Improving Talent Management System Processes by Social Capital

                                    
   
Introduction 
  Talent management can be described as the process through which organizations anticipate their talents needs and is the critical factor in the success of organizations, as well as one of the most important and challenging challenges of human capital in the 21^st century which has a growing trend. Despite many studies of talent management in recent years, existing knowledge about this concept is limited in countries with developing economies. On the other hand, most articles in this field are conceptual and few empirical studies have been done. As a result, investigating talent management using an empirical approach, in a private context (Khash Cement Company), is important that the present study addresses.
One of the ways organizations use to maintain talents is the creation of cohesion and communication between the workforce in the organization, which in the field of management and social science is called social capital. In the absence of social capital, other assets of the organization, including human capital, loses its effectiveness and rushes through the paths of cultural and economic development, ruggedness and difficulty.
In this research, the role of social capital in improving talent management is experimentally investigated.
Theoretical framework
Social capital values people's relationships and often this logical relationship is lost in HR management. It is expected that social relationships will enhance identity and cognition, and network structures will facilitate the talent management process in the organization. Considering the existing models related to talent management, in this research, the following dimensions are considered as components of talent management (Ahmadi et al., 2012): Recruitment, The system of evaluation and discovery of talents, Development and improvement, Retention. Moreover, social capital consisted three dimensions named: relational, cognitive, and structural  (Nahapiet & Ghoshal, 1998)
Methodology
The research method is applied and the technique of data analysis is descriptive-correlational. The statistical population of this research included all employees of Khash Cement Factory, which was 274 people. According to Morgan table, 160 people were selected by simple random sampling.
Data Analysis and Findings
In this research, the second-generation method of structural equation modeling (e.g. partial least squares method (PLS)) is used.
Fittness of measurement model
Fit the model is examining the difference between the proposed model and the actual model. In order to measure the fittness of the model, factor loadings of the measurement model were first calculated and all were more than 0.5 that are suitable.
1) Cronbach's alpha  and Combined Reliability
The PLS method uses a more modern benchmark than alpha, known as Combined Reliability (CR). The values for both criteria in this study was higher than 0.7, which indicates that the model is reliable.
2) Convergent Validity
The second criterion is fitness of measurement model. This criterion (AVE) represents the average of the variance shared between each structure with its own indices, and the greater the correlation, the greater is the fitness. The mean value of the extracted variance (AVE) is always greater than 0.5 and the combined value of the combined reliability (CR) is greater than AVE, so convergent validity is also confirmed.
3) Divergent Validity
The third measurement criterion of fitting of the measurement model is discriminant validity. It uses correlation matrix (Fornell-Larcker). This matrix is the ratio of a structure to its indexes in comparison of its structure with other structures. This matrix has acceptable divergent validity if the numbers in the original diameter are greater than the values below its original diameter. The results of this matrix confirmed the suitable divergent validity.
Fittness of the structural model
1)Z significant coefficients (t-values)
The first criterion for measuring the relationship between structures is in the significant t-values. Significant coefficients indicate whether the research hypotheses are meaningful or not? Path coefficients of the variables are outside the absolute magnitude range of 1.96, so at the level of 95% confidence, the relationships between the variables is significant and research hypotheses are confirmed.
2)R² criterion
A measure used to connect the measurement part and the structural part of the structural equation modeling and indicates the effect that an exogenous variable has on an intrinsic variable. Except two dimensions (evaluation and absorption), R² values were above 0.67 that are considered strong.
3)Q² criterion
This criterion, determines the predictive power of the model in the model's intrinsic structures. The values of this criterion for most variables are higher than 0.35 and represent the strong predictive power of the model.
Overall Model fit (The goodness of Fit (GOF))
In Partial Least Squares, this criterion is the most important index of model fit, and by this criterion, the fit of the general model of structural equations is controlled.
The results show that the value of this criterion is higher than 0.36 which indicates the overall and robust fit of the model (the success of the model tested in the prediction of latent variables).
Testing research hypotheses
After reviewing the fitting of the structural and overall measurement model, the researcher is allowed to test the hypotheses and report the research findings.
Discussion and Conclusion
The purpose of this study was to investigate the role of social capital in talent management systems in Khash Cement Company. The role of social capital as a influential factor for describing the relative success of the companies is confirmed by a large number of scientists. The results of this research also show that the improvement of social capital in the organization influences the development of talent management process. The existence of social capital in an organization and the effective interaction among individuals will improve talent management (attraction, assessment and discovery of talent, development, training, and retention).
However, to generalize the results of talent management studies from one country or context to another, it is necessary to pay attention to conditions such as economical and social circumstances.