A supervised machine learning approach to trace doctorate recipients’ employment trajectories

Abstract

Only scarce information is available on doctorate recipients’ career outcomes ( BuWiN, 2013 ). With the current information base, graduate students cannot make an informed decision on whether to start a doctorate or not ( Benderly, 2018 ; Blank et al., 2017 ). However, administrative labor market data, which could provide the necessary information, are incomplete in this respect. In this paper, we describe the record linkage of two data sets to close this information gap: data on doctorate recipients collected in the catalog of the German National Library (DNB), and the German labor market biographies (IEB) from the German Institute of Employment Research. We use a machine learning-based methodology, which (a) improves the record linkage of data sets without unique identifiers, and (b) evaluates the quality of the record linkage. The machine learning algorithms are trained on a synthetic training and evaluation data set. In an exemplary analysis, we compare the evolution of the employment status of female and male doctorate recipients in Germany.

Citation analysis of Ph.D. theses with data from Scopus and Google Books

“…The DNB has a mandate to collect all publications originating from Germany, including Ph.D. theses. This source of dissertation data has been found useful for science studies research previously (Heinisch and Buenstorf 2018;Heinisch et al 2020). We downloaded records for all Ph.D. dissertations from the German National Library online catalog in April 2019 using a search restriction in the university publications field of "diss*", as recommended by the catalog usage instructions, and publication year range 1996-2018.…”

Are Employment Trajectories of STEM Doctoral Degree Holders Gender-Specific? Evidence from a Large German Technical University

“…We use rich data for TU Berlin to explore potential gender-specific patterns in the career trajectories of more than 1,800 STEM DDHs covering a 10-year period starting five years before doctorate completion and running up to five years afterwards. Our dataset was built by refining record linkage techniques developed by Heinisch et al (2020) while linking administrative information provided by TU Berlin with the Integrated Employment Biographies (IEB) dataset of the Institute for Employment Research (IAB). It covers more than 80 % of the respective TU Berlin graduation cohorts for whom we have detailed data on employment sectors -university, non-university research or other sectors-and also employment volume for the whole period.…”

Postdoctoral employment and future non-academic career prospects

“…In total, approximately 80% of the workforce in Germany is covered, whereby self-employed persons and civil servants are not included. Both datasets were linked using machine learning methods [ 48 ].…”