Structure of the Variables that Affect Job Satisfaction Study in Cuban Organizations

This paper aims to establish a model of structural equations that summarizes the degree of incidence of variables related to job satisfaction in organizations. It starts from the identification of the variab les to analyze according to the literature on the subject and those that were corroborated through the consultation of experts, later proceed to the construction of instruments for their measurement and application of these to 4,235 employees in 150 entities from different productive sectors. With the results obtained in the surveys, a confirmatory factor analysis was developed, which verified the incidence of the variab les evaluated in the work satisfaction and later proceeded to the construction of the model. The model obtained shows the relat ionship between the different variables that affect job satisfaction and confirming its multifactorial character as a construct. The data that facilitated the generation of the model allow to establish a logic of management of the job satisfaction from man aging in an integrated way the latent variables that condition them, ensuring their correct management, such as: the design of work content, working conditions, interpersonal relat ionships with peers in general and with managers in particu lar, policies and methods of stimulation and the group work.


Intro d uctio n
The studies of job satisfaction (JS) are as old as they are different (1976, quoted by Alegre, Mas -Machuca andBerbegal-Mirabent, 2016), which is analy zed by varied sciences like Psychology and Organizational management. For this reason, among others, the variety of research on the subject is abundant. In some investigations they conceive it as an end, placing man in the center of attention (Park, Seo, Park, Bettini, and Smith, 2016). Others, as a means to impact on performance indicators, such as: Productivity, quality service, organizational co mmit ment, staff turnover and absenteeism; where it is assumed as a general hypothesis that satisfied workers have greater co mmit ment and create the conditions to raise quality, customer satisfaction and productivity and decrease staff turnover and absenteeism (Chen, Sparrow, and Cooper, 2016;A lvarez-García, Del Río -Rama, Saraiva, and Ramos-Pires, 2016;A masonah, Gamor, Deri, and Amissah, 2016;Jafari Navimipour and Soltani, 2016;Chiang and Lin, 2016;Hsu and Liao, 2016;A mponsah -Tawiah, Annor, and Arthur, 2016;Hsu and Liao, 2016).Several investigations, on the other hand, s eek to identify or understand which variables influence JS and how they interact (Alegre, et al., 2016;Amponsah -Tawiah, et al., 2016), while others seek to measure or evaluate the level of JS (Chang and Cheng, 2014;Ro w, Paik, Griffeth, and Allen, 2014;Macintosh and Krush, 2014;Spagnoli, Caetano, and Santos, 2012). Th is research aims to know, under an integral conception, what variables are involved in JS, and the extent to which they interrelate with each other. conception and evaluation of the work environ ment, job experiences and a perception of all elements of work and the workp lace...", Pires (p. 3). Severa l concepts of this theme are presented by Amissah et al. (2016) in wh ich it is clear that in spite of the variety and common or conflicting points of view, they all result in ideas similar to those set forth by Locke (1976). These authors conclude and accept the idea that JS is not influenced by the same factors in all organizations, and that the variables to study and measure it must change from one organization to another.
In general, research on the subject presents a partial view of JS frequently focus ing on the influence of a limited group of variab les; observing the existence of five g roups related to the content of work, working conditions, group work, the reward system and general variables that are associated with workers welfare conditions. These are not necessarily directly related to the organization or its way of operating but that affect the SL.In the group of variab les related to the work content, researchers like Hsu and Liao (2016) emphasize the incidence of variables related to this aspect, considering variables such as the variety of skills, personal autonomy, task identity and the importance of the task, and the feedback of work that favors learning.
The working conditions, safety, and spirituality that must be ensured in the workplace are evaluated by researchers such as Beehner andBlackwell (2016), Consiglio, Borgogni, Di Tecco, andSchaufeli (2016). In this sense, specific analyzes have been developed like the incidence of non -territorial work in JS (Kim, Candido, Tho mas, and de Dear, 2016). In terms of interpersonal relationships or group work, it was found leadership style analyzes, empowerment of workers, participation in decision-making, where positive relationships have generally been observed (Chiang and Lin, 2016;Edú-Valsania, Moriano, and Molero, 2016;Lee, Kim, and Perdue, 2016;Kim, Knutson, and Choi, 2016). For their part, Kanyurhi and Bugandwa Mungu Akonkwa (2016) analy ze as conditioning variables: leadership, organizational structure, relationship with boss, supervisors and colleagues and the work itself.
In the same way, the incidence of the variables related to the reward of emp loyees in JS has been deepened, with general investigations in this area like Baule andSoost (2016) andMiller et al (2016). Further research is focused on the development of concrete rewards schemes and the incidence of hierarchy in the organizat ion as a form o f reward (Ogwueleka and Marit z, 2016;Otto and Bolle, 2016). There are more particular studies that delve into specific variables with in reward forms such as the effects of equity, stimu lation methods and perception of payment system (Oxley and Pandher, 2016;Park and Sturman, 2016;Sieweke, Köllner, and Süß, 2016;Wickramasinghe and Wickramasinghe, 2016).
Welfare conditions are analyzed by Park et al. (2016), who introduced in the study variables that are exogenous to the organization such as personal experiences, vocational preparation, and satisfaction with life in general, but limit the study to a very particular field : people with disabilities. Similarly, Hu and Ho (2016) argue that a family-friendly environment favors JS.
There is another group of researchers (Alegre et al., 2016) who analyze variab les fro m d ifferent groups in a mixed way, although they exclude variables of considerable importance, such as those who deepen the incidence of work organization, relationships with co-workers and supervisors, group work, personal autonomy and mastery of the strategy. Also, the research of A missah et al. (2016) presents a particular analysis o f the hotel industry and proposes to evaluate the following variables: salary, pro motion, supervision, training, work content. Siengthai and Pila -Ngarm (2016) present a fairly general view of the study by identifying variables such as Possibilities for dev elopment, achievement and recognition, salary, relat ionships with supervisors and co -workers, status, participation in decisionmaking, working conditions and some welfare conditions, but their greater emphasis is on the variables related to the work content like personal autonomy, task identification, variety of skills, meaning of tasks and feedback. Another integral vision of the analysis is the one made by Álvarez Santos (2106), who proposes the integration of all variables treated in the literature but does not delve into whether they affect JS in general.
Similarly, in the previous literature, the techniques used to delve into JS are varied. For example, Baldschun, Töttö, Hämäläinen, and Salo (2016)

Selection of the Variables That Condition the Js
In order to establish the study variables, the variables analy zed in the literature were reviewed and classified, which were submitted to the evaluation of a group of experts (11 university professors with PhD degrees related to business management who have carried out studies on this subject and have at least more than 12 years of research experience) were chosen, according to their competence coefficient, showing co mpetences higher than 0.84. The in itial review of the literature evidenced the existence of 33 variables with some degree of incidence in the JS, and the evaluation of the experts led to reducing the listing to 25, as shown in Tab le 1. It was included the variables selected by al less than 75% of experts.

Design of the Method of Evaluation of the Variables
A survey was designed to know the state of the 25 variables chosen in the entities studied. A question was asked to assess the overall satisfaction of the entity. In all cases, the variables should be evaluated using a Likert scale fro m 1 to 5 (from "totally unsatisfied" to "totally satisfied").

Sample Definition
The survey was applied to a varied group of representative organizations fro m d ifferent sectors, with a variety of years of founding and size of the organization, as shown in Table 2. Within the organizations a non -sampling was applied probabilistic for stratified convenience, covering 30% of workers and ensuring representation of the different

Analysis of the Relationship between the Variables and Their Influence on the JS
With the information co llected, a confirmatory factor analysis (CFA) was performed in order to verify the existence of a construct that would exp lain the relationship of the variables under st udy. Likewise, the existence of dimensions and the reliability of the scales used were evaluated. The SPSS 23 statistical package was used to establish the factorial structure of the set of observed variables, allowing to evaluate if there is a relationship between the observed variables and the existence of underlying latent constructs (Aktepe, Ersöz, and Toklu, 2015; Kesari and Atulkar, Schumacher and Lomax, 1996).

Structural Equation Modeling
Structural

Evaluation of the Model
To determine the validity of the analyzed model, the indicators used for this purpose will be used, as shown in Table 3.

Results
With the application of the instruments in the organizations research objects it was possible to perform a CFA, which was valid when a KM O coefficient of 0.901 was observed, a general value of exp lained variance of 78.21 and the Bart lett sphericity test was highly significant, confirming the existence of five dimensions that group a total of 25 variables that influence the overall satisfaction. In addition, reliability was evaluated, obtaining a Cronbach's alpha in all instruments above 0.822 (Table 4). As can be seen, the latent variables found to correspond to those analyzed previously in the literature review and those selected by the experts, corroborating that the variables studied can be grouped into five latent variables related to work content, working team, stimulation, working conditions and welfare conditions. Fro m the previous results, we proceeded to verify the structure of the relationships between variables and dimensions and between them and the observed variable "general satisfaction". Figure 2 shows the result of the analysis of the structural equation model obtained. As can be seen, the totality of the analyzed variables shows some degree of in fluence on general job satisfactio n, resulting in a more pronounced influence on variables related to stimulat ion and work content, and to a lesser extent those related to working conditions. Table 5 shows the adjustment indicators of the observed model, based on which it is possible to s tate that the relationships between the variables studied constitute a good model, based on verifying that the adjustment indicators satisfy the established parameters except for GFI (0.872) and RMSEA (0.082).

Discussion and Conclusions
The observed model was based on the variables substantiated in the literature related to the JS. Through this model, the mu ltifactorial character o f the JS as a construct is verified. Based on the above, it is considered that the results obtained correspond to the results of previous studies.
The variables of the JS studied correspond in general with those contemplated by Álvarez Santos (2016), Siengthai and Pila-Ngarm (2016) and are representative of the analyzed under a vision biased by the rest of the researchers considered in the rev iew of the preced ing literature.Both the results of the factorial analysis and those of the analysis of the structural equations model allowed to corroborate that in the JS there are five groups of variab les, co mmonly addressed in the literature in a biased or isolated way.
The working conditions dimension is one of the most studied (Beehner and Blackwell, 2016;Consiglio, Borgogni, Di Tecco, and Schaufeli, 2016;Kim, Candido, Tho mas, and de Dear, 2016), both regarding the ergonomics of the post and hygiene, and to a lesser extent, the aesthetics and material -technical supply. The group work dimension is also frequently approached and the results obtained are consistent with those presented by previous authors such as Chiang and Lin (2016) Specifically, in the analysis developed through structural equation models to evaluate which variables influence the JS and in what form they do, the results found in this research coincide with those of Álvarez -García et al. (2016), finding incidence in the JS of the variables: Leadership, personal development, and material -technical assurance. However, this study does not contemplate the rest of the variables analy zed here. The results of Hsu and Liao (2016) are also closely related to the analysis of one of the dimensions that influence the JS, and within it the same variables; On the other hand, this author in his research considers the influence in the organizational co mmit ment and of this one on the JS. About the research of Park et al. (2016) did not find similarities since the study above concentrates, essentially, on evaluating the influence of variables external to the organization but that can certainly impact the JS.
As a fundamental difference of the results of this research concerning the precedents, it can point out the integral evaluation of the influence of the different variab les and dimensions of the JS, wh ich have been investigated in a biased way or for very specific contexts. However, in this research it is possible to present and evaluate an integral vision of how the various variables analyzed in the literature affect the JS and in several organizational contexts, knowing through the model of structural equations the degree of incidence of variables and dimensions in the general JS.
The results obtained in this research are limited in the sense that it is not deepened in the incidence of variables external to the organization such as the living and family environment o r the h istory prior to work in the organization as the origin of the workers, nor it is analyzed if there are variables related to demographic characteristics such as sex, age, seniority that could somehow mark d ifferences in the JS. Similarly, it is not investigated to what extent the JS that is