Maryland Population Research Center

Assistant Professor
Joint Program/Survey Methodology
University of Maryland
College Park, MD 20742

Email: fkreuter@umd.edu
Phone: (301) 405-0935
Office: 1218Q LeFrak Hall

CV

Publications

Grants

Recent Scientific Accomplishments

Kreuter’s research focuses on sampling and measurement errors in complex surveys. She worked in particular on survey nonresponse and interviewer effects, as well as on measurement errors due to variation in question format or data collection modes. In a 2005 paper on design and interviewer effects published in the Journal of Official Statistics, Kreuter and co-author R. Schnell showed the strong homogenizing effect of interviewers on answers to in-person interviews. For many survey items, this effect is stronger than the homogenizing effect of geographical clustering, a finding that is important for practical implementations of adjustments to account for the effects of survey design features on estimated variances. Her dissertation, published in 2002 under the auspices of the German Science Foundation, focused on the measurement of the fear-of-crime. Her results, based on qualitative and quantitative interviews, as well as experimental studies, demonstrated the increased and biased measurement error attached to currently used fear-of-crime indicators compared to alternative crime and location specific measures. During a research fellowship in 2004 at the Max-Planck-Institute for Human Development, Kreuter collaborated with the German representatives for the OECD Programme for International Student Assessment to investigate the effects of measurement errors in children’s proxy-reports of various indicators for parents?social economic status. A 2006 publication with K. Maaz and R. Watermann showed the higher reliability of proxy-reports about parent’s education compared to reports on parents?occupation and vocational training history. At the 2005 Annual Conference of the American Association for Public Opinion Research, Kreuter presented the effects of these measurement errors on the estimation of students?achievement in Germany. She demonstrated that subgroup specific error rates lead to an underestimation of the effect of parent’s social economic status on student achievement in the 9-year schooling track compared to the 13-year track. Her current research with R. Tourangeau and S. Presser is focused on measurement errors attached to data collection modes. In their presentation at the Joint Statistical Meeting in 2006, they showed the interaction effects of perceived item sensitivity and mode on reporting errors.

Funded Research

In 2006, The National Science Foundation (NSF Awards 0549916 and 0550002) funded the project entitled, "Collaborative Research on Latent Class Models of Measurement Error" under the direction of Frauke Kreuter, in collaboration with Roger Tourangeau (University of Michigan & Joint Program in Survey Methodology). Together with a postdoctoral student at the University of Michigan (Ting Yan) they will conduct a series of studies to investigate the use of latent class analysis (LCA) in questionnaire development. Kreuter also has received support from the Charles Cannell Fund for her research on interviewer effects. With her backing, several Ph.D. students with whom Kreuter has been working (jointly with T.E. Raghunathan and R. Groves) have received grants from the Rensis Likert Fund to support their participation in research on nonresponse weighting adjustments. Kreuter also was instrumental in securing support for a JPSM PhD student working on a nationwide nonresponse study related to media use.

Future Research Plans

Over the next few years, Kreuter will continue to investigate interviewer effects, in particular the effect of interviewers on survey participation, as well as the effect of interviewer behavior and vocal characteristics such as rate of speech on the cognitive processes involved in answering survey questions. Being a member of the International Workshop of Household Survey Nonresponse she will lead the workshops?discussion on paradata. Kreuter is currently investigating the potential value of observational data, such as interviewer observations regarding the neighborhood in which a sample unit is located, for nonresponse adjustments to reduce nonresponse bias. She has a pending R03 on these issues.

Maryland Population Research Center
0124N Cole Student Activities Building (#162)
College Park, MD 20742
Phone: 301-405-6403
Fax: 301-405-5743