[Study on the applicability of the Brief Job Stress Questionnaire based on classical test theory and item response theory]
Y P Zhang,T L Yan,J Li,H N Wang
DOI: https://doi.org/10.3760/cma.j.cn121094-20231206-00144
2024-05-20
Abstract:Objective: To evaluate the quality and item characteristics of the Brief Job Stress Questionnaire (BJSQ) among employees in manufacturing and service industries. Methods: From December 2021 to December 2022, a total of 2077 employees from 4 manufacturing and service enterprises in Beijing were selected by the method of combining purpose sampling and convenience sampling. The Chinese version of BJSQ was used to carry out a survey on occupational stress from 57 items in 4 dimensions including job stress factors, stress response, social support factors and satisfaction. Classical test theory (CTT) was used to analyze the validity and reliability of the questionnaire. The Semejima hierarchical response model in item response theory (IRT) was used to analyze the differentiation a, difficulty coefficient b and information content of each item. Results: Among the 2077 subjects, the age M (Q(1), Q(3)) was 33 (28, 37) years old, and the working age M (Q(1), Q(3)) was 4 (2, 8) years. There were 723 (34.8%) people in manufacturing industry and 1354 (65.2%) people in service industry. Eleven common factors were extracted by exploratory factor analysis, with a cumulative variance contribution rate of 62.823%, the variance of the common factors of each entry ranging from 0.451 to 0.865, and the range of factor loading values from 0.413 to 0.825, with 5 items having cross-loadings. The results of the validation factor analysis showed that the model fit indexes of root means square error of approximation was 0.055, comparative fit index was 0.950, Tucker Lewis index was 0.948, and standardized root mean square was 0.066. Content validity results showed that the total scores of the Chinese version of the BJSQ were positively correlated with the scores of the 4 dimensions (r(s)=0.487-0.936, P<0.05) . The results of the reliability analysis showed that the Cronbach's alpha coefficient for the total questionnaire was 0.945, and the Cronbach's alpha coefficients of job stress factors, stress response and social support factors were 0.775, 0.957, and 0.830, respectively. The Spearman-Brown coefficient for the total questionnaire was 0.866, and for the 3 dimensions, it was 0.572, 0.882 and 0.772. The results of IRT analysis showed that only 1 of the 57 items had a differentiation a-value<0.30, the difficulty coefficient b-value of each item ranged from -12.02 to 11.09, of which 8 items had a difficulty coefficient that was too high, 3 items had a difficulty coefficient that was too low, and 3 items did not meet the requirements, and the average amount of information in each item ranged from 0.022 to 2.566, and there were 47 items with average amount of information>0.088 (5/57) . Conclusion: The Chinese version of BJSQ has good reliability and validity in the typical occupational groups of manufacturing and service industries in China, and most of the items have good performance, but some items still need to be further improved and optimized or deleted.