home | vita | publications | talks | research | teaching

Ongoing Research Projects

Research Project for SNSF Professorship: Coincidence Analysis

Coincidence Analysis (CNA) is a Boolean method of causal data analysis first presented in Baumgartner (2009a; 2009b). CNA has been developed on the methodological drawing board and against the background of various idealizing assumptions. As of now, its applicability to real-life data is hence severely restricted. Moreover, in its current state it can only model dichotomous variables. Nonetheless, as shown in Baumgartner (2014), CNA has some distinguished advantages over the currently dominant approach to Boolean causal discovery, viz. Qualitative Comparative Analysis (QCA) (cf. Ragin 1987; 2008). Contrary to QCA, CNA can process data that are generated by causal chains and common cause structures, it can eliminate all redundancies from causal models without having to resort to counterfactual simplifying assumptions, and CNA searches for causal structures without the analyzed variable set antecedently being partitioned into exogenous and endogenous variables.
In light of these promising prospects, this project aims to bring Coincidence Analysis from the drawing board to effective, flexible, and computer-assisted applicability in real-life contexts of causal discovery. In collaboration with researchers working with Boolean causal models—e.g. from the fields of social and political sciences or biology—, CNA shall be adapted to the demands of its users. Furthermore, the theoretical and conceptual foundation of Boolean causal reasoning shall be clarified by spelling out the details of the theory of causation relative to which Boolean models must be interpreted, by analyzing the required background assumptions of Boolean data analysis, by scrutinizing the divide between Boolean and non-Boolean dimensions of causal structures, and by comparing CNA with non-Boolean methods. In this vein, a precise understanding of the domain of applicability and the inferential potential of CNA shall be gained. The output of the project will be a fully worked out, ready-to-use, maximally general, and theoretically grounded method of Boolean causal data analysis that is applicable to data not processable by other methods and that, thus, constitutes a valid alternative for researchers interested in Boolean dimensions of causality.
Collaborators on this project:

Further Ongoing Projects

The Extension of Cognition 
This projects investigates the implications of the currently flourishing theories of mechanistic constitution (e.g. Craver 2007) for the hypothesis of extended cognition, which states that cognitive processes can and do have constituents that occur outside of the head (Clark & Chalmers 1998; Rowlands 2009; Wheeler 2010; Drayson 2010).
Collaborator: Wendy Wilutzky.

An Abductive Theory of Constitution  The starting point of this project is the recent result of Baumgartner & Gebharter (2015) showing that Craver's popular mutual manipulability account of constitution is unsuited to ground a viable method for the empirical identification of constitutional relations. As an alternative, we develop an abductive theory of constitution, which exploits the fact that phenomena and their constituents are unbreakably coupled via common causes. The best explanation for this common-cause coupling is the existence of an additional dependence relation, viz. constitution.
Collaborator: Lorenzo Casini.

A Bayesian Theory of Constitution The goal of this project is to develop a Bayesian theory of constitution that identifies as constituents those spatiotemporal parts of a phenomenon whose causal roles contain the phenomenon's causal role. By drawing on the conceptual resources of Bayesian networks, the project should pave the way for a Bayesian methodology for constitutional discovery.
Collaborator: Lorenzo Casini.

Model Ambiguities in QCA This project explores the problem of model ambiguities in Qualitative Comparative Analysis (QCA). Data analyzed by QCA can often be accounted for by multiple causal models that fare equally well with respect to all parameters of fit. The degree of ambiguity sometimes reaches such extreme proportions that no causal conclusions are possible. Yet, due to severe deficiencies in popular QCA software, researchers are typically unaware of these ambiguities.
Alrik Thiem.

Evaluating Configurational Comparative Methods To date, hundreds of researchers have employed Qualitative Comparative Analysis (QCA) as a configurational method of empirical social research. However, the correctness of QCA qua causal inference procedure has not been carefully scrutinized in the literature so far. This projects aims to fill that glaring gap. We first lay out the criteria an adequate method of configurational data analysis has to satisfy, and second, implement a battery of inverse-search trials to test how QCA performs with respect to these criteria in different discovery contexts.
Collaborator: Alrik Thiem.

Is it Possible to Generate Empirical Evidence for the Existence of Macro-To-Micro Causation? In recent years, numerous non-reductive physicalists (e.g. Shapiro, Sober, Raatikainen, Menzies) have argued that, by adopting a variant of Woodward's (2003) popular interventionist theory of causation, it becomes possible to provide empirical evidence in favor of the existence of macro-to-micro downward causation. This projects intends to show that all of these proposals are bound to fail, for it is impossible, in principle, to generate evidence for downward causation. The question as to the existence of macro-to-micro causation is of inherently pragmatic nature.