A structural equation modeling to understand the determinants of employee turnover: An empirical study

Abstract

With the Indian economy growing rapidly, the rate of employment has seen a tremendous upward rise. The expansion of boundaries and the credible usage of internet have heightened the demands for highly skilled managers and professionals. An opinion poll of HR professionals from various countries reveals a unanimous view that rampant job hopping has become a culture across industries (Khatr, Chong & Budhwar, 2001: 54). A number of earlier studies suggest that the length of service in a job can be used to estimate the levels of worker job satisfaction. The assumption is that the less satisfied workers tend to resign while the more satisfied workers tend to remain in a job. This situation should result in a higher average level of satisfaction reported by employees who have been in a given organization for a greater period of time. Kim, Price Mueller & Watson (1996) Price and Mueller’s Model is a comprehensive employee turnover model which captures a 360 degree view of the determinants of employee turnover. It comprises of Individual determinants, Structural determinants, Environmental determinants, Process variables and correlates. In this paper, we initially studied the impact of age, salary and marital status of the correlates on structural and process determinants, and we further consider studying the moderating effect of eight variables of structural and two variables of process determinants in the Indian context by using Structural equation modeling.

Keywords

Structural determinants process determinants job satisfaction organization commitment autonomy job stress

1. Introduction

One of the biggest challenges for the human resource professional is to retain its employees. Employee turnover has grown in complexity and many new theories and concepts have evolved over a period of time, beginning from March & Simon (1958), who believed that “under nearly all conditions the most accurate single predictor of labor turnover is the state of economy” (March & Simon, 1958. P. 100; Hom & Griffeth, 1995, P. 53). They also believed that enticements provided through an employer could influence employee job satisfaction and the desire to remain with the present employer as turnover theory and practice progressed. Price (1977) linked structural or organizational determinants to the turnover process. He suggested that employees develop activities to enhance job satisfaction rather than be dissatisfied, as employees are more likely to stay when job satisfaction outweighs dissatisfaction. Structural determinants play a vital role in creating job satisfaction and organizational commitment in employees. The author would like to study few co-relates such as Age, Cost to Company, Marital Status, Number of Companies Worked for, Years of Experience, and Number of Years in the same company and understand the impact of the same on structural determinants and process variables. In this paper we intend to study the moderating effect of six correlates on structural and process determinants in the Indian context.

2. Literature review

Employee turnover in organizations is a very serious issue; it is costly for organizations to search, select, train and replace leavers (Mobley, 1982; Price 1977; Staw, 1980). When low or average performers leave an organization it doesn’t create much of impact, whereas when high performers leave an organization, (especially those who are managers and professionals), this leaves a negative impact on the overall organizational effectiveness and firm success (Holtom, Mitchell, Lee & Inderrieden, 2005; Mitchell, Holtom, Lee, Sablynski & Erez, 2001). However literature review and analysis of the effects of employee turnover on companies is quite limited; most research on employee turnover in the past focuses on understanding and explaining employee turnover at individual levels (Griffeth, Hom & Gaertner, 2000; Hom & Griffeth, 2000; Hom & Griffeth, 1995; Mowday, Porter & Steers, 1982; Peterson 2004). Barrick & Zimmerman (2005) researched 1500 studies in the past 50 years on employee turnover. Most studies focused on understanding reasons for employee departure; few focused on what employers should do to hold people back, in order to increase stickiness in their organizations. Terborg and Lee (1984) were the first ones to stress that without the study of organizational determinants, the concepts of employee turnover would reduce understanding. They also questioned the existence of homology between functional relations found at the individual level of analysis and those that might exist at the organizational level of analysis. Among the most frequent causes of turnover in the European context is the status of employees in the typical organization. Turnover is negatively correlated & commitment is positively correlated to age, seniority (tenure), skill, job prestige, earnings, and chances of promotion. Women have higher turnover than men. Other determinants, which may themselves be a result of status in an organization, include job satisfaction, job challenge, job achievement, and cohesiveness with fellow employees. These determinants are again negatively correlated to turnover and positively related to commitment to an organization (Kerr 9147; Smith and Kerr, 1953; Silcock, 1954; Herzberg et al., 1957; Hedberg 1960; Hall and Schneider, 1972; Hrebiniak and Alutto, 1972; Buchanan, 1974). Although the relationship between pay and employee turnover has generated a tremendous amount of interest and research over the years, it has not been able to produce much interest in this area as most researchers have downplayed the importance of pay to turnover. Hom & Griffeth (1995) maintain that both employers and employees perceive pay playing a central role in turnover. Prevailing motivational theories have tended to emphasize intrinsic motivation, maintaining that lower needs are of lesser importance in modern society (Brief & Aldag, 1989). Research in management literature has generally underscored the limited role of pay, with findings indicating only modest associations between compensation and employee departure. Age has been found to be positively correlated with what has been defined as commitment, but its effect on structural determinants and individual determinants are not clearly mentioned (Hackett et al., 1994; Hrebiniak and Alutto, 1972, Meyer Allen and Smith 1993). With the Indian economy growing rapidly, the rate of employment has seen a tremendous upward rise. The expansion of boundaries and the credible usage of internet have heightened the demands for highly skilled managers and professionals. An opinion poll of HR professionals from various countries reveals a unanimous view that rampant job hopping has become a culture across industries (Khatr, Chong & Budhwar, 2001: 54). A number of earlier studies suggest that the length of service in a job could be used to estimate the levels of job satisfaction of workers. The assumption is that less satisfied workers tend to resign while more satisfied workers tend to remain in a job. Consistent with this thinking, a negative relationship between job satisfaction and turnover has been reported by several researchers (Atchinson and Lefferts, 1975; Karp et al., 1973; Locke, 1976; Mobley et al., 1979). This situation should result in a higher average level of satisfaction reported by employees whose have been in a given organization for a greater period of time.

The structural determinants under study have been taken from the Price and Mueller model (Kim, Price Mueller & Watson, 1996). In this model, Kim differentiated three types of determinants named as individual determinants, structural determinants, and environmental determinants. These in turn affect process variable, and correlates influences the determinants. The author in this study would like to understand the moderating effects of correlates (Age, Cost to Company, and Marital Status). A brief definition of structural determinants, process variable, and correlates are mentioned below.

Structural Determinants: Structural determinants are also termed as organizational determinants; the factors which are under the control of an organization (Mueller and Price, 1990). Mueller and Price defined structural determinants of employee turnover as the relationship of structural attributes of the work environment as well as the pattern of social relations within the work environment. Such structural determinants find their roots in internal causes such as characteristics of the work performed, the reward system, the promotion system, and the interaction of organizational members. There are eight determinants of structural determinants as follows: autonomy, distributive justice, job stress, pay, professional growth, promotional chances, routinization, and social support.

Autonomy: Autonomy is one of the most important determinants that influences turnover. Autonomy plays a major role in employee growth and development. Autonomy also increases the risk-taking ability of an employee. Price and Mueller (1981 & 1986) defined autonomy as the degree to which an employee exercises power relative to his/her job, and found it negatively correlated to turnover through job satisfaction and organizational commitment.

Distributive Justice: Distributive Justice is the perception of employees towards their organization(s). It covers all aspects such as promotion, rewards, salary hikes, fringe benefits, etc. It is also used to compare employees within in the system. Price & Mueller (1986) & Mueller et al. (1994) define distributive justice as the degree to which rewards and punishments are related to the amount of an employee’s input into an organization. They also found that distributive justice is negatively correlated to turnover through job satisfaction and organizational commitment.

Job Stress: Job stress is one of the most critical determinants to be studied. A key point is whether an employee has the sufficient skill set to execute a job, or is in an organization that expects more from an employee than his/her abilities can permit for a given job profile. The basic definition of job stress is the extent to which job duties cannot be fulfilled. Job stress is positively correlated to turnover via job satisfaction and organizational commitment (Price & Mueller, 1981 & 1986 & Mueller et al., 1994).

Pay: With the Indian economy in a vibrant state there is a demand and supply mismatch within professionals. Many instances about job hopping are quoted in literature. Literature review has shown that pay has been neglected while distributive justice has been given greater importance. The definition of pay is money and its equivalent which employees receive for their services to an employer. Price and Mueller (1990) argue that distributive justice, as opposed to pay, affects an employee’s attitude and ultimately turnover.

Professional Growth: How do employees see themselves in a defined span of time? Is there enough growth opportunity? March and Simon (1958) argued that if an organization is large enough then employees will continue to stay. The definition of professional growth is chances provided by an employer to increase job-related knowledge and skills. Kim et al. (1996) however found a negative relationship between turnover via job satisfaction and organizational commitment.

Promotional Chances: An employee’s perception of his or her career growth in an organization; how fast an employee can grow in an organization. Promotional chances are defined as the degree of potential vertical occupational mobility within an organization. Price & Mueller (1981), Price & Mueller (1986), and Kim et al. (1996) found a negative relationship to turnover via job satisfaction and organizational commitment.

Routinization: Repetitive jobs coupled with a good educational background can increase the chances of an employee quitting a job. Organizations normally hire overqualified personnel for doing repetitive work. Mueller et al. (1994), Kim et al. (1996) found that routinization is positively related to turnover via job satisfaction and organizational commitment.

Social Support: Social support has three aspects: social support from coworkers, family, & supervisors. All three support systems are very important for an employee. If an employee is deprived of assistance on job related problems, then this acts as a catalyst for quitting. Mueller et al. (1994) and Kim et al. (1996) found that social support is negatively related to turnover via job satisfaction and organizational commitment.

Process variables: Process variables are the intermediary linkages between turnover and the structural, environmental, and individual determinants of turnover. There are primarily two: job satisfaction and organizational commitment. Job satisfaction is defined as the degree to which employees have a positive affective orientation towards employment by the organization (Price 1997). Organizational commitment is defined as loyalty to an organization (Price 1997). These two variables are viewed as essential components of turnover models because their empirical relationships with voluntary turnover have been firmly established through numerous Meta-Analyses (e.g., Cohen, 1993; Cohen & Hudecek, 1993; Cotton & Tuttle, 1986; Hom & Griffeth, 1995; Hom, Caranikas-Walker, Prussia & Griffeth, 1992; Steel & Ovalle, 1984; Tett & Meyer, 1993).

3. Objective of the study

1.
Impact of age on Structural and Process determinants.
2.
Impact of salary on Structural and Process determinants.
3.
Impact of marital status on Structural and Process determinants.
4.
Moderating effect of six correlates on structural and process determinants.

4. Research methodolgy

This research is based on primary and secondary data. Primary data was used in the form of questionnaires along with turnover interviews to understand reasoning. The turnover intentions questionnaire by Kim and Muller was used to collect data from the respondents.

Measurement of Instrument: A total of 149 questionnaires were administered for each participant, of which 10 questionnaires were from the demographic point of view and 10 questions were from the information point of view. A total of 33 questions were administered on Likert scale. Structural determinants have seven variables including Autonomy, Distributive Justice, Job Stress, Professional Growth, Promotional Chances, Routinization and Social Support. A total of 36 questions are administered in these determinants. A process variable has four variables including Job Satisfaction, Organizational Commitment, Search Behaviour and Intention to Quit. Few of the process variables such as Comparison of Alternative with Present Job, Expected Utility of Search, Expected Utility of Quitting, Thinking of Quitting, and Intention to Search the data for these variables were also collected for future analysis.

Descriptive and inferential research was conducted with the help of a structured questionnaire. Employees from different industries were questioned and observed. Questions were framed using Likert Scale (five and seven point). The data is collected from binary solutions. Binary solutions are the biggest distribution of Dell, HP, and Lenovo in India. They also have a back office for client servicing. The survey sampled mainly consists of customer touch point employees. The data was collected from 386 employees.

Data Collection: A total of 1302 participants were surveyed, making this one of the biggest samples size for testing this model to close the literature gap. Data samples were collected from four different companies from four different sectors. The entire sample collected was from customers facing executives who dealt with customer directly. Organisation One is a large and reputed software company in India that deals with software development and has got a Pan India presence. Out of 900 odd customers facing executives, 750 people participated in this survey and filled up 100% of the questionnaire correctly. The questionnaire was administered across all levels in the organisation and across all job designations.

The second company we picked was a large distributor for hardware products specialised in dealing computers and printers and allied hardware products with a national presence. 100% of the staff from the sales and customer service was administered. This company is one of the largest of its kind in the industry and caters to B2B Segment.

Table 1

Structural determinants Descriptive statistics

Variables $P$ -value Decisions ( $\alpha=$ 0.025) Mean

Autonomy 0.000 $P<\alpha$ ; reject H0 Younger Older

Distributive justice 0.000 $P<\alpha$ ; reject H0 3.97 2.63

Job stress 0.000 $P<\alpha$ ; reject H0 3.99 2.81

Professional growth 0.000 $P<\alpha$ ; reject H0 4.08 3.31

Promotional chances 0.000 $P<\alpha$ ; reject H0 2.97 2.74

Routinization 0.000 $P<\alpha$ ; reject H0 4.10 3.39

Social support 0.000 $P<\alpha$ ; reject H0 4.10 3.38

4.04 3.37

Structural determinants	Descriptive statistics
Variables	$P$ -value	Decisions ( $\alpha=$ 0.025)	Mean
Autonomy	0.000	$P<\alpha$ ; reject H0	Younger	Older
Distributive justice	0.000	$P<\alpha$ ; reject H0	3.97	2.63
Job stress	0.000	$P<\alpha$ ; reject H0	3.99	2.81
Professional growth	0.000	$P<\alpha$ ; reject H0	4.08	3.31
Promotional chances	0.000	$P<\alpha$ ; reject H0	2.97	2.74
Routinization	0.000	$P<\alpha$ ; reject H0	4.10	3.39
Social support	0.000	$P<\alpha$ ; reject H0	4.10	3.38
			4.04	3.37

The third company we chose is service based. It is one of the world’s largest companies into managing customer Loyalty and Engagement programs. Their business is to manage back end for organisations into technology development and redemption and catalogue management. Their business has a very unique model as they Design, Develop, Co-create and implement Loyalty and Engagement solution to enhance business productivity and create satisfaction. There are more than 90 customers facing staff in the organisation, however, 84 officially participated in the survey with 100% accuracy and completion.

The fourth company is an emerging leader in the Automobile sector and is part of a big automobile giant. It manufactures and sells two wheelers in the Indian Market. Its model consists of a 60 odd dealership across India managed by 100 odd people. The questionnaire was sent to all 103 officials, however only 83 participated and returned 100% completed data.

All four companies were from different sectors and verticals in their industries. They were selected due to ease in assessing the employee data for administration of the questionnaire. The sales and marketing team was very cooperative, and the data was collected from the companies over one and a half years. The questionnaire was administered through pen and paper for the Software Company and for the other three companies; it was administered through survey portal survey monkey.com. The sample selected is convenient sampling.

5. Data anlaysis and findings

5.1 Impact of age on structural and process determinants

5.1.1 Structural determinants

$H_{0}$ : Age is independent of all the variables of Structural determinants

$H_{1}$ : Age is dependent of all the variables of Structural determinants

From the above table, we concluded that at 95% level of significance the null hypothesis is rejected for all factors of structural determinants. As an alternative hypothesis accepted, this means that age is dependent on all variables of Structural determinants. So further, we have to find out which age group perception was more or less for that we would use descriptive statistics. From the descriptive statistics, we concluded that the employees of older age groups have negative perceptions for the variables of structural determinants such as autonomy, distributive justice, job stress, promotional growth, routinization and social support as compared to younger age groups.

There is no significant relationship between age and turnover intentions. At any age people may tend to leave jobs or may not leave jobs based on structural determinants and process determinants.

5.1.2 Process determinants

$H_{0}$ : Age is independent of all the variables of process determinants

$H_{1}$ : Age is dependent of all the variables of process determinants

Table 2

Process determinants Descriptive statistics

Variables $P$ -value Decisions ( $\alpha=$ 0.025) Mean

Job satisfaction 0.000 $P<\alpha$ ; reject H0 Younger Older

Organization 0.001 $P<\alpha$ ; reject H0 5.33 2.12

commitment 5.63 2.46

Process determinants	Descriptive statistics
Variables	$P$ -value	Decisions ( $\alpha=$ 0.025)	Mean
Job satisfaction	0.000	$P<\alpha$ ; reject H0	Younger	Older
Organization	0.001	$P<\alpha$ ; reject H0	5.33	2.12
commitment			5.63	2.46

Table 3

Structural determinants
Variables	$P$ -value	Decisions ( $\alpha=$ 0.025)
Autonomy	0.000	$P<\alpha$ ; reject H0
Distributive justice	0.002	$P<\alpha$ ; reject H0
Job stress	0.000	$P<\alpha$ ; reject H0
Professional growth	0.161	$P>\alpha$ ; accept H0
Promotional chances	0.000	$P<\alpha$ ; reject H0
Routinization	0.002	$P<\alpha$ ; reject H0
Social support	0.001	$P<\alpha$ ; reject H0

From the above table, we concluded that at 95% level of significance the null hypothesis rejected for all factors of process determinants. As alternative hypothesis accepted, this means age is dependent on all variables of process determinants. When we try to relate the age group of managers to structural determinants it was found that the reasons for youngsters quit were lack of professional growth, autonomy and distributive justice; for the elders, reasons included promotional chances, routinization and social support were also the reasons for turnover intentions. The young managers are still keen to work and variables like promotional chances, routinization and social support are a few of the reasons. They are also able to positively cope with the job stress component.

5.2 Impact of salary on structural and process determinants

5.2.1 Structural determinants

$H_{0}$ : Income of employee is independent of all the variables of Structural determinants

$H_{1}$ : Income of employee is dependent of all the variables of Structural determinants

From the above table, we concluded that at 95% level of significance the null hypothesis is rejected for all factors except professional growth of structural determinants. As the alternative hypothesis accepted, this means the salary of an employee is dependent on all the variables of Structural determinants except professional growth.

5.2.2 Process determinants

$H_{0}$ : Income is independent of all the variables of process determinants

$H_{1}$ : Income is dependent of all the variables of process determinants

Table 4

Process determinants Descriptive statistics

Variables $P$ -value Decisions ( $\alpha=$ 0.025) Mean

Job satisfaction 0.003 $P<\alpha$ ; reject H0 High Medium Low

Organization commitment 0.001 $P<\alpha$ ; reject H0 4.67 3.67 2.63

5.23 3.84 2.12

Process determinants	Descriptive statistics
Variables	$P$ -value	Decisions ( $\alpha=$ 0.025)	Mean
Job satisfaction	0.003	$P<\alpha$ ; reject H0	High	Medium	Low
Organization commitment	0.001	$P<\alpha$ ; reject H0	4.67	3.67	2.63
			5.23	3.84	2.12

From the above table, we conclude that at 95% level of significance the null hypothesis is rejected for all the factors of process determinants. As the alternative hypothesis accepted, this means that income is dependent on all the variables of process determinants. Therefore, we further have to find out which age group perception was more or less for that we will use descriptive statistics. From the descriptive statistics, we conclude that employees of the lower income group have adverse perception towards the variables of process determinants such as job satisfaction and organization commitment as compared to medium and high-income group. Therefore, the organization should concentrate on the lower income group.

5.3 Impact of marital status on structural and process determinants

5.3.1 Structural determinants

$H_{0}$ : Marital status of employee is independent of all the variables of Structural determinants

$H_{1}:$ Marital status of employee is dependent of all the variables of Structural determinants

Table 5

Structural determinants Descriptive statistics

Variables $P$ -value Decisions ( $\alpha=$ 0.025) Mean

Autonomy 0.000 $P<\alpha$ ; reject H0 Married Unmarried

Distributive justice 0.002 $P<\alpha$ ; reject H0 3.77 2.75

Job stress 0.000 $P<\alpha$ ; reject H0 4.00 4.81

Professional growth 0.607 $P>\alpha$ ; accept H0 4.35 3.36

Promotional chances 0.003 $P<\alpha$ ; reject H0 4.20 3.09

Routinization 0.000 $P<\alpha$ ; reject H0 4.23 3.56

Social support 0.002 $P<\alpha$ ; reject H0 4.12 3.44

Structural determinants	Descriptive statistics
Variables	$P$ -value	Decisions ( $\alpha=$ 0.025)	Mean
Autonomy	0.000	$P<\alpha$ ; reject H0	Married	Unmarried
Distributive justice	0.002	$P<\alpha$ ; reject H0	3.77	2.75
Job stress	0.000	$P<\alpha$ ; reject H0	4.00	4.81
Professional growth	0.607	$P>\alpha$ ; accept H0	4.35	3.36
Promotional chances	0.003	$P<\alpha$ ; reject H0	4.20	3.09
Routinization	0.000	$P<\alpha$ ; reject H0	4.23	3.56
Social support	0.002	$P<\alpha$ ; reject H0	4.12	3.44

From the above table, we conclude that at 95% level of significance the null hypothesis is rejected for all factors except professional growth of structural determinants. As the alternative hypothesis accepted, this means that marital status of the employee is dependent of all variables of Structural determinants except professional growth. We can therefore conclude that the married and unmarried group perceptions were the same for professional growth. So we further have to find out which group perception was more or less; for this we will use descriptive statistics for the remaining factors. From the descriptive statistics, we conclude that employees of the married status group have a greater positive perception towards factors such as autonomy, job stress, promotional chances, routinization and social support whereas for distributive justice is higher in unmarried status group.

5.3.2 Process determinants

$H_{0}$ : Marital Status is independent of all the variables of process determinants

$H_{1}$ : Marital Status is dependent of all the variables of process determinants

Table 6

Process determinants Descriptive statistics

Variables $P$ -value Decisions ( $\alpha=$ 0.025) Mean

Job satisfaction 0.000 $P<\alpha$ ; reject H0 Married Unmarried

Organization 0.001 $P<\alpha$ ; reject H0 5.33 2.12

commitment 5.63 2.46

Process determinants	Descriptive statistics
Variables	$P$ -value	Decisions ( $\alpha=$ 0.025)	Mean
Job satisfaction	0.000	$P<\alpha$ ; reject H0	Married	Unmarried
Organization	0.001	$P<\alpha$ ; reject H0	5.33	2.12
commitment			5.63	2.46

Table 7

Overall selected fit indices
Fit measures	Proposed	Comments
	SEM model
RMR	0.06	RMR $<$ 0.08; Model fit
RMSEA	0.05	RMSEA $<$ 0.07; Model fit
GFI	0.95	GFI $\geqslant$ 0.95; Model fit
NFI	0.78	NFI $\geqslant$ 0.95; Model fit
CFI	0.90	CFI $\geqslant$ 0.95; Model fit
Chi-Square	Sig.	df
96	0.468	21

From the above table, we conclude that at 95% level of significance the null hypothesis is rejected for all the factors of process determinants. As the alternative hypothesis accepted, this means marital status is dependent on all variables of process determinants. So we further have to find out which age group perception was more or less; for that we will use descriptive statistics. From the descriptive statistics, it is observed that in the unmarried status group the perception of job satisfaction and organization commitment is greater compared to the married status group. So we can conclude that married employees are more satisfied with their job and are more committed towards their organization.

5.4 Moderating impact of the variables

(a) (a)
Model evaluation or testing.
(b)
CFA model can be evaluated on basis of two major aspects (Hair et al. 1998): Overall goodness of fit. The interpretability, size and statistical significance of the model parameter estimates. The model is shown in Fig. 1.

Absolute Fit Measures: Provides the most basic assessment of how well a researcher’s theory fits sample data. It does not explicitly compare the goodness of fit (GOF) of a specified model to any other model. The limitation of general goodness of fit can be overcome with the help of absolute fit measures, which include RMR, RMSEA and GFI.

RMR is Root Mean Residual: The average of residuals between individual observed and estimated covariance and variance terms. If the value is $<$ 0.08 the model is better (Hair et al. 1998). However, values as high as 0.08 are deemed acceptable (Hu and Bentler, 1999).

RMSEA is Root Mean Square Error of Approximation: Measures attempts to correct for tendency of the chi square, RMSEA is less than $<$ 0.07 indicates that the model is absolutely fit (Hair et al. 1998).

Goodness of fit index (GFI): The possible range of GFI value is $\geqslant$ 0.95, which indicate better fit (Hair et al. 1998).

Incremental Fit Indices: Assesses how well a specified model fits relative to some alternative baseline models.

Normed Fit Index (NFI): One of the original incremental fit indices. The value $\geqslant$ 0.95 indicates better fit (Hair et al. 1998).

Comparative Fit Index (CFI): An incremental fit index that is the improved version of the normed fit index (NFI). The value $\geqslant$ 0.95 indicates the better fit (Hair et al. 1998).

Hair et al. (1998) argued that there was not yet consensus among scholars regarding the standard of the goodness of fit test. Thus, he suggested that researchers pick any one or two indices to evaluate the goodness of fit towards the model. In this research, the results have shown overall significant values of goodness of fit to accept the proposed model. The Goodness of fit test which indicates Chi Square value is 96; and sig. value is 0.468. Hence lower the chi square value, better the model, sig. value is $>$ 0.05, i.e. 0.468, test cannot reject null hypothesis. In other words, the test showed that the model is fit for the data. The critical ratio is calculated as Chi square value divided by degree of freedom). For our model the critical ratio value is (96/21 $=$ 4.57) which is less than 5. The values of the fit indices indicate a reasonable fit of the measurement model with data (Byrne, 2001, pp. 79–86).

The model fit the data perfectly, since the estimation of this model required precisely at all degrees of freedom that was available. The completely standardized parameter estimates of this model are shown in the above table.

Table 8

Completely standardized

parameters estimates for model

Independent Job Organization

variables satisfaction commitment

Autonomy 0.42 0.22

Distributive justice 0.39 0.22

Job stress 0.31 0.47

Promotional chances 0.34 $-$ 0.40

Routinization 0.25 0.53

Social support 0.14 0.74

Professional growth $-$ 0.11 $-$ 0.40

Table 9

Completely standardized

significance parameters

for job satisfaction

Independent variables $t$ value $p$ value

Autonomy $\to$ Job satisfaction 5.75 0.000

Distributive justice $\to$ Job satisfaction 5.18 0.00

Job stress $\to$ Job satisfaction 5.09 0.00

Promotional chances $\to$ Job satisfaction 5.37 0.00

Routinization $\to$ Job satisfaction 5.37 0.00

Social support $\to$ Job satisfaction 0.03 0.00

Professional growth $\to$ Job satisfaction $-$ 1.40 0.161

Figure 1.
Results of the model.

The above table shows the standard regression weights of the latent variables manifesting the construct overall job satisfaction and organization commitment. These regression weights represent the degree of association between the construct and manifesting variables. For example, if overall job satisfaction increases by 1 standard deviation, the standard deviation of autonomy, Distributive Justice, Job Stress, Promotional Chances, Routinization, Social Support and Professional Growth will increase by 0.42, 0.39, 0.31, 0.34, 0.25, 0.14 and 0.113 respectively. Similarly, if organization commitment increases by 1 standard deviation, the standard deviation of autonomy, Distributive Justice, Job Stress, Promotional Chances, Routinization, Social Support and Professional Growth will increase by 0.22, 0.22, 0.47, $-$ 0.40, 0.53, 0.74 and $-$ 405 respectively. In addition, if job satisfaction level increases by 1 standard deviation, the organization commitment increases by 0.839. There is the negative impact of job satisfaction and organization commitment on employee turnover.

Critical ratio is similar to $t$ values in conventional statistics; it is calculated by regression weights divided by standard error, the value 1.96 or more than 2 considered as significant. Out of seven variables with job satisfaction, 6 variables are significant at $<$ 0.001 level and with organization commitment 5 variables are significant at $<$ 0.001 level; this shows that the relation between variables and their sub variables concerned are statistically significant. Both variables job satisfaction and organization commitment are statistically significant.

Table 10

Completely standardized

significance parameters for

organization commitment

Independent variables $t$ value $p$ value

Autonomy $\to$ Organization commitment 3.03 0.000

Distributive justice $\to$ Organization commitment 3.03 0.00

Job stress $\to$ Organization commitment 7.7 0.00

Promotional chances $\to$ Organization commitment 6.23 0.00

Routinization $\to$ Organization commitment 8.63 0.00

Social support $\to$ Organization commitment 5.96 0.00

Professional growth $\to$ Organization commitment $-$ 5.01 0.002

Table 11
Completely standardized parameters estimates and completely standardized parameters for the moderating effect of job satisfaction and organization commitment with employee turnover

Model estimates

Variables Estimates $t$ value & $p$ value

Job satisfaction $\to$ 0.839 30.18, 0.00

Organization commitment

Job satisfaction $\to$ $-$ 0.902 40.12, 0.15

Employee turnover

Organization commitment $-$ 0.785 45.11, 0.25

$\to$ Employee turnover

6. Conclusions and recommendations

	Completely standardized
	parameters estimates for model
Independent	Job	Organization
variables	satisfaction	commitment
Autonomy	0.42	0.22
Distributive justice	0.39	0.22
Job stress	0.31	0.47
Promotional chances	0.34	$-$ 0.40
Routinization	0.25	0.53
Social support	0.14	0.74
Professional growth	$-$ 0.11	$-$ 0.40

	Completely standardized
Autonomy $\to$ Job satisfaction	5.75	0.000
Distributive justice $\to$ Job satisfaction	5.18	0.00
Job stress $\to$ Job satisfaction	5.09	0.00
Promotional chances $\to$ Job satisfaction	5.37	0.00
Routinization $\to$ Job satisfaction	5.37	0.00
Social support $\to$ Job satisfaction	0.03	0.00
Professional growth $\to$ Job satisfaction	$-$ 1.40	0.161

	Completely standardized
Autonomy $\to$ Organization commitment	3.03	0.000
Distributive justice $\to$ Organization commitment	3.03	0.00
Job stress $\to$ Organization commitment	7.7	0.00
Promotional chances $\to$ Organization commitment	6.23	0.00
Routinization $\to$ Organization commitment	8.63	0.00
Social support $\to$ Organization commitment	5.96	0.00
Professional growth $\to$ Organization commitment	$-$ 5.01	0.002

	Model estimates
Organization commitment
Job satisfaction $\to$	$-$ 0.902	40.12, 0.15
Employee turnover
Organization commitment	$-$ 0.785	45.11, 0.25
$\to$ Employee turnover

Employee turnover is a very complex, unstructured, nonlinear and dynamic phenomena. Organizations spend huge amounts of time, money and effort to retain talented employees for success of their business, however till today researchers have been unable to completely decode the process of turnover. This study is an attempt to conduct a multilevel investigation of determinants of employee turnover in Indian organizations. The structural modeling findings further state that there is a negative correlation between professional growth and job satisfaction. The findings state that an organization needs to work on framing proper policies of promotion; as a result, the chances of employee turnover will reduce. Promotional policy will also have a positive impact on organization commitment. Job satisfaction and job commitment are the most important variables for reduction in employee turnover. There is close relationship between job satisfaction and organization commitment and vice versa.

From the findings employees of the older age group negatively perceive attributes such as autonomy, distributive justice, job stress, promotional chances, routinization and social support as compared to the younger age group. Employees in the older age group are dissatisfied with their jobs and lack commitment to their organization. Empowerment policies, freedom to take decisions and proper reward managements will help reduce employee turnover.

Income of the employee is dependent on all variables of structural determinants except professional growth. We can conclude that the chances provided by an employer to increase job-related knowledge and skills for different income groups i.e. lower, medium and higher are the same. For the attributes autonomy, distributive justice, job stress, promotional chances, routinization and social support the perception of higher income is positive as compared to the medium income group. So the higher income groups are more satisfied with their jobs and they are more committed towards the organization; the organization, however, should think about innovative practices to design right compensation policies to increase job satisfaction amongst employees.

Married employees have greater positive perception towards factors such as autonomy, job stress, promotional chances, routinization and social support whereas distributive justice is greater in the unmarried status group.

There is a significant relationship between the variables i.e. autonomy followed by distributive justice, promotional chances, job stress routinization and social support of structural determinates on process variable i.e. job satisfaction. Finally, there is a significant relationship between the variables i.e. autonomy followed by distributive justice, job stress routinization and social support of structural determinates on process variable i.e. organization commitment.

References

Allen

Shore

, & Griffeth

R. W.

(2003). The role of perceived organizational support and supportive human resource practices in the turnover process. Journal of Management.

Barrick

M. R.

, & Zimmerman

R. D.

(2005). Reducing voluntary avoidable turnover through selection. Journal of Applied Psychology.

Budhwar

(2004). Managing human resources in Asia-Pacific. London: Routledge.

Hom

P. W.

, & Griffeth

R. W.

(1995). Employee turnover. Cincinnati: South/Western.

Hotlem

B. C.

Mitchell

T. R.

Lee

T. W.

, & Inderrieden

E. J.

(2005). Shocks as causes of turnover: What they are and How organizations can manage them. Human Resource Management, 44(3), 337-352.

Guthrie

J. P.

(2000, Dec). Alternative pay practices and employee turnover: An organization economics perspective. Group and Organization Management, ABI/Inform Global, 25(4), 419.

Mobley

W. H.

(1977). Intermediate linkages in the relationship between job satisfaction and employee turnover. Journal of Applied Psychology, 62, 237-240.

Mobley

W. H.

(1982). Employee turnover: Causes, consequences and control: Reading; Addison-Wesley.

Price

J. L.

(1977). The study of turnover. Ames: Iowa state university press.

10.

Gaertner

(1999). Structural determinants of job satisfaction and organizational commitment in turnover models. Human Resource Management Review, 9(4), 479-493.

	Completely standardized
	significance parameters
	for job satisfaction
Independent variables	$t$ value	$p$ value
Autonomy $\to$ Job satisfaction	5.75	0.000
Distributive justice $\to$ Job satisfaction	5.18	0.00
Job stress $\to$ Job satisfaction	5.09	0.00
Promotional chances $\to$ Job satisfaction	5.37	0.00
Routinization $\to$ Job satisfaction	5.37	0.00
Social support $\to$ Job satisfaction	0.03	0.00
Professional growth $\to$ Job satisfaction	$-$ 1.40	0.161

	Completely standardized
	significance parameters for
	organization commitment
Independent variables	$t$ value	$p$ value
Autonomy $\to$ Organization commitment	3.03	0.000
Distributive justice $\to$ Organization commitment	3.03	0.00
Job stress $\to$ Organization commitment	7.7	0.00
Promotional chances $\to$ Organization commitment	6.23	0.00
Routinization $\to$ Organization commitment	8.63	0.00
Social support $\to$ Organization commitment	5.96	0.00
Professional growth $\to$ Organization commitment	$-$ 5.01	0.002

	Model estimates
Variables	Estimates	$t$ value & $p$ value
Job satisfaction $\to$	0.839	30.18, 0.00
Organization commitment
Job satisfaction $\to$	$-$ 0.902	40.12, 0.15
Employee turnover
Organization commitment	$-$ 0.785	45.11, 0.25
$\to$ Employee turnover