An empirical evaluation of job classification using online job advertisements

Abstract

Information system (IS) jobs have continually evolved over time along with the fast changes of information technology. Many scholars and organizations have tried to create various IS job classifications and have tried to classify job advertisements into specific job categories. However, we do not know how well the job classification represents real-world jobs. Since most companies post their job advertisements on the Web, it is valuable to use them for examining the limitation of the current IS job classification. By developing a crawling system to search job advertisements on the Web, we collected 139,573 job advertisements for about one year from eleven countries. We attempted to classify them to the pre-defined job categories suggested by Niederman et al. [1] using an exact term matching approach. We found that only small portion of the job advertisements were classified into the defined categories. This result implies that the current job classification scheme is not a sufficient tool to analyse IS jobs and a improved classification approach is needed for further job analysis.