Structural Analysis of the Correlation between Occupational Stress and Quality of life in University Administrators Using a PLS Approach
Ahad Khalilzadeh1, Behnam Talebi 2*, Assadollah Khadivi 3
1- PhD Student in Educational Administration, Dept. of Educational Administration, Tabriz Branch, Islamic Azad University, Tabriz, Iran.
2- Assistant Prof., Dept. of Educational Administration, Tabriz Branch, Islamic Azad University, Tabriz, Iran.
3- Assistant Prof., Dept. of Educational Administration, Tabriz Branch, Farhangian University, Tabriz, Iran.
* Corresponding author: Behnam Talebi; E-mail: btalebi1972@gmail.com
Abstract
Background: University administrators play a key role in providing educational services to students and faculty members. This study aims to explain occupational stress (OS) in university administrators based on health-related quality of life (QoL) using a PLS approach.
Materials and Methods: The statistical population of this correlational study included all 70 managers in the field of education at Islamic Azad University of Tabriz in 2020. Based on Krejcie and Morgan's sampling table, 60 people were selected by simple random sampling. Data were collected using the revised OS questionnaire of the United Kingdom Health Safety Executive (2004) and the World Health Organization's QoL (WHOQOL-BREF) questionnaire. The questionnaire of QoL included dimensions of mental health, physical health, social health, environmental health, and general health. OS questionnaire is not categorized into components. Composite reliability was 0.96 for the OS questionnaire and 0.93 for the QoL questionnaire. To test the hypotheses, structural equation modelling was employed with SPSS and PLS software.
Results: The results showed a significant negative association between QoL and its dimensions, including general health, physical health, mental health, environmental health, and social health with OS (P < 0.01). In addition, OS could be predicted only based on the dimensions of physical health (P ₌ 0.02) and environmental health (P ₌ 0.01).
Conclusion: Paying attention to QoL in university administrators could reduce their OS and promote positive working consequences at the university.
Keywords: Occupational Stress, Quality of Life, Mental Health.
Introduction
The central role of the educational system in developing human resources and human societies in various dimensions has been well established. Although many factors are effective in developing an educational system, education managers of the university, as the most effective factors in planning for the higher education system, play a leading role in improving the quality of the education system and achieving its goals [1]. Education managers of the university are among the most important factors in providing educational services, facilitating research, planning for educational affairs, and offering guidance services to students in order to achieve goals of the higher education system in any country. In an environment in which education managers experience OS, students and professors will be stressed, which will affect performance of the higher education system of that environment [2].
Stress is one of the factors exerting destructive physical and psychological effects [3]. Besides, OS is recognized as one of the most effective factors in maintaining physical and mental health among employees [4]. OS is an adaptive response to an external situation, which leads to physical, psychological, or behavioral problems among employees in an organization [5].
According to experts, general health of university staff is essential [6], being one of the most important work-related issues in modern societies [7]. However, some factors, such as working conditions, overwork, heavy responsibilities, and changing ergonomic conditions could be harmful to employees, thereby causing OS [8]. QoL is one of the important personal factors having a significant effect on OS [9]. In the meantime, QoL is a multidimensional concept affected by many factors, such as physical conditions, mental health, personal beliefs, and social relationships [10]. According to the World Health Organization, "QoL is a comprehensive concept that describes an individual's perception of various aspects in their personal status, including physical health, mental health, self-care, social relationships, and adaptation to their environment" [11]. According to the analysis performed by the World Health Organization, QoL includes the four main dimensions of physical health, mental health, social relationships, and environmental health [12].
Research results in Iran and other countries indicate that QoL and OS are interrelated. The results of the study by Jessica (2017) show that stress is a negative predictor of QoL [13]. Several studies on employees in various organizations confirm this interrelationship as well [14-20, 8, 9, 4]. Research by Yao et al (2015) [21] as well as Khan and Khurshidi (2017) [22] on university staff showed a negative relationship between these two variables. Upon reviewing available research on these two variables among university and faculty staff, one could say that such research is rare (Alves et al, 2019) [23], which deserves to receive more attention especially in Iran. Accordingly, this study aims to answer the question of ‘What is the role of QoL in predicting OS in university education managers?' Given that this study aims to investigate the relationship between QoL and OS and considering the rareness of research in university education managers in terms of these two factors, the results of this study could be effective in identifying the relationship, reducing job stress, and improving QoL among university education managers.
Materials and Methods
This correlation research is based on structural equation modeling. The statistical population of this study were all 70 education managers of Islamic Azad University of Tabriz, including the Vice Chancellor for Education, Director General of Education, heads of education departments, heads of faculties, and heads of departments in academic year 2019-2020. The research sample consisted of 60 education managers who were selected through random sampling, with its size determined based on Bewick, Cheek, and Ball's (2003) table [24]. Based on this table, to estimate the sample size according to previous correlation studies, the expected correlation rate of (R ₌ 0.25) (P ₌ 0.05), and a two-tailed hypothesis were selected. According to values mentioned above and based on values of the critical correlation coefficient distribution Table, the sample size was set at 60. To test the hypotheses, structural equation modeling and the partial least squares (PLS) method were used. The inclusion criteria included being a faculty or staff member at Islamic Azad University of Tabriz and having a managerial position in 2020. In addition, the subjects were free to leave the study at any time. Two questionnaires used for data collection in this study included:
QoL Questionnaire: To measure QoL, the short form of the questionnaire presented by the World Health Organization (1989) (WHOQOL-BREF) was used, which included 4 dimensions of physical health (7 items), mental health (6 items), social relationships (3 items), environmental health (8 items). In addition general health measures by 2 items. This questionnaire has been used in many studies on QoL, including those of Hosseini Amin et al (2016) [25] and Khoshnoodi et al (2017) [26]. Nejat et al (2006) [27] in their research, by translating and standardizing this questionnaire as well as Iranian psychometrics, concluded that structural factors of this tool had acceptable validity and reliability in healthy and sick groups. They estimated that reliability of this questionnaire was 0.77, 0.77, 0.75, and 0.84 for the dimensions of physical health, mental health, social health, and environmental health, respectively, and added that the discriminate validity of this questionnaire was acceptable in all areas and general health [27]. Hosseinkhani and Talebi (2020) confirmed the validity of this tool for the same dimensions using first and second order confirmatory factor analysis. Accordingly, the factor value was higher than 0.4 for all four dimensions as well as general health [28]. The composite reliability of this questionnaire
was 0.93 in the present study using PLS software, and its reliability was 0.91 according to the Cronbach's alpha formula. This 5-point Likert scale questionnaire was answered by managers in a self-reported manner. For each item, point 5 indicates high health, and point 1 indicates low health.
OS Questionnaire: The OS questionnaire has 35 questions. This questionnaire has been compiled by the United Kingdom Health Safety Executive (2004) [29] and translated by Azad Marzabadi and Gholami Fesharaki (2010) [30]. In addition, its reliability was estimated to be 0.78, and its validity was confirmed by content and structure validity. Accordingly, results of their research showed that reliability and validity of the questionnaire were acceptable, and its reliability coefficient was 0.78 based on Cronbach's alpha [30]. In the present study, the composite reliability of this questionnaire using PLS software was 0.96, and its reliability based on the Cronbach's alpha formula was 0.96. Managers answered this questionnaire on a 5-point Likert scale. Accordingly, points 5 and 1 indicated the highest and lowest amounts of stress for each item, respectively.
This study has been extracted from a doctoral thesis at Islamic Azad University of Tabriz under code 1024851978569711398162259049; accordingly, ethical requirements were met, informed consent forms were obtained from the research sample, and confidentiality of respondents' information was ensured in this study.
Results
In terms of the education level, 73.3% of the research sample had a doctoral degree that was the highest frequency. Accordingly, the frequencies of a master's degree and a bachelor's degree were 20% and 6.7%, respectively. In terms of age, the 41-50 year age group accounted for 56.7% of the sample size with the highest frequency, the age group of over 50 had the frequency of 31.7%, the 31-40 year age group had the frequency of 10%, and the 20-30 year age group accounted for 1.7% of the sample size. In addition, in terms of gender, 76.7% of the respondents were male, and 23.3% of them were female (Table 1).
Table 1. Demographic characteristics of education managers
Demographic characteristics |
N |
P% |
|
Age |
20-30 year age group |
1 |
1.666 |
31-40 year age group |
6 |
10 |
|
41-50 year age group |
34 |
56.667 |
|
Over 50 year age group |
19 |
31.667 |
|
Gender |
Male |
46 |
76.667 |
Female |
14 |
23.333 |
|
Education level |
PhD |
44 |
73.333 |
MA |
12 |
20 |
|
BA |
4 |
6.667 |
According to data in Table 2, the mean QoL of education managers is 110.03 ± 15.55, and they have high QoL. Besides, the mean OS is 65.91 ± 23.64, being lower than average. In addition, the status of the education managers in terms of Qol components is higher than average.
Table 2. Descriptive indicators of quality of life and occupational stress in education managers
Variable |
M ± S.D |
Min |
Max |
Quality of life |
110.03 ± 15.55 |
82 |
130 |
General health |
8.3 ± 1.58 |
4 |
10 |
Physical health |
29.28 ± 4.52 |
19 |
35 |
Mental health |
25.35 ± 3.93 |
16 |
30 |
Environmental health |
34.11 ± 5.07 |
23 |
40 |
Social health |
12.98 ± 2.3 |
8 |
15 |
Occupational stress |
65.91 ± 23.64 |
35 |
102 |
Upon examining the relationships between the variables, the results of the Pearson's correlation test showed a significant negative relationship between QoL and its dimensions, including general health, physical health, mental health, social health, and environmental health, with OS (P < 0.01). Upon testing the research hypotheses, structural equation modeling showed that OS could be highly predicted only based on physical health and environmental health. The fit of structural equation modeling, based on convergent validity indices, the factor load of the items in the measurement model, the corrected coefficient of determination, the predictive power of the model in evaluating structural modeling, and suitability of the overall model, was considered and approved. Upon using PLS software to analyze the data, if the value of the factor load for the items was higher than 0.4, it would confirm validity of the items in measuring the variable; if the value of the significance level of the path coefficients was less than 0.05, it would indicate significance of the relationship between the variables. Q2 shows the predictive power of the model. The coefficient of determination indicates the prediction power of the criterion variable based on the predictor variable. A coefficient of determination over 0.67 indicates high predictability. In addition, goodness of fit (GoF) was developed as an overall measure of the model fit for PLS-Structural Equation Modeling. Accordingly, a prediction power value (Q2) over 0.35 indicates a high predictive power, and a model fit (GOF) over 0.25 indicates a favorable model fit [31]. The SRMR is defined as the difference between the observed correlation and the model's implied correlation matrix. It makes the assessment of the average magnitude of discrepancies between observed and expected correlations possible, as the absolute measure of the fit criterion for the model. A value less than 0.10 or 0.08 is considered a good fit. Henseler et al (2014) introduced the SRMR as a goodness of fit measure for PLS-SEM that would be used to avoid model misspecification [32]. The value of T in structural equation modeling indicates the significance of the path coefficient between the variables, the value of which being at least 1.96 in the test of the two-tailed hypothesis. If the value of T was higher than 1.96, the relationship between the two variables would be confirmed. Accordingly, the factor load of all items is higher than 0.4, being valid; the value of the adjusted coefficient of determination is 0.91, being strong; the prediction value is equal to 0.39, being above 0.35 and strong; and the good of fitness value of the model is 0.3, being above 0.25 and above average, indicating the model has a good fit. The results are presented in Table 2 and Figs. 1 and 3.
Based on the results of the PLS regression test (Table 2), the model fit indices show that both measurement models, i.e. structural and total models of the research, have a good fit.
Table 3. Results of partial least squares regression regarding occupational stress prediction and the model fit
Model fit |
Variable |
EH |
GH |
MH |
PH |
SH |
QL |
Measurement |
Composite reliability |
0.86 |
0.81 |
0.82 |
0.84 |
0.93 |
0.87 |
Path coefficient |
0.47* |
0.21 |
0.22 |
0.34* |
0.23 |
0.04 |
|
Structural |
Adjusted R-squared |
0.91 |
|||||
Q2 |
0.39 |
|
|
|
|
|
|
T |
2.43* |
1.18 |
0.92 |
1.97* |
0.27 |
0.31 |
|
Total |
GOF |
0.3 |
|||||
SRMR |
0.07 |
α < 0.05*
Fig. 1 shows p-values for each item in relation to the variables, as well as the path coefficients. Accordingly with table 3, the p-value in EH and PH is lower than 0.05 and significant. Based on Fig. 1, EH and PH have an effect on OS (P < 0.05). In the study of the standard path coefficients, the direct effect of PH on OS with a path coefficient of 0.34 and a significance level of less than 0.05, as well as the direct effect of EH on OS with a path coefficient of 0.47 and a significance level of less than 0.05, were significant. The path coefficient of QoL, GH, SH, and MH with OS was not significant due to a significance level of higher than 0.05.