed in this study is there is still issue which the organization did not give sufficient safety training to the employees. Refer to the studies reviewed so far, they pointed out that the improving of training practices is required to improve safety climate. Hence, there is a need to further explore the influence of safety training practices towards safety climate. (Arcury et al., 2012; Hinze, 2005; Hassanein and Hanna, 2008; Chen and Jin, 2012; Wameedh et al., 2011).

Another weakness which needs to be addressed is the ignorance of safety leadership in improving safety performance. In Malaysia, organizations would believe that it is enough to outlined safety plan without any leadership in making it become success. Although there are many researches study on leadership in safety performance (Carmeli et al., 2010; Adamshick, 2007; Petersen, 2004; Zohar, 2002; Ginsburg et al., 2010), there are lack of research studied using safety leadership. (Wu et al., 2006; Wu et al., 2007; Wu et al., 2008; Wu et al., 2010). The commitment of leaders towards safety climate that involved safety leadership need to be explained in future studies. (Wu et al., 2008; Zohar and Luria, 2010; Zohar, 2002). Rowley (2009) proposed that future research should be conducted to examine the influence of safety leadership to safety performance with safety climate act as mediating variables as safety leadership have great impact on safety performance. (Hinze, 2005; Huang and Hinze, 2006; Lingard et al., 2010a). To date, there had been little discussion about safety leadership although transformational and transactional leadership had been widely discussed.

A serious issue with however, is the lack of attention on safety climate that are critical in influencing safety performance. There are several studies (Brondino et al., 2012; Martinez-Corcoles et al., 2012; Zohar, 2008; Choudhry et al., 2009; Huang et al., 2012) measured the significant of safety climate on safety performance. However, Zohar, (2010), argued that after 30 years of many research on safety climate, there is still an ambiguity of the effect of safety climate on safety performance.

In general, therefore, it seem that there is a need to conduct a study on the effect of safety training practices, safety leadership and safety climate to safety performance in manufacturing sector Malaysia and therefore this study set out to investigate the effects of safety training practices, safety leadership and safety climate to safety performance.

1.4 Research Question

Does safety training practice affect safety performance?

Does safety leadership affect safety performance?

Do the effect of safety climate mediate the relationship of safety training practices on safety performance?

Do the effect of safety climate mediate the relationship of safety leadership on safety performance?

1.5 Research Objective

To identify the relationship between safety training practices and safety performance.

To identify the relationship between safety leadership and safety performance.

To identify the effect of safety climate mediate the relationship of safety training practices on safety performance.

To identify the effect of safety climate mediate the relationship of safety leadership on safety performance.

1.6 Conceptual Framework

Figure 1.1 shows the conceptual framework of this research. The research question seeks to find the link between safety training practices and safety performance. In respect to this, safety training practice is appointed to be the independent variable in this study. Additionally, safety leadership is also being assigned as the other independent variable in finding the link between safety leadership and safety performance. Safety climate is believe to be vital in mediating both safety training practices and safety leadership in the efforts of organizations to improve safety performance. In addressing this, safety climate is assigned as the mediator of this research.

Figure 1.1: Conceptual framework.

1.7 Scope of the study

This study will be a cross-sectional study that will be carrying out upon 400 respondents from thirteen iron and steel base manufacturing companies that are chosen from Federation of Malaysian Manufacturer (FMM) directory. The data collection method of this study is by questionnaires as this is a quantitative study. There are two independent variables in this study, namely safety training practices and safety leadership. The dependent variable in this study is safety performance. Apart from independent and dependent variables, there is a mediating variable in this study, which is safety climate.

Next, safety training practice refers to the efforts of organization in preparing safety training for the employees. It is measured by using Wu et al.???s (2008) Safety Training Practice Scale. Furthermore, safety leadership consists of safety coaching, safety caring, and safety controlling that will be measured by the adoption of Wu et al.???s (2008) Safety Leadership Scale. Meanwhile, safety performance that measured by using Wu et al.???s (2008) Safety Performance Scale comprising proactive and reactive safety performance. Wu et al.???s (2007) Safety Climate Scale will be used in measuring safety climate which compromises of five dimensions, CEO???s safety commitment and action, manager???s safety commitment and action, employee???s safety commitment, perceived risk, and also emergency response.

The data collected will be analyzed in Structural Equation Modelling (SEM). According to Hoyle (1995), SEM is a comprehensive approach that measures the direct as well as indirect effects between variables statistically. Byrne (2009) further proposes that SEM test the causal relationship between various variables and determine if the model fit with the concept.

1.8 Significant of the study

The present study will be signified in the contribution to the understanding towards workplace safety issues in Malaysia. Firstly, this study will provide the evidences of the relationship between safety training practices and safety performance. The results will be showing the interaction and the reaction of safety performance towards safety training practices. This interaction will help the related parties in improving safety performance in the workplace.

Secondly, this study will also provide the evidences on the influence of safety leadership towards safety performance. This study can be served as a supportive reference of the leaders??? commitment towards safety in improving the safety performance as proposed by Wu et al., (2008). As the employees will always follow the leaders, the implementation of safety leadership can help in enhancing the safety performance of the employees or the workmen.

Finally, the finding of this study will enrich the body of knowledge on workplace safety in Malaysia. Despite the high awareness on occupational safety in western countries, there is still a lack on the understanding and focal on safety issues in Malaysia. Due to the lack of awareness on safety issues, there has been less research regarding occupational safety and health being conducted in Malaysia. Thus, this study will provide some empirical knowledge regarding safety issues for the understanding of safety issues in the context of Malaysia.

Chapter 2: Literature review

2.1 Introduction

This chapter discusses on the accident causation models, and the safety performance as the dependent variable. Moreover, the independent variables which includes of safety training practices and safety leadership will also be discussed in this chapter. At the end of the chapter, mediator, which is safety climate, was discussed.

2.2 Accident causation model

Accidents can be defined as the unexpected and undesired events that occur due to human activities, directly and indirectly (Hollnagel, 2004). Safety Institute of Australia (2012) proposes that understanding the causation of accidents is the first step in preventing the occurrence of accidents as there is a need to understand a problem before curing it. The evidence shown by Culvenor et al. (2007) suggested that there are various types of accidents causations, for instance, lack of hazard control planning, unsafe working conditions, lack of safety training, unsafe equipment, carelessness of employees, and poor workplace layout. Surprisingly, Qureshi (2007) argues that the causation of accidents may not be relate to human factor, whereas, it relate to the system errors. Nevertheless, both human errors and system errors can be explained by the accident causation models.

2.2.1 Swiss Cheese Model

Figure 2.1 showed the Swiss Cheese Model developed by Reason (1990). Swiss cheese model illustrates the accidents caused by the system failure (Reason 1998; National Council of State Boards of Nursing 2012; Perneger, 2005). Shappell (2000), suggests that there are four layers of errors due to human failures, which at the end leads to the harms and losses. Each layer of ???cheese??? indicates each type of defense or prevention towards the errors, which at the end form four types of defenses or prevention in the model. In Reason???s model, each of the failure will influence the failure of the next layer. (Shappell, 2000; Reason, 2006, Perneger, 2005). Reason (1998), argues that although there are layers of defense, each layer of defense is actually containing some gaps or ???holes??? that allows human failure to occur.

Refer to figure 2.1, there are two types of failures in the model, active failures and latent failures. Active failures refer to the ???software??? failures. It refers to the failure of human systems due to the incompliance of safety rules or violation of safety regulations. (Reason, 1998; Shappell, 2000; Aini and Fakhru???l-Razi, 2013). Controversially, the failures and errors on the machines themselves which initiated by the failure of designers, architects and maintenance had been labeled as the latent failures (Reason, 1998; Reason, 2006; Aini and Fakhru???l-Razi, 2013).

C:\Users\Allan\Desktop\1472-6963-5-71-1.jpg

Figure 2.1: Swiss Cheese Model (Source: Reason, 2006)

Both the ???software failures??? and ???hardware failures??? contribute to the existence of the gaps or ???holes??? on each slice of cheese. In another word, the active failures and latent failures cause the incomplete of the defenses and preventions in each stage with the ???life-span??? of active failures shorter than latent failures (Reason, 1998; Reason, 2006, Shappell, 2000). This is because as human will change their behaviors from time to time, therefore active failures that arose from the errors made by human may eliminated as the employees changed their behaviors to put safety as top priority.

On the other hand, latent failures were made to the system itself and it cannot be eliminated unless the system totally changed. Therefore, the latent failures are permanently. Finally, at the end, when the holes and gaps of each slice of cheese come to a straight line, the accidents will happen and losses incurred. (Perneger, 2005; NCSBN, 2012).

In light of the Swiss Cheese Model, it can be seen that there are different types of failures that lead to the occurrence of accidents although prevention had been done. It also indicates that in order to improve safety performance, what should be done is to improve the employees??? safety behaviors in order to eliminated the so called ???software failures???. Considering to the different situations and causations of accidents, it can be said that workplace accidents are actually a recycling process that is explained in the Risk Thermostat Model.

2.2.2 Risk Thermostat Model

Figure 2.2 is the Risk Thermostat Model that developed by Adams (1999). Risk thermostat model illustrates the interaction of five components, propensity to takes risks, perception of risks, balancing behavior, rewards as well as accidents. The risk thermostat model is actually formed by the cycles which it is actually an ongoing process.

By developing this risk thermostat model, Adams (1999), explains that every individual within the organization is proclivity to take risks. However, each of the individual or employees has different possibility of risk taking based on the rewards. (Adams, 1999; Adams, 2003). As for the perception of risks, Adams (2003) argues that it refers to the interpretation and their feeling towards the risk. Adams further offers that the interpretation of the risks is relatively influenced by the experiences of accidents, not only the accidents experienced by the individual own self but also accidents experienced by their peers.

C:\Users\Allan\Desktop\Picture1.png

Figure 2.2: Risk Thermostat Model (Source: Adams, 1999, Page 9.)

The cycles of Risk Thermostat Model actually starts by the performance, either safe performance (rewards) or unsafe performance (accidents). Safe performance (rewards) triggers the high propensity to take risks while unsafe performance (accidents) will influence the perception of risks at the same time. In the process of balancing behavior, employees will perform either comply and participate or did not with safety rules and regulations outlined by the organizations. When safety rules and regulations were complied and participated, the outcome will be rewards and otherwise (Adams, 1999; Adams, 2003). However, each of the performance outputs will contributes to the input and the cycles begin again and again.

In short, it is clearly seen that the cycle of Risk Thermostat Model is actually started from the safety performance, either high or low. Therefore, there is a need for safety performance to be discussed.

2.3 Safety performance

Several attempts have been made to measure safety performance and it has revealed that safety performance consists of two dimensions, safety compliance and safety participation. (Zohar, 2008; Yang et al., 2010; Martinez-Corcoles et al., 2012; Brondino et al., 2012). Tharaldsen et al. (2010) as well as Kwon and Kim (2013), addressed safety compliance refers to the compliance towards the rules and regulations while safety participation can be defined as the engagement towards safety issues.

Hsu et al. (2012), argues that in order to measure safety performance, proactive measurement should be used. In addition, Hsu et al. (2012), further proposed the dimensions of the safety performance measurement which includes employees??? perceptions of leaders??? commitment in safety, the effectiveness of safety and health training, and the degree of influence the leaders have. However, Wu and Li (2006), claims that apart from accidents rates, there are other measurements that can be used in measuring the safety performance. Wu and Li (2006), further suggests that safety management, safety equipment, and safety training practices can also be used to measure safety performance. Thus, safety training practice is vital in influencing the safety performance of the employees.

2.4 Safety training practices

According to Health and Safety Executive (HSE) (2011), training carries the meaning of providing information needed and helping the employees in developing required skills and knowledge in general. Training is a form of investment which prepare contents that are beneficial to the employees in order to change their unsafe behavior into safe behavior (Choudhury, 2006; Yang, 2010; Wrede, 2009; Mohammad Fam et al., 2012). Safety training practices equipped the employees with capacities needed in handling their task safely and competently (Sari, 2009; Antonio et al., 2013). Thus, safety training practice is one of the major practices that served as a vital and important factor in reducing workplace incidents, which in other word, increase safety performance. (Wameedh et al., 2011 ; HSE, 2011; Burke et al., 2011).

Siu et al. (2004) have conducted a study on safety performance which provides the argument that the increased risk of workplace incidents contributed by the insufficient of job knowledge, skills, and also training practices. In response to this issue, Burke et al., (2007), and also Cheng et al, (2012), proposed that the occurrence of workplace accidents contributed by the lack of knowledge and skills while safety training practices fulfil the lack. Recent studies by Burke et al. (2011) and Burke et al. (2006), suggested that the level of employees??? engagement in safety training practices relatively reflect the level of safety performance. However, in order to increase the level of safety performance, safety leadership is also required.

2.5 Safety leadership

According to Mullen and Kelloway (2009), safety leadership can be defined as the combination of different leadership styles that share the notion of safety is in top priority. Previous studies have investigated the relationship of safety leadership and safety performance and it reported that there is significant influence of safety leadership on safety performance (Lingard et al., 2012; Yang et al., 2010; Zohar, 2002; Rowley, 2009). Rowley (2009), further defined safety leadership as a practice that used to predict and prevent the future incidents that had not been happened yet.

In recent studies (Oltedal, 2011; Ginsburg et al., 2010; Nahrgang et al., 2007; Ankrah, 2007), scholars proposed that safety leadership acts as a vital role in proceeding the occurrence of favourable safety performance by safety climate. According to Mullen et al. (2011), safety leadership is far more effective in shaping positive safety behaviour and attitudes through inspiring and promoting.

Wu et al. (2008) introduced that there are three dimensions of safety leadership, namely safety coaching, safety caring, and safety controlling as well. Wiegand (2007) explained that safety coaching refers to the efforts of leaders in managing the safety performance that involve interpersonal interaction and communication. As for safety caring, it refers to the level of concern of leaders towards safety issues. Safety caring involves the efforts to ensure the quality of safety in the workplace (Wu et al., 2010). Whereas, Wu et al. (2008) proposes that safety controlling is the use of power in outlining the safety rules and regulations to be complied by the employees in order to achieve safe performance.

2.6 Safety climate

Looking back to the emerging history of safety climate, the first serious discussion and analyse of safety climate emerged during year 1980 with the research conducted by Zohar (1980). Since then, a considerable amount of literature has been published on safety climate (Arcury et al., 2012; Mohammad Khandan et al., 2011; Cheyne et al., 2003; and Siti Fatimah Bahari and Clarke, 2013). Safety climate is actually part of safety culture that can be measured. It refers to the manifestation of day-to-day perceptions of the safety culture in term of belief, values, practices, and policies towards safety in the organization (Huang et al, 2012; Chenhall, 2010; Zohar and Luria, 2010; Yule, 2008; Mearns et al., 2003). Recent studies (Lingard et al., 2012; Li et al., 2009; Burke et al., 2008: Johnson, 2007) found out that there is a strong evidence of safety climate to affect safety performance.

Refer to a major study conducted by Li et al., (2009), the evidences of the study have shown that as an important indicator, safety climate can reflect safety performance very well. This view is supported by Colley et al. (2013) who writes a paper on safety climate. Colley et al. (2013) argues that the way employees decoding or interpreting the policies, practices, and the commitment towards safety is highly related to their perception of the values within the organization. It is believe that the attitudes and perceptions of workers shall influence their behavior in safety and after that intervene the performance (Shang and Lu, 2009; McCaughey et al., 2013). Therefore, while safety climate is the perception of the organizational values, it is highly influenced the interpretation of safety related issues within the organization.

Recently, Lingard et al. (2010a) suggested four types of safety climate (Figure 2.3) theoretically according to the climate level and also climate strength, which comprising contradictory, indifferent, obstructing, and also strongly supportive safety climate. A contradictory safety climate refers to the high climate level with weak climate strength. Contradictory safety climate involves with the low consensus towards the priority of safety in getting work done, there is an exist of mixed messages to which each of the employees and supervisors may have different perception towards the important of safety. However, the strongly supportive safety climate is totally different with contradictory safety climate. In strongly supportive safety climate, the employees have a highly consensus which they put the safety as the top priority. The employees share the same value that safety is most important.

C:\Users\Allan\Desktop\Note1.jpg

Figure 2.3: Safety climate types (Source: Lingard et al., 2010a)

On the other hand, indifferent safety climate argues that there is a low consensus which the employees did not or hardly share the same view on the priority of safety. In this indifferent safety climate, the safety was not put into priority. However, in obstructing safety climate, although there is a low level of safety climate level, the consensus is very strong. It means that, the all of the employees do shared the same view, however, in this consensus, they put safety as secondary or lower priority, rather. They put the productivity as top priority.

In a recent studies (Choudhry et al., 2009; Yeung and Chan, 2012), safety climate formed by the combination of various dimensions. Choudhry et al. (2009), and Yeung and Chan (2012) identifies six dimensions of safety climate, including of status of safety officer, status of safety committee, level of risk at workplace, effects of safe conduct on promotion, effect of safe conduct on social status, and effect of required work pace on safety.

2.7 Summary of Literature Review

In short, this chapter apparently addresses some key safety issues and gaps of this research. The accident causation model includes Swiss Cheese Model and Risk Thermostat Model are thoroughly discuss in this chapter to show the causation of workplace accidents. Turning into the variables, this chapter seeks to discuss how safety leadership and its dimensions influence safety performance. Meanwhile, the linkage between safety training practice together with its dimensions and safety performance. In explaining the relationship between safety leadership and safety performance as well as the linkage between safety training practice and safety performance, there is a fact to be recognized that the relationship and linkage is strictly influence either directly or indirectly by safety climate.

Chapter 3: Research Methodology

3.1 Introduction

This chapter discusses the research design, as well as the population and sampling. Moreover, the development of the research instrument will be explained in this chapter. Furthermore, the process of pilot study is elaborated. At the end of the chapter, data analysis is discussed.

3.2 Research Design

This research is a quantitative research. According to Creswell (2002), a quantitative research refers to the research that measures the causal relationship, hypotheses testing, and theories testing which uses survey as data collection instrument. Creswell further proposes that quantitative research involved with statistical data. Furthermore, the design of this research is descriptive and correlational study. Elifson (1998) proposes that descriptive study describe the characteristics of the trend or situations that are desired. Descriptive study helps the researcher understand the phenomena and intercorrelation between the variables. (Sekaran and Bougie, 2009). Whereas, correlational study had been defined as a technique that is able to describe and measure the link and relationship between two variables statistically (Gravetter and Wallnau, 2002).

3.3 Population and Sampling

This study employs probability sampling method, whereby the researcher knew every possibility of each member of population to be chosen (Fairfax County Department of Neighbourhood and Community Services, 2012). Fairfax County Department of Neighbourhood and Community Services (2012) confirms that the results of probability samples can generalize the whole population eventhough it is costly and difficult to conduct. To be exact, a simple random sampling is adopted as the sampling strategy that ensure each of the individual of the population equal chances to be chosen as the samples (Teddlie and Yu, 2007).

According to Israel (1992), sample size can be determined by using published tables. By refering to the tables in Appendix A, sample size of this study can be determined by the referring precision levels ??5%. Due to the total population of this study that exceeding 100000, a sample size of 400 respondents will be used in this study.

Respondents of this study will be the employees from the iron and steel based manufacturing companies chosen from Federation of Malaysian Manufacturers (FMM) directory. It is because the number of accidents happened in iron and steel industries achieved the highest among all of the other manufacturing industries, with every two working hours there will be one accident happened in iron and steel based companies in Malaysia by year 2011 (Social Security Organization, 2011). By reviewing the statistical data by state, it indicates that Perak had the highest numbers of accidents in 2011(Department of Safety and Health, 2011), therefore the iron and steel based companies chosen in this study are from Perak, which is a total of thirteen iron and steel based companies.

3.4 Development of Research Instrument

Within this research, the research design was cross-sectional survey, by questionnaire as the data collection tool. Saunders dan Thornhill (2009), point out that questionnaire is an instrument that is most economical in collecting data from the large sample size accurately. As proposed by Sekaran and Bougie, (2009), by using questionnaire as data collection tool, the data can be collected efficiently.

In this research, it is apparent that the questionnaire is the adoption of questionnaires from Wu et al. (2007) and Wu et al. (2008). In order to measure the safety leadership, the Safety Leadership Scale developed by Wu et al. (2008) was adopted. At the same time Safety Climate Scale from Wu et al. (2007) is adapted for measurement of safety climate. Meanwhile, adoption of Wu et al. (2008) is used in this questionnaire for the Safety Performance Scale as well as Safety Training Practices Scale. The adoption of Wu et al.???s questionnaires in the measurements of independent variables, dependent variable, and also mediation is due to the proven high reliability of the questionnaires (Alpha Cronbach: 0.84 to 0.97) (Shah Rollah, 2011; Wu et al., 2007; Wu et al., 2008).

A five-points likert scale as shown in Figure 3.1 is used in the questionnaire to enable the respondents to rank or choose the most appropriate mark for the specific items. The interval scale was used in the questionnaire as it allows the data collected to be calculated in the analysis process (Greener, 2008).

Figure 3.1: Likert Scale (Source: Boone and Boone, 2012; Page 2.)

3.5 Pilot Study

As Aisah Md. Nor et al. (2007) proposes, the reliability of an instrucment can be known by the measurement of the outcomes of pilot study using Alpha Cronbach., In which she proposes that the instrument of questionnaire will only be verified as reliableif the measurement of the pilot study???s outcomes reached the minimum of 0.7 from the range of 0 to 1 in the Alpha Cronbach index.

Tavakol and Dennick (2011), offers that the usage of Alpha Cronbach test is to determine the internal consistency of the questionnaires. Tavakol and Dennick (2011) further explains that internal consistency refer to the interrelations between an item towards other items within the questionnaires.

Nonetheless, construct validity will also being test in pilot study in order to ensure the construct of the questionnaire will measure the exact desired data. Bhattacherjee (2012) reported that construct validity is vital in a positivist study and it shall be measured by factor analysis or correlation.

3.6 Data Analysis

The collected data of this research will be analyzed in Structural Equation Modeling (SEM). SEM is a statistical technique that is powerful and useful in measuring quantitative data. SEM is suitable in hypothesis testing, causal relationship testing, as well as model measurement (Hoe, 2008). In recent years, as suggests by Schumacker and Lomax (2004), SEM is widely used in measuring the relationship between variables.

Moreover, multiple mediation analyses were also used in analyzing the data in this research. Multiple mediation analyses exist to test the mediating effects of the mediator between multiple independent variables and dependent variable (Kenny, 2011). Below show the equations that used in various studies (Zhao et al., 2009; Iacobucci et al., 2007; Preacher and Hayes, 2004):

Equation 1:

Equation 2:

Equation 3:

Where: Y = dependent variable

X =independent variables

M=mediator

i =constant value for independent variables.

c= coefficient for independent variables and mediator

e= regression residual

Furthermore, multiple regressions were used in this research. Multiple regressions are used usually when there are multiple repressors variables (Bartholomew et al., 2008). As this research consists of two independent variables, therefore multiple regressions was used to analyze the data.

3.7 Summary of Research Methodology

This chapter explained the research design, methodology and data analysis of this research. As discussed earlier in this chapter, the research design of this research is descriptive and correlation survey in identify the linkage between safety training practice and safety performance, as well as linkage between safety leadership and safety performance. The employees from iron and steel base industries in manufacturing sector were chosen as the respondents of this survey. Data will be collected by using questionnaire with the adoption of Safety Training Practice Scale (Wu et al., 2008), Safety Leadership Scale (Wu et al., 2008), Safety Climate Scale (Wu et al., 2007), and Safety Performance Scale (Wu et al., 2008). The data collected will be analyzed using Structural Equation Modeling, Multiple Mediation Analysis, and also Multiple Regression Analysis.

Figure 3.2: Research objectives and method of data analysis

No.

Research Objectives

Methodology

1

To identify the relationship between safety training practices and safety performance.

Simple Linear Regression Analysis

2

To identify the relationship between safety leadership and safety performance.

Simple Linear Regression Analysis

3

To identify the effect of safety climate mediate the relationship of safety training practices on safety performance.

Path Analysis

4

To identify the effect of safety climate mediate the relationship of safety leadership on safety performance.

Path Analysis