Investigation of Factors Affecting Employee Satisfaction of IT Sector

Job satisfaction or employee satisfaction has various definitions, but we can generalize it by how gratified an individual is with his or her job. Happy employees help to strengthen the company by lowering turnover and increasing loyalty. Job satisfaction also promotes a healthy working environment that helps to attract talent and increase productivity. However, little research has been done that focuses specifically on the IT sector. The goal of this research is to measure the level of satisfaction among Kuwaiti IT workers and discover tangible and intangible factors affecting their job satisfaction. To highlight factors contributing to positive satisfaction in the IT jobs in Kuwait, we propose a six-factor structural model, including compensation, workplace, intangible benefits, support, communication, and satisfaction. A targeted snowball descriptive survey was distributed via WhatsApp messages to Information Technology workers; 209 responses were collected after data cleaning. SPSS statistical software was used to analyze the data, with results indicating IT employees felt an average level of satisfaction. Additionally, several workrelated variables were significantly associated with job satisfaction. Work position showed a statistically significant association with work satisfaction. Finally, individuals in a leading position reported higher satisfaction compared to individuals in non-leading positions. Keywords—Job satisfaction; IT sector; productivity; intangible benefits; communication

It is hard to measure, yet we all differ in nature, and just as tangible benefits could be crucial to some employees, intangible factors could be more important to others, especially in an economy where there are a lot of cutbacks or layoffs.
Prior to moving forward with our study, we conducted a review of the related research conducted over the past twenty years. The next section will summarize these studies. Based on the findings we designed our survey with consideration of cultural differences.
The collection of the literature was directed towards the IT sector, as we lack such studies in Kuwait. As information technology departments became the backbone of every company, it became hard to find any organization that does not have an IT department. If we are allowed to generalize, as using technology became a required skill for every worker, we might consider all workers as IT workers.
This study would like to contribute and enrich studies in that subject, in a middle eastern country such as Kuwait. The concluding points will help decision-makers in improving workplace environments.
The following sections are arranged as follows: Section 2 will visit various previous studies focusing on job satisfaction for IT workers. Our approach will be discussed in Section 3. In Section 4, our results will be given. And finally, in Section 5, we will sum up our findings.

A. Defining Job Satisfaction
Hoppock defined job satisfaction as "any combination of psychological, physiological and environmental circumstances that cause a person to truthfully say I am satisfied with my job" [11]. Yet, the most widely used definition of job satisfaction was made by Locke, who defines it as "a pleasurable or positive emotional state resulting from the appraisal of one's job or job experiences" [14]. According to Vroom, job satisfaction is positive feedback from individual workers towards their current job [15]. Wanous and Lawler state job satisfaction as the "sum of job facet satisfaction across all facets of a job" [16]. This is very similar to Spector, who defines it as "how people feel towards their job from different aspects" [17] and Schermerhorn," as the emotional response towards various aspects of an employee's work" [18]. More definitions supported the same meaning.
Reilly describes job satisfaction as the feeling that a worker has towards his job, influenced by the perception of one's job [19]. Mansoor, Muhammad, Fida, Nasir, and Ahmad suggested a similar definition: how positively people feel about their job [20]. Ellickson and Logsdon defined job satisfaction as the degree to which employees like their work [21].
Phillips and Connell defined it as "the degrees to which employees are content with the job they perform" [22]. More attempts to define the concept of satisfaction have resulted in the definition being the final state of the psychological process [23]. Many studies have suggested many definitions, with the majority focusing on how the employee feels about his job in general.

B. Job Satisfaction Factors
There are, according to Arnold and Feldman, a number of factors that make people feel positive or negative about their jobs [24]. Researchers have contributed heavily to prioritize these factors based on their influence on job satisfaction.
Nwagwu conducted a Nigerian study to observe job satisfaction among IT artisans. The study's main discovery showed 300 IT artisans surveyed were dissatisfied with their jobs; however, high expectations of a breakthrough and the trend of IT were key reasons for staying in their jobs [25].
However, other studies have shown that financial factors and promotions are the leading factors for job satisfaction [6,26,27,28]. Studies have shown that low financial income leads to high insecurity [29]. In addition to financial factors, Akbar et al. explored additional factors such as prospects for the working environment, training, career growth and improvement [28].
Frontczak and Else focused on the indoor work environment's quality and building design, defining a good workplace space as when workers are granted a private office space with windows close by [8]. Lottrup, Stigsdotter, Meilby, and Claudi supported this claim in their research, empathizing on the importance of having buildings with green surroundings and window views [10].Additional factors such as flexible working hours, work relations, family insurance, allowance, promotion, and benefits were discussed by Alam and Shahi [9].
In addition, they highlighted the significance of positive reviews from an employee's superior. Other researchers found that work relationships and higher morale significantly influence the level of satisfaction [30,31]. Furthermore, high ethical expectations in the workplace lead to greater satisfaction [12]. Additional studies have concentrated on gender and how it can play an important role in work satisfaction [32,33]. However, other studies have denied this claim [20,34]. Kowal and Roztocki have argued that women are less satisfied with their jobs [35].
A study by Clark discovered that although females occupy a lower position in their average job and get lower income than their male counterparts, the expectations of females have been contended to be lower in comparison to males. Therefore, females tend to report greater job satisfaction levels [36].
A study by Bordin, Carina, Bartram, and Casimir conducted in Singapore amongst IT workers shows that psychological empowerment can increase job satisfaction and organizational commitment. Additionally, the study revealed that similarly supervisory support is an important factor for the same reasons [37].
When examining other factors, other studies revealed that employees with flexible working hours had been seen to have higher job satisfaction than those without [9]. They tend to have more time in their private lives and harmonize with their profession [9]. They also found that forcing ethical work standards increased job satisfaction [14].
Lim discovered that wage, degree, a sense of belonging, faith in wanting to belong, a feeling of acceptance, job autonomy, and promotion opportunities were related to job satisfaction while evaluating it for library Information Technology staff [38].
Lumley, Coetzee, Tladinyane, and Ferreira carried out a cross-sectional analysis on a group of IT workers in companies in South Africa to investigate the connection between job satisfaction and employee organizational commitment. It suggested a significant relationship between job satisfaction and affective and normative commitment [39].
A study was conducted in India on IT workers has concluded that there is a strong link between job satisfaction and employee loyalty. And the main determinants of job satisfaction and employee loyalty are supervisory support, career growth, and job security [40].
Another research conducted in Singapore showed that personal accomplishment intercedes the relationship between emotional intelligence and job satisfaction for IT workers [41].
Wong, in a study from Hong Kong, argued that the effects of organizational culture on knowledge sharing leads to job satisfaction, which leads to an improvement in organization performance [42]. www.ijacsa.thesai.org Another study conducted by Adebiaye found that work attitude, cordial working relationships, and management support affect job satisfaction [44]. Sunil Misra and Kailash B. L. Srivastava, found that team building between bank employees generates competencies that positively affect employee effectiveness and job satisfaction [45].
Spann designed a study to investigate the relationship between the conflict and ambiguity role and job satisfaction for non-managerial IT. They concluded that there is a direct relationship of job satisfaction with both role conflict and role ambiguity [46].
Indian research conducted on IT professionals examined the relationship between work exhaustion and job satisfaction and discovered a negative correlation, additionally a positive correlation between work exhaustion and turnover intention. The study also considers the impact of emotional dissonance, role ambiguity, role conflict, the fairness of rewards, autonomy, and the perceived workload on IT professionals [47].

III. METHODS
A survey of a descriptive nature was used [48] to achieve our study goals, answering the following questions: RQ1: What is the average job satisfaction score for Kuwaiti IT workers? RQ2: What are the tangible and intangible factors influencing Kuwaiti IT workers job satisfaction? RQ3: Which job characteristics are significantly associated with job satisfaction?
A targeted snowball survey was distributed via WhatsApp to Information technology workers, of which 209 responses were collected after data cleaning. The survey contained fivepart sections completed by all respondents. The first-part is the demographic questions that consists of four questions, followed by the job characteristics the job characteristics which comprises of seven questions. The next sections were organized as follows: tangible benefits, intangible benefits, work relations questions, and general satisfaction related questions.
Continuous variables were summarized using means and standard deviations and categorical variables such as demographic and work characteristics were summarized using counts and percentages.
Histograms were used to assess the presence of univariate outliers. Scaled variables were also examined for points above or below three standard deviations from the mean. Data was explored for missing observations prior to the analysis. Histograms were also inspected for normality. Mahalanobis distance was used to check for multivariate normality.
Exploratory factor analysis was performed using maximum likelihood. Oblimin rotation (with Kaiser Normalization) was used. Variables were removed if they loaded on more than 1 latent variable (>0.4 on more than 1 latent variable) or did not load significantly on any of them (< 0.5).
Confirmatory factor analysis was performed to assess whether the data fit the hypothesized measurement model previously defined. Six, five, and four factor solutions were tested to assess the most appropriate factor structure to use. Reliability of the constructs was assessed using Cronbach's alpha and composite reliability. A value greater than 0.7 was considered satisfactory. The convergent validity of the constructs was assessed using the average variance extracted which should be greater than 0.5 for all constructs. Divergent validity was assessed by comparing the correlations between latent variables to square root the average variance extracted √ . Divergent validity was met if none of the correlations between latent variables was higher than square root the . Individual indicators were allowed to load on only one factor and the latent variables were allowed to freely co-vary. The overall model fit was assessed using the following indices:  The root mean square error of approximation (RMSEA) and the corresponding 90% Confidence interval.
 The comparative fit index (CFI).
 The standardized root mean square residual (SRMR).
The lower bound of good fit for the TLI and the CFI is considered to be 0.90. For the RMSEA and the SRMR, the upper bounds for good fit are considered to be 0.08 and 0.10, respectively. C min /df less than 5 was considered an indication of good model fit (Table I). These cut off criteria for model fit were used as previously defined [49].
Hypotheses were tested using structural equation modelling (SEM).
Scale reliability analysis was performed using Cronbach's alpha. Cronbach's alpha is a measure of internal consistency which assesses how closely related a set of items are as a group. Cronbach's alpha is a function of the number of test items and the average inter-correlation among the items. The acceptable value for Cronbach's α is > 0.7. SEM was performed to assess the association of the independent latent variables with the main DV (satisfaction with work). Model fit was assessed using the same previously mentioned fit measures. The R 2 was also calculated for the DV. R 2 represents the proportion of variance in the DV that is explained by IVs. Hypothesis testing was performed at 0.05 significance level.
Standardized coefficients were used to compare the effects of the independent variables included in the SEM. The standardized coefficients divide the size of the effect by the relevant standard deviations. So instead of being in terms of the original units of X and Y, the standardized regression coefficients are in terms of standard deviations which facilitates comparing regression coefficients. The R 2 is the squared multiple correlation and was used to assess the proportion of variance in the dependent variables that is explained by the independent variables. Statistical analysis was performed using SPSS v25 and R studio v1.1.463.

A. Satisfaction Across Kuwaiti IT Workers
Means and standard deviations were used to summarize the distribution of job satisfaction across various demographic and work characteristics. Scores for latent variables were computed by averaging the scores for the items included in the final CFA and SEM. One-way ANOVA was used to assess the association of various demographic and work factors with job satisfaction. One-way ANOVA was used since the DV (job satisfaction) is continuous in nature. Moreover, it can accommodate IVs with two or more levels unlike independent t-test which can only accommodate IVs with only two levels.

IV. RESULTS
The initial data included 218 responses (n = 218). Nine responses were identified as outliers using Mahalanobis distance and were removed from the analysis (n = 209). Table II shows the characteristics of the study sample. Table III shows the final factor structure. Six factors were identified: compensation (2 variables), workplace (6 variables), intangible benefits (2 variables), communication (2 variables), support (3 variables), and satisfaction (3 variables).
After excluding variables that did not meet the criteria, 18 items were used in the final analysis. These items formed a sixfactor structure and none of the items loaded on more than a factor (latent variable).

1) Model choice:
Results for CFA show that the six-factor solution provided an appropriate fit for the data. Workplace and compensation were used as two separate latent variables although both of them represent one aspect of the tangible benefits. This was done since model fit showed that combining them as one latent variable (five-factor model 1) resulted in poor model fit compared to the six-factor structure. Poor fit was also observed when communication and support were forced to load as one latent variable (five-factor model 2).
Results show that the six-factor model fits the data better compared to all remaining models as indicated by the AIC, and RMSEA. The TLI and CFI were also higher for the six-factor model. Likelihood ratio test showed that the six-factor model was significantly better compared to the remaining three models (Table IV). Thus, the six-factor solution was deemed appropriate since all fit measures were within the acceptable range. In addition, the C min /df and the SRMR were 0.511 and 0.05 for the six-factor model, respectively.

B. Convergent and Divergent Validity
Results show that reliability was acceptable for all constructs (~0.7 or higher for all constructs). Convergent validity was confirmed by the fact that AVE was greater than 0.5 for all constructs (Table V). Divergent validity was assessed by comparing √ of the construct to the correlation with the remaining latent variables (√ should be higher than any corresponding correlation). This assumption was met for all constructs except for workplace that showed a strong correlation with communication (0.89). Factor loadings were greater than 0.7 for all variables (Fig. 1).

C. Structural Equation Modelling
A structural model was assessed in which satisfaction was used as the DV while all remaining five constructs were used as IVs. The proposed structural model (Fig. 2) was a good fit for the data as shown by CFI (0.964), TLI (0.954), RMSEA (0.065), upper 90% RSMEA confidence interval (0.078), and SRMR (0.054). All the proposed relations were statistically significant (Table IV).
Results show that the five IVs explain 72.1% of the variance in the DV (satisfaction of IT workers) as shown by the R 2 . All five variables showed a statistically significant association with satisfaction with work. Compensation showed a statistically significant positive association with satisfaction (Std. β = 0.263, P < 0.05). This means that satisfaction increases by 1 standard deviation (SD) for each 1 SD increase in compensation which indicates that IT workers are more likely to be satisfied with work if they report satisfaction with payment. Effect of workplace showed a statistically significant negative association with job satisfaction (Std. β = -0.433, P = 0.002). This indicates that workers who are more affected by the workplace are less likely to be satisfied with the job.
Intangible benefits showed a statistically significant positive association with job satisfaction (Std. β = 0.413, P < 0.001). A similar result was observed with communication (Std. β = 0.323, P = 0.019) and support (Std. β = 0.278, P = 0.003). These results indicate that better communication with co-workers, support, as well as intangible benefits are associated with higher satisfaction with work. Comparing the standardized coefficients show that intangible benefits were the www.ijacsa.thesai.org strongest positive influencing factor. For each 1 SD increase in intangible benefits, satisfaction with work increases by 0.413 SD.

D. Job Satisfaction among Kuwaiti IT Employees
The average satisfaction with work was 3.04 (1.04) among Kuwaiti-IT workers which indicates a neutral state of satisfaction among the IT employees (Table VI). Table VII shows that several work-related variables were significantly associated with job satisfaction. Position showed a statistically significant association with satisfaction with work (F = 3.514, P < 0.1). Individuals in a leading position reported higher satisfaction compared to individuals in non-leading positions (3.26 vs. 2.96).
Number of previous jobs showed a statistically significant association with satisfaction (F = 5.47, P < 0.05). The mean satisfaction score was also lower among participants with two or more previous jobs compared to individuals who had 1 previous job or less (2.78 vs. 3.1). Job location also showed a statistically significant association with satisfaction (F = 4.987, P < 0.05). Individuals who reported having a job near home reported higher satisfaction compared to those who did not (3.3 vs. 3).

V. CONCLUSION
Job satisfaction is one of the main challenges facing the administration of all organizations. The average satisfaction score in the current analysis indicates a moderate level of satisfaction for Kuwaiti IT workers. The proposed six-factor structural model (compensation, workplace, intangible benefits, support, communication and satisfaction) was a good fit for the data as indicated by fit measures, convergent and divergent validity. Analysis results supported the pre-defined hypotheses. Compensation (tangible benefits), communication, support, intangible benefits showed a statistically significant positive association with job satisfaction. Higher levels of these variables result in higher job satisfaction. The perception of workplace (tangible benefits) showed a statistically significant negative association with job satisfaction. Individuals who are more affected by the workplace environment were less likely to report job satisfaction which supports the association between workplace and job satisfaction. The five IVs explained 72.1% of the variance in the DV (job satisfaction).
Our findings suggest that managers need to review current pay policies in order to build a satisfactory working atmosphere and offer fair pay, provide clear job instructions, and facilitate positive co-worker relationships. www.ijacsa.thesai.org Three characteristics related to work have shown a statistically significant association with job satisfaction: job position, number of previous jobs and location of work. Participants in a leading position are more likely to be satisfied with the job than those who are not. Participants with two or more previous jobs were less likely to be satisfied with the job than those with one or less previous job. Finally, participants who work in a job near their home were more likely to be satisfied than others who live far from their work.
As the present study was confined to participants working in IT field, it is not possible to generalize the findings to other professional contexts and regions. Furthermore, the sample is very narrow with limited factors, including more factors and a broader sample, to be considered in future studies.