Three Corrected Score Tests for Generalized Linear Models with Dispersion Covariates

We develop three corrected score tests for generalized linear models with dispersion covariates, thus generalizing the results of Cordeiro , Ferrari and Paula (1993) and Cribari-Neto and Ferrari (1995) . We present, in matrix notation, general formulae for the coefficients which define the corrected statistics. The formulae only require simple operations on matrices and can be used to obtain analytically closed-form corrections for score test statistics in a variety of special generalized linear models with dispersion covariates. They also have advantages for numerical purposes since our formulae are readily computable using a language supporting numerical linear algebra. Two examples, namely, iid sampling without covariates on the mean or dispersion parameter oand one-way classification models, are given. We also present some simulations where the three corrected tests perform better than the usual score test, the likelihood ratio test and its Bartlett corrected version. Finally, we present a numerical example for a data set discussed by Simonoff and Tsai (1994) .     

Rat race, redistribution, and growth

Introducing locally negatively interdependent preferences into a simple AK growth model easily explains the often observed insignificant or positive correlation between distortionary redistribution and growth rates. Positive capital income taxes and lump sum transfers are harmful for growth, but people rationally vote for them in order to reduce “rat race” overaccumulation. A “neutrality proposition” holds if the pivotal voter is the mean voter, as in a representative agent case, but it fails if the pivotal voter is poorer than the average citizen.     

Qualitative valuation of environmental criteria through a group consensus based on stochastic dominance

This paper introduces a qualitative valuation method to elicit stakeholders' intensities of preferences for a complex environmental issue and multiple social groups. Environmental valuation studies have shown that in any complex environment with a diversity of environmental services, stakeholders have difficulties using a monetary valuation to make trade-offs between different environmental services. Stated preference methods such as the Contingent Valuation Method (CVM) have been criticised for their individualistic format and assumptions of commensurability between environmental criteria. To alleviate both of these criticisms, we propose a qualitative valuation method. The method contains a discursive step to allow stakeholders to discuss and construct a list of environmental criteria and alternative plans. The list of criteria and plans is subsequently used by a group of experts to formulate an Impact Matrix (IM), which is to be used in the succeeding individualistic steps of the methodology. The first individualistic step consists of asking the stakeholders to rank Alternative Impacts (AIs) in the IM for each single criterion. The stakeholders are then asked to express intensities of their preferences through pairwise comparisons between the AIs of the constructed rank order on each single criterion. These intensities are expressed on a qualitative scale. Subsequently, to provide social intensities of preferences, a social preference (social rank order) is first determined for each single criterion. We propose to use the median value among the intensities of preferences as the social intensity of preference by assuming interpersonal comparability and taking into account stochastic monotonocity. This is a pre-processing step, which allows us to reach social intensities of preferences in the Lar rangeland (Iran), where several social groups have conflicting interests on rangeland services, leading to conflicting preferences on environmental criteria.     

Impostazione Bayesiana di un problema di analisi discriminatoria nell'ambito di un modello non parametrico

Seguendo l'impostazione predittiva delineata da J. Aitchison e I. R. Dunsmore (1975) viene proposto un modello non-parametrico di analisi discriminatoria dal punto di vista bayesiano.Si ricorre a tal fine ad una versione modificata dello schema di scambiabilità parziale e si utilizza il processo «mistura di prodotti di processi di Dirichlet» introdotto da D. M. Cifarelli ed. E. Regazzini (1978). Si presenta infine un'applicazione del modello ad un tipico problema di diagnosi medica.

---

J. Aitchison and I. R. Dunsmore (1975) have stressed the importance of the predictive distribution for the solution of classical statistical problems. Following this approach we show initially how a modified version of the model of partial exchangeability can be usefully applied to derive the so called diagnostic distribution. Subsequently a nonparametric model of discriminatory analysis is derived, wherein a crucial role is played by a particular process named «mixture of products of Dirichlet processes» introduced by D. M. Cifarelli and E. Regazzini (1978) following a paper by C. E. Antoniak (1974) which in turn generalized the well known Dirichlet process developed by T. S. Ferguson (1973). Finally a numerical application to a medical problem is provided.

---

---

Versione definitiva pervenuta il 24-9-81

---

Lavoro eseguito nell'ambito del GNAFA-CNR. L'argomento di questo lavoro mi è stato suggerito dal professor Eugenio Regazzini. A quest'ultimo ed al prof. Donato Michele Cifarelli sono grato per avermi segnalato alcune oscurità nel testo originario.     

Project evaluation: Externalities must not be disregarded

While technology assessment is now usually performed at government level, methods proposed in the literature for evaluating single R and D projects by industrial concerns do not generally dedicate sufficient attention to possible externalities.
The authors propose a model for evaluating R and D projects based on the concept of 'limiting constraints' imposed by the different organizations or groups of interest with which the company interfaces.