The empirical social & behavioral sciences (ESBS) generally—and psychology as the largest ESBS field particularly —have made significant advances in mathematical modeling. All these fields, however, feature comparatively soft empirical structures and unsophisticated theoretical superstructures (mere directional hypotheses). These fields as a whole thus appear as a soft science. The ongoing “replication/confidence crisis” rightly aggravates this impression.
Strengthening the theoretical structures requires hardening the empirical structures. This demands new descriptive and normative insights—insights, that is, into how researchers should describe (“model”) and evaluate empirical data, how they should construct and evaluate hypotheses and theories. In brief, this requires insights into how to arrange an ESBS-research program. Only if this central need is met can the ESBS become progressive, trustworthy, and useful.
The project addresses this need with insights from the philosophy of science and the philosophy of statistics to improve ESBS-research. We deploy inductive considerations (pertaining to what we can learn from data), deductive considerations (pertaining to how we construct theories for data), and abductive considerations (pertaining to the next best action, given our pragmatic stance). This exemplifies an efficient route to trustworthy and useful ESBS-research.
Through descriptive statistical, statistical-inferential, and science-philosophical research, the project:
- Explains the replication/confidence crisis in the ESBS as the result of having arranged research efforts in ways that misapply and over-interpret our best statistical inference methods;
- Develops the research program strategy (RPS) as a superior methodology to overcome the crisis, by embedding RPS into qualitative research, by furthering “induction analysis” as an alternative to current meta-analysis, and by providing a toolbox for theory construction;
- Disseminates the explanation and the remedy via publications, presentations, and teaching materials among such key-players as researchers, funding agencies, and science-journalists.
Compared to what ESBS-research offers today, project results improve how researchers learn from data, and how they construct, and validate, sophisticated (point-predicting) empirical theories.
Join the Project
MTR: Models, Theories, Research Programs
Project No: 118C257
PI: Dr. Frank Zenker
BOĞAZİÇİ UNIVERSİTY Department of Philosophy