Mathematical models in construct validity. Measurement is a core component in applied and research psychology. One of the main questions in this field is “How should scientists and psychologists proceed to make valid inferences about attributes such as intelligence, creativity and personality?” There are two steps in this process: 1. Using appropriate instruments (e.g. a high quality intelligence test), and 2. Using appropriate mathematical functions for aggregating test results (e.g. using the right function to compute an intelligence score). Our project concerns the second step. Specifically, we investigate the kinds of mathematical functions that should be used for attributes that have a compositional structure. E.g. if the premise is that intelligence is composed of verbal and mathematical ability, how do we compute an intelligence score from measurements of these abilities? In contrast, previous research has focused on mathematical functions for causal relations (i.e. answering the question given the premise that verbal and mathematical ability cause intelligence).

Artificial expert systems for complex decision making. This research is about the best way to formally represent psychological theories, e.g. the causes and characteristics of psychiatric disorders, in order to capture the complexity of human psychology and to make valid inferences. Specifically, we examine of how propositional and first order logic can be used to formally describe psychology, and how these languages can be used in artificial expert systems to aid complex decision making (e.g. making the correct psychiatric diagnosis in an applied setting).