Reward Systems
Analysis of Results and Hypotheses Testing
In this fourth section of the current study, descriptive statistics will be explained, and Construct reliability is assessed using Cronbach's Alpha. The dynamic between the three main variables in this research and the four developed hypotheses are analyzed and tested using established statistical analysis tools such as SEM, Confirmatory Factor Analysis (CFA) using IBM SPSS AMOS 22.0.
Descriptive Statistics
Descriptive statistics presented in the first section of the questionnaire (Table 1), show that the respondents, was dominated (>50%) by those in the 30–39 years age group, which is representative of the Kingdom's population's median age (GASTAT, 2021). The majority of respondents have 6–15 years of professional experience, which provides a considerable and meaningful perspective of employment within the industry. In addition, >67% of respondents hold undergraduate degrees or higher. In terms of concentration within the industry, oil and gas (primary and secondary industrial sectors) dominate the sample population with >61% representation, followed by general industrial services and manufacturing with 13% and 10%, respectively. Other subsectors, tertiary industrial sectors' representation varied between 1–3% such as the banking & finance, construction, science and tech, health, hospitality industries.
Table 1.
Socio-demographic characteristics.
| Variable (n = 216) | Frequency | % |
|---|---|---|
| Age (years) | ||
| <30 | 31 | 14.4 |
| 30–39 | 110 | 50.9 |
| 40–49 | 27 | 12.5 |
| 50+ | 48 | 22.2 |
| Education | ||
| High school | 38 | 17.6 |
| Diploma | 33 | 15.3 |
| Undergraduate | 97 | 44.9 |
| Masters | 44 | 20.4 |
| Doctorate | 4 | 1.9 |
| Gender | ||
| Male | 203 | 94.0 |
| Female | 13 | 6.0 |
| Industry | ||
| Banking & Finance | 3 | 1.46 |
| Construction | 1 | 0.46 |
| Science and tech | 6 | 2.74 |
| Health | 8 | 3.74 |
| Hospitality | 1 | 0.46 |
| Industrial services | 29 | 13.4 |
| Manufacturing | 23 | 10.6 |
| Mining | 2 | 0.92 |
| Oil & Gas | 133 | 61.5 |
| Project management | 3 | 1.46 |
| Other | 7 | 3.26 |
| Experience | ||
| 1–5 years | 27 | 12.5 |
| 6–10 years | 62 | 28.7 |
| 11–15 years | 102 | 47.2 |
| 16+ years | 25 | 11.6 |
Reliability of Constructs and Correlation Test
Correlation coefficients and Cronbach's Alpha (Table 2) for all study variables and latent constructs were derived and tabulated (including the modified latent constructs). Construct reliability was assessed using Cronbach's Alpha, which was found to be above 0.6 - an accepted cutoff point - for each construct in our study. Pearson's correlation coefficient magnitude revealed that there is a strong correlation between the three constructs "IR", "PM", and "JS". No weak or moderate correlation was found between latent constructs at the 0.000 significant levels.
Table 2.
Reliability of constructs & correlation coef.
| Latent Variables | Mean | SD | Cronbach's Alpha (α) | EJS | PM |
|---|---|---|---|---|---|
| Incentive Rewards (10-items) | 3.81 | 0.63 | 0.81 | 0.87 *** | 0.76 *** |
| Incentive Rewards-modified (7-items) | 3.60 | 0.78 | 0.84 | 0.89 *** | 0.85 *** |
| Performance Measures (4-items) | 4.05 | 0.56 | 0.60 | 0.84 *** | |
| Employee Job Satisfaction (5-items) | 3.49 | 0.82 | 0.85 |
Measurement Model
Confirmatory Factor Analysis (CFA) was computed using AMOS 22.0 to test the measurement models. As part of confirmatory factor analysis, factor loadings were assessed for each item. The indicators of all three constructs, Performance Measure (PM), Job Satisfaction (JS), and Incentives and Rewards (IR) loaded adequately on their respective factors with no loading below the 0.6 cutoff point (Table 3, Table 4, Table 5 and Table 6).
Table 3.
Dimensionality of Performance Measure.
| Items (Total Variance Explained 47.63%) | Factor Loadings |
|---|---|
| PM 1 | 0.74 |
| PM 2 | 0.72 |
| PM 3 | 0.64 |
| PM 4 | 0.60 |
Table 4. Dimensionality of Employee Job Satisfaction.
| Items (Total Variance Explained 47.63%) | Factor Loadings |
|---|---|
| PM 1 | 0.74 |
| PM 2 | 0.72 |
| PM 3 | 0.64 |
| PM 4 | 0.60 |
Table 5.
Dimensionality of Incentives/Rewards.
| Items (Total Variance Explained 39.60%) | Factor Loadings |
|---|---|
| IR 1 | 0.23 |
| IR 2 | 0.24 |
| IR 3 | 0.65 |
| IR 4 | 0.60 |
| IR 5 | 0.76 |
| IR 6 | 0.76 |
| IR 7 | 0.75 |
| IR 8 | 0.52 |
| IR 9 | 0.67 |
| IR 10 | 0.74 |
Table 6.
Dimensionality of Incentives/Rewards (adjusted for modified model).
| Items (Total Variance Explained 52.24%) | Factor Loadings |
|---|---|
| IR 3 | 0.65 |
| IR 4 | 0.63 |
| IR 5 | 0.78 |
| IR 6 | 0.76 |
| IR 7 | 0.74 |
| IR 9 | 0.68 |
| IR 10 | 0.78 |
Pertaining to the Incentives and Rewards construct, three items loaded relatively inadequately, and after reevaluating the survey and looking at the literature to reconcile these low loadings (below 0.5, see Table 5 and Table 6), it was decided to drop them. The new factor loadings with the total variance explained (the latent construct IR captured variance of the seven remaining IR items) were tabulated, demonstrating how much the parameters and fit were enhanced.
The model-fit measures were used to assess the model's overall goodness of fit (Chi-Square, CMIN/df, GFI, CFI, TLI, SRMR, and RMSEA) and all of the values were within the appropriate acceptance levels. The three factors of the modified model, Figure 2, (Performance Measure, Incentives and Rewards, and Job Satisfaction) all yielded an adequate fit with the following parameter values: Chi-Square = 147.93, p = 0.000, CMIN/df = 1.59, GFI = 0.93, CFI = 0.96, TLI = 0.95, SRMR = 0.0447, and RMSEA = 0.052.
Figure 2.
The Model of the Study.
The Scale fit indices show that Performance Measure is an adequate fit (Chi-Square = 177. 31, CMIN/df = 4.15, GFI = 0.91, CFI = 0.92, TLI = 0.90, SRMR = 0.032, and RMSEA = 0.071), the scale fit indices show that Incentives and Rewards is an adequate fit (Chi-Square = 254.32, CMIN/df = 5.54, GFI = 0.92, CFI = 0.94, TLI = 0.91, SRMR = 0.07, and RMSEA = 0.064) and the scale fit indices show that Job Satisfaction is an adequate fit (Chi-Square = 148.46, CMIN/df = 4.52, GFI = 0.9, CFI = 0.91, TLI = 0.92, SRMR = 0.04, and RMSEA = 0.048).
Hypotheses Testing Results
The model was run using bootstrapping procedure in AMOS (performing 5000 resamples). Additionally, statistical significance for the indirect effect was determined at 95 per cent bias and accelerated confidence intervals which is a standard in social sciences. According to our data and the calculation parameters applied, our results demonstrated the following: (see Table 7).
Table 7.
Mediation Analysis.
| Latent Variables | Direct Effect | Indirect Effect | p |
|---|---|---|---|
| (H1) IR → PM | 0.10 | 0.77 | |
| (H3) IR → JS → PM | 0.67 | 0.04 | |
| (H2b) JS → PM | 0.74 | 0.04 | |
| (H2a) IR → JS | 0.91 | 0.00 |
Hypothesis 1 is not supported. There is an insignificant direct effect between our independent variable Incentive Reward and our dependent variable Performance Measures. The standardized coefficient for direct effect (H1) of IR on PM (excluding the effect of the mediator JS) is insignificant with (β = 0.1; p > 0.05 = 0.77), a finding that indicates that hypothesis 1 is not supported, and the IR → PM relationship is insignificant.
Hypothesis 2a is supported. There is a significant direct effect between our independent variable of focus, Incentives and Rewards, and our mediating variable of interest, Job Satisfaction. The standardized coefficient for the direct effect (H2a) of IR on JS is significant with (β = 0.91; p < 0.05 = 0.00). This means that hypothesis 2a is supported, and the IR → JS relationship is indeed significant.
Hypothesis 2b is supported. There is a significant direct effect between our mediating variable, Job Satisfaction, and our dependent variable, Performance Measures. The standardized coefficient for the direct effect (H2b) of JS on PM is significant with (β = 0.74; p < 0.05 = 0.04). This means that hypothesis 2b is supported, and the JS → PM relationship is indeed significant.
Hypothesis 3 is supported. There is a significant indirect effect (mediation effect by Job Satisfaction) between IR and PM (β = 0. 67; p < 0.05 = 0.04). We can say that incentives and rewards have a significant indirect effect on performance measure through the mediator job satisfaction. So, the mediation effect here is said to be a full-mediation effect because the direct effect IR → PM is insignificant, nevertheless, the indirect effect IR → JS → PM is significant.