Content uploaded by Ruoyu Jin
Author content
All content in this area was uploaded by Ruoyu Jin on Jan 10, 2020
Content may be subject to copyright.
1
A Structural Equation Modeling Approach to Studying the Relationships among Safety1
Investment, Construction Employees’ Safety Cognition, and Behavioral Performance
2
Yu Han1, Jie Li2, Xiulan Cao3, Ruoyu Jin4
3
Abstract
4
This study aimed to investigate the internal relationships between safety investments and
5
construction employees’ behavioral performance with safety cognition as the mediating
6
factor. A comprehensive methodology was adopted, including theoretical modeling of safety
7
investments, questionnaire survey, and Structural Equation Modeling (SEM). In the
8
theoretical model, four factors (i.e., personal protection equipment (PPE), safety education,
9
insurance purchased for site employees, and safety incentives) were adopted as safety
10
investment categories. These four categories were studied of their correlation to the overall
11
safety investment, which was tested of its contribution to employees’ behavioral safety
12
performance in both direct and indirect ways. Indirectly, safety cognition was introduced as a
13
mediator to bridge safety investments and behavioral performance. A questionnaire
14
consisting of 28 indicators was adopted to describe safety investment, safety cognition, and
15
behavioral performance. A random sampling approach and the top-down method were
16
implemented to recruit construction site employees from the south-eastern region of China.
17
The follow-up SEM analysis revealed that all the four investment categories positively
18
contributed to the overall safety investment, which was found significantly correlated to
19
employees’ safety cognition and behavioral performance. Safety incentive was identified as
20
the most significant factor contributing to the overall investment. The current study extends
21
1Associate Professor, Faculty of Civil Engineering and Mechanics, Jiangsu University, 301 Xuefu Road,
Zhenjiang, 212013, Jiangsu, China. Email: hanyu85@yeah.net
2Graduate research assistant, Faculty of Civil Engineering and Mechanics, Jiangsu University, 301 Xuefu Road,
Zhenjiang, 212013, Jiangsu, China. Email: lijie_win@yeah.net
3Graduate research assistant, Faculty of Civil Engineering and Mechanics, Jiangsu University, 301 Xuefu Road,
Zhenjiang, 212013, Jiangsu, China. Email: 1244722912@qq.com
4Senior Lecturer, School of Environment and Technology, University of Brighton, Cockcroft Building 616,
Brighton, UK. BN24GJ. Phone: +44(0)7729 813 629, Email: R.Jin@brighton.ac.uk
2
prior studies of safety investments by adopting a quantitative approach from employees’22
perspective. It provides insights for construction employers regarding how safety investments
23
could affect behavioral performance. Employers are suggested to balance the tangible (e.g.,
24
incentive) and intangible (e.g., safety insurance) investment categories. This study also
25
contributes to establishing the internal links among safety investments, safety cognition, and
26
behavioral safety performance. Based on the current findings, future work could investigate
27
how to optimize safety investments to achieve higher behavioral performance. The current
28
study based in China could be applied in a different geographic context by testing the
29
correlations between safety investments and behavioral safety performance.
30
Keywords: Construction employee; safety behavior; safety cognition; safety investment;
31
Structural Equation Modeling (SEM)
32
Introduction
33
Construction is one of the most risky industries due to its comparatively lower safety
34
performance measured by injury rates (Lingard and Rowlinson 2015). An earlier study by
35
Zou et al. (2007) found that safety was one of the main risks in China’s construction industry,
36
including insurance not purchased for employees, no insurance for major equipment,
37
inadequate safety measures or unsafe operations, and poor competency of construction
38
workers, etc. In China, construction workers are largely from rural and less economically-
39
developed regions. It is common that they learn basic construction skills from their family
40
members who are on the same team, and they are likely to mimic unsafe behaviors from
41
peers (Zhang 2017). More than half of construction workers in China have not completed or
42
barely finished middle school education (Zhang and Li 2016). In more recent years, high
43
occurrences of construction accidents have caused public concerns. Safety requirements are44
being enforced and monitored, such as mandatory usage of personal protection equipment45
(PPE). Although it is expected of the 100% adoption rate of mandatory PPEs in all projects,46
3
the safety attitude, perception, and awareness of construction workers could vary crossing47
projects. Construction workers might behave in a more risky way to gain more income or to
48
save time especially under a tight project schedule. There is a lack of empirical evidence of
49
how certain investment categories (e.g., insurance) affect the behavioral safety performance.
50
Safety performance could be evaluated by different measurements, including the reactive
51
and proactive measurements. The reactive measurements include accident or injury related
52
occurrences. The proactive measurements highlight the preventive actions to avoid harms, for
53
example, behavior-based safety performance. Safety performance could be affected by
54
multiple factors related to safety investments, employees’ safety behavior, safety awareness,
55
and safety monitoring (Flin and Mearns 1994; Choudhry et al. 2007; Chen and Jin 2013).
56
Support at the organizational level to employees’ health and safety generally leads to higher
57
safety performance (Mearns et al. 2010). Safety investment, as one of the main ways of
58
organizational support, is affected by multiple factors, such as the organizational capacity to
59
control risks and management skills (Yoon et al. 2000). Safety investment could be divided
60
into different categories such as education and PPE (Qiang et al. 2004). So far, more studies
61
have focused on safety investments at the organizational level, with limited research targeting
62
the individual level. Specifically, there has been limited investigation quantifying the effect
63
of safety investment categories on employees’ behavioral safety performance. There has also
64
been limited in-depth research focusing on how the overall safety investments affect safety
65
performance through safety culture (Feng, 2013). Individual awareness and perception
66
towards different safety investment categories (e.g., insurance) could affect the behavioral
67
safety performance in either a direct manner, or an indirect way through the mediation of
68
safety cognition. Investigating the effects of various safety investment categories on69
behavioral safety performance is critical based on the facts that: it provides the guides for70
construction employers to properly allocate their budget related to safety; it also contributes71
4
to the body of knowledge in construction safety management by establishing the theoretical72
framework incorporating safety investments, behavioral safety performance, and other
73
human-based safety factors (e.g., safety cognition).
74
Prior studies (e.g., Yong et al. 2000; Zou et al. 2007; Wang et al. 2014; Man et al. 2017)
75
either investigated the importance of safety investments at the organizational level, or
76
analyzed the formation of unsafe behaviors in a qualitative approach. Workers are direct
77
participants in all construction activities and are most vulnerable to be victims of accidents. A
78
further study from the employees’ perspectives in the context of safety culture (Guldenmund
79
2007) would be needed to investigate the correlations among safety investments for site
80
employees, their safety cognition, and behavioral performance. Aiming to address these
81
aforementioned limitations, this study investigates the effects of safety investments on
82
behavioral performance with safety cognition as the vehicle. The objectives of this study
83
include: (1) initiating a theoretical model incorporating safety investments, safety cognition,
84
and behavioral safety performance. Safety investment is measured in four main categories
85
related to safety education, PPE, safety incentive, and safety insurance defined by Cao (2018).
86
Behavioral performance is divided into behavioral compliance and behavioral participation
87
suggested by Neal (1995); (2) investigating the effects of safety investment categories on
88
behavioral performance; and (3) discussing the mediating effect of safety cognition as the
89
vehicle to bridge safety investments and employees’ behavioral performance. This study
90
contributes to the body of knowledge in construction safety management both practically and
91
academically. Practically, the current study offers insights of how various safety investment
92
factors could impact behavioral safety performance. Academically, it leads to further research
93
in optimizing safety investment categories towards enhanced safety culture and improved94
safety performance.95
Literature Review96
5
Investments in construction safety97
Investments in safety must be formulated as preventive measures against fatal accidents
98
(Shohet et al. 2018). According to Shohet et al. (2018), safety investments cover costs of
99
equipment, training, insurance, and other personal costs related to construction activities. The
100
investments in safety would lead to enhancement in safety performance (Lu et al. 2016).
101
Safety education, safety incentives, safety insurance, and PPE, as listed by Cao (2018), are
102
critical factors or categories in construction safety investments. Safety investment, according
103
to Feng (2013), could be divided into different categories such as basic investment and
104
voluntary investment. Basic safety investments are defined as accident prevention activities
105
that are required by industry or governmental regulations, including staffing cost, safety
106
equipment and facility cost, and mandatory training cost (Feng et al. 2014). Voluntary
107
investments are generally determined by individual organizations or projects (Feng et al.
108
2014). They include costs related to in-house safety training, safety inspection and meeting,
109
safety incentives and promotion, and safety innovation (Laufer 1987; Tang et al. 1997; Hinze
110
2000; Feng et al. 2014). Different types of safety investments could have various effects on
111
safety performance (Feng 2013), and are affected by other internal and external factors such
112
as safety culture and site hazard levels (Feng 2015). Safety performance is improved with a
113
higher level of safety investments, but could be mediated by safety culture (Feng et al. 2014).
114
Studying the effects of different safety investment categories on safety performance is hence
115
considered important (Cao 2018).
116
Safety cognition in the context of safety culture
117
Personal cognition reflects how an individual selects, organizes, and explains information
118
from external sources (Chen et al. 2011). Social cognition is not separated from safety119
climate, which forms safety culture as indicated by Marquardt et al. (2012). Multiple studies120
(e.g., Guldenmund 2000; Rowatt et al. 2006; Parker et al. 2006) indicate that safety cognition121
6
would significantly affect employees’ safety behavior, which further influences safety122
performance. Individual safety cognition is crucial to construction safety performance (Chen
123
et al. 2011). Safety cognition could be linked to employees’ implicit assumptions of safety,
124
their prior safety scenarios, and their own safety knowledge (Liu 2018). Marquardt et al.
125
(2012) further divided safety cognition according to the implicit and explicit levels. In the
126
construction industry, employees’ implicit safety cognition is formed from their prior work
127
scenarios which establish their own safety knowledge (Han et al. 2019c). The prior work
128
scenarios and safety knowledge affect individuals’ safety perceptions (Marquardt et al. 2012).
129
Safety perception is a core part of explicit safety cognition (Han et al. 2019c), which is
130
largely equal to safety climate in terms of the measurement criteria (Guldenmund 2000;
131
Rowatt et al. 2005). These measurement criteria include perceptions towards jobsite hazards
132
(Han et al. 2019c), individuals’ perceptions of self-capability to identify, evaluate, and
133
control site hazards (Han et al. 2019b), as well as their awareness and knowledge of safety
134
behaviors of themselves and their peers (Chen and Jin 2012).
135
Behavioral safety performance
136
It was found that employees’ behavior in the forms of acts or omissions contributed to up
137
to 80% of work-related injuries (Health and Safety Executive 1999). IOSH (2015)
138
emphasized that one way to improve safety performance was to introduce a behavioral safety
139
process and to reduce unsafe behaviors. These unsafe behaviors (e.g., improperly wearing
140
PPEs) could result in accidents, including falls, electrocution, struck-by, and caught-in–
141
between which are defined as Focus 4 Hazards (OSHA 2011). Construction safety
142
management should highly target workers’ unsafe behaviors (Chen and Jin 2012). Studies
143
from Lingard and Rowlinson (1998) and Cooper (2003) indicated that the behavior-based144
safety (BBS) program could enhance safety performance. Nevertheless, critical factors within145
safety climate are key to successful implementation of BBS, including employee engagement,146
7
safety training, and management capabilities (DePasquale and Geller 1999). Griffin and Hu147
(2013) defined two key safety behavioral measurements, namely safety participation and
148
safety compliance. It was recommended by Griffin and Hu (2013) that future research could
149
explore individual and organizational mediators influencing safety behaviors. The social
150
psychology theory of Baron and Kenny (1986) and the construction safety cognition
151
framework defined by Han et al. (2019c) inferred that safety cognition could serve as the
152
mediator influencing individuals’ safety behaviors.
153
Methodology
154
Research design
155
This study was based on the research hypotheses regarding the impacts of safety
156
investments on site employees’ behavioral performance. A total of 14 hypotheses were
157
originally proposed as illustrated in Fig.1.
158
<Insert Fig.1 here>
159
The details of these hypotheses are explained in details below:
160
H1a: investments in PPE significantly affect employees’ behavioral participation;
161
H1b: investments in PPE significantly affect employees’ behavioral conformance;
162
H1c: investments in PPE significantly affect employees’ safety cognition;163
H2a: investments in safety education significantly affect employees’ behavioral164
participation;165
H2b: investments in safety education significantly affect employees’ behavioral166
conformance;
167
H2c: investments in safety education significantly affect employees’ safety cognition;
168
H3a: investments in safety incentives significantly affect employees’ behavioral
169
participation;
170
8
H3b: investments in safety incentives significantly affect employees’ behavioral171
conformance;172
H3c: investments in safety incentives significantly affect employees’ safety cognition;173
H4a: investments in safety insurance significantly affect employees’ behavioral
174
participation;
175
H4b: investments in safety insurance significantly affect employees’ behavioral
176
conformance;
177
H4c: investments in safety insurance significantly affect employees’ safety cognition;
178
H5a: employee’s safety cognition significantly influences their behavioral
179
participation;
180
H5b: employee’s safety cognition significantly influences their behavioral
181
conformance.182
It is further noticed that the four investment categories can be combined as one overall183
safety investment, which could have significant effects on behavioral safety performance as
184
indicated by Lu et al. (2016). It is seen in Fig.1 that this research aims to explore the role of
185
safety cognition as the mediating factor between safety investments and behavioral safety
186
performance. Han et al. (2019c) defined the framework of safety cognition, which could be
187
divided into implicit and explicit cognitions. The implicit social cognition refers to
188
employees’ assumptions which influence individual behaviors (Schein 1992). The implicit
189
cognition affects the explicit cognition, which could be equated to safety climate in
190
measuring individual attitudes, awareness, and perceptions towards safety (Guldenmund
191
2000; Rowatt et al. 2005). Safety cognition reflects a construction employees’ awareness and
192
perception of potential site hazards, as well as the capability of decision making to behave
193
properly. Behavioral safety performance is defined as safety participation and safety
194
compliance in this study following Neal (1995) and Neal et al. (2000). According to Neal et
195
9
al. (2000), safety participation refers to employees’ involvement in safety-related activities in196
the workplace; safety compliance mainly refers to employees’ conformance to safety
197
regulations.
198
Safety investment generally refers to funds spent on preventing accidents, and on
199
protecting the health/physical integrity of construction workers (Tang et al. 1997; Zou et al.
200
2010). The overall safety investment could be divided into various categories which could
201
have varied influences on safety performance (Feng 2013). These investment categories listed
202
by Feng (2014) can be labelled as tangible or intangible factors from the perspective of site
203
employees. Tangible investments refer to those categories that are easily seen or physically
204
sensed by employees. They are generally visible hardware devices or products, such as PPE
205
which can be seen and physically used by employees. The intangible investments are
206
generally progressive actions or processes which are not in a physical form of products or
207
hardware. For example, employers invest on safety insurance and training for their employees,
208
but employees may ignore these intangible investments because they do not directly see the
209
cost of insurance or education as they would physically sense their PPE. The safety incentive
210
is defined as a tangible investment because employees can directly see the extra income
211
awarded for their good safety performance.
212
It is hypothesized that these safety investments aiming to prevent injuries or other
213
accidents could be mediated by employees’ safety cognition which further affects the
214
behavioral performance. Employees with highly positive safety cognition would be more
215
likely to appreciate the safety investments of their employers, to more actively participate in
216
safety education, and to conform to safety regulations. Therefore, the research framework in
217
Fig.1 can be further induced to the adjusted theoretical model shown in Fig.2.218
<Insert Fig.2 here>219
The social psychology theory proposed by Baron and Kenny (1986) stated that there was220
10
a mediator that intervened the effects of a stressor or external scenario on the outcome. In the221
context of construction safety behavior, these four safety investment categories serve as
222
external scenarios which could affect employees’ behavioral outcomes. But the degree of
223
effect, as inferred by Baron and Kenny (1986) and Han et al. (2019c), could be intervened by
224
safety cognition as the mediator. Therefore, Fig.2 is deduced following the theories of social
225
psychology and safety cognition for the follow-up quantitative analysis.
226
Questionnaire survey
227
This research started from a review of existing literature (e.g., Hinze 1997; Glendon and
228
Litherland 2001; Newaz et al. 2016; Tholén et al. 2013) in safety investments, employee’s
229
safety cognition, and behavioral performance. According to the literature review and the
230
researchers’ earlier work (i.e., Cao et al 2018), the indicators of safety investments, safety
231
cognition, and behavioral performance were defined. A questionnaire survey to China’s
232
construction site employees was planned incorporating these indicators. The initiated
233
questionnaire was peer reviewed by both academics and construction safety professionals in
234
China. A total of 36 peer reviewers were invited to provide feedback to the initialized
235
questionnaire to ensure that the statements were clear without vagueness, and easily
236
understood by construction employees especially workers. These peer reviewers included
237
graduate students in the construction management program of Jiangsu University, academic
238
staff, and industry professionals in the local construction industry. Their feedback was
239
collected during August and September in 2017, and discussed within the research team. The
240
finalized questionnaire corresponding to the 28 indicators is provided in Table 1.
241
<Insert Table 1 here>
242
These 28 indicators were statements asked to employees during the site questionnaire243
survey. Each statement was generated from references listed in Table 1. From October 2017244
to January 2018, questionnaire surveys were conducted from a total of 39 construction sites245
11
in the south-eastern region of China. Site employees were guided to rank each indicator with246
a Likert-scale score, from “1” meaning “strong disagree with the statement” to “5” indicating
247
“strongly agree”.
248
Sampling
249
Since 2010, along with the national promotion of digitalization in construction (Jin et al.
250
2015), China has been promoting the digital strategies in construction site management, for
251
example, virtual reality (VR) and other video technologies used in construction safety
252
education. In this study, the consistent random and unbiased sampling procedure described by
253
Li et al. (2017) was conducted in the south-eastern coastal region of China, which represented
254
the country’s economically active region where the video-based safety education had been
255
more commonly adopted in building construction projects. Site employees recruited in the
256
questionnaire survey were from the high-rise residential building sector. It was expected that
257
site employees had either undergone or at least been aware of video-based safety education.
258
The consistent top-down method described by Chen et al. (2018) for site survey was adopted.
259
Basically, the research team initially contacted the top management personnel (e.g.,
260
executives) of ongoing construction projects. If the top management personnel agreed on site
261
visits and showed interests on the research, they would then schedule the questionnaire
262
survey to their site employees. Afterwards, administering of questionnaire surveys was
263
coordinated between three research team members and project management staff for each site
264
visit. At the beginning of each site survey, all employees were explained with the purpose of
265
the study and ensured that no personal or company information would be included. Each
266
question was explained to survey participants to ensure no vagueness or confusion. For
267
example, the high intensity of incentives described in the indicator of X7 in Table 1 meant268
the frequency and amount of cash award for employees’ excellent safety performance. A269
larger amount of cash award or a more frequent award would mean a higher intensity. During270
12
the site survey, participants were further encouraged to ask for clarification if anything in the271
questionnaire was unclear to them. They were also made aware that they could withdraw the
272
survey at any time.
273
Among the totally 380 questionnaires received through site surveys, 326 of them were
274
found valid after excluding incomplete questionnaires or those with the same Likert-scale
275
scores for all indicators within the same category (e.g., safety education investment). About
276
55% of the survey population was construction workers and the remaining 45% came from
277
crew foremen or other site management personnel (e.g., safety manager, superintendent, etc.).
278
Nearly 60% of them had over 10 years’ site experience. The detailed demographic
279
information of the survey participant sample is provided in Table2.
280
<Insert Table 2 here>
281
Structural Equation Modeling Approach
282
Following the site questionnaire surveys, Cronbach’s alpha analysis was applied to check
283
the reliability of indicators. According to Bland and Altman (1997) and DeVellis (2003), a
284
Cronbach’s alpha value close to or above 0.70 would suggest acceptable internal
285
consistencies among indicators. The Structural Equation Modeling (SEM), which had been
286
widely used in behavioral sciences based on a combination of factor analysis and path
287
analysis (Hox and Bechger 1998), was adopted in this study to test these correlations among
288
safety investment, safety cognition, and behavioral safety performance described in Fig.1 and
289
Fig.2. The sample size for SEM was suggested to be not lower than 10 times the number of
290
variables (Bentler and Chou 1987; Bollen 2014; Nunnally 1967). In this study, the ratio of
291
sample size at 380 to the number of indicators at 28 met the requirement. The exploratory
292
factor analysis (EFA) was adopted to identify the underlying factor structure of a dataset as293
demonstrated by Shan et al. (2018). EFA is the proper approach for SEM to hypothesize an294
underlying construct and to estimate factors that influence responses on observed variables295
13
(Suhr 2006). EFA has been traditionally adopted to explore the possible underlying factor296
structure of a set of measured variables without preconceived structure on the outcome (Child
297
1990). EFA KMO (i.e., Kaiser-Mayer-Olkin) and Bartlett sphere test were introduced in EFA
298
for the validity analysis. KMO measures the amount of a variance shared among the
299
indicators which are designed to measure a latent variable (Shan et al. 2018). The KMO value
300
higher than 0.5 would be considered acceptable (Kaiser 1974). The SEM was later conducted
301
to analyze the loading factors and path coefficients between different factors. The model-fit
302
test following the guide provided by Wu (2009) was performed to evaluate the SEM
303
outcomes. These measurements for Goodness-of-fit of SEM are defined in Table 3, where the
304
ideal numerical range of each measurement is provided. More detailed explanations of these
305
indices in Table 3 can be found in Hox and Arnhem (1998), Kaplan (2001), and Shadfar and
306
Malekmohammadi (2013).
307
<Insert Table 3 here>
308
Results
309
Initial validation of data collected from site questionnaire surveys
310
The reliability test based on Cronbach’s alpha analysis is presented in Table 4.
311
<Insert Table 4 here>
312
All Cronbach’s alpha values for each category as well as the overall value close to or313
over 0.70 indicated that the reliability was generally acceptable. The KMO and Bartlett314
spherical tests were then conducted for the further validity analysis. The KMO value at 0.837315
and the Bartlett spherical test significance at 0.000 indicated satisfactory correlations among316
indicators. Therefore, the further factor analysis could be conducted. The initial structural317
model is illustrated in Fig.3.318
<Insert Fig.3 here>319
Following the SEM procedure using AMOS (Division of Statistics + Scientific320
Computation 2012) for the initial model shown in Fig.3, the Goodness-of-fit test displayed in321
14
Fig.2 was conducted and presented in Table 5.322
<Insert Table 5 here>
323
The values of AGFI, GFI, and NFI below 0.90 indicated that the initial model should be
324
modified in order to meet the SEM requirements according to Table 3.
325
Model modification
326
The modification of the initialized model in Fig.3 should not only meet the statistical
327
requirements shown in Table 3, but should also make the theoretical sense in construction
328
safety management. These two criteria (i.e., statistical and theoretical aspects) were both
329
considered in the modification process. When the Goodness-of-fit test did not yield
330
satisfactory outcomes, either model building or model trimming should be applied to modify
331
the model. As guided by David Garson and Statistical Associates Publishing (2015), the
332
model building approach by adding paths based on the theoretical sense and the MI (i.e.,
333
Modification Indices) was implemented to improve the Goodness-of-fit. According to Wu
334
(2009), a path could be added for a pair of indicators whose MI value is over 4.0. Following
335
this initial test, several pairs of indicators shown in Fig.3 were found with relatively large MI
336
values, such as e12 and e13 with the MI value at 21.584, as well as e22 and e23 (MI value at
337
16.408). From the theoretical sense according to the researchers’ prior construction safety
338
research (e.g., Cao et al. 2018), using PPE could increase construction workers’ safety
339
awareness towards unsafe behaviors of co-workers. Similarly, workers’ active demonstration
340
of safe operation was correlated to their participation in safety meetings. Therefore, similar341
pairs of indicators with higher MI values validated from the theoretical sense were added342
with paths in the modified model as seen in Fig.4.343
<Insert Fig.4 here>344
The further Goodness-of-fit test for the modified model shown in Fig.4 is summarized in345
Table 6.346
<Insert Table 6 here>347
15
All the indices in Table 6, e.g., CMIN/DF value below 3, GFI over 0.90, and RMSEA348
lower than 0.05, indicated the satisfactory test results for processing the modified model.
349
Other measurements such as AGFI, CFI, NFI, and IFI values not lower than 0.90 showed that
350
the modified model met the statistical requirements shown in Table 2. The modified model
351
was hence considered suitable for further evaluation. Finally, the path coefficient and
352
significance tests were performed to evaluate the modified model. As seen in Table 7, the
353
standard error, critical ratio, as well as pvalue measuring the significance were applied to
354
investigate the correlations among safety investments, safety cognition, and behavioral
355
performance illustrated in Fig.2.
356
<Insert Table 7 here>
357
All path coefficients higher than 0 and pvalues below 0.05 indicated that all the four
358
safety investment factors were significantly correlated to the overall safety investment, which
359
further significantly contributed to safety cognition, and finally behavioral safety
360
performance. The path coefficients displayed in Fig.4 quantified the significance level of
361
each investment category to the overall safety investment. Safety incentives are found with
362
the strongest correlation to the overall safety investment with the path coefficient at 0.98,
363
followed by PPE investment (0.92), and safety education investment (0.89). Safety insurance
364
was identified as the least significant investment category, with the path coefficient at 0.75.
365
The modified model displayed in Fig.4 and Table 7 inferred that although safety investments
366
had directly significant effects on behavioral safety performance, these direct effects were367
less significant (pvalues at 0.047 and 0.001 respectively) compared to the significance levels368
of other paths in Table 7. In comparison, safety investments turned out with stronger369
correlation with safety cognition with the path coefficient at 0.90. Safety cognition was370
further significantly connected to behavioral performance. Specifically, safety cognition had371
a stronger correlation to behavioral participation with the path coefficient at 0.67 compared to372
16
its correlation with behavioral conformance (0.52). It was inferred that safety cognition373
worked as a vehicle that bridged safety investments and behavioral performance. All the four
374
investment categories were found with significant correlations to safety cognition, which was
375
found significantly affecting the two main behavioral performance factors.
376
Discussion
377
Man et al. (2017) suggested that safety incentives and safety education were key drivers
378
to reduce construction workers’ unsafe behaviors. Besides safety education and safety
379
incentives, PPE investment and safety insurance, as mentioned by Zou et al. (2007) within
380
the Chinese construction culture, were other key factors for organizations and stakeholders to
381
consider in safety investments. This study investigated the effects of safety investments on
382
employees’ behavioral safety performance with safety cognition as the mediator. Adopting a
383
three-step research methodology (i.e., theoretical modeling, questionnaire survey, and
384
Structural Equation Modeling (SEM)), it was found that the overall safety investment was
385
significantly correlated to employees’ safety cognition, and further affecting the behavioral
386
performance. Overall, this study provided a quantitative approach to verify the statement of
387
Lu et al. (2016) that safety investments contributed to enhanced behavioral performance. As a
388
step forward, this study divided the safety investment into four major categories and
389
evaluated each category’s effect on employees’ behavioral performance.
390
The social psychology theory described by Baron and Kenny (1986) indicated that the
391
stressor was input variables that could affect individuals’ behavioral outcomes. Applying the
392
social psychology theory into construction safety management, the stressor could be site
393
conditions (e.g., tight project schedule) that affect employees’ decision of whether or not to
394
behave riskily in order to achieve certain desires. Man et al. (2017) and Feng (2019) stated395
that these desires included saving time and effort, or gaining more income. Gaining more396
income in less working time was identified as one of the major causes of construction397
17
workers’ unsafe behaviors (Feng, 2019). Therefore, safety incentive was defined as one398
investment category in this study to address employees’ desire to gain more income. It was
399
verified that incentive had the highest correlation to the overall safety investment compared
400
to three other categories of investments (i.e., insurance, education, and PPE).
401
The social behavioral theories proposed by Deci and Ryan (1985) and Ryan and Deci,
402
(2000) revealed that human behaviors were driven by a variety of motivations and the
403
motivation-initiated behaviors aimed to satisfy the innate psychological desire. This desire
404
was a necessary but not a sufficient condition for employees to conduct risky behaviors.
405
Construction employees might have different motivations to behave unsafely, such as being
406
social and demonstrating self-capability (Choudhry and Fang 2008; Man et al. 2017). Lack of
407
safety knowledge or biased attitudes towards safety could drive these motivations towards
408
unsafe behaviors among newer employees. But for more experienced employees, over-
409
confidence of their own capability could also cause risky behaviors (Han et al. 2019a). It is
410
hence suggested that periodic safety training and education be carried out to construction
411
employees at different experience levels (Han et al. 2019b). Intervening construction workers’
412
motivation (e.g., gaining more income) towards unsafe behavior through education is part of
413
safety investment. Investments in safety education is needed besides incentives to correct
414
employees’ biased safety perceptions or attitudes, and to enhance their safety knowledge (e.g.,
415
proper use of PPEs). Examples of safety education investments include organizing periodic
416
safety workshops, implementing safety programs, and hiring safety professionals for site
417
monitoring, etc. Therefore, investments in safety education or training is another critical
418
factor affecting the behavioral performance of site employees.
419
Besides safety incentives and education/training, safety insurance and PPE costs are two420
other investment categories affecting employees’ behavioral performance. The Risk421
Homeostasis Theory (Wilde 1982) stated that individuals tend to take more risks if they had a422
18
stronger sense of safety. Klen (1997) further showed that workers behaved more riskily with423
PPEs. However, researchers in this study do not aim to deny the importance of PPE, but
424
emphasize that the stressor (e.g., PPE) does not necessarily lead to improved behavioral
425
performance. Instead, the mediating effect through safety cognition could bridge the
426
investment in PPE and employees’ behavioral outcomes. Individuals’ safety cognition could
427
be enhanced through proper safety education.
428
Safety incentive, as one tangible benefit from employees’ perspective, is identified as the
429
most significant contributor to the overall safety investment. The direct financial gain through
430
incentives becomes the strongest motivation for employees to behave safely. In contrast,
431
safety insurance that employers invest on site employees, is a less significant contributor to
432
behavioral performance. It is implied from the path coefficient analysis shown in Fig.4 that
433
construction employees tend to perceive tangible safety investments (i.e., incentives and PPE)
434
as stronger motivations to work safely. However, this does not mean employers should invest
435
more in safety incentives or PPE, but a more balanced and comprehensive coverage of safety-
436
related investments between tangible and intangible factors.
437
Insurance, as one intangible category from the employees’ perspective, is found with the
438
lowest effect on the overall safety investment, the importance of insurance should not be
439
downplayed. More studies could be performed to explore the effects of different types of
440
insurance on employees’ safety cognition and behavioral performance. The different types of
441
insurance include but are not limited to the legally required minimum coverage of injuries,
442
and a more comprehensive package with a wider coverage of employees’ health and safety.
443
It should be noticed that the tangible and intangible features of these four investment
444
categories are defined from the perspective of site employees, depending on whether the445
investment items could be directly sensed by employees. This study implies the gap between446
employees’ safety climate and the organizational safety culture. From the employer or the447
19
organization’s perspective, all of the four investment categories are actually tangible, as the448
organization can directly see the financial expenditure for purchasing PPE, insurance for
449
employees, incentives, and training. Nevertheless, employees would have different
450
perceptions towards the four investment categories. They would generally view incentives as
451
a more tangible category because they could gain extra income. In contrast, insurance that
452
their employer purchase for them might not be well noticed or even ignored. This gap
453
between individual employees and the organization leads to further research on bridging
454
individual needs and organizational strategies through mediators such as safety cognition.
455
Conclusion
456
This study adopted four main safety investment factors (i.e., categories), namely safety
457
education, personal protection equipment (PPE), safety incentive, and safety insurance.
458
Through site questionnaire surveys and Structural Equation Modeling approach, these four
459
categories were investigated of their correlation to site employees’ safety cognition and
460
behavioral performance. All the four investment categories were found positively
461
contributing to the overall safety investment, which was found significantly affecting site
462
employees’ safety cognition and behavioral performance. Safety cognition was also found
463
positively contributing to the behavioral performance, especially behavioral participation.
464
Among the four investment categories, the more tangible safety investment (i.e., incentives)
465
was found with the highest correlation to the overall safety investment. In contrast, the
466
intangible investment categories (e.g., insurance) were perceived by employees with lower
467
significance. The current findings indicate that there is a mediator (i.e., safety cognition) to
468
bridge investments on employees’ safety and the behavioral performance. This study
469
contributes to the body of knowledge both practically and academically. Practically, it470
provides insights for construction enterprises on the effects of safety investments on471
enhancing employees’ behavioral safety performance, as well as the significance of different472
20
investment categories towards employees’ behavioral performance. Specifically, employers473
need to realize that these investment categories (e.g., education) which are all tangible at the
474
organizational level, may be perceived differently by individual employees. Employers are
475
suggested to have balanced safety investments between tangible (e.g., incentives) and
476
intangible (e.g., insurance) categories. Academically, the current findings lead to further
477
research on how different categories of safety investments would affect employees’
478
behavioral safety performance with safety cognition as the vehicle. A positive safety
479
cognition embedded in the site safety climate and organizational safety culture is a key
480
mediator to bridge safety investments and behavioral performance.
481
Further research could focus on how to optimize the different investment categories in an
482
effective safety program aiming to establish proper site safety climate and to enhance
483
behavioral safety performance. The effects due to different arrangements of incentives can be
484
compared, for example, the effects between more frequent but smaller amounts of cash
485
awards (e.g., $100 cash award monthly per awardee) and less frequent but larger amounts of
486
incentives (e.g., $300 cash award quarterly per awardee). Currently, the initial model
487
established is limited to jobsites in south-eastern region of China. Future studies could apply
488
this model in a different geographic region worldwide, and quantify the mediating effect of
489
safety culture as the vehicle to bridge safety investments and employees’ behavioral safety
490
performance.
491
Data Availability Statement
492
Data generated or analyzed during the study are available from the corresponding author
493
by request.
494
Acknowledgement495
This research is supported by the National Natural Science Foundation of China (Grant496
No. 51408266), MOE (Ministry of Education in China) Project of Humanities and Social497
21
Sciences (Grant No.14YJCZH047), Foundation of Jiangsu University (Grant No. 14JDG012),498
and Writing Retreat Fund provided by University of Brighton.
499
References
500
References
501
Baron, R.M., and Kenny, D.A. 1986. “The moderator-mediator variable distinction in social
502
psychological research: conceptual, strategic, and statistical considerations.” Journal of
503
Personality and Social Psychology.51 (6): 1173-1182.https://sci-hub.tw/10.1037/0022-
504
3514.51.6.1173.
505
Bentler, P. M., and Chou, C. P. 1987. “Practical issues in structural modeling.” Sociological
506
Methods & Research, 16(1), 78-117.
507
Bland, J., and D. Altman. 1997. “Statistics notes: Cronbach’s alpha.” BMJ314 (7080): 572.
508
https://doi.org/10.1136/bmj.314.7080.
509
Bollen, K. A. 2014. “Structural equations with latent variables (Vol. 210).” John Wiley &
510
Sons. Chapel Hill, North Carolina, U.S.
511
Cao, X. 2018. “Research on the impact of construction workers' safety investment on
512
workers' safety performance.” Master’s thesis. Zhenjiang, Jiangsu, China. In
513
Chinese.https://doi.org/CNKI:CDMD:2.1018.837718.
514
Cao, X., Han, Y., Zhou, Y., Feng, Z., and Sun, H. 2018. “Impact of construction workers’
515
safety input to workers’ safety performance.” Journal of Civil Engineering and
516
Management, 35(3): 151-156. In Chinese.https://doi.org/10.13579/j.cnki.2095-
517
0985.2018.03.025.
518
Chen, Q., and Jin, R. 2012. “Safety4Site commitment to enhance jobsite safety management
519
and performance.” J. Constr. Eng. Manage., 138(4): 509–
520
519.https://doi.org/10.1061/(ASCE)CO.1943-7862.0000453.
521
Chen, Q., and Jin, R. 2013. “Multilevel safety culture and climate survey for assessing new
522
safety program.” J. Constr. Eng. Manage., 139(7): 805-
523
817.https://doi.org/10.1061/(ASCE)CO.1943-7862.0000659.
524
Chen, W.T., Lu, C,S., and Huang, Y.H. 2011. “Investigating the Safety Cognition of
525
Taiwan’s Construction Personnel.” Journal of Marine Science and Technology. 19(4):
526
398-408.
527
http://ss.zhizhen.com/detail_38502727e7500f26194a759be8f1f032d82f6e3911fd47b01921
528
b0a3ea255101c944b624736f9e8535ca7549d85c4cf42800f6598bc55676fd31f067dc4cf93d
529
f6accec00b714b9aa4a1d6e502b8694e?
530
Chen, Y., McCabe, B., and Hyatt, D. 2018. “A resilience safety climate model predicting
531
construction safety performance.” Saf. Sci. 109, 434-
532
445.https://doi.org/10.1016/j.ssci.2018.07.003
533
Child, D. 1990. “The essentials of factor analysis, second edition.” London: Cassel
534
Educational Limited.https://doi.org/10.2307/2061984.
535
Choudhry, R. M. 2014. “Behavior-based safety on construction sites: A case study.” Accident
536
analysis & prevention, 70: 14-23.537
https://doi.org/10.1016/j.aap.2014.03.007.538
Choudhry, R.M., and Fang, D. 2008. “Why operatives engage in unsafe work behavior:539
Investigating factors on construction sites.” Saf. Sci. 46: 566-540
584.https://doi.org/10.1016/j.ssci.2007.06.027.541
Choudhry, R.M., Fang, D., and Mohamed, S. 2007.“The nature of safety culture: A survey of542
the state-of-the-art.” Saf. Sci.45(10): 993-1012.https://doi.org/10.1016/j.ssci.2006.09.003.543
22
Cooper, M. D. 2003. “Behavior-based safety still a viable strategy.” Safety & Health,544
National Safety Council, Franklin, IN, 46–48.
545
http://ss.zhizhen.com/detail_38502727e7500f261ac46f6fc49726136913bd64630e70ec1921
546
b0a3ea255101c944b624736f9e8507dcd3e84e4901938a59c78cedb7de8c9917d3d48ddbc1
547
dc4925b1cdc8ba5cc42d75bbe2b00eef3d?&apistrclassfy=1_5_7.
548
David Garson and Statistical Associates Publishing 2015. “Structural Equation Modeling.”
549
2015 Edition, available via http://www.statisticalassociates.com/sem_p.pdf , accessed on
550
23 Apr 2018.
551
Deci, E. L., and Ryan, R. M. 1985. “Intrinsic motivation and self-determination in human
552
behavior.” New York: Plenum.https://doi.org/10.1007/978-1-4899-2271-7.
553
DePasquale, J. P., and Geller, E. S. 1999. “Critical success factors for behavior-based safety:
554
A study of 20 industry-wide applications.” J. Saf. Res. 30(4): 237–
555
249.https://doi.org/10.1016/s0022-4375(99)00019-5.
556
DeVellis, R. F. 2003. “Scale development: Theory and applications. ”2nd ed. Thousand Oaks,
557
CA: Sage.
558
http://xueshu.baidu.com/usercenter/paper/show?paperid=71b4014f14e0f87c09eee30b22af7c3
559
5&site=xueshu_se&hitarticle=1.
560
Division of Statistics + Scientific Computation 2012. “Structural Equation Modeling using
561
AMOS: an introduction.” The University of Texas at Austin. Available via <
562
https://stat.utexas.edu/images/SSC/Site/AMOS_Tutorial.pdf >, accessed on 18 Mar 2018.
563
Feng, Y. 2013. “Effect of safety investments on safety performance of building projects.” Saf.
564
Sci. 59: 28-45.https://doi.org/10.1016/j.ssci.2013.04.004.
565
Feng, Y. 2015. “Mathematical models for determining the minimum level of voluntary safety
566
investments for building projects.” J. Constr. Eng. Manage., 141(7):
567
4015015.https://doi.org/10.1061/(ASCE)CO.1943-7862.0000987.
568
Feng, Y., Teo, E.A.L., Ling, F.Y.Y., and Low, S.P. 2014. “Exploring the interactive effects
569
of safety investments, safety culture and project hazard on safety performance: An
570
empirical analysis.” Int. J. Proj. Manage., 32(6): 932-
571
943.https://doi.org/10.1016/j.ijproman.2013.10.016.
572
Feng, Z. 2019. “Study on Decision-making Deviation Mechanism of Unsafe Behavior of
573
Construction Workers from the Perspective of Motivation Conflict.” Master’s thesis.
574
Zhenjiang, Jiangsu, China. In Chinese.
575
Flin, R. H, and Mearns, K.J. 1994. “Risk Perception and Safety in the Offshore Oil Industry.”
576
SPE Health, Safety and Environment in Oil and Gas Exploration and Production
577
Conference.https://doi.org/10.1016/0950-4230(95)00032-V.
578
Glendon, A.I., and Litherland, D.K. 2001. “Safety climate factors, group differences and
579
safety behaviour in road construction. ”Saf. Sci. 39(3): 157-
580
188.https://doi.org/10.1016/S0925-7535(01)00006-6.
581
Griffin, M. A., and Hu, X. 2013. “How leaders differentially motivate safety compliance and
582
safety participation: The role of monitoring, inspiring, and learning.” Saf. Sci. 60: 196-
583
202.https://doi.org/10.1016/j.ssci.2013.07.019.
584
Guldenmund, F. W. 2000. “The nature of safety culture: A review of theory and research.”
585
Saf. Sci. 34: 215–257.https://doi.org/10.1016/S0925-7535(00)00014-X.
586
Guldenmund, F.W. 2007. “The use of questionnaires in safety culture research–an evaluation.”587
Saf. Sci. 45(6): 723-743.https://doi.org/10.1016/j.ssci.2007.04.006.588
Han, Y., Feng, Z., Zhang, J., Jin, R., and Aboagye-Nimo, E. 2019a. “Employees’ safety589
perceptions of site hazard and accident scenes.” J. Constr. Eng. Manage., 145(1):590
04018117. https://doi.org/10.1061/(ASCE)CO.1943-7862.0001590.591
23
Han, Y., Jin, R., Wood, H., and Yang, T. 2019b. “Investigation of demographic factors in592
construction employees’ safety perceptions.” KSCE Journal of Civil Engineering. 23(7):
593
2815-2828. https://doi.org/10.1007/s12205-019-2044-4.
594
Han, Y., Yin, Z., Liu, J., Jin, R., Gidado, K., Painting, N., Yang, Y., and Yan, L. 2019c.
595
“Defining and Testing a Safety Cognition Framework Incorporating Safety Hazard
596
Perception.” J. Constr. Eng. Manage. 145(12), 04019081.
597
https://doi.org/10.1061/(ASCE)CO.1943-7862.0001718.
598
Hao, X. 2015. “Statistical analysis on personal protective equipment accidents in construction
599
industry.” Safety Science &Technology, 11: 167-171. In
600
Chinese.http://kns.cnki.net//KXReader/Detail?TIMESTAMP=637031453484313750&DB
601
CODE=CJFQ&TABLEName=CJFDLAST2015&FileName=LDBK201503038&RESUL
602
T=1&SIGN=IXckwHIYGdIH5sFQkvlcexSjMMA%3d.
603
Health and Safety Executive 1999. “Reducing error and influencing behaviour.” (HSG48).
604
HSE Books, Norwich, UK. https://doi.org/10.1108/ss.1999.11017cae.007.
605
Hinze, J. 1997.“Construction Safety”. New Jersy: Prentice Hall.
606
Hinze, J. 2000. “Incurring the costs of injuries versus investing in safety” in Coble, R.J.,
607
Hinze, J. and Haupt, T.C. (eds.)Construction Safety and Health Management, Prentice-
608
Hall, New Jersey.
609
Hox, J. J., and Bechger, T.M. 1998. “An introduction to structural equation modeling.”
610
Family Science Review. 11, 354-373.https://doi.org/10.1109/MC.1987.1663443.
611
Hu, B., and Tao, K. 2015. “The Development Countermeasure of Security Mechanism of
612
Industrial Injury Insurance for Migrant Workers.” Social Security Studies. 05: 67-72.
613
http://kns.cnki.net//KXReader/Detail?TIMESTAMP=637031293949573750&DBCODE=C
614
JFQ&TABLEName=CJFDLAST2016&FileName=SHBY201505008&RESULT=1&SIG
615
N=YU%2b5VCAW7ezzKCazQcYN%2b%2bA%2fgnU%3d.
616
Hu, W., Tang, M., and Liu, Z. 2017. “Work-Related Injury Insurance Regulation and Firm
617
Productivity Growth.” Insurance Studies. 07: 101-
618
114.https://doi.org/10.13497/j.cnki.is.2017.07.009.
619
Huang, H. 2017. “How to correctly identify and control accident problem.” Construction
620
Technology. 46: 637-639.
621
http://kns.cnki.net//KXReader/Detail?TIMESTAMP=637031454708532500&DBCODE=C
622
JFQ&TABLEName=CJFDLAST2017&FileName=SGJS2017S1168&RESULT=1&SIGN
623
=lt%2fI1F5%2f2UWNLN0YvrVm3k0Wvow%3d.
624
Huang, Q., Qi, S., Zhang, Y., and Cheng, J. 2018. “SD model of intervention strategies for
625
habitual unsafe behavior of construction workers.” Journal of China Safety Science.28:
626
22-31. In Chinese.https://doi.org/10.16265/j.cnki.issn1003-3033.2018.07.005.
627
IOSH 2015. “Looking for higher standards: Behavioural safety-improving performance” pp:
628
1-5, Leicestershire, UK.
629
Jin, R., L. Tang, and K. Fang.2015. "Investigation into the current stage of BIM application
630
in China’s AEC industries." WIT Transactions on The Built Environment, 149: 493-
631
503.https://doi.org/10.2495/BIM150401
632
Kaiser, H.F., Rice, J., Little, J.,and Mark, I. 1974. “Educational and psychological
633
measurement.Am. Psychol. Assoc. 34: 111-
634
117.https://doi.org/10.1177/0013164484444028.635
Kaplan, D. 2001. “Structural Equation Modeling.” International Encyclopedia of the Social636
&Behavioral Sciences. 15215-15222.https://doi.org/10.1016/B0-08-043076-7/00776-2.637
Klen, T. 1997. “Personal protectors and working behavior of loggers.” Saf. Sci. 25(1-3): 89-638
103.https://doi.org/10.1016/S0925-7535(97)00036-2.639
24
Laufer, A. 1987. “Construction safety: economics, information and management640
involvement.” Constr. Manag. Econ. 5: 73–
641
90.https://doi.org/10.1080/01446198700000007.
642
Li, H., and Li, S. 2017. “Research on Relationship Between Safety Cognition and Unsafe
643
Behavior of Coal Miners.” Safety in Coal Mines. 11, 233-236. In
644
Chinese.https://doi.org/10.13347/j.cnki.mkaq.2017.11.062.
645
Li, Q., Ji, C., Yuan, J., and Han, R. 2017. “Developing dimensions and key indicators for the
646
safety climate within China’s construction teams: A questionnaire survey on construction
647
sites in Nanjing.” Saf. Sci.: 79, 11-18.https://doi.org/10.1016/j.ssci.2016.11.006
648
Lingard, H., and Rowlinson, S. 1998. “Behavior-based safety management in Hong Kong’s
649
construction industry: The results of a field study.” Constr. Manage. Econ. 16(4): 481–
650
488.https://doi.org/10.1080/014461998372259.
651
Lingard, H., and Rowlinson, S. 2015. “Occupational Health and safety in construction
652
management.” Spon Press, Oxon, UK.
653
Liu, J. 2018. “Study on the safety cognition deviation characteristics and formation
654
mechanism of construction workers.” Master Thesis. Jiangsu University, Zhenjiang,
655
Jiangsu, China, 12-16. In Chinese.
656
https://kreader.cnki.net/Kreader/CatalogViewPage.aspx?dbCode=CMFD&filename=1018845
657
710.nh&tablename=CMFD201901&compose=&first=1&uid=WEEvREcwSlJHSldRa1Fh
658
dXNXaEdzWU14R2IvbldQQk1CMzh4S0Z6WkkrVT0=$9A4hF_YAuvQ5obgVAqNKP
659
CYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!!
660
Lu, M., Cheung, C.M., Li, H., and Hsu, S.C. 2016. “Understanding the relationship between
661
safety investment and safety performance of construction projects through agent-based
662
modeling.”Accident Analysis & Prevention. 94: 8-
663
17.https://doi.org/10.1016/j.aap.2016.05.014.
664
Lv, H. 2014. “On Migrant Workers' Work Injury Accidents and Influence Factors——
665
Evidences from Pearl River Delta.” Social Science&Philosophy.36: 53-61. In Chinese.
666
http://kns.cnki.net//KXReader/Detail?TIMESTAMP=637031318610042500&DBCODE=CJ
667
FQ&TABLEName=CJFDLAST2015&FileName=JNXB201411011&RESULT=1&SIGN
668
=SUf6FD40l6s7SZA7stTK1WtIpD0%3d.
669
Man, S.S., Chan, A.H.S., and Wong. H.M. 2017. “Risk-taking behaviors of Hong Kong
670
construction workers – A thematic study.” Saf. Sci. 98: 25-
671
36.https://doi.org/10.1016/j.ssci.2017.05.004.
672
Marquardt, N., Gades, R., and Robelski, S. 2012. “Implicit social cognition and safety
673
culture.” Human Factors and Ergonomics in Manufacturing & Service Industries, 22 (3):
674
213–234.https://doi.org/10.1002/hfm.20264.
675
Mearns, K., Hope, L., Ford, M.T., and Tetrick, L.E. 2010. “Investment in workforce health:
676
Exploring the Implications for workforce safety climate and commitment.” Accident
677
Analysis and Prevention, 42: 1445-1454.https://doi.org/10.1016/j.aap.2009.08.009.
678
Mitropoulos, P. and Memarian, B. 2012. “Team processes and safety of workers: Cognitive,
679
affective, and behavioral processes of construction crews.” J. Constr. Eng. Manage.
680
138(10): 1181-1191.https://doi.org/10.1061/(ASCE)CO.1943-7862.0000527.
681
Neal, A., Griffin, M.A., and Hart, P.M. 2000. “The impact of organizational climate on safety
682
climate and individual behaviour.” Saf. Sci. 34, 99-109.683
Neal, D. 1995. “Industry-specific human capital: evidence from displaced workers.” Journal684
of Labor Economics. 13(4): 653-677.https://doi.org/10.2307/2535197.685
Newaz, M.T., Davis, P., Jefferies, M., Pillay, M., 2016. “The psychological contract of safety:686
The missing link between safety climate and safety behaviour in construction sites.” in:687
Arezes, P. (Ed.) International Conference on Safety Management and Human Factors,688
2016. Springer Verlag, pp: 199-210.https://doi.org/10.1007/978-3-319-41929-9_19.689
25
Nunnally, J.C., 1994. Psychometric theory 3E. Tata McGraw-Hill Education.690
OSHA. 2011. “Construction Focus Four training.” U.S. Dept. of Labor,
691
Washington,DC,〈http://www.osha.gov/dte/outreach/construction/focus_four/index.html〉
692
(Apr. 13, 2017).693
Parker, D., Lawrie, M., and Hudson, P. 2006. “A framework for understanding the694
development of organisational safety culture.”Saf. Sci. 44: 551–695
562.https://doi.org/10.1016/j.ssci.2005.10.004.696
Qiang, M., Fang, D., Xiao, H., and Chen, Y. 2004. “Research on safety investment and safety697
performance of construction projects.” China Civil Engineering Journal, 37(11): 101-107.698
In Chinese. https://doi.org/10.15951/j.tmgcxb.2004.11.018.699
Rowatt, W.C., Franklin, L.M., and Cotton, M. 2005. “Patterns and personality correlates of700
implicit and explicit attitudes toward Christians and Muslims.” Journal for the Scientific701
Study of Religion, 44: 29–43.https://doi.org/10.2307/3590517.702
Ryan, R.M., and Deci, E.L. 2000. “Intrinsic and extrinsic motivations: classic definitions and703
new directions.” Contemporary Educational Psychology, 25: 54–704
67.https://doi.org/10.1006/ceps.1999.1020.705
Schein, E. 1992. “Organisational culture and leadership”(2nd ed.). San Francisco: Jossey-706
Bass.https://doi.org/10.1057/9780230625426_15.707
Shadfar, S., and Malekmohammadi, I. 2013. “Application of Structural Equation Modeling708
(SEM) in restructuring state intervention strategies toward paddy production development.”709
International Journal of Academic Research in Business and Social Sciences, 3(12): 576-710
618.711
Shan, W., Zhang, C., and Wang, J. 2018. “Internal Social Network, Absorptive Capacity and712
Innovation: Evidence from New Ventures in China.” Sustainability. 10(4):713
1094.https://doi.org/10.3390/su10041094.714
Shao, J., Han, X., and Liu,W. 2018. “The influence and mechanism of external environment715
resource scarcity on employees remuneration preference.” Acta Psychological Sinica. 50:716
1428-1437. In Chinese. https://doi.org/10.3724/SP.J.1041.2018.01428717
Shohet, I.M., Luzi, M., and Tarshish, M. 2018. “Optimal allocation of resources in718
construction safety: Analytical-empirical model.” Saf. Sci. 104: 231-719
238.https://doi.org/10.1016/j.ssci.2018.01.005.720
Suhr, D. 2006. “Exploratory or Confirmatory Factor Analysis.” SAS Users Group721
International Conference (pp. 1 - 17). Cary: SAS Institute, Inc.722
Tang, S.L., Lee, H.K., and Wong, K. 1997. “Safety cost optimization of building projects in723
Hong Kong.” Constr. Manag. Econ. 15: 177–724
186.https://doi.org/10.1080/01446199700000005.725
Tholén, S.L., Pousette, A., and Törner, M. 2013. “Causal relations between psychosocial726
conditions, safety climate and safety behaviour - A multi-level investigation.” Saf. Sci. 55:727
62-69.https://doi.org/10.1016/j.ssci.2012.12.013.728
Toole, T. M. 2002. “Construction site safety roles.” J. Constr. Eng. Manage. 128(3): 203-729
210.https://doi.org/10.1061/(asce)0733-9364(2002)128:3(203).
730
Wang, C., Lv, S., and Yang, K. 2015.“Experimental research on safety education of
731
construction enterprises based on behavioral psychology.” Journal of Safety Science and
732
Technology.11: 158-165. In Chinese.https://doi.org/10.11731/j.issn.1673-
733
193x.2015.04.026.
734
Wang, W., Yang, Q., Lu, M., and Yu, Q. 2014. “Research on the construction safety
735
investment regulation evolutionary game and application.” Advances in Intelligent Systems
736
and Computing, 280: 145-153.https://doi.org/10.1007/978-3-642-55182-6_13.
737
26
Wang, Y., Wu, L., and Chen, F. 2014. “On the evolutionary game of the airline hazard events738
punishment mechanism based on sd. Journal of Safety and Environment.” Journal of
739
safety and Environment.16: 169-173. In Chinese.https://doi.org/10.13637/j.issn.1009-
740
6094.2016.02.034.
741
Wilde, G.J. 1982. “The theory of risk homeostasis: implications for safety and health.” Risk
742
analysis, 2(4): 209-225.https://doi.org/10.1111/j.1539-6924.1982.tb01384.x.
743
Wirth, O., and Sigurdsson, S. O. 2008. “When workplace safety depends on behavior change:
744
Topics for behavioral safety research.” J. Saf. Res., 39(6): 589-
745
598.https://doi.org/10.1016/j.jsr.2008.10.005.
746
Wu, M. 2009. “Structural Equation Modeling: conducting and implementing AMOS.”
747
Chongqing University Publisher. Chongqing, China. 52-58, 159-170. In Chinese.
748
http://ss.zhizhen.com/detail_38502727e7500f2620c097755dc9d94735c47471da6826df1921b
749
0a3ea25510134114c969f2eae5c2a766e49c915a8a04da699dbe30ebdc390e52aa49ae43412598
750
699ff6dfcc888b7bc4d071878f220?&apistrclassfy=0_3_9.
751
Yoon, H., Lee, H., and Moon, L.I. 2000. “Quantitative business decision-making for the
752
Investment of Preventing Safety accidents in chemical plants.” Computers and Chemical
753
Engineering, 24:1037-1041.https://doi.org/10.1016/s0098-1354(00)00531-7.
754
Zeng, J., Liu, P., and Liu, Y. 2009. “Institution, implementation, and compliance—Coal Mine
755
Safety Control Analysis.” Coal Technology. 28: 187-189. In Chinese.
756
Zhang, J. 2017. “Research on difference of hazard perception ability for construction
757
workers.” Master’s thesis. Jiangsu University, Zhenjiang, Jiangsu,
758
China.https://kns.cnki.net/KCMS/detail/detail.aspx?dbcode=CMFD&dbname=CMFD201
759
801&filename=1017719197.nh&uid=WEEvREcwSlJHSldRa1FhdkJkVG1EREREY2F4S
760
1BZbWxTUEpGOTMzanFrcz0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovw
761
HtwkF4VYPoHbKxJw!!&v=MDQyOTZyV00xRnJDVVJMT2ZidVpxRmlEblViL0tWRjI
762
2R2JTNUY5REZxSkViUElSOGVYMUx1eFlTN0RoMVQzcVQ=
763
Zhang, J., and Li, J. 2016. “Applying SEM in studying factors affecting construction workers’
764
safety behavior.” Journal of Safety and Environment,Vol. 3,pp. 182-
765
187.https://doi.org/10.13637/j.issn.1009-6094.2016.03.036.
766
Zohar, D., and Luria, G. 2004. “Climate as a social-cognitive construction of supervisory
767
safety practices: scripts as proxy of behavior patterns.” Journal of applied psychology,
768
89(2): 322.https://doi.org/10.1037/0021-9010.89.2.322.
769
Zou, P. X., Zhang, G., and Wang, J. 2007. “Understanding the key risks in construction
770
projects in China.” Int. J. Proj. Manage. 25(6): 601-
771
614.https://doi.org/10.1016/j.ijproman.2007.03.001.
772
Zou, P.X.W., Sun, A.C.S., Long, B., Marix-Evans, P. 2010. “Return on Investment of Safety
773
Risk Management System in Construction.” In: Proc., CIB World Congress, Salford
774
Manchester UK, 11–13 May, 2010.
775
776
777
778
779
780
781
782
783
784
785
786
787
27
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
Table 1. A total of 28 indicators in the questionnaire survey
805
Category
Indicator in the questionnaire
References
PPE
investment
X1:My employer provides me with good personal protection equipment that
motivates me to participate actively in safety-related activities.
Feng (2013);
Feng et al.
(2014); Hao
(2015); Lv
(2014)
X2:The specific personal protective equipment that is related to my job
duties makes me behave safely in my work.
X3:The adequate personal protective equipment improves my understanding
of the site hazards (e.g., working at height).
Investment in
safety
education
X4: The experiential safety education, for example, watching video,
experiencing jobsite operation conditions with Virtual Reality and other
safety education approaches, motivates me to more effectively participate in
safety activities.
Huang et
al.(2018);
Qiang et al.
(2004);
Wang et
al.(2015)
X5: The specific safety education related to my work makes me well comply
with safety rules and regulations.
X6: The diversified and varied safety education makes me better understand
the occupational safety risks.
Safety
incentive
X7: The high intensity of safety incentive motivates me to more effectively
participate in setting safety plans and objectives.
Shao et al.
(2013);
Wang et al.
(2014)
X8: Compared to verbal or certificate-based safety awards, the cash
incentive better motivates me to comply with company's safety rules.
X9: Compared to multiple small safety incentives, a single but larger amount
of safety incentive improves my awareness of site hazard sources.
Safety
insurance
X10: Work-related injury insurance motivates me to proactively correct the
unsafe behavior of peers.
Hu et al.
(2017); Hu
and Tao
(2015)
X11: Medical insurance makes me work in the safest way.
X12: The comprehensive safety insurance that my employer purchases for
me, has led to a higher level of awareness that I have towards unsafe
behavior of my peers.
Safety
cognition
X13: I can fully realize the hazards during work.
Huang
(2017); Li
and Li
(2017);
Mitropoulos
and
Memarian
(2012);
Zohar and
Luria (2004)
X14: I can fully understand the occupational hazards corresponding to
different types of site duties.
X15: I know well different unsafe behavior types and the consequences at
work.
X16: I have developed my knowledge and understanding of the safety rules
and regulations.
X17: I have developed my strong awareness of hazard sources and
occupational risks.
X18: I am fully aware of my peers’ unsafe behaviors and relevant safety
regulations
28
Safety
behavioral
participation
X19: I actively participate in the development of site safety plans.
Choudhry et
al. (2007);
Wirth and
Sigurdsson
(2008);
Choudhry
(2014)
X20: I will stop the unsafe behavior of my peers during work.
X21: I participate actively in the improvement of site safety.
X22: I actively demonstrate safe operation and behaviors to other
employees.
X23: I actively participate in safety meetings.
Safety
behavioral
conformance
X24: I always wear the right and appropriate safety protection equipment
during work.
Neal (1995);
Toole
(2002); Zeng
et al. (2009)
X25: I always follow the company's safety rules and regulations during
work.
X26: I always work in the safest way as I can on-site.
X27: I always behave according to the correct safety procedures on-site.
X28: I often remind my peers of the importance of safety on-site.
806
807
808
809
Table 2. Demographic summary of survey participants (N=326)
810
Category
Sample size
Percentage (%)
Gender
Male
282
86.5
Female
44
13.5
Education level
Primary school or below
53
16.3
Middle School
140
42.9
High School
53
16.3
College or university
80
24.5
Job position
Workers
178
54.6
Crew foremen
73
22.4
Management personnel
75
23.0
Years of site
experience
0-10
138
42.3
10-20
165
50.6
20-30
23
7.1
811
812
813
814
815
816
817
818
819
820
821
822
29
823
824
825
826
827
828
829
830
831
832
833
Table 3. Definitions of Goodness-of-fit indices (source from Wu, 2009)
834
Measurement
Definition
Numerical
range
Satisfactory range
Ideal range
CMIN/DF
Ratio of normed chi-
square to degree of
freedom
＞0
≤5
≤3
RMSEA
Root Mean Square Error
of Approximation
0-1
≤0.08
≤0.05
p
Level of significance
0-1
≤0.05
≤0.05
RMR
Root mean Square
Residual
/
The lower value
indicates a higher
degree of goodness
The lower value the
better
GFI
Goodness of Fit
0-1
≥0.80
≥0.90
AGFI
Adjusted Goodness of Fit
0-1
≥0.80
≥0.90
NFI
Normed Fit Index
0-1
≥0.90
≥0.90
IFI
Incremental
Fit Index
0-1
≥0.90
≥0.90
CFI
Comparative Fit Index
0-1
≥0.90
≥0.90
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
30
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
Table 4. Reliability test results of the factors based on 28 indicators
870
Factor
Cronbach’s Alpha
Number of indicators
PPE investment
0.686
3
Safety education
0.668
3
Safety incentives
0.702
3
Safety insurance
0.751
3
Safety cognition
0.817
6
Safety behavioral participation
0.823
5
Safety behavioral conformance
0.828
5
Overall Cronbach’s alpha value
0.947
28
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
31
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
Table 5. Goodness-of-fit test for the initial model
912
Model type
CMIN/DF
RMSEA
P
RMR
AGFI
GFI
NFI
IFI
CFI
Initial model
1.645
0.045
0
0.024
0.870
0.891
0.870
0.944
0.938
Standard model
1
1
1
1
Independent
model
11.492
0.180
0
0.216
0.132
0.192
0
0
0
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
32
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
Table 6. Goodness-of-fit test for the modified model
958
Model type
CMIN/DF
RMSEA
P
RMR
AGFI
GFI
NFI
IFI
CFI
Initial model
1.311
0.031
0
0.021
0.900
0.916
0.901
0.975
0.970
Standard model
1
1
1
1
Independent
model
11.492
0.180
0
0.216
0.132
0.192
0
0
0
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
33
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
Table 7. Path coefficient analysis and significance tests of the initial model
1004
Path
Estimate
Standard
Error
Critical
Ratio
p
Standardized
Estimate
Safety investment =>Safety cognition
0.844
0.092
9.206
***
0.899
Safety investment =>Safety behavioral
participation
0.343
0.173
1.986
0.047*
0.321
Safety investment =>Safety behavioral
conformance
0.544
0.165
3.290
0.001**
0.479
PPE investment <=Safety investment
0.964
0.105
9.172
***
0.920
Safety incentives <=Safety investment
0.979
0.101
9.172
***
0.981
Safety insurance investment <=Safety
investment
0.925
0.099
9.315
***
0.753
Safety education investment =>Safety
investment
1.000
0.892
Safety cognition =>Safety behavioral
participation
0.760
0.195
3.901
***
0.668
Safety cognition =>Safety behavioral
conformance
0.629
0.178
3.542
***
0.521
Note：1.* denotes that p＜0.05; **denotes p＜0.01; ***means p＜0.001; 2. Following the guide of Wu1005
(2009), the estimate for safety education investment correlating to safety investment is standardized as 1to1006
run the significance tests for other paths in Table 7.1007
1008
