Purpose: We try to find out differences between personal and job-related features to know which better explain job satisfaction. This study is made in a year of economic growth and in two years of economic crisis, in order to determine if the economic crisis affects to previous results.Design/methodology: The data are from the Quality of Labour Life Survey by the Ministry of Employment and Social Security in Spain, in 2007, 2009 and 2010. We use linear models (ANOVA), principal component analysis and stepwise multiple regression. The variables are degree of satisfaction with the current job and a group of personal variables (gender, age and education level) and job-related variables (with a maximum of 14 variables depending on the method).Findings: Using linear models get the variables related to work which provide better results to explain job satisfaction, and after a stepwise regression made with factors of principal component analysis, we find out that salary is one of the last factors in this explanation. The variables that influence on job satisfaction do not depend on the economic cycle, although the hierarchies are different among them.Social implications: During the crisis, the demands of workers are lower because they prefer to have a job with low working conditions and low salary than lose their job. Reducing the degree of satisfaction with stability and wages is due to the economic situation, because labour contracts are less stable and remunerated.Originality/value: We have compared the results of stepwise regression made with the original variables and the factors of principal component analysis. The combination of these methodologies is new in studies of job satisfaction, as well as the original combination of 14 variables related to work.
A Principal Component Analysis (PCA)-based framework for automated variable selection in geodemographic classification
