Description du projet
Progrès en matière de méthodes qualitatives et mixtes dans le domaine des sciences politiques
Financé par le Conseil européen de la recherche, le projet ENHANCEDQMMR entend contribuer aux progrès des méthodes qualitatives autonomes et de la recherche multiméthode (RMM). Il s’appuiera sur des techniques qualitatives de pointe et cherchera à faire progresser les méthodes utilisées en sciences politiques de quatre manières: premièrement, il examinera comment les chercheurs choisissent entre les points de vue relationnels et corrélationnels sur la causalité; deuxièmement, il élaborera des outils pour renforcer l’inférence fondée sur l’analyse comparative qualitative (ACQ); troisièmement, il comparera les performances et les résultats dérivés de l’ACQ et des analyses de régression; et quatrièmement, il formulera des normes pour la RMM bayésienne. Les résultats du projet seront intégrés dans un logiciel disponible gratuitement, ce qui permettra une utilisation et un impact plus larges.
Objectif
Over the last 20 years, qualitative methods in political science have developed rapidly on three dimen-sions. First, set theory was formulated as an alternative to what is called the quantitative worldview. Second, process tracing for the analysis of mechanisms evolved as a complement to the estimation of marginal effects. Process tracing has also been tied to Bayesianism as opposed to frequentism. Third, process tracing became an element of multi-method research (MMR), integrating it with frequentist statistics or Qualitative Comparative Analysis (QCA).
An important element of the development of qualitative methods is its contrast with quantitative methods. At the same time, quantitative researchers critically commented on that development. The constant exchange has contributed to the progress of qualitative methods, but the debate has reached an impasse in some respects, or has pursued one line of development while neglecting others.
Building on state-of-the-art qualitative methods, ENHANCEDQMMR seeks to overcome this impasse and to explore new ground.It makes four significant contributions to the progress of standalone qualitative methods and as part of MMR. First, it examines experimentally how researchers decide between a set-relational and correlational view on causation and whether they can realize designs in accord with their initial decision. Second, it develops tools for sensitivity analyses, diagnostics, and the modeling of diverse data structures via QCA for strengthening QCA-based inference. Third, it com-pares the performance of QCA and regression analyses under simulated data-generating processes with the goal of generating comparative diagnostics, possibly allowing one to adjudicate between both methods in observational research. Fourth, it formulates standards for Bayesian MMR by combining Bayesian process tracing with Bayesian statistics and Bayesian QCA, respectively. The insights of the project will be implemented in freely available software.
Champ scientifique
Programme(s)
Thème(s)
Régime de financement
ERC-STG - Starting GrantInstitution d’accueil
50931 Koln
Allemagne