Measurements in the social and behavioral sciences are often discrete (e.g., highest degree earned, response option selected on a survey or test, career choice). My research lies at the intersection of statistical models for multivariate discrete data and psychometrics. My current focus is on models with latent variable interpretations, including item response theory models, discrete choice models, and their formulations as generalized linear and non-linear (mixed) models (i.e., HLM, and GLMMS).

Key Professional Appointments

Professor Educational Psychology, Psycholgy & Statistics, Univeristy of Illinois, 2008 - present

Professor Educational Psychology, Psycholgy & Statistics, Univeristy of Illinois, 2008 - present

Associate Professor Departments of Educational Psychology, Psychology, and Statistics, University of Illinois at Urbana-Champaign, 2000 - 2008

Assistant Professor Educational Psychology and Psychology, University of Illinois at Urbana-Champaign, 1993 - 1995


Ph.D., Quantitative Psychology, University of Illinois at Urbana-Champaign, 1993

M.S., Statistics, University of Illinois at Urbana-Champaign, 1986

B.A., Psychology and Economics, University of California at Berkeley, 1982

Awards, Honors, Associations

Incomplete List of Teachers Ranked as Excellent by Their Students Spring and Fall Semesters, University of Illinois at Urbana-Champaign, 2001 - 2001

Distinguished Senior Scholar College of Education, 2009 - 2009

Faculty Fellow Bureau of Educational Research, 2005 - 2006

Faculty Fellow National Center for Supercomputing Applications, 2005 - 2006

Dissertation Award Division 5, American Psychological Association, 1995 - 1995

Dissertation Award Psychometric Society, 1993 - 1993

Research & Service

In recent work, I have shown that two mathematically distinct models yield nearly the same results. One model is the standard formulation of latent variables models and the other is an alternative formulation. The implication of this work is that models where responses to items reflect a latent variables behave the same as model built on the notion that items define the latent variable. A further implication is that estimation of measurement models can be done with much less complicated routines and these alternative appear to be more stable. I have made major progress in model estimation (research board grant) and extended it to more complex latent structure. I have also extended this work to more complex latent structures

I have also collaborate on research. I am currently and actively collaborating with a diverse team (Edps, ECS, CS and CITL) that studying online learning. In particular, whether an online context has affodances for underrepresented groups in STEM. We have submitted 2 grants, received 1 internal grant, and submitted 2 papers in the 6 months that we have worked together.


Anderson, C., Kim, J., & Keller, B. (2013) Multilevel Modeling of Discrete Response Data. A Handbook of International Large-Scale Assessment: Background, Technical Issues and Methods of Data Analysis  link >

Kim, J., Anderson, C., & Keller, B. (2013) Multilevel analysis of assessment data. Handbook of International Large-Scale Assessment  link >

Anderson, C. (2013) Multidimensional item response theory models with collateral information as Poisson regression models Journal of Classification 30 (2), 276-303  link >

Allen, N., Todd, N., Anderson, C., Davis, S., & Javdani, S. (2013) Council-based approaches to intimate partner violence:  Evidence of distal change in the system American Journal of Community Psychology 52, 1-12

Anderson, C., Verkuilen, J., & Peyton, B. (2010) Simultaneously estimated multinomial logistic regressions for discrete response data Journal of Educational and Behavioral Statistics 35, 422-452  link >

Rutkowski, L., Vasterling, J., Procter, S., & Anderson, C. (2010) Posttraumatic stress disorder and standardized test-taking ability. Educational Researcher 102, 223-233,.

Anderson, C. (2009) Categorical data analysis with a psychometric twist. Handbook of quantitative methods in psychology Sage: Thousand Oaks, CA

Anderson, C., & Yu, H. (2007) Log-multiplicative association models as item response models Psychometrika 72, 5-23  link >

Anderson, C., Li, Z., & Vermunt, J. (2007) Estimation of models in a Rasch family for polytmous items and multiple latent variables Journal of Statistical Software 20 (6), 1-37  link >

Rooij, d., M., A., C. J., ., & Anderson, C. (2007) Visualizing, summarizing, and comparing odds ratio structures European Journal of Methodology 3, 139-148

Tettegah, S., & Anderson, C. (2007) Pre-service teacher's empathy and cognitions: Statistical analysis of text data by graphical models Contemporary Educational Psychology 32 (1), 48-82

Kroonenberg, P., & Anderson, C. (2006) Additive and multiplicative models for three-way contingency tables: Darroch (1974) revisited. Multiple correspondences analysis and related methods Chapman & Hall/CRC: London

Vermunt, J., & Anderson, C. (2005) Joint correspondence analysis (JCA) by maximum likelihood Methodology 1, 18-26

Anderson, C. (2002) Analysis of multivariate frequency data by graphical models and generalizationsof the multidimensional row-column association model Psychological Methods 7, 446-467

Anderson, C., & Bockenholt, U. (2000) Graphical regression models for polytomous variables Psychometrika 65, 497-509  link >

Anderson, C., & Vermunt, J. (2000) Log-multiplicative association models as latent variable models for nominal and/or ordinal data Sociological Methodology 30, 81-121

Anderson, C., Wasserman, S., & Crouch, B. (1999) A p* primer:logit models for social networks Social Networks 21, 37-66


Co-Principal Investigator Using Technology and Vignette Technique in Educational Research: From Qualitative Text to Statistical Modeling, Bureau of Educational Research, 2005 - 2006

Co-Principal Investigator From Narratives, Multimedia, and Empathy to Statistical Modeling, Campus Research Board, 2005 - 2006

Principal Investigator Multivariate Multinomial Logistic Regressions Models as Item Response Theory Models with Covariates, National Science Foundation, 2004 - 2006

Co-Principal Investigator Visualization of Vignette and Statistical Models: An Integrated Approach, National Center for Supercomputing Applications, 2005 - 2005


Editorial board member Psychological Methods, 2004 - 2014

Member of editoral board of Psychological Methods Psychological Methods, 2003 - 2014

Chair of Membership committee American Psychological Association, Division 5 (Evaluation, Measurement & Statistics), 2009 - 2012

Member Board of Trustees Psychometric Society, 2005 - 2009

Book Review Editor Psychometrika, 2005 - 2008

Associate Editor of Psychometrika Psychometrika, 2005 - present


I teacher statistics courses to students in the social and behavioral sciences, ACES and business. I regularly teach categorical data analysis, multivariate analysis, and multilevel modeling. I am currently developing a graduate seminar on Bayesian Statistics.

I encourage students to develop their own research ideas and interests. One way is by allowing students to do project in lieu of taking an exam. I mentor students both within my own department and from other departments.

Students in categorical data analysis and multilevel models have the option of doing a project, which are often published or parts of larger research projects. For example, a project spanned both courses, developed into a dissertation, and the student won the Seymour Sudman dissertation award.

I met at least weekly with one student working on her dissertation and another who is developing his proposal for his early research project.


Hierarchical Linear Models This course provides an overview of the use of multilevel models. Students will learn the techniques and theory of hierarchical linear models and apply the methods to data from studies in education, psychology and social sciences. Topics covered include multilevel analyses, random intercept and slope models, 2- and 3-level models, hypothesis testing, model assessment, longitudinal (repeated measures) data, and generalized hierarchical models for categorical variables. Course Information: Same as PSYC 587 and STAT 587. Approved for letter and S/U grading. Prerequisite: EPSY 581 and EPSY 582, or PSYC 406 and PSYC 407.

Categorical Data in Ed/Psyc Concepts and methods for analyzing categorical data with an emphasis placed on building and applying models in education, sociology and psychology. Generalized linear models covered including logistic and Poisson regression models, loglinear, logit, and probit models, and models for ordinal data. Course Information: Same as PSYC 589 and SOC 579. Approved for letter and S/U grading. Credit is not given for EPSY 589 and STAT 426. Prerequisite: EPSY 581 or PSYC 507.

Carolyn Anderson

Professor, Educational Psychology



236C Education Building
1310 S. Sixth St.
Champaign, IL 61820

Edit Refresh

Refreshing data...please wait