Validating the Staff Satisfaction Index and the Happy Career for In-Service Firefighters

Community Health Department, Faculty of Medicine, Universiti Kebangsaan Malaysia, Jalan Yaacob Latiff 56000 Cheras, Kuala Lumpur, Malaysia Headquarters Fire and Rescue Department of Malaysia, Lebuh Wawasan, Presint 7, Putrajaya 62250, Malaysia School of Management, Universiti Sains Malaysia, Minden 11800, Penang, Malaysia Daffodil International University, Dhaka, Bangladesh Faculty of Economics and Business, Universiti Malaysia Sarawak (UNIMAS), Kota Samarahan, Malaysia Fakulti Ekonomi dan Pengurusan (FEP), Universiti Kebangsaan Malaysia (UKM), Bangi, Malaysia Fakulti Pengurusan dan Perniagaan, Universiti Teknologi MARA (UiTM), Shah Alam, Malaysia


Introduction
Firefighter's duties are unique and unpredictable, depending on their assigned tasks in an undetermined work environment in times of crisis. erefore, firefighter's satisfaction and feelings of happiness while performing their duties are worth exploring as a platform for continuous improvement in managing employee safety, health, and welfare issues. e typical employee satisfaction questionnaire is not suitable to be administered to firefighters because the developed items are generic and lack transcultural adaptation for in-service firefighters.
erefore, a tailor-made tool to measure the firefighter's level of satisfaction and happiness at work is needed.
Developing a satisfaction and happiness tool is critical for understanding the firefighter's cognitive, emotional, and physical views/experiences and the work environment they interact with. Protection against safety hazards at work and the provision of welfare frequently emerge as important factors based on various shared experiences pertaining to their tasks. For example, wearing full gear while engaged in firefighting activities to protect them against hazardous chemicals arising from the combustion of on-site materials. To wear full gear, one must be physically and mentally fit to ensure the heavy equipment does not cause harm while handling other heavy firefighting equipment. Wearing full gear and hauling hoses during the fire response requires 42.77 ml/kg/min of oxygen consumption [1]. is demands high cardiovascular endurance and mental resilience, especially when the time to task completion is prolonged. Like employees in other organizations, firefighters seek recognition. Rewards for firefighters who do a good job include salary and special allowances, career advancement, injury compensation, and an organizational climate that takes care of their needs. e pooling and creating of satisfaction measurement items (Table 1) were guided by the Occupational Safety and Health Act of 1994 [2]. Two theoretical models (i.e., Maslow's hierarchy of needs, as explained by a contemporary researcher [3] and Herzberg's motivator-hygiene theory [4]) and experiential input from firefighters were used. e items were adopted from the Job Descriptive Index (JDI) [5] to measure the staff's job satisfaction. Originally, the JDI was a "facet" measure of job satisfaction, meaning that respondents are asked to think about specific facets of their job, such as coworkers, the work itself, pay, opportunities for promotion, and supervision. e respondents were asked to rate their satisfaction with these specific facets. However, our satisfaction measurement omitted teamwork among coworkers because of firefighters' strong bond in providing fire protection and suppression services and responding to other types of crises. Hence, this facet was considered inappropriate to include.
In view of the multidimensionality of the satisfaction measurement items, the cluster of items is an index rather than a simple scale giving one composite score-in this case, a satisfaction score. e index was called the Staff Satisfaction Index (SSI). A Happy Career (HC) scale was developed to measure positive emotions while performing assigned firefighter duties. Furthermore, the scale is needed to complete the statistical analysis. It consists of five multidimensional items capturing meaning, personality, fit, work environment, and skill utilization. e items were derived from the literature [6], data from unstructured interviews with in-service firefighters, and the authors' personal observations. is study aims at using confirmatory factor analysis (CFA) to validate that the generated items fit the hypothetical structure.

Study Design and Respondents.
is study aims at evaluating the newly developed SSI and HC scales and determining whether the previous hypothetical structure fits the items via CFA. Data were collected from in-service firefighters in Malaysia between 6/24/2021 and 7/24/2021 via self-administered online questionnaires. e filtered questionnaires were customized depending on respondents' answers pertaining to the type and duration of their service. Respondents stating that they were volunteer/commonservice firefighters or had been in service for 2 years or less were excluded. In this study, data from 6970 out of 8581 respondents were included for further analysis. e total number of respondents exceeded the minimum sample size requirement (n � 977). is was calculated using G * power 3.1.9.2 software [7]. e a priori sample size was calculated with the F-test family with linear multiple regression (fixed model, R 2 deviation from zero) with the settings as follows: f 2 � 0.02 (small effect size), α � 0.05, number of predictors � 16, and power set at 80% (Figure 1).

Instruments.
is study used postexploratory factor analysis of the generated SSI and HC scales. A total of 70 SSI items, excluding two global rating items, were arranged under the dimensions of protection against hazards and welfare. Subdimensions of protection included personal protection suit (PPS, 5 items), workspace (WS, 3 items), equipment used for operation or work (EQUIPMENT, 7 items), documentation related to standard operating procedures or work manual (DOC, 4 items), addressing occupational safety and health issues (OSH, 5 items), workload (WL, 5 items), psychological care (PSYCARE, 6 items), physical fitness (FIT, 6 items), health surveillance (HSURV, 4 items), and supervision (SV, 3 items). Subdimensions of welfare included salary (SALARY, 4 items), special allowance (SP ALLOW, 2 items), compensation for occupational injury or death (COMP, 3 items), career development (CAREER, 4 items), caring (CARE, 5 items), and humanity (HUMANITY, 4 items). All SSI items began with "I am satisfied with . . .." e HC scale consists of five multidimensional items and was used to measure the firefighter's feelings of happiness related to their job in terms of meaning, personality fit, work environment, and skill utilization. A total of three HC items started with "I am happy to work in the Department because. . ."; the other two items omitted the initial wording because they would have made them too lengthy exceeding 15 words per statement. e respondents were expected to rate their agreement level for SSI and HC using five-point Likert scale (1 � Strongly disagree, 2 � Disagree, 3 � Slightly agree, 4 � Agree, and 5 � Strongly agree).
ere was no undecided/natural response, and the respondents were forced to evaluate their own level of agreement rather than "sitting on the fence."

Data Preparation.
e data were screened for blank responses, outliers, and missing values before converting to a .csv file for confirmatory factor analysis (see supplementary material named FRDM SSI HC For PLS SEM.txt). We employed partial least squares structural equation modelling (PLS-SEM) using SmartPLS 3.3.3 [8] as the statistical tool to evaluate the measurement and structural model. PLS-SEM was used because it is able to handle complex models with multidimensional and formative items. It also has the ability to manage nonparametric data, as survey research is not typically distributed normally [9]. We chose hierarchical component modelling, specifically the reflective formative disjoint two-stage approach, because the SSI has two hypothetical dimensions (i.e., protection against hazards and welfare) and 16 subdimensions. e number of items in each subdimension was not similar. In the usual manner, the SSI was regressed on the HC scale, which consisted of five multidimensional items ( Figure 2). Confirmatory factor analysis: Reflective formative disjoint two-stage approach.
We followed the suggestions of Anderson and Gerbing [10] to test the model developed using a two-step approach. First, the measurement model was tested for validity and to I am satisfied with the implementation of a health examination after staff is exposed to a health-hazardous operation. H4 I am satisfied with the department's collaboration with health organizations to monitor the health of staff.

I1
I am satisfied with the professionalism of my teammates at work. I2 I am satisfied with my supervisor who always gives reminders regarding individual safety.

KB) Compensation of Rescuers (Personal and Family) KB1
I am satisfied with the management of the application for compensation for staff who were injured while on duty. I am satisfied with the way the management shows concern to the injured staff.
KE2 (KP3) I am satisfied with the Department's concern taking over the routine task of managing the schooling of the children of officers and staff when outstation (e.g., providing transportation to school). I am satisfied with the current workplace distance from the residential location.

KE6 (KH2)
I am satisfied with the permission given by the Department to officers and staff to do work outside of duty hours (such as lawn mowing work, driving e-hailing, associations, teaching etc.). KE7 (KH3) I am satisfied with the Department allowing me to take sick leave that has been certified by a recognized medical practitioner.

KE9
(KH4) I am satisfied with the permission to perform light duty duties to officers and staff who have health problems.

WG
Overall, I am satisfied with the welfare element provided by the Department to the members. Happiness at work S1 I am happy to work in the Department because this job gives meaning and purpose. S2 I am happy to work in the Department because the given tasks suit my personality.

S3
I am happy to work in the Department because I feel proud to be a part of its staff. S4 e Department has created a work environment where I can deliver the best possible service. S5 I can give additional efforts and contributions to achieve the Department's mission. SG Overall, I am happy to work in the Department.
Note. PG, WG, SSG, and SG are global subdimensions. Italics indicate a new item label postexploratory factor analysis. KB4, KC5, KD2, KD3, and KE8 were excluded after the exploratory factor analysis using Promax rotation with Kaiser normalization.
determine the reliability of the instruments. Next, a structural model was used to test the validity, collinearity, significance, and relevance of the formative subdimensions using the guidelines given by Hair et al. [11] and Ramayah et al. [12].

Measurement Model Analysis
Step 1. Drawing specified hypothetical subdimensions. In the SmartPLS workspace, a model, as shown in Figure 2, was drawn. All 16 hypothetical subdimensions were regressed to the HC endogenous construct. We then used Calculate > PLS Algorithm to obtain factor loading and average variance extracted (AVE) to assess convergent validity, composite reliability (CR) to assess each subdimension's internal reliability, heterotrait-monotrait ratio of correlation (HTMT) to assess discriminant validity, and latent variable for assessing the structural model of the reflective formative disjoint two-stage analysis at a later stage. We copied and pasted each subdimension's latent variables in the original file and saved it as a new .csv file for normality testing. In the file, we added the column

Advances in Public Health
RANDOM with the formula RAND() followed by "Enter" and "Tab," double-clicked the plus sign, and saved the file for assessing common method bias using full collinearity testing (see supplementary material named FRDM SSI HC LV.txt).
Step 3. Checking for common method bias Since the data were collected from a single source, we first tested for common method bias by following the suggestions of Kock and Lynn [13], namely, to test full collinearity. In this method, all 17 constructs were regressed on a RANDOM variable using IBM SPSS version 26. ere is no bias from the single source data if the variance inflation factor (VIF) ≤ 5 [14]. e analysis yielded a VIF consistently less than 5 (Table 2); thus, single source bias is not a serious issue with our data.
Step 4. Measurement model assessment On the one hand, all 16 subdimensions of the SSI are reflective; on the other hand, the HC scale is a formative model (Figure 2). For the reflective measurement model, we assessed the loading, AVE, and CR. e value of the loading should be ≥ 0.5, the AVE should be ≥ 0.5, and the CR should be ≥ 0.7. As shown in Table 3, the AVEs are all higher than 0.5 and the CRs are all higher than 0.7. All loadings are higher than 0.708, indicating that a latent variable is able to explain at least 50% of the subdimension's variance [11]. No item was deleted. For the formative measurement model, we assessed the bootstrapped outer weight significance and outer VIF. All bootstrapped outer weights are significant at a    (Table 4). We then assessed discriminant validity using the HTMT criterion suggested by Henseler et al. [15] and updated by Franke and Sarstedt [16]. e HTMT values should be ≤ 0.90. As shown in Table 5, the HTMT values of each construct are ≤0.90. Hence, we can conclude that the respondents understood that the 16 reflective constructs were distinct. e measurement model is both valid and reliable.
Step 5. Structural model assessment For the structural model, we assessed the convergent validity by redundancy analysis. In assessing convergent   validity, a global indicator was designed for each formative construct. We used three global rating items, namely, for the construct of protection against hazard at work (PG), welfare (WG), and HC (SG), respectively (Figure 3). e redundancy results range between almost 0.7 and more than 0.8 ( Figure 3). erefore, the formative subdimensions of the construct do contribute to its intended content at a sufficient level of convergent validity.
A total of 16 latent variables from Step 1 was used to draw the reflective formative disjoint two-stage approach (Figure 4). e significance and relevance of formative subdimensions were assessed by looking at the bootstrapping values of the outer weights. Based on the results shown in Table 6, all subdimensions in the formative construct satisfy the VIF values of less than 5 [17]. It can be concluded that collinearity is not an issue for the estimation of the PLS path model. e significance and relevance of formative subdimensions were examined. e results show that all formative subdimensions are significant except for health surveillance and psychological care. Prior literature provides evidence for the relevance of these subdimensions for capturing the attributes of protection against hazards at work [18,19]. Hence, these subdimensions are retained in the formative protection construct even though their outer weights are not significant.
Step 6. SSI and HC path model assessment Prior to testing the SSI and HC path model, the issue of lateral collinearity was examined to reveal the robust cause and effect between SSI and HC. In this model, the enhancement of the hypothetical SSI leads to happiness among firefighters. A stricter VIF value of 3.3 or up to 5 was employed to indicate the absence of potential collinearity issue [20]. Since the data were not normal, we followed the bootstrap procedure, using 7000 subsamples to avoid inflation or deflation of the standard errors due to non-normality issues. We reported a bias-corrected confidence interval for path coefficients (β). We considered a β of 0.21 and above as a significant parameter [21]. Later, we assessed the coefficients of determination (R 2 ), effect size (f 2 ), and predictive relevance (Q 2 ). R 2 is a measure of the model's predictive accuracy (i.e., the combined effect of exogenous variables on endogenous variables). R 2 represents the amount of variance in the HC scale explained by all formative subdimensions in protection against hazards and welfare and the exogenous constructs linked to it. Past and contemporary scholars provide a different acceptable R 2 based on the area of research. Cohen [22]; a Professor of Psychology, suggested 0.26, 0.13, and 0.02 to describe substantial, moderate, and weak levels of predictive accuracy, respectively. Almost three decades later, Hair et al. [17] suggested 0.75, 0.50, and 0.25 to describe substantial, moderate, and weak levels of predictive accuracy, respectively. Interestingly, Falk and Miller [23] suggested the lowest acceptable R 2 value, 0.10 or higher, to explain the variance of a particular endogenous construct.
To ensure complete reporting, the effect size of the protection and welfare constructs were further evaluated using Cohen f 2 . According to Cohen [24], f 2 values of 0.35, 0.15, and 0.02 are considered large, medium, and small effect sizes, respectively. Lastly, the Q 2 via the blindfolding procedure was calculated, which removes data from the dataset based on the predetermined distance value, called D. e D value can be any number from 5 to 12. In this procedure, a D value of 7 was chosen so that the 6970 observations in the dataset were divided by 7, giving 995.7, which is not an integer. Chin [25] concluded that a Q 2 larger than 0 indicates that exogenous constructs have predictive relevance for an endogenous construct.
Based on the assessment of the path coefficient, shown in Table 7, both relationships were found to have a t-value ≥2.33, which is significant at a 0.01 level of significance. e predictors of protection against hazards at work (β � 0.370, p < 0.001) and welfare (β � 0.375, p < 0.001) are positively related to happiness, explaining 46.6% of the variance in happiness at work. e R 2 value of 0.466 is above the 0.26 value suggested by Cohen [22], indicating a substantial model.
e predictors have small-to-medium effect sizes (the Cohen f 2 is near 0.15) and relevance in predicting happiness (Q 2 � 0.339).

Discussion
is study aims at demonstrating the validation process of SSI and HC as a final phase of a systematic scale development and scale validation. We chose to perform an upto-date empirical PLS-SEM analysis using hierarchical component modelling, namely, the reflective formative disjoint two-stage approach.
is is because the SSI is complex, with a total of 70 items that are not uniformly distributed in their respective subdimensions. A total of 16 solidified SSI subdimensions unevenly distributed under two dimensions were examined. ese dimensions included protection against hazards at work and welfare. Out of the 16 SSI subdimensions, only two did not show significant relevance to assessing firefighter's satisfaction (health surveillance and psychological care). However, one cannot assume that these subdimensions have no impact on one's assessment of personal job satisfaction pertaining to the content and context of the said job, especially among firefighters.
Health surveillance, especially in regard to the respiratory system, is vital to firefighters, who are constantly exposed to smoke and various by-products of combustion, including polycyclic aromatic hydrocarbons, volatile organic compounds, di-(2-ethylhexyl) phthalate, and polybrominated diphenyl ethers, which are far more toxic than the contaminates released from fires decades ago [26].  Advances in Public Health 9   Figure 4: Reflective formative disjoint two-stage structural model. While the use of self-contained breathing apparatus during fire suppression operations has increased, the breathing apparatus is not consistently used prior to or during overhaul activities because of habitual practices or mismatching demands and supplies, which leads to adverse health effects among firefighters. erefore, there is caused to be concerned about the respiratory health of firefighters. Although Malaysia has never experienced any deadliest fire before [27], Malaysia had an average rate of fire incidents of around 1025 fires per million inhabitants per year from 2006 to 2014 [28]. While a residential fire may take 30 minutes to put out, a bushfire may take up 14,400 minutes or more. is is considered a significant exposure to the firefighter's respiratory system. In some states, the Fire and Rescue Department routinely performs medical examinations to assess general physical health and includes a routine panel of blood tests for chronic diseases, vision and hearing tests, electrocardiograms, and a lung-function test [29]. Unfortunately, these examinations are not adjusted based on what firefighters' respiratory systems have been exposed to in an operation period. Matching biomarkers of exposure (e.g., 1-hydroxypyrene or protein adducts) or biomarkers of effect (e.g., eosinophils levels) to a specific exposure should be regularly tested. Furthermore, no pattern analysis was performed to determine whether there is a trend in lung function over time. Firefighters may be deemed "not fit for duty" after years of service, which may lead to a reduction of functional crews. Mathias et al. [26] found that firefighters had decreased lung function at rates two to four times greater than expected over 5 years. In addition to a history of smoking, exposure to smoke and various combustion by-products over years of service may have increased the risk for interstitial lung disease and autoimmunity in firefighters [30].
is may cause premature death. erefore, the health surveillance subdimension of the SSI should be maintained. e environment of firefighters is physically and psychologically demanding. Some often think of the mind and body as being separate, but mental and physical health are closely interrelated. In a firefighter's world, responding to a dirty, dangerous, demeaning, and chaotic environment is rarely discussed. erefore, it is not thought to be an issue; in fact, it is a taboo topic. is is because firefighters are real-life heroes. ey run toward danger. ey are one of the first officials to reach the scene in times of crisis. ey know that there is a chance that specific tactics might not go as planned. Nevertheless, they are motivated to do more, to go beyond their pay rate. It is likely that in these moments, firefighters might have a glimpse of an undesirable memory that needs to be dealt with professionally. Being heroes, they are unlikely to vent their stress and trauma with those they are close to, even their wives. If they are not being asked, no one knows what they are going through internally [31]. A possible mechanism to execute such a delicate enquiry is the Employee Assistance Program (EAP), which can prevent late detection of psychological issues when someone has attempted suicide.
e EAP aims at performing crisis interventions and at offering education on stress management and other shortterm assistance with various life challenges [32].
is is necessary because every working adult may experience workfamily and work-life conflicts that lead to various health consequences, such as burnout [33], and firefighters are no exception [34]. It is important to help affected firefighters become resilient in every possible opportunity. Although the EAP has been repeatedly proven to be an essential element in nurturing resilience, especially within critical uniformed organizations, the EAP is an "extraterrestrial program" in Malaysia. Since the EAP is not properly incorporated into the Fire and Rescue Department of Malaysia, it is an open question how the department manages psychological issues at the individual level. Firefighters may not face as high emotional burden as healthcare workers, but as human beings, they need some psychological care. erefore, the psychological care subdimension of the SSI should be maintained.
is article presents 16 culturally tailored subdimensions of the SSI and the five formative items, within the HC scale that should be used to measure satisfaction and happiness among firefighters. Each formative subdimension of the SSI has its own weight, simulating real phenomena in determining the final SSI score of in-service firefighters. is work has practical implications for other uniformed organizations. ey can use the subdimensions as a solid framework to measure their own employees. We believe that this article gives a comprehensive framework to researchers for analyzing a complex model with multiple first-and secondorder independent variables and regressing on a formative dependent variable. In the light of hope, this article will be able to guide researchers to develop and validate new or adopted constructs or to test theories successfully. e SSI and HC scales are a vital contribution to assessing firefighters' satisfaction and happiness levels at work. Both tools are statistically reliable and valid. Now they are ready to fly high and be heroes.

Data Availability
e data used to support the findings of this study are included within the supplementary information file(s).

Conflicts of Interest
e authors declare that there are no conflicts of interest.

12
Advances in Public Health