Dr. Kristian Kleinke

  • Dr. Kristian Kleinke ist jetzt Mitarbeiter an der Universität Siegen und unter kristian.kleinke@uni-siegen.de zu erreichen.
  • Diese Seite wird nicht weiter aktualisiert.

Academic Positions

  • 2014: Staff Scientist, University of Hagen, Institute of Psychology
  • 2009–2014: Scientist, DFG project “Robust and Efficient Multiple Imputation of Complex Data Sets”, University of Bielefeld, Faculty of Sociology (Projectwebsite)
  • 2008–2011: Lecturer & Research Assistant (Psychological Methods and Evaluation), University of Bielefeld, Department of Psychology
  • 2008–2009: Teaching Assignment (Personality Psychology), University of Applied Sciences Bielefeld

Education and Qualifications

  • 2014, University of Bielefeld, Germany: Doctor rerum naturalium. Dissertation: “Efficient Multiple Imputation of Complex Data Structures (Like Multilevel Data, Zero‐Inflated Count Data, and Multilevel Count Data)”
  • 2008, University of Erlangen‐Nürnberg, Germany: Diplom in Psychology (Dipl.‐Psych. (Univ.))
  • 2008, University of Waikato, New Zealand: Bachelor of Social Sciences with Honours

My research interests

Missing data and multiple imputation

My primary research interests are missing data and multiple imputation. My focus are multiple imputation solutions for complex data structures like panel data and “non-normal” missing data problems, i.e. when convenient distributional assumptions of the standard MI procedures are violated.
Missing data are a nuisance: Unfortunately, it is the rare exception rather than the rule that empirical researchers can analyse complete data sets.
Missing data are problematic (i.e. they could lead to biased parameter estimates and standard errors) when (a) they are numerous, (b) not missing completely at random (MCAR) in the sense of Rubin (1976) and (c) a suboptimal missing data procedure is used.
If missing data percentages are not trivial and the MCAR assumption is not likely to hold, statistical inferences should be based on state-of-the-art missing data procedures, like for example multiple imputation.
Currently there are two popular MI approaches, joint modelling like for example Schafer's (1997) normal model MI approach and conditional modelling, which is also known under the name "multiple imputation by chained equations". In joint modelling, a joint model is specified for all the variables to be imputed, whereas in conditional modelling, a separate imputation model is specified for all the incomplete variables in the data set and the missing data problem is handled on a variable to variable basis. This approach is quite flexible, as data of different types can be imputed under various distributional assumptions.

In my previous research I focussed on missing data (conditional modelling) solutions for count data (ordinary count data, overdispersed count data, zero-inflated count data and multilevel count data). We have published an R-package called "countimp", which is available from www.uni-bielefeld.de/soz/kds/software.htmlOpens in a new window.
Currently, I focus on (robust) MI solutions for situations, when empirical data do not exactly follow the convenient parametric assumptions of standard statistical models.

My workshops and short courses

I regularly give workshops about missing data, multiple imputation, longitudinal data analysis / structural equation modelling and the statistical software R.

Recent workshops:

  • "Missing data / Multiple Imputation", two-day workshop, University of Hamburg (with Martin Spiess)
  • "Introduction to Multiple Imputation", one-day workshop, Justus Liebig University Giessen
  • "R basics", two-day workshop, University of Hagen
  • "Multiple Imputation", one-day workshop, TU Dortmund
  • "Linear Causal Modelling with Missing Data", three-day seminar, University of Bielefeld & TU Dortmund (with Jost Reinecke)

Missing data software

The R package "countimp" - Multiple imputation of incomplete count data (Kleinke & Reinecke, 2013) can be obtained from:


Publications (Selection)

Scientific Journals (peer review)

Mehr Informationen

Kleinke, K. (in press). Multiple imputation under violated distributional assumptions – A systematic evaluation of the assumed robustness of predictive mean matching. Journal of Educational and Behavioral Statistics.

Kleinke, K., & Reinecke, J. (2013a). Multiple imputation of incomplete zero-inflated count data. Statistica Neerlandica, 67(3), 311–336. doi: 10.1111/stan.12009
http://onlinelibrary.wiley.com/doi/10.1111/stan.12009/abstract

Kleinke, K., Stemmler, M., Reinecke, J., & Lösel, F. (2011). Efficient ways to impute incomplete panel data. Advances in Statistical Analysis, 95(4), 351–373. doi: 10.1007/s10182-011-0179-9
http://link.springer.com/article/10.1007/s10182-011-0179-9

Book Chapters

Mehr Informationen

Kleinke, K., & Reinecke, J. (2015a). Multiple imputation of multilevel count data. In U. Engel, B. Jann, P. Lynn, A. Scherpenzeel, and P. Sturgis (Eds.), Improving Survey Methods: Lessons from Recent Research (pp. 381–396). New York: Routledge, Taylor & Francis Group.
http://www.psypress.com/books/details/9780415817622/Opens in a new window

Kleinke, K., & Reinecke, J. (2015b). Multiple imputation of overdispersed multilevel count data. In: Uwe Engel (Ed.), Survey Measurements. Techniques, Data Quality and Sources of Error (pp. 209–226). Frankfurt a. M.: Campus/The University of Chicago Press.
http://press.uchicago.edu/ucp/books/book/distributed/S/bo22196267.html

Technical Reports

Mehr Informationen

Kleinke, K., de Jong, R., Spiess, M., & Reinecke, J. (2011). Multiple imputation of incomplete ordinary and overdispersed count data (Technical Report). Bielefeld: University of Bielefeld, Faculty of Sociology. doi: 10.13140/RG.2.1.2185.3927 .

Kleinke, K., & Reinecke, J. (2013b). countimp 1.0 - A multiple imputation package for incomplete count data (Technical Report). Bielefeld: University of Bielefeld, Faculty of Sociology. doi: 10.13140/RG.2.1.3889.3286

Presentations and Posters (Selection)

Mehr Informationen

Kleinke, K. & Reinecke, J. (2016, September). Multiple imputation of incomplete count data - The countimp package in R. Paper presented at the 16th annual conference of the European Society of Criminology (ESC), Münster, Germany.

Kleinke, K. (2016, September). Multiple Imputation of Multilevel Data by "Two-Level Predictive Mean Matching". Paper presented at the 50th Congress of the German Psychological Society (DGPs), Leipzig, Germany.

Kleinke, K. (2016, July). Multiple imputation by predictive mean matching when sample sizes are small. Paper presented at the 7th European Congress of Methodology, Palma de Mallorca, Spain.

Kleinke, K. (2015, December). Multiple imputation by predictive mean matching – Advantages, pitfalls, and new developments. Paper presented at the Colloquium "Quantitative methods and statistics", University of Bielefeld, Bielefeld.

Kleinke, K. (2015, October). Multiple imputation by predictive mean matching when sample sizes are small – some caveats. Paper presented at the Colloquium of the Research Center for the Psychological Study of Individual and Community Change, University of Hagen, Hagen.

Kleinke, K. (2015, June). Predictive Mean Matching – The missing data swiss army knife. Paper presented at the Colloquium of the Research Center for the Psychological Study of Individual and Community Change, University of Hagen, Hagen.

Kleinke, K. (2014, October). Multiple Imputation of Incomplete Panel Data. Paper presented at the Colloquium of the Research Center for the Psychological Study of Individual and Community Change, University of Hagen, Hagen.

Kleinke, K. & Reinecke, J. (2013, September). countimp 1.0 - A multiple imputation package for incomplete count data. Poster presented at the Conference "Survey methods in future research.", University of Bremen, Bremen.

Kleinke, K. & Reinecke, J. (2012, July). Multiple imputation of multilevel count data. Paper presented at the RC33′s 8th International Conference on Social Science Methodology, Sydney, Australia.

Kleinke, K., & Reinecke, J. (2011, July). Multiple imputation of zero-inflated count data. Paper presented at the 4th Conference of the European Survey Research Association (ESRA), Lausanne.

Kleinke, K., & Reinecke, J. (2011, May). Multiple imputation of incomplete count data. Paper presented at the American Sociological Association (ASA) Spring Methodology Conference, Tilburg.

Kleinke, K. (2010, September). Multiple imputation of incomplete count data. Paper presented at the Methodology Center Brown Bag Seminar Series, Penn State University, Pennsylvania.

Kleinke, K., & Stemmler, M. (2009, September). Can we have simplicity and quality? Is it safe to use normal model multiple imputation for non-normal longitudinal data? Paper presented at the 9th METHEVAL conference (Fachgruppentagung Methoden und Evaluation der DGPs), Bielefeld.

Kleinke, K., Stemmler, M. & Lösel, F. (2006, September). Dropout Analyse und Missing Data Schätzungen bei einem längsschnittlichen Datensatz. Posterbeitrag auf dem 45. Kongress der Deutschen Gesellschaft für Psychologie, Nürnberg.

Psychologische Methodenlehre | 12.08.2021