Advertisement

Drohende Studienabbrüche durch Frühwarnsysteme erkennen: Welche Informationen sind relevant?

  • Kerstin SchneiderEmail author
  • Johannes Berens
  • Julian Burghoff
Schwerpunkt
  • 5 Downloads

Zusammenfassung

Um abbruchgefährdete Studierende früh im Studienverlauf zu unterstützen, werden auch an deutschen Hochschulen verstärkt Frühwarnsysteme entwickelt und eingesetzt. Dabei unterscheiden sich die Systeme sowohl in den eingesetzten Verfahren als auch in den zugrundeliegenden Studierendendaten. In dem vorliegenden Beitrag dient das Frühwarnsystem FragSte als Benchmark. FragSte nutzt alle Studierendendaten nach §3 HStatG und verwendet Methoden des maschinellen Lernens. Es wird geprüft, ob vergleichbar genaue aber weniger datenintensive Frühwarnsysteme entwickelt werden können. Die Ergebnisse zeigen, dass insbesondere in den ersten beiden Semestern, die für einen erfolgreichen Studienverlauf von großer Bedeutung sind, die Datenanforderungen für ein Frühwarnsystem sehr hoch sind. Nach dem zweiten Studiensemester kann ein reines ECTS-Monitoring ausreichen.

Schlüsselwörter

Studienabbruch Frühwarnsystem Maschinelles Lernen Relevante Informationen 

Early detection of student dropout: what is relevant information?

Abstract

In order to support students who are at risk of dropping out of university, early detection systems are being increasingly developed and used. This also applies to Germany. The systems differ both in the methods used and in the underlying student data. Our benchmark is FragSte, an early detection system for German universities that uses all available administrative student data according to §3 HStatG and applies machine learning methods. In the paper we test, whether less data demanding systems can be developed without significant losses in forecasting quality. It turns out that the data requirements for an early detection system are very high. This is in particular true for the first two semesters, which are of great importance for a successful course of studies. After the second semester, however, an ECTS monitoring might be sufficient.

Keywords

Early detection system Machine learning Relevant information Student drop out 

Literatur

  1. Ahles, L., Köstler, U., Vetter, N., & Wulff, A. (2016). Studienabbrüche an deutschen Hochschulen. Stand der Thematisierung und strategische Ansatzpunkte. Baden-Baden: Nomos.CrossRefGoogle Scholar
  2. Arcidiacono, P., Aucejo, E., Maurel, A., & Ransom, T. (2016). College attrition and the dynamics of information revelation (NBER Working Papers - National Bureau of Economic Research).Google Scholar
  3. Arnold, K., & Pistilli,  M. (2012). Course signals at Purdue: using learning analytics to increase student success (Proceedings of the 2nd International Conference on Learning Analytics and Knowledge, S. 267–270).Google Scholar
  4. Arulampalam, W., Naylor, R. A., & Smith, J. P. (2005). Effects of in-class variation and student rank on the probability of withdrawal: cross-section and time-series analysis for UK university students. Economics of Education Review, 24, 251–262.CrossRefGoogle Scholar
  5. Asif, R., Merceron, A., Ali, S. A., & Haider, N. G. (2017). Analyzing undergraduate students’ performance using educational data mining. Computers & Education, 113, 177–194.CrossRefGoogle Scholar
  6. Baker, R. S. J. D., & Yacef, K. (2009). The state of educational data mining in 2009: A review and future visions. Journal of Educational Data Mining, 1(1), 3–17.Google Scholar
  7. Barefoot, B. O., Gardner, J. N., Cutright, M., Morris, L. V., Schroeder, C. C., Schwartz, S. W., Siegel, M. J., & Swing, R. L. (2005). Achieving and sustaining institutional excellence for the first year of college. San Francisco: Jossey-Bass.Google Scholar
  8. Berens, J., Schneider, K., Görtz, S., Oster, S., & Burghoff, J. (2019). Early detection of students at risk - predicting student dropouts using administrative student data from German universities and machine learning methods. Journal of Educational Data Mining, 11(3), im Erscheinen.Google Scholar
  9. Blüthmann, I., Lepa, S., & Thiel, F. (2008). Studienabbruch und -wechsel in den neuen Bachelorstudiengängen.CrossRefGoogle Scholar
  10. Bourn, J. (2007). Staying the course: The retention of students in higher education.Google Scholar
  11. Bowers, A. J., Sprott, R., & Taff, T. A. (2013). Do We Know Who Will Drop Out? A Review of the Predictors of Dropping out of High School: Precision, Sensitivity and Specificity. The High School Journal, 96, 77–100.CrossRefGoogle Scholar
  12. Bradley, A. P. (1997). The use of the area under the ROC curve in the evaluation of machine learning algorithms. Pattern recognition, 30(7), 1145–1159.CrossRefGoogle Scholar
  13. Brahm, T., Jenert, T., & Wagner, D. (2016). The crucial first year: a longitudinal study of students’ motivational development at a Swiss Business School. Higher Education, 73(3), 459–478.CrossRefGoogle Scholar
  14. Brandstätter, H., Grillich, L., & Farthofer, A. (2006). Prognose des Studienabbruchs. Zeitschrift für Entwicklungspsychologie und Pädagogische Psychologie, 38(3), 121–131.CrossRefGoogle Scholar
  15. Dekker, G. W., Pchenenizkiy, M., & Vleeshouwers, J. M. (2009). Predicting students drop out: a case Staudy (Proceedings of the 2nd International Conference on Educational Data Mining (EDM´09), S. 41–50).Google Scholar
  16. Frank, E., Hall, M. A., & Witten, I. H. (2016). The WEKA workbench. Online appendix for “data mining: practical machine learning tools and techniques”. Burlington: Morgan Kaufmann.Google Scholar
  17. Gaebel, M., Hauschildt, K., Mühleck, K., & Smidt, H. (2012). Tracking learners’ and graduates’ progression paths. TRACKIT. Brüssel, Geneva: EUA Publications.Google Scholar
  18. Gold, A., & Souvignier, E. (2005). Prognose der Studierfähigkeit - Ergebnisse aus Längsschnittanalysen. Zeitschrift für Entwicklungspsychologie und Pädagogische Psychologie, 37, 214–222.CrossRefGoogle Scholar
  19. Hall, M. (1998). Correlation-based feature selection for machine learning. Google Scholar
  20. Heublein, U., Ebert, J., Hutzsch, C., Isleib, S., König, R., Richter, J., & Woisch, A. (2017). Zwischen Studienerwartung und Studienwirklichkeit. Ursachen des Studienabbruchs, beruflicher Verbleib der Studienabbrecherinnen und Studienabbrecher und Entwicklung der Studienabbruchquote an deutschen Hochschulen. (Forum Hochschule, 1|2017). Hannover: DZHW.Google Scholar
  21. Jetter, T. (2017). Membrain NN. https://www.membrain-nn.de/. Zugegriffen: 3. Apr. 2019.Google Scholar
  22. Kemper, L., Vorhoff, G., & Wigger, B. U. (2018). Predicting student dropout: a machine learning approach. Google Scholar
  23. Kotsiantis, S. B., Pierrakeas, C. J., & Pintelas, P. E. (2003). Preventing student dropout in distance learning using machine learning techniques. In V. Palade, R. J. Howlett & L. Jain (Hrsg.), Knowledge-based intelligent information and engineering systems. KES 2003 (Lecture notes in computer science Bd. 2774). Berlin: Springer.Google Scholar
  24. Larsen, M. L., Kornbeck, K. P., Kristensen, R. M., Larsen, M. R., & Sommersel, H. B. (2013). Dropout phenomena at universities: what is dropout? Why does dropout occur? What can be done by the universities to prevent or reduce it? A systematic review. Danish Clearinghouse for Educational Research.Google Scholar
  25. Quinlan, J. R. (1986). Induction of decision trees. Machine Learning, 1, 81–106.Google Scholar
  26. Schapire, E., & Freund, Y. (1997). A decision-theoretic generalization of on-line learning and an application to boosting. Journal of Computer and System Science, 55, 119–139.CrossRefGoogle Scholar
  27. Schapire, R. E., & Freund, Y. (2012). Boosting - foundations and algorithms. Google Scholar
  28. Seidmann, A. (1996). Spring retention revisited: RET = E ID + (E + I + C). IV. College and University, 71, 18–20.Google Scholar
  29. Stinebrickner, T., & Stinebrickner, R. (2008). The effect of credit constraints on the college drop-out decision: a direct approach using a new panel study. American Economic Review, 98, 2163–2184.CrossRefGoogle Scholar
  30. Stinebrickner, T., & Stinebrickner, R. (2014). Academic performance and college dropout: using longitudinal expectations data to estimate a learning model. Journal of Labor Economics, 32, 601–644.CrossRefGoogle Scholar
  31. Swets, J. A. (1988). Measuring the accuracy of diagnostic systems (S. 1285–1293). Washington: American Association for the Advancement of Science.Google Scholar
  32. Trapmann, S., Hell, B., Weigand, S., & Schuler, H. (2007). Die Validität von Schulnoten zur Vorhersage des Studienerfolgs− eine Metaanalyse. Zeitschrift für pädagogische Psychologie, 21, 11–27.CrossRefGoogle Scholar
  33. Wilson, K., Murphy, K., Pearson, A., Wallace, B., Reher, V., Buys, N., & Buys, N. (2016). Understanding the early transition needs of diverse commencing university students in a health faculty: informing effective intervention practices. Google Scholar
  34. Witten, I. H., Eibe, F., & Hall, M. A. (2016). Data mining: practical machine learning tools and techniques. Burlington: Morgan Kaufmann.Google Scholar
  35. Xenos, M. (2004). Prediction and assessment of student behaviour in open and distance education in computers using Bayesian networks. Computers & Education Journal, 43, 345–359.CrossRefGoogle Scholar
  36. Zafar, M. B., Valera, I., Rdriguez, M. G., & Gummadi, K. P. (2017). Fairness beyond disparate treatment & disparate impact: learning classification without disparate mistreatment (WWW ’17 Proceedings of the 26th International Conference on World Wide Web, S. 1171–1180).Google Scholar
  37. Zhang, Y., Qiang, F., Quddus, M., & Davis, C. (2014). An examination of the impact of early intervention on learning outcomes of at-risk students. Google Scholar
  38. Zweig, M. H., & Campbell, G. (1993). Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. Clinical Chemistry, 39, 561–577.Google Scholar

Copyright information

© The Editors of the Journal 2019

Authors and Affiliations

  1. 1.Fakultät für WirtschaftswissenschaftSchumpeter School of Business and EconomicsWuppertalDeutschland

Personalised recommendations