A machine-learning based model to identify PhD-level skills in job ads

Abstract Around 60% of doctoral graduates worldwide ended up working in industry rather than academia. There have been calls to more closely align the PhD curriculum with the needs of industry, but an evidence base is lacking to inform these changes. We need to find better ways to understand what industry employers really want from doctoral graduates. One good source of data is job advertisements where employers provide a ‘wish list’ of skills and expertise. In this paper, a machine learning-natural language processing (ML-NLP) based approach was used to explore and extract skill requirements from research intensive job advertisements, suitable for PhD graduates. The model developed for detecting skill requirements in job ads was driven by SVM. Our preliminary results showed that ML-NLP approach had the potential to replicate manual efforts in understanding job requirements of PhD graduates. Our model offers a new perspective to look at PhD-level job skill requirements. 1 Introduction Abundant evidence shows that industry employers are often dissatisfied with key aspects of PhD graduates’ workplace performance (e.g., Cumming, 2010; G08, 2013; Australian Department of Education, 2014; Hancock, 2019), particularly in relation to professional skills like communications (McCarthy & Wient, 2019). PhD graduates themselves also indicate the education they received during their candidature did not address job market needs outside of academia (Golde & Dore, 2001). Reports indicate that employers were often dissatisfied with doctoral employees’ demonstration of soft skills at work (e.g., Cumming, 2010; Cyranoski et al., 2011). Such frustration from stakeholders leads us to question the fitness for purpose of doctoral degrees. PhD was originally designed to help people into academic careers, but its fitness for purpose has been questioned for over 80 years (Dale, 1930). As the number of PhD graduates increase tremendously (Auriol et al., 2013; Gould, 2015) and the academic job market remains relatively static in scale (Larson et al., 2014), many PhD graduates will be unable to secure academic positions. Despite the fact that most PhD graduates will be working outside academia, universities are still training their candidates based on research competencies desired in academia. One such example of a popular research training framework is the Vitae Researcher Development Framework (RDF) based on research by Bray and Boon (2011). To enhance PhD employability, Mewburn et al. (2018) argue it is not enough to understand only academic workforce requirements: non-academic professions may have different needs. Some initiatives have been undertaken to understand the so called ‘transferrable skills’ needed from PhD job seekers. Consequently, many add-on courses based on the long list of skill terms have been put in place at universities (Barnacle & Mewburn, 2010). However, scholars such as Neumann and Tan (2011) and Platow (2012) have expressed concerns about the generic quality of these initiatives. Take professional skills such as teamwork and empathy: these may mean very different things in different workplaces and industry domains. Concerns are reasonable about the ambiguity of previous initiatives suspicious of over- generalisation of skills and neglect of the context in which skills are deployed. However, …

Content truncated to 500 words