Direkt zum InhaltDirekt zur SucheDirekt zur Navigation
▼ Zielgruppen ▼

Humboldt-Universität zu Berlin - Lebenswissen­schaftliche Fakultät - Institut für Psychologie

Abschlussarbeiten

Studierende mit einem ausgeprägten Interesse an aktuellen Themen der Psychologischen Methodenlehre, sind herzlich eingeladen, sich jederzeit mit Themenvorschlägen an MitarbeiterInnen des Lehrstuhls Methodenlehre zu wenden. Ebenfalls (ko-)betreuen wir regelmäßig Bachelor- und Masterarbeiten zu inhaltlichen Themen im Bereich entwicklungspsychologischer Dynamiken in affektiven und kognitiven Funktionen, sofern die Arbeit einen quantitativ-methodischen Schwerpunkt aufweist. Kontaktieren Sie uns einfach, und wir besprechen gemeinsam die Wahl eines geeigneten Themas. Ausgewählte Vorschläge zu aktuellen Abschlussarbeiten und Ansprechpartnern finden Sie untenstehend.

 


 

Masterarbeit: Unobserved heterogeneity in IRT models

Martin Hecht

Schlagwörter: IRT, heterogeneity

Unobserved heterogeneity has been shown to impact interpretability of logistic regression results (e.g., Mood, 2010). The goal of this thesis is to explore and illustrate the effects of unobserved heterogeneity in IRT models.

Mood, C. (2010). Logistic Regression: Why we cannot do what we think we can do, and what we can do about it. European Sociological Review, 26, 67–82. doi:10.1093/esr/jcp006

 

 

Masterarbeit: Position effects and person characteristics

Martin Hecht

Schlagwörter: position effects, large-scale assessment, multi matrix designs

Position effects are commonly observed methodological phenomena that stem from using multi matrix designs in large-scale assessments (e.g., Hecht et al., 2015; Weirich et al., 2014). The goal of this thesis is to explore the impact of certain person characteristics on position effects.

Hecht, M., Weirich, S., Siegle, T., & Frey, A. (2015). Effects of design properties on parameter estimation in large-scale assessments. Educational and Psychological Measurement75, 1021–1044. doi:10.1177/0013164415573311

Weirich, S., Hecht, M., & Böhme, K. (2014). Modeling item position effects using generalized linear mixed models. Applied Psychological Measurement, 38, 535–548. doi:10.1177/0146621614534955

 

 

Masterarbeit: Test designs and IRT-ML estimation

Martin Hecht

Schlagwörter: large-scale assessment, multi matrix designs, IRT, ML estimation, JML, CML, MML

The goal of this thesis is to explore whether certain maximum likelihood estimation procedures (joint, conditional, marginal ML) are more suitable for certain large-scale assessment test designs.

 

 

Masterarbeit: Using response times for optimal test assembly

Martin Hecht

Schlagwörter: large-scale assessment, test assembly, response time

The target testing time of a large-scale test is usually defined by a relative criterion (e.g., a certain percentage of students is supposed to completely finish the test within the time limit). However, when assembling large-scale tests from an item pool response time information is available at the item level. The goal of this thesis is to explore how target testing time can be met by combining item-level response times. 

 

 

Masterarbeit: Booklet effects in large-scale assessments

Martin Hecht

Schlagwörter: booklet effects, large-scale assessment, multi matrix designs

Booklet effects are commonly observed methodological phenomena that stem from using multi matrix designs in large-scale assessments (e.g., OECD, 2012; Hecht et al., 2015). The goal of this thesis is to explore the impact of booklet effects on statistics relevant for educational monitoring.

 

OECD. (2012). PISA 2009 technical report. Paris, France: Author. Retrieved from
https://www.oecd.org/pisa/pisaproducts/50036771.pdf

 

Hecht, M., Weirich, S., Siegle, T., & Frey, A. (2015). Modeling booklet effects for nonequivalent group designs in large-scale assessment. Educational and Psychological Measurement, 75, 568–584. doi:10.1177/0013164414554219

 

 

Masterarbeit: Optimization of Bayesian samplers for IRT and CT models

Martin Hecht

Schlagwörter: bayesian statistics, IRT, CT

The estimation of IRT models may become cumbersome when typical data settings of large-scale assessments are reached (i.e., large number of persons and items). The goal of this thesis is to explore and evaluate the optimization potential of Bayesian samplers for this heavy data scenario.

 

 

Bachelor- oder Masterarbeit: Kalman filter for GLMM

Martin Hecht

Schlagwörter: time-series, Kalman filter, IRT, GLM, GLMM

Kalman filters are popular techniques to weigh measurement and process error and thus to optimize person point estimates (e.g., Faragher, 2012). The goal of this thesis is to review and compare software implementations of Kalman filters suitable for IRT/GLM data situations.

 

Faragher, R. (2012). Understanding the Basis of the Kalman Filter Via a Simple and Intuitive Derivation. IEEE Signal Processing Magazine, 29, 128–132. doi:10.1109/MSP.2012.2203621

 

 

Bachelor- oder Masterarbeit: Missing data on background variables

Martin Hecht

Schlagwörter: missing data, large-scale assessment, imputation, plausible values (PVs)

„Background“ variables (e.g., demographic person characteristics) are used to calculate plausible values (PVs). In case of incomplete background data, some kind of imputation is needed (e.g., Weirich et al., 2014). The goal of this thesis is to review and compare strategies, methods, and software for imputation of background variables.

Weirich, S., Haag, N., Hecht, M., Böhme, K., Siegle, T., & Lüdtke, O. (2014). Nested multiple imputation in large-scale assessments. Large-Scale Assessments in Education2(1). doi:10.1186/s40536-014-0009-0

 

 

Bachelorarbeit: Automated test assembly

Martin Hecht

Schlagwörter: large-scale assessment, multi matrix designs, test assembly

The goal of this thesis is to review and compare strategies, methods, and software for automated test assembly.