Age differences significantly affected safety climate across 54 Hong Kong construction sites, with older employees having more favourable views than younger workers [11] . Older employees may have a lower risk tolerance due to hazard awareness and near miss experiences than younger workers. Likewise, [12] concluded that age significantly affects the construction safety climate in Saudi Arabia. Both [11] [12] corroborate the critical role of ageing in shaping safety perceptions across different geographical locations. Conversely, a systematic literature review revealed insignificant differences in work performance between old and young workers [13] . These mixed results indicate the importance of considering potential differences in cognitive and cultural contextual factors contributing to the observed age-related discrepancies in safety climate.

It was concluded that gender differences influenced worker safety perceptions in Ghana [14] . Similarly, [15] found significant effects of gender differences on safety climate across nine construction sites in China. Another study concluded that gender impacted risk perception in Saudi Arabia [16] . These studies revealed that female workers in emerging economies have a better safety climate than their male counterparts. However, [17] concluded that gender did not affect worker safety behaviour among construction workers in China. This may indicate the presence of safety climate moderating factors such as personality traits, cognitive abilities, and organisational context. Therefore, further investigation across different regions and industries can provide a comprehensive insight into how gender affects safety climate.

Marital status was found to influence the safety climate in construction projects in Hong Kong [18] . Also, [19] concluded that married workers had a favourable safety climate across 22 construction projects in Hong Kong. Workers with family responsibilities may comprehend potential risks better and exhibit voluntary behaviour to maintain job security, leading to better attitudes and behaviours towards safety. However, these findings are specific to the Hong Kong context and may not be universally applicable. The cultural and social dynamics associated with the construction industry in Hong Kong could influence the relationship between marital status and the safety climate. Hence, further research incorporating diverse geographical and cultural contexts can provide important insights.

Work experience was observed to significantly affect the safety climate in Saudi Arabia construction projects [12] . This finding could be attributed to more appreciation of safety risk with experience, thus shaping safety behaviour. In contrast, [20] demonstrated that work experience had little influence on the safety climate in a Chinese construction company. Recently, [17] also found that years worked in a particular occupation did not affect worker safety behaviour among Chinese construction workers. The literature review above shows inconsistent conclusions. Hence, there is a need to explore the underlying reasons for these discrepancies further, including the country-specific characteristics of the construction industry. These inconsistencies underscore the importance of considering contextual influences when implementing safety programmes to enhance the safety climate in construction settings.

A comprehensive systematic review of construction literature and a quantitative study in a construction environment in the United States of America showed that education significantly influenced safety risk perceptions [21] . Likewise, [12] found that the education level of construction workers influenced the safety climate in Saudi Arabia. Similarly, [22] showed a positive association between education level and the safety climate of industrial workers in Ghana. The evidence by [12] [21] [22] indicates a positive link between the level of education and safety climate among construction workers across North America, Asia, and Africa. This suggests that highly educated workers may have better occupational hazard awareness and risk mitigation knowledge. This finding implies that managers could consider developing targeted safety training and awareness programmes to enhance safety understanding among less educated workers.

On the basis of the literature review above, the following research hypotheses have been formulated as follows:

H1: There is a significant difference in the perception of safety climate between gender (male and female) groups.

H2: There is a significant difference in the perception of safety climate between marital status (married and single) groups.

H3: Education is significantly related to perceptions of the safety climate, with more educated employees having a more positive perception than less educated employees.

H4: Age is significantly related to the perception of the safety climate, with older employees having a more positive perception than younger employees.

H5: More experienced employees have a more favourable safety climate perception than less experienced ones.

The first portion of the questionnaire comprised demographic questions relating to the respondents, including their age, gender, marital status, education level, and work experience. This information was used to assess the influence of these personal characteristics on safety climate. Part two of the questionnaire consisted of 31 questions relating to safety climate dimensions namely management commitment, supervision commitment, worker involvement, safety commitment, rules compliance, tolerance to risk, communication and competence. The 31 sample questions were created from Fang’s 87 questions [11] . Employees marked the appropriate responses in the Likert scale that ranged from strong disagreement to agreement. A pilot study of 22 safety practitioners reviewed and improved the questionnaire.

The Statistical Package for the Social Sciences (SPSS) software was used to perform inferential statistics namely t-test and Analysis of Variance (ANOVA).

This quantitative research examined demographic factors influencing safety climate in construction projects in the case study organisation using a survey questionnaire.

The project construction population was 1000 employees. A total of 206 employees participated in the study and were selected through stratified random sampling to represent the employment roles equally. This 206 sample is above 169 samples to allow for generalisation from a random sample assuming a 7% sampling error [23] .

The study participants were given blank questionnaires before commencing the work shift and requested to return completed forms at the start of the next shift. The respondents confidentiality was safeguarded by ensuring that they completed the questionnaire anonymously. The study participants provided verbal consent prior to data collection.

A t-test was performed to evaluate the influence of gender and marital status (demographic variables) on safety climate. The t-test is essential for testing differences in mean scores among diverse groups [24] . The hypothesis was accepted for values less than 0.05. Furthermore, the ANOVA statistical method was applied to determine if the level of education, age, and working experience (demographic variables) affected safety climate. The t-test is suitable for comparing statistical significance between means of two groups e.g married and single while ANOVA is ideal for measuring statistical significance between means of more than two groups e.g diverse years of work experience.

A t-test was undertaken to determine if the differences in mean scores for gender and marital status were statistically significant (See Table 1 ).

Source: SPSS.

For gender, t statistic (204) = −3.363 and significance level (p < 0.001), suggests a statistically significant difference in safety climate perceptions between the two groups. Hence, the hypothesis (H1) suggesting significant differences in the perception of safety climate between gender groups is accepted. Furthermore, marital groups had the following results: t (204) = −1.404 and (p < 0.081) for the one tailed test and (p < 0.162) for the two-tailed. The p-values for marital status are greater than 0.05, suggesting insignificant differences in means between the two groups. Hence, hypothesis (H2) indicating significant difference in the perception of safety climate between marital status groups is rejected.

The ANOVA test examined the interaction between safety climate and demographic variables namely education, age, and work experience (See Table 2 ).

Source: SPSS.

The significance level (p-value) for level of education (0.350) and age (0.091) are higher than the significance level of 0.05, indicating statistically insignificant differences in safety climate scores amongst the various levels of education and age groups respectively. Hence hypotheses H3 and H4 suggesting that education and age differences significantly affect perceptions of the safety climate, are rejected. The associated significance level for work experience of 0.019 is less than 0.05, suggesting that the interaction effect among the different work experience groups is statistically significant. Hence hypothesis H5 that more experienced employees have a better safety climate perception than newer employees is accepted.

The ANOVA analysis showed that age did not significantly impact opinions of safety climate. These results contrast past and recent research, which suggests that age differences significantly affect safety climate perceptions, with older employees having a positive view compared to younger workforce [11] [12] [17] . Young workers may underestimate safety risks more than old employees [15] . The insignificant differences between perceptions of safety climate among the different age groups may point to older employees influencing the younger workforce’s safety culture. This could also reflect an organisational culture creating similar employee behaviours which can be confirmed qualitatively.

The ANOVA test showed that age had a statistically insignificant impact on safety climate. These results are inconsistent with most safety climate research suggesting that more educated construction workers exhibited positive safety behaviour [12] [21] [22] . Therefore, context-specific and other factors could be influencing the significance of education level on the safety climate. Thus, future research could focus on understanding these contextual issues.

The ANOVA results showed that years of experience in the organisation significantly influenced the safety climate. This suggests that years of experience in the company significantly shape these perceptions. These statistical findings are consistent with [12] assertion that years of experience significantly affect the perception of the safety climate amongst the construction workforce. The practical implication of this finding is that work experience needs consideration when designing and implementing safety initiatives. For example, more experienced employees can be assigned to high-risk workstations, while the inexperienced workforce can be transferred to low-risk areas. Again the more experienced workers can be appointed as mentors, coaches and role models for the inexperienced employees to assist them in attaining the desired safety maturity.

The t-test proved that gender has a considerable role in worker perceptions of the safety climate, with females having a better perception than males. These findings are consistent with historical and recent research indicating that gender differences influence worker safety perception in construction projects [14] - [16] . This highlights the need for gender-specific safety interventions such as tailoring safety programmes to address the specific needs and perceptions of different gender groups. Again this finding supports the call for deliberate inclusion of females into a construction sector predominantly dominated by males.

The t-test revealed insignificant mean differences between married and single employees. The t-test results contrast research findings demonstrating that marital status influences the construction safety climate [18] [19] . A qualitative deep-dive analysis of the reasons behind these perceptions can provide more insight into the findings.

The study was conducted in a specific time frame and hence future studies could be longitudinal to assess changes in perceptions over time. Secondly, while the sample size was adequate to generalise at 7% level of error, future research could further reduce the error level to below 5% by further increasing the sample size.