2.1. Research Background
In the consideration of a worker’s health, his or her individual characteristics, specifically BMI by heart rate, height, and weight, constitute important biometric information. Previous studies have linked heart rate variability (HRV) to physiological responses and physical activity [
21]. Other studies have used the heart rate as an index to confirm physical health and activity status [
22,
23]. Another study used biological information, such as heart rate, to examine the relationship between office workers’ job satisfaction and work productivity [
24]. Few studies have examined these relationships.
Workload is influenced by a variety of factors, including individual physical (e.g., age, exercise, and nutrition) and environmental factors (e.g., temperature, humidity, and noise) [
25,
26]. Various factors are known to contribute to workload [
27]. However, it is difficult to determine working conditions by taking all factors into account. Therefore, measuring physical loads, based on workers’ heart rate-based measurements, can be an alternative means of understanding these factors [
28,
29,
30,
31]. Continued engagement in activities with a high physical workload results in physical fatigue among workers, causing reduced productivity and motivation, carelessness, and poor judgment, which affects work quality and job satisfaction. In addition, it is widely recognized in the construction industry that high physical workloads can cause accidents and injuries [
32]. In regions such as Japan, where the working population is aging, an increase in the proportion of older workers may affect a decline in the workforce in the construction industry [
33,
34].
2.2. The Current State of the Japanese Construction Industry
The number of construction workers in Japan was at its peak in 1997, reaching 6.85 million, and has been declining since, to 5.30 million in 2018. The number of workers on construction sites decreased to 4.55 million in 1997 and 2.97 million in 2018 [
35]. In addition, construction workers are mostly at an advanced age, with 34.1% found to be >55 years old in 2018, and 11.0%, <29 years old. There is an increasing need to pass on technology to the next generation and secure long-term leaders [
36,
37].
Figure 1 and
Table 1 show the composition of construction workers in Japan and in our study by age. The two-sample Kolmogorov–Smirnov test (two-sided
p = 0.8928) was performed to determine whether the two groups were comparable according to age. The results showed no difference between the two samples at the significance level of 5%. In addition, it could be surmised that the study participants were in the same age group as that of construction workers in Japan. The age composition and average age of the current study participants were very similar to those of workers in the Japanese construction industry.
Older workers had the best perceptions on safety, indicated the highest level of job satisfaction, and recorded the lowest accident involvement rate [
16]. From a practical perspective, understanding age-related perceptions of the workplace would benefit management’s decisions regarding workers’ adaptability, work effectiveness, implementation of safety management policies, and handling of age-related accident characteristics [
16,
18,
19]. The main objective of this study was to compare the effects of psychological factors on construction workers’ perceptions of their work across age groups. In the study, the workers were divided into two groups. Workers under the average age (≤45) were categorized as young workers and workers over the average age (≥46) were categorized as older workers.
2.3. The Hypothetical Model of Construction Workers
A literature review was conducted to investigate issues related to productivity through team performance within construction companies. In this study, employees’ perceptions of occupational safety were evaluated using the Work Safety Scale (WSS) [
38]. Questions from the 2011 Workplace Employment Relations Study (WERS) were used to measure workers’ workplace awareness, the relationship between supervisors and workers, job satisfaction, attitudes toward work, and skills [
39,
40]. Through an interview process with experienced site supervisors of the surveyed construction company, we devised a study questionnaire that consisted of items on psychological factors that may affect construction workers’ perception of work, with five of these items designed to address questions relating to the work environment.
Our self-developed questionnaire was used to assess construction workers’ perceptions regarding the suitability of their working environment. Below, we present the hypotheses that this study intended to test; we also explain the link between these and previous research, as discussed above.
2.3.1. Hypothesis 1 (H1). Health Risk Indicators Affect the Feeling of Safety and Proactive Work Behavior in the Workplace
In this study, the health risk indicators of workers were measured by the resting heart rate and BMI. The construction industry is subject to strict work schedules, irregular working hours, a need for a controlled air temperature in the work environment to ensure workplace safety, and workers being subject to poor work performance as a result of work stress [
41]. To investigate the effects of a hot working environment in construction sites and the health of workers, and to reduce workers’ heat stress in the workplace, strict management of the working environment and the management of workers’ health are necessary. Previous studies have shown the importance of reducing the impact on workers’ health [
42,
43].
An elevated resting heart rate has an impact on mortality risk, heart disease, and cardiovascular disease [
44,
45]. In a study of a cohort of East Asians [
35], people with a BMI (weight divided by height squared) in the range of 22.6 to 27.5 had the lowest risk of death. The higher the BMI, the higher the risk of death from cancer, cardiovascular disease, and other causes.
Construction workers care about their health and well-being at work [
46] and recognizing their daily health risks has increased expectations for working conditions, which is consistent with increased sensitivity to safety and positive work. As part of an annual health check [
47], the construction company regularly measures employees’ physical characteristics, such as heart rate, blood pressure, height, and weight, in order to increase their health and health care awareness, including providing workers with their heart rate, weight, and BMI measurements. Workers’ perceptions of health risks can lead to expectations and positive behaviors about their working conditions.
2.3.2. Hypothesis 2 (H2). A Feeling of Safety in the Workplace Affects Job Satisfaction
Overload and poor working conditions reduce workers’ job satisfaction [
48]. Job satisfaction depends on a variety of factors, but strategies to improve it depend on work conditions [
49]. Mclain focused on workers’ stress related to their perceptions of work environments and showed that individuals are regularly exposed to health and safety threats, and that subjective interpretations of risk can affect job satisfaction [
50]. Analysis of job satisfaction and perceptions of safety show that workers with high job satisfaction have a positive view of safety in the work environment and that those who are dissatisfied have a negative view of safety in that environment [
51]. Other studies have shown the relevance of both work satisfaction and perceived environmental safety [
52,
53]. Workers can work without worry and anxiety when they have a sense of security in the work environment and high work satisfaction.
2.3.3. Hypothesis 3 (H3). Job Satisfaction in the Workplace Affects Work Skills
There is no general agreement on defining job satisfaction or occupation yet [
54]. Job satisfaction is a complex and multifaceted concept that has different meanings according to different authors. Job satisfaction is usually related to motivation, but it is not the same as motivation and may be related to personal feelings of achievement [
55]. Workers’ job satisfaction is correlated with the breadth of their work abilities, which play an important role in the workplace [
56], the congruency between workers’ required and actual skills [
57], and workers’ skill development [
58].
In most of the literature on job satisfaction, there are dealings with the effects of explanatory variables such as wages, working hours, education, health status, contract type, and work skill [
57]. Matching worker expectations with the actual situation at work affects job satisfaction. This study considers matches in job satisfaction, and job skills which is one of the work aspects. Bos et al. noted that inconsistencies between actual work skills and work requirements can have a negative impact on work life and therefore affect the relationship between work skills and job satisfaction [
59]. Workers with high job satisfaction are likely to have higher professional and technical skills related to construction machines and specialized equipment at the workplace, resulting in more productivity.
2.3.4. Hypothesis 4 (H4). Team Performance in the Workplace Is Affected by Workers’ Work Skills, Feeling of Safety, and Proactive Behavior
Hypotheses 4-1–4-3
As a result, work skills (H4-1), feeling of safety (H4-2), and proactive work behaviors (H4-3) increase workers’ productivity through teamwork. The awareness of team performance means that workers’ awareness leads to team effectiveness and organizational performance [
69]. In assessing team performance, although there are available many models of team performance, there is no one-size-fits-all approach, and it must be tailored to its environment [
71]. It is important for a highly productive team to engage in activities among team members, rather than relying solely on task operations such as interacting with tools and systems [
72]. In this study, team activities, such as cooperation and communication among team members, are important in recognition of team performance.
2.3.5. Structural Model and List of Variables
For this study, we developed a hypothetical structural model for a workers’ perceptions model involving six latent variables, based on previous studies, past empirical knowledge, theoretical expectations, and evaluations by a co-researcher. To analyze the relationship between workers’ health risk indicators and awareness, the hypothetical model of this study is shown in
Figure 2. The definition of the latent variables is shown in
Table 2.
2.4. Data Collection
We analyzed the effects of, and relationships between, self-awareness, occupational safety awareness, productivity awareness, and work awareness in construction work, using data obtained from questionnaires completed by construction workers, as well as the workers’ resting heart rate and BMI. Structural equation modeling (SEM) was used to analyze the questionnaire data, while confirmatory factor analysis (CFA) was used to analyze the relationship between the structure of the hypothetical model and observed variables. R Language 3.5.1 was used as the programming language and development environment for statistical analysis; R Studio was used as the development environment.
In the construction company that participated in the study, the workers regularly undergo medical check-ups and physical assessments and they also have an opportunity to take note of their own measured values. Other than age, participants were required to provide physical information such as height and weight, and biological information regarding resting heart rate. A 5-point Likert scale was used for responses on the Work Environment Awareness Questionnaire to enable assessment of the impact of work environment awareness on latent variables in construction workspaces. Participants were required to respond to questions relating to feelings of safety, proactive work behavior, job satisfaction, work skills, and team performance (
Appendix A).
We explained the purpose of this research to all participants before the questionnaire. The participant's rights (participation in the questionnaire did not cause any disadvantage to the participants, ensuring that the collected data was anonymous, and that the participants could be interrupted during the questionnaire response) were explained to the participants. Thirty-three participants who did not respond to any of the questionnaire items were not included in the analysis. The total number of participants used in the analysis was 324, with 165 young workers aged ≤45 years, and 159 older workers aged ≥46 years. Participants’ demographics, such as gender, age, employment status, and employment level are shown in
Table 3. The composition of either worker group is shown in
Table 4.
2.5. Data Analysis
We compiled the questionnaire results related to the observed variables that affected each latent variable in the virtual model in
Figure 2 and performed statistical analysis on Group A (young workers) and Group B (older workers).
Prior to analysis of the questionnaire responses through SEM, the reliability of the responses was analyzed using CFA [
73]. In a hypothetical structural model, observed variables with a weak factor loading for the latent variables are excluded [
74]. Groups A and B initially consisted of 17 observed variables, but after analysis, seven observed variables did not show a strong factor load on the latent variables in both groups. As a result, the number of variables was reduced to 10, as weak factor loadings were eliminated [
75]. In both groups, the deleted observed variables in feeling of safety is FS3; proactive work behaviors are PWB3 and PWB4; job satisfaction is JS3; work skills are WS3; and team performance are TP3 and TP4 (
Appendix A). Before further analysis, the measurement properties were examined to ensure the model’s reliability and validity [
76]. First, to check the reliability and validity of the data, we obtained Cronbach’s alpha [
75]. A value greater than 0.7 is generally considered ideal to ensure data reliability [
77]. The reliability values for all constructs are shown in
Table 5. All values were above 0.7, indicating good confidence for both datasets; these datasets were subsequently used for analysis.
The BMI and resting heart rate obtained from the worker’s answer were used as the observed variables for investigating the effect of the latent variable health risk index (HRI). In Groups A and B, HRI analyzes the relationships between the other five latent variables by SEM.
The descriptive statistics and correlation results for the young workers are shown in
Table 6. The latent variables with the highest ratings were productive work behaviors (PWB; mean = 3.99, standard deviation (SD) = 0.86), followed by team performance (TP; mean = 3.94, SD = 0.90), and feelings of safety (FS; mean = 3.65, SD = 0.94). The average scores for job satisfaction (JS) and work skills (WS) were 3.31 and 3.08, respectively. HRI did not yield any statistically significant correlations with other latent variables, and WS did not show significant correlation with FS and PWB. With the exception of these, the remaining variables were positively correlated with each other. Correlations were found between JS and WS (r = 0.21,
p < 0.01), between FS and TP (r = 0.26,
p < 0.01), and between WS and TP (r = 0.27,
p < 0.01). The correlation between FS and JS (r = 0.49,
p < 0.01), and that between PWB and TP (r = 0.41,
p < 0.01) constituted the largest correlations among the latent variables.
The descriptive statistics and correlation results for older workers are shown in
Table 7. The latent variables with the highest ratings were PWB (mean = 3.84, SD = 0.82), followed by TP (mean = 3.80, SD = 0.89), and FS (mean = 3.59, SD = 0.87), similar to the results for young workers. The average scores for WS and JS were 3.44 and 3.21, respectively. HRI showed a negative correlation with FS and PWB, however the correlation between HRI further PWB was not significant. In addition, other latent variables that showed a positive correlation with HRI were not significant. The relationship between JS and PWB was not significant. With the exception of these, all variables were positively correlated and significant. The correlation between FS and JS (r = 0.39,
p < 0.01) and that between JS and WS (r = 0.48,
p <0.01) constituted the largest correlations among the latent variables. Correlations were found between WS and TP (r = 0.22,
p < 0.01), between FS and TP (r = 0.29,
p < 0.01), and between PWB and TP (r = 0.25,
p < 0.01).