1,721,154 research outputs found

    Modelling University students’ final grades by ordinal variables

    No full text
    The final assessment received by graduates in Italy is expressed on a 66-110 scale points and this quantity is determined by the marks average of all exams and an extra value depending on the judgement of the Commission with regard to the thesis defense. Although the final grade is expressed on a quantitative scale, what it is really important for the future career of a graduated student is the qualitative ordinal evaluation of the final grade (from a very low assessment up to excellent, and first class honors). Empirical evidence and academic rules confirm that finale grades are related both to subjects’ covariates and latent variables connected to students’ background and career. In this paper, we will apply models for ordinal data as this framework seems more reliable with respect to the final judgment. Specifically, we introduce a class of ordinal models where a direct evaluation of the probability of a qualitative result is fitted and then modelled by means of gender, duration of studies and secondary school diploma marks. An empirical case study have been analyzed for a large data set of graduated in Political Sciences, University of Naples Federico II, and the main results are discussed

    Observed information matrix for MUB models

    No full text
    In this paper we derive the observed information matrix for MUB models, without and with covariates. After a review of this class of models for ordinal data and of the E-M algorithms, we derive some closed forms for the asymptotic variance covariance matrix of the maximum likelihood estimators of MUB models. Also, some new results about feeling and uncertainty parameters are presented. The work lingers over the computational aspects of the procedure with explicit reference to a matrixoriented language. Finally, the finite sample performance of the asymptotic results is investigated by means of a simulation experiment. General considerations aimed at extending the application of MUB models conclude the pape

    Un prototipo statistico per la valutazione comparata dell'efficacia didattica dei corsi di studio universitari

    No full text
    In this work we propose a statistical method for the comparative evaluation of the University performances, on the basis of the CFUs (=Crediti Formativi Universitari) gained by the students. The method relies on the analysis of the CFU distribution function. It conveys immediate and simple interpretations. Moreover, its graphical version allows to define success and failure areas, and to relate these measures to the average CFU achieved by the students enrolled in a given course. Some examples and considerations about further developments conclude the paper

    A general approach for modelling individual choices

    No full text
    In this article, we generalize the approach for deriving the Inverse Hyper- Geometric (IHG) random variable proposed by Ridout (1999). After reviewing some preliminary results, the paper focuses on a general elationship between a statement on the sequential odds of the choices and the probability distribution of preferences. Finally, some statistical consequences of these findings are discussed

    Some approximations for the asymptotic variance of the maximum likelihood estimator of the parameter in the inverse Hypergeometric random variable

    No full text
    We derive some approximations for the asymptotic variance of the Maximum Likelihood estimator for the parameter of the Inverse Hypergeometric random variable. For most statistical models, the asymptotic variance is usually derived after some algebraic manipulations. In this paper, we show that this lengthy calculations can be overcome by simple and accurate linear approximations. The interest for this result arises from a statistical model for preferences that has been recently proposed for evaluation studies, preferences analyses and marketing researches
    corecore