Adaptive sequential design for regression on multi-resolution bases

This work investigates the problem of construction of designs for estimation and discrimination between competing linear models. In our framework, the unknown signal is observed with the addition of a noise and only a few evaluations of the noisy signal are available. The model selection is performed in a multi-resolution setting. In this setting, the locations of discrete sequential D and A designs are precisely constraint in a small number of explicit points. Hence, an efficient stochastic algorithm can be constructed that alternately improves the design and the model. Several numerical experiments illustrate the efficiency of our method for regression. One can also use this algorithm as a preliminary step to build response surfaces for sensitivity analysis.     

Estimation in nonlinear mixed-effects models using heavy-tailed distributions

Nonlinear mixed-effects models are very useful to analyze repeated measures data and are used in a variety of applications. Normal distributions for random effects and residual errors are usually assumed, but such assumptions make inferences vulnerable to the presence of outliers. In this work, we introduce an extension of a normal nonlinear mixed-effects model considering a subclass of elliptical contoured distributions for both random effects and residual errors. This elliptical subclass, the scale mixtures of normal (SMN) distributions, includes heavy-tailed multivariate distributions, such as Student-t, the contaminated normal and slash, among others, and represents an interesting alternative to outliers accommodation maintaining the elegance and simplicity of the maximum likelihood theory. We propose an exact estimation procedure to obtain the maximum likelihood estimates of the fixed-effects and variance components, using a stochastic approximation of the EM algorithm. We compare the performance of the normal and the SMN models with two real data sets.     

Empirical likelihood for partially linear additive errors-in-variables models

This study investigates the empirical likelihood method for the partially linear additive models in which certain covariates are measured with additive errors. An empirical log-likelihood ratio for the parametric component is proposed based on the profile procedure, and a nonparametric version of the Wilk’s theorem is derived. Then, the confidence regions of the parametric component with asymptotically correct coverage probabilities are constructed by the obtained results. Furthermore, a simulation study is conducted to illustrate the performance of the proposed method.     

Missing values: sparse inverse covariance estimation and?an?extension to sparse regression

We propose an ℓ ₁-regularized likelihood method for estimating the inverse covariance matrix in the high-dimensional multivariate normal model in presence of missing data. Our method is based on the assumption that the data are missing at random (MAR) which entails also the completely missing at random case. The implementation of the method is non-trivial as the observed negative log-likelihood generally is a complicated and non-convex function. We propose an efficient EM algorithm for optimization with provable numerical convergence properties. Furthermore, we extend the methodology to handle missing values in a sparse regression context. We demonstrate both methods on simulated and real data.     

Design of computer experiments: space filling and beyond

When setting up a computer experiment, it has become a standard practice to select the inputs spread out uniformly across the available space. These so-called space-filling designs are now ubiquitous in corresponding publications and conferences. The statistical folklore is that such designs have superior properties when it comes to prediction and estimation of emulator functions. In this paper we want to review the circumstances under which this superiority holds, provide some new arguments and clarify the motives to go beyond space-filling. An overview over the state of the art of space-filling is introducing and complementing these results.     

Development of Science Anxiety Scale for Primary School Students

The principal aim of the study is to develop a new scale Science Anxiety Scale and to examine its the psychometric properties and construct validity of the Science Anxiety Scale in a sample of 797 primary school students. Exploratory factor analysis was applied and found to have a two-dimensional structure. Confirmatory factor analyses provide evidence for a solid two-dimension structure reflecting the theorized construct: personal and environmental. Also, for convergent validity, item reliability, construct reliability and average variance values were calculated and it was found that item reliability and construct reliability values generally provided the required criteria. The average variance extracted value had a value below the criteria value for both sub-dimensions. Discriminant validity was above the criteria value in all dimensions. The reliability of the Science Anxiety Scale scales appeared to be satisfactory, with good internal consistency. The Science Anxiety Scale was found to be reliable and valid for the Turkish primary student context with a few exceptions.     

Individual- and Neighbourhood-Level Indicators of Subjective Well-Being in a Small and Poor Eastern Cape Township: The Effect of Health,Social Capital,Marital Status,and Income

Our study used multilevel regression analysis to identify individual- and neighbourhood-level factors that determine individual-level subjective well-being in Rhini, a deprived suburb of Grahamstown in the Eastern Cape province of South Africa. The Townsend index and Gini coefficient were used to investigate whether contextual neighbourhood-level differences in socioeconomic status determined individual-level subjective well-being. Crime experience, health status, social capital, and demographic variables were assessed at the individual level. The indicators of subjective well-being were estimated with a two-level random-intercepts and fixed slopes model. Social capital, health and marital status (all p < .001), followed by income level (p < .01) and the Townsend score (p < .05) were significantly related to individual-level subjective well-being outcomes. Our findings showed that individual-level subjective well-being is influenced by neighbourhood-level socioeconomic status as measured by the Townsend deprivation score. Individuals reported higher levels of subjective well-being in less deprived neighbourhoods. Here we wish to highlight the role of context for subjective well-being, and to suggest that subjective well-being outcomes may also be defined in ecological terms. We hope the findings are useful for implementing programs and interventions designed to achieve greater subjective well-being for people living in deprived areas.     

Demography and cross-country differences in savings rates: a new approach and evidence

This paper includes both dependency thesis and pension motive for savings to explain the large differences in cross-country savings rates. The two demographic factors are incorporated into an overlapping generation model, and the steady-state savings rates for a sample of 109 countries are computed. Both demographic factors can explain up to 68% of the dispersion in the cross-country savings rates. Furthermore, if the expenditure burden is sufficiently high, fertility has a greater impact on cross-country savings rate differences than longevity does. This study also satisfactorily explains the large gap in savings rates between the high- and low-income countries.     

A politico-economic model of aging,technology adoption and growth

This paper provides a politico-economic theory that explains how an economy evolves when the longevity of its citizens is jointly determined with the process of economic development. We propose a three-period overlapping generation model where agents’ decisions embrace two dimensions: a private choice about education and a public one on innovation policy. We find that (a) poverty traps can emerge in human capital accumulation, (b) higher life expectancy increases the incentive to innovate for both young and adults, (c) different political configurations can arise depending on endogenous demographic structures and (d) the steady state can entertain both innovation and its absence.     

Social security reform and the support for public education

Old age pensions and public education account for a large share of public budgets. We link both programs through a tax-transfer system that is also sensitive to labor market distortions. We analyze the impact that alternative pension reforms have, through the political process, on publicly financed education. We explain how changes in the pension system design affect the link between the two programs and also labor market incentives. These effects, if they exist, act in opposite directions. Overall, we find that most proposals that entail a partial privatization of pensions reduce the willingness of the society to fund public education.