Structural Equation Modelling

The same ‘goodness-of-fit’ guidelines that were used in the previous section to discuss the goodness of fit for the CFA are used in this section to discuss the SEM results for the various models that were tested. Table 4.56 depicts the SEM fit statistics for the two models depicted in Figures 4.4 and 4.5. These will be discussed under each model in this section. In addition, SEM estimates for the relevant model parameters are also reported, Standardised Regression Weight (SRW), Squared Multiple Correlation (SMC) and p-values are reported and discussed.

TABLE 4-56:SEMGOODNESS-OF-FIT STATISTICS (RED INDICATES ACCEPTABLE FIT)

Category Code Target SEM1 SEM2

Sample size n 381 381

No. of items m 60 60

Sample size; No. of items Category n;m.Cat. 250 < n < 1000; m ≥ 30

Absolute/ predictive fit Abbr. Target Observed Observed

Chi-square (Maximum likelihood) χ² 2188.04 3081.36

df 1369 1467

p ≥ .050 < .0005 < .0005

χ²/df ≤ 3 1.60 2.10

Comparative Fit Indices

Bentler-Bonnet normed fit index NFI ≥ .90 .88 .82

Bentler comparative fit index CFI ≥ .90 .95 .90

Other

Joreskog adjusted GFI AGFI ≥ .95 .80 .75

95%Lo .037 .051

Root mean square error of approximation RMSEA ≤ .08 .040 .054

95%Hi .043 .056

4.8.2.1 MODEL 1

The conceptual model from Chapter 2 is illustrated in Figure 4.4. Here, all of the variables were treated as independent variables that lead to the DV: Happiness of South African engineers.

FIGURE 4-4:MODEL 1(PREVIOUSLY ILLUSTRATED AS FIGURE 2.4-CONCEPTUAL MODEL)

The “goodness-of-fit” results for Model 1, illustrated in Table 4.56, indicate that the χ²/df (1.60), CFI (0.95) and RMSEA (0,040) were within the required ranges. However, the p- value (< 0.005), NFI (0.88) and AGFI (0.80) did not fall into the acceptable ranges. Furthermore, the SEM estimates for Model 1, illustrated in Table 4.57, indicate that multic- ollinearity exists between the IV’s and that only IV3: Work-Life Balance has a significant relationship with the DV: Happiness of South African engineers. Therefore, Model 1 is not a feasible model.

TABLE 4-57:SEMESTIMATES FOR MODEL 1 Relationship SRW p-value SMC IV01--->DV -0.045 .581 0.810 IV02--->DV 0.278 .047 0.810 IV03--->DV 1.150 <.0005 0.810 IV04--->DV -0.495 .259 0.810 IV05--->DV 0.432 .061 0.810 IV06--->DV -0.274 .130 0.810 IV07--->DV -0.033 .692 0.810 IV08--->DV -0.079 .419 0.810 IV09--->DV -0.018 .663 0.810 IV10--->DV -0.022 .760 0.810 4.8.2.2 MODEL 2

Model 2, illustrated in Figure 4.5, was statistically constructed as Model 1 was not a feasible model. Here, IV6: Work Satisfaction was regarded as an intermediate variable for IV1: Workplace Influence, IV2: Workplace relationships, IV3: Work-life balance and IV4: Purpose. IV6: Work Satisfaction, IV5: Optimism, IV7: Leisure, IV8: S.A. Pride, IV9: Classism and IV10: Trust were directly linked to the DV: Happiness of South African engineers.

FIGURE 4-5:MODEL 2

The “goodness-of-fit” results for Model 2 were also illustrated in Table 4.56 and shows the χ²/df (2.10), CFI (0.90) and RMSEA (0.0542) were within the target ranges. However, the p- value (< 0.005), NFI (0.82) and AGFI (0.75) did not yield acceptable results. In addition, Table 4.58 illustrates that the IV’s IV1: Workplace Influence, IV2: Workplace Relationships, IV3: Work-Life Balance and IV4: Purpose have significant relationships with the intermediate variable IV6: Work Satisfaction and that IV5 to IV10 have significant relationships with the dependent variable DV: Happiness.

TABLE 4-58:SEMESTIMATES FOR MODEL 2

Relationship SRW p-value SMC IV01--->IV06 0.083 .277 0.865 IV02--->IV06 0.682 <.0005 0.865 IV03--->IV06 0.062 .824 0.865 IV04--->IV06 0.176 .564 0.865 IV05--->DV 0.770 .243 0.593 IV06--->DV -0.102 .609 0.593 IV07--->DV -0.739 .314 0.593 IV08--->DV 6.868 .168 0.593 IV09--->DV -0.728 .279 0.593 IV10--->DV -6.534 .194 0.593 4.9 Summary

The primary aim of Chapter 4 was to address the RQM: How happy are engineers working

in the engineering industry in South Africa? which corresponds to ROM – To measure the

level of happiness of engineers working in the South African engineering industry. To

achieve this, the results of the primary research study were analysed and discussed. Three hundred and eighty one respondents participated in this study. Exploratory Factor Analysis, descriptive statistics and various inferential statistics were conducted and the result of the factors was deemed acceptable with all of the Cronbach’s Alpha scores measuring above 0.69.

Statistical relationships between the independent variables and dependent variable were explored through Pearson’s correlation analysis. Furthermore, relationships between selected demographic information and both the dependent variable and selected

independent variables were explored through Chi-square tests, descriptive data analysis methods, ANOVA and t-Tests. The measurement instrument for the Dependent Variable was reviewed and deemed fit by means of Confirmatory Factor Analysis, although room for improvement was identified for the measurement instrument of the Independent Variables.

Finally, Structural Equation Modelling was conducted to determine whether the conceptual model from Chapter Two was fit, however, it was found that although all of the variables were initially regarded as independent variables, SEM indicated that some variables were better suited as intermediate variables and, as such, the model was adapted accordingly. Chapter Four, therefore, concludes with a new recommended model (Model 2) for measuring happiness among South African engineers.

FIGURE 4-6:MODEL 2(PREVIOUSLY ILLUSTRATED AS FIGURE 4.5)

The first five research questions and research objectives have been addressed in the first four chapters. In Chapter 5, a conclusion to the study will be made and RQ5: What are the

factors of the proposed model measuring the happiness of South African engineers? which

correlates to RO5: Identify the key factors to improve happiness in the South African

engineering profession will be addressed. Chapter 5 further addresses the RQM: How happy are engineers working in the engineering industry in South Africa? which will

address ROM: To measure the level of happiness of engineers working in the South African engineering industry.

CHAPTER 5: Findings, Conclusion and Recommendations

5.1 Introduction

In Chapter 4, the results of the empirical study were presented, analysed and discussed. The chapter concluded with a conceptual model for measuring happiness amongst South African engineers. The chapter further addressed RQ5: What are the factors of the

proposed model measuring the happiness of South African engineers? This corresponds

with RO5: Identify the key factors to improve happiness in the South African engineering

profession.

This chapter serves as the final chapter of the study and presents the findings, managerial recommendations and conclusion to this study. The RQM: How happy are engineers working in the engineering industry in South Africa? This corresponds with ROM: To

measure the level of happiness of engineers working in the South African engineering industry is addressed.

The Chapter outline is illustrated in Figure 5.1.

FIGURE 5-1:CHAPTER FIVE OUTLINE

Chapter 1: Introduction and Problem Statement

Chapter 2: Literature Review

Chapter 3: Research Design and Methodology

Chapter 4: Results and Analysis

•5.1 Introduction •5.2 Summary of Treatise •5.3 Key Findings of the study

•5.4 The tested model for measuring happiness among South African engineers •5.5 Managerial Recommendations

•5.6 Limitations and future research •5.7 Conclusion

Chapter 5: Findings, Conclusion and Recommendations