Modeling individual differences

Interindividual differences in personality attributes and cognitive performance are at the core of many domains of psychological research. The assessment of interindividual differences in personality traits can be distorted, however, due to individual response styles, method effects of the assessment instruments and situational fluctuations. In a similar vein, the measurement of cognitive abilities can be biased by guessing tendencies and varying cognitive processing strategies. An in-depth analysis of personality traits and cognitive abilities therefore requires theory-based statistical models that allow researchers to disentangle target traits from other response processes and to separate multiple dimensions of cognitive processing in a given task.

In personality research, responses to questionnaire items or vignette scenarios are confounded with response styles of the individual (e.g., general preference of extreme, moderate or affirmative response options) and method effects of the assessment instrument (e.g., reversed item effects; self vs. other judgments). Additional sources of error variance or systematic bias can come from situation effects and from interactions of response styles with method effects (e.g., moderating effects of item features on the impact of response styles). Effects of the target trait, response styles, assessment method and situation can be investigated and controlled through statistical analysis with multidimensional IRT models, structural equation methods or generalized multilevel models.

In cognitive research, theory-driven measurement models can be used to decompose global performance indices into measures of specific cognitive processes. The decomposition allows researchers to analyze interindividual differences in abilities (e.g., memory, attention), response tendencies (e.g., guessing) and cognitive strategies (e.g., information integration versus exemplar-based categorization) as person variables that jointly determine performance in a given task. Statistical approaches to achieve multidimensional cognitive assessment include multilevel extensions of multinomial processing-tree and mixture distribution models that are tailored to the specific task setting.


Dissertation projects can focus on interindividual differences in personality or cognition, or on the interplay between individual differences in one domain (e.g., personality) and other fields of psychological research (e.g., processing strategy in judgment situations or memory tasks). Some exemplary dissertation topics are listed below, but additional project ideas are welcome and can be developed with the responsible SMiP researchers.

Possible model development projects

  • IRT models for unravelling trait assessment from response styles and method effects (Advisors: Meiser, Lischetzke, Hilbig).
  • Exemplar-based and rule-based processes in judgment (Advisors: Bröder, Hilbig).

Possible model application projects

  • The structure of cognitive speed (Advisors: Voss, Meiser).
  • Memory-task response biases as cognitive traits and states (Advisors: Kuhlmann, Hütter, Meiser).
  • A life-span perspective on controllable and uncontrollable learning in evaluative conditioning (Advisors: Hütter, Klauer, Kuhlmann).
  • Modeling individual differences in judgment and decision making (Advisors: Hilbig, Erdfelder, Meiser).