Matthias Schonlau

Professor

Matthias SchonlauContact Information:
Matthias Schonlau

Matthias Schonlau's personal website

Research interests

Professor Schonlau's research interests include applied survey sampling and survey methodology, machine learning from text data such as open-ended questions as well as software implementation of his research.

Web surveys are often conducted with a non-probability sample. One approach to adjust for selection bias is to ask so-called webographic or lifestyle questions on both a web survey and a smaller RDD phone survey and to employ propensity scoring to adjust for selection. Professor Schonlau found that such adjustments usually attenuate but do not necessarily eliminate bias.

Professor Schonlau spent a sabbatical with the SOEP department at DIW in Berlin, Germany. SOEP is the largest and oldest German longitudinal household panel. During this time he investigated selectivity when collecting genetic samples as part of a survey, compared two approaches to cross-sectional weights in longitudinal household surveys, and investigated the effect of so-called following rules in household panels on sample size.

Respondent Driven Sampling (RDS) is a referral sampling method. After reaching a sampling equilibrium, population estimates can be derived based on bias-adjusted sampling proportions. RDS is only useful when more conventional methods are not practical (no sampling frame is available, screening for a rare population is too expensive, confidentiality concerns).

As such, this method is starting to become the standard method for sampling people with HIV and is gaining popularity for sampling in underdeveloped countries where sampling frames are often lacking. Professor Schonlau's implementation of RDS in Stata/C++ has generated interest from five continents. Follow-up work is investigating RDS in the context of web surveys.

Professor Schonlau actively collaborates with the International Control Project (ITC) at Waterloo which has survey panels in more than 20 countries covering more than 50% of the world's population.

Education/biography

Professor Schonlau joined the faculty in 2011. From 1999-2011 he was a statistician at RAND corporation and head of the RAND Statistical Consulting Service. He was initially located at RAND's Santa Monica (Los Angeles) headquarters and then moved to RAND's Pittsburgh office. Professor Schonlau spent the academic year 2009/2010 on sabbatical at the German Institute for economic analysis (DIW) in Berlin, Germany. The sabbatical was made possible in cooperation with the Max Planck Institute for Human Development (MPIB). From 1997-1999 Professor Schonlau held a joint appointment with the National Institute of Statistical Sciences and AT&T Labs - Research. He obtained his PhD from the University of Waterloo in 1997 and his master's from Queen's university in 1993. Professor Schonlau grew up in Germany.

Selected publications

  • Schonlau M, Liebau E. Respondent-driven sampling, The Stata Journal, 2012; 12(1): 72-93.
  • Schonlau M, Watson N, Kroh M. Household survey panels: how much do following rules affect sample size? Survey Research Methods. 2011; 5(2):53-61.
  • Schonlau M, Martin LT, Haas A, Derose K, Rudd R.  Patients' Literacy Skills: More than just reading ability? Journal of Health Communication, 2011; 16(10): 1046-1054.  DOI: 10.1080/10810730.2011.571345.
  • Schonlau M, Reuter M, Schupp J, Montag C, Weber B, Dohmen T, Siegel, NA, Sunde U,  Wagner GG., Falk A. Collecting Genetic Samples in Population Wide (Panel) Surveys: Feasibility, Nonresponse and Selectivity. Survey Research Methods, 2010; 4(2): 121-126.
  • Schonlau M., Van Soest A, Kapteyn A, Are `Webographic' or attitudinal questions useful for adjusting estimates from Web surveys using propensity scoring?   Survey Research Methods, 2007; 1(3): 155-163.
  • Schonlau M., Welch WJ. Screening the Input Variables to a Computer Code Via Analysis of Variance and Visualization. In Screening: Methods for Experimentation in Industry, Drug Discovery and Genetics. Eds: A. M. Dean and S. M. Lewis, Springer Verlag, New York; 2006:308-327.
  • Schonlau M. Boosted Regression (Boosting): An introductory tutorial and a Stata plugin. The Stata Journal, 2005; 5(3):330-354.
  • Schonlau M.  Visualizing Hierarchical and Non-Hierarchical Cluster Analyses with Clustergrams. Computational Statistics: 2004; 19(1):95-111.
  • Schonlau M, Zapert K, Payne Simon L, Sanstad K,  Marcus S, Adams J,  Spranca M, Kan H-J,  Turner R, Berry S. A comparison between a propensity weighted web survey and an identical RDD survey.  Social Science Computer Review.2004; 22(1):128-138.
  • Schonlau M, DuMouchel W, Ju W, Karr A, Theus M, Vardi Y. Computer intrusion: Detecting masquerades. Statistical Science. 2001;16(1):58-74.
Affiliation: 
University of Waterloo
Contact information: