We study the identification and consistency of Bayesian semiparametric IRT-type models, where the... more We study the identification and consistency of Bayesian semiparametric IRT-type models, where the uncertainty on the abilities’ distribution is modeled using a prior distribution on the space of probability measures. We show that for the semiparametric Rasch Poisson counts model, simple restrictions ensure the identification of a general distribution generating the abilities, even for a finite number of probes. For the semiparametric Rasch model, only a finite number of properties of the general abilities’ distribution can be identified by a finite number of items, which are completely characterized. The full identification of the semiparametric Rasch model can be only achieved when an infinite number of items is available. The results are illustrated using simulated data.
La recherche des causes et des effets est pratique courante dans les sciences de la nature. Cette... more La recherche des causes et des effets est pratique courante dans les sciences de la nature. Cette demarche repose le plus souvent sur l’experimentation. Mais peut-on deduire des liens de causalite dans les sciences sociales, ou l’experimentation est souvent impossible ou contraire a l’ethique ? Cet ouvrage tente de repondre par l’affirmative, en examinant successivement la demarche causale dans les sciences, l’experimentation et les methodes d’observation, la modelisation et la validite des resultats scientifiques. La modelisation structurelle d’un point de vue statistique est mise en exergue comme une des methodes permettant de proposer des liens de cause a effet dans les sciences sociales. L’ouvrage se termine par des reflexions sur les relations entre recherche scientifique et action politique.
... Introduction Guillaume Wunsch " To be or not to be?" mankind has probably raised th... more ... Introduction Guillaume Wunsch " To be or not to be?" mankind has probably raised this question since the first men and women confronted life ... sense that it does not require specific assumptions to be made about the underlying distribution of the survival times (D. Collett, 1994). ...
In this paper, we evaluate from a Bayesian point of view how much information is lost when the sa... more In this paper, we evaluate from a Bayesian point of view how much information is lost when the sampling process for the 2x2 contingency table is speci ed conditionally on the two margins as in the exact test of Fisher. We rst analyse the general problem of admissible conditioning ...
The objective of this paper is to present a short overview of the Structural Causal Modelling (SC... more The objective of this paper is to present a short overview of the Structural Causal Modelling (SCM) framework developed by the present authors in a series of articles spanning the last decade or so. The text is based on a presentation given at Statistics Netherlands in Heerlen on December 4, 2018 (Russo, Wunsch, Mouchart, 2019). The purpose is to explain how the SCM framework provides the tools to hypothesize, model, and test explanatory mechanisms. Our framework proves particularly useful in social science contexts, since it allows us to adopt an explicit causal perspective even when analyzing observational data. Social science experiments are notoriously difficult to carry out for ethical or practical reasons, and our approach allows social scientists to go beyond mere description and to propose a causal explanation even in the absence of experiments and interventions
We study the identification and consistency of Bayesian semiparametric IRT-type models, where the... more We study the identification and consistency of Bayesian semiparametric IRT-type models, where the uncertainty on the abilities’ distribution is modeled using a prior distribution on the space of probability measures. We show that for the semiparametric Rasch Poisson counts model, simple restrictions ensure the identification of a general distribution generating the abilities, even for a finite number of probes. For the semiparametric Rasch model, only a finite number of properties of the general abilities’ distribution can be identified by a finite number of items, which are completely characterized. The full identification of the semiparametric Rasch model can be only achieved when an infinite number of items is available. The results are illustrated using simulated data.
La recherche des causes et des effets est pratique courante dans les sciences de la nature. Cette... more La recherche des causes et des effets est pratique courante dans les sciences de la nature. Cette demarche repose le plus souvent sur l’experimentation. Mais peut-on deduire des liens de causalite dans les sciences sociales, ou l’experimentation est souvent impossible ou contraire a l’ethique ? Cet ouvrage tente de repondre par l’affirmative, en examinant successivement la demarche causale dans les sciences, l’experimentation et les methodes d’observation, la modelisation et la validite des resultats scientifiques. La modelisation structurelle d’un point de vue statistique est mise en exergue comme une des methodes permettant de proposer des liens de cause a effet dans les sciences sociales. L’ouvrage se termine par des reflexions sur les relations entre recherche scientifique et action politique.
... Introduction Guillaume Wunsch " To be or not to be?" mankind has probably raised th... more ... Introduction Guillaume Wunsch " To be or not to be?" mankind has probably raised this question since the first men and women confronted life ... sense that it does not require specific assumptions to be made about the underlying distribution of the survival times (D. Collett, 1994). ...
In this paper, we evaluate from a Bayesian point of view how much information is lost when the sa... more In this paper, we evaluate from a Bayesian point of view how much information is lost when the sampling process for the 2x2 contingency table is speci ed conditionally on the two margins as in the exact test of Fisher. We rst analyse the general problem of admissible conditioning ...
The objective of this paper is to present a short overview of the Structural Causal Modelling (SC... more The objective of this paper is to present a short overview of the Structural Causal Modelling (SCM) framework developed by the present authors in a series of articles spanning the last decade or so. The text is based on a presentation given at Statistics Netherlands in Heerlen on December 4, 2018 (Russo, Wunsch, Mouchart, 2019). The purpose is to explain how the SCM framework provides the tools to hypothesize, model, and test explanatory mechanisms. Our framework proves particularly useful in social science contexts, since it allows us to adopt an explicit causal perspective even when analyzing observational data. Social science experiments are notoriously difficult to carry out for ethical or practical reasons, and our approach allows social scientists to go beyond mere description and to propose a causal explanation even in the absence of experiments and interventions
A vector stochastic process may be decomposed in to its expectation and a residual process. A lin... more A vector stochastic process may be decomposed in to its expectation and a residual process. A linear dynamic model is defined by a set of dynamic linear relations constraining the 's given some conditioning variables and by the distribution of the process. This paper presents a strategy for the specification of this class of models providing computable posterior distributions for a suitable class of prior measures. Some conditional independence properties characterizing exogeneity conditions through global or sequential cuts, innovation property or non causality relations are studied and are shown to allow reductions by conditioning of the model.
Uploads