Cognitive Models of Quantitative Estimation and Seeding Effects in Real-World Contexts

• Applicant: David Izydorczyk
• Subject Area: General, Cognitive and Mathematical Psychology
• Term: since 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 526349149

How severe is the condition of this patient? What is the monetary value of this real estate? How much insulin should I take for this meal? Estimating quantities and numbers such as these, in different contexts and situations, is essential to our everyday life. The answers to such questions often guide our decisions and our behavior and hence, the accuracy of our estimates matters. In many domains, people’s quantitative estimates are deficient. Together with their framework of quantitative estimation, Brown and Siegler (1993) developed the so called seeding paradigm as an intervention to improve the accuracy of people’s numerical judgments. The simple intervention consists of presenting correct values for a small subset of items, the so-called “seeding” stimuli. This will also improve the estimates for non-presented transfer items from the same judgment domain. Despite the large positive and potential long lasting effects, the seeding paradigm has only been applied in few studies and currently lacks a well-defined theoretical basis With this project, we aim at combining laboratory-based computational models of quantitative judgments with the hitherto vaguely formulated verbal theory of naturalistic estimation by Brown and Siegler (1993) to develop a theory- and evidence-based efficient and easy-to-implement training method for improving the accuracy of people’s judgments. More specifically, our goal is to model the underlying cognitive mechanisms of estimation and seeding effects in real-world contexts using the formal models used in the multiple-cue judgment literature. These models have hitherto been confined to applications in controlled laboratory settings using mostly artificial stimuli. We will assess the most promising candidate by formal model comparisons. In order to do this, our proposed project involves collecting a large amount of similarity judgments to extract functional attributes of natural stimuli in various knowledge domains (e.g., food items). These attributes, and the data they are based on, will be an important contribution for the research community and allow us to actually model the knowledge updating effects observed in the seeding paradigm using well-established computational models of people`s judgments. We implement a series of studies using various knowledge domains and experimental designs. Providing a theoretical basis for the seeding effectallows us to derive and test hypothesis about the most efficient way to improve judgments in to two highly relevant applied domains, namely the estimation of nutritional information and carbon footprints of food items. Using this knowledge we aim at investigating the potential of incorporating a theory-based seeding procedure as part of a disease management program to help patients with newly diagnosed Type 1, as well as to test the potential of the seeding procedure as an easy-to-implement and large-scale intervention in an ecologically valid everyday context (social media).

• DFG Programme: Research Grants
• Co-Investigator: Professor Dr. Arndt Bröder