Fixed T dynamic panel data estimators with multifactor errors

This article analyzes a growing group of fixed T dynamic panel data estimators with a multifactor error structure. We use a unified notational approach to describe these estimators and discuss their properties in terms of deviations from an underlying set of basic assumptions. Furthermore, we consider the extendability of these estimators to practical situations that may frequently arise, such as their ability to accommodate unbalanced panels and common observed factors. Using a large-scale simulation exercise, we consider scenarios that remain largely unexplored in the literature, albeit being of great empirical relevance. In particular, we examine (i) the effect of the presence of weakly exogenous covariates, (ii) the effect of changing the magnitude of the correlation between the factor loadings of the dependent variable and those of the covariates, (iii) the impact of the number of moment conditions on bias and size for GMM estimators, and finally (iv) the effect of sample size. We apply each of these estimators to a crime application using a panel data set of local government authorities in New South Wales, Australia; we find that the results bear substantially different policy implications relative to those potentially derived from standard dynamic panel GMM estimators. Thus, our study may serve as a useful guide to practitioners who wish to allow for multiplicative sources of unobserved heterogeneity in their model.     

基于神经网络的测井评价油层污染研究

利用测井资料,在剖析传统的测井评价油(气)层污染方法的基础上,介绍了利用BP神经网络这一现代数理统计新技术来实现测井多参数评定油气层污染程度的方法和技巧;以华北二连油田实际资料为例,建立了测井判释油气层污染的数学物理模型,并进行了模拟预测。结果表明:该法简便可行,对污染评价参数预测精度较高,对污染程度评价合理准确,可作为油(气)层污     

Self-exciting threshold models for time series of counts with a finite range

In this article, an integer-valued self-exciting threshold model with a finite range based on the binomial INARCH(1) model is proposed. Important stochastic properties are derived, and approaches for parameter estimation are discussed. A real-data example about the regional spread of public drunkenness in Pittsburgh demonstrates the applicability of the new model in comparison to existing models. Feasible modifications of the model are presented, which are designed to handle special features such as zero-inflation.     

Multiple imputation for missing edge data: A predictive evaluation method with application to Add Health

Recent developments have made model-based imputation of network data feasible in principle, but the extant literature provides few practical examples of its use. In this paper, we consider 14 schools from the widely used In-School Survey of Add Health (Harris et al., 2009), applying an ERGM-based estimation and simulation approach to impute the network missing data for each school. Add Health's complex study design leads to multiple types of missingness, and we introduce practical techniques for handing each. We also develop a cross-validation based method – Held-Out Predictive Evaluation (HOPE) – for assessing this approach. Our results suggest that ERGM-based imputation of edge variables is a viable approach to the analysis of complex studies such as Add Health, provided that care is used in understanding and accounting for the study design.     

Sparse group lasso for multiclass functional logistic regression models

Sparsity-inducing penalties are useful tools for variable selection and are also effective for regression problems where the data are functions. We consider the problem of selecting not only variables but also decision boundaries in multiclass logistic regression models for functional data, using sparse regularization. The parameters of the functional logistic regression model are estimated in the framework of the penalized likelihood method with the sparse group lasso-type penalty, and then tuning parameters for the model are selected using the model selection criterion. The effectiveness of the proposed method is investigated through simulation studies and the analysis of a gene expression data set.     

Task-based learning approaches for supporting the development of social science researchers’ critical data skills

Conducting social science research is a complex process, and social science students face a number of learning challenges in order to develop their skills. Moreover, new sources and types of data including so-called big data are providing new opportunities for research, but also pose methodological challenges. In this article, we explore a task-based learning approach for teaching social science research methods. We draw on evidence from two case study learning tasks: (i) the collection and analysis of Twitter data; and (ii) designing and conducting a face-to-face and online survey. The students were guided to co-lead the tasks, apply their knowledge and then to critically reflect. The pedagogical framework of task-based learning provides opportunities to embed learning in new ways by integrating knowledge, practice and critical reflection. Task-based learning can create a dynamic learning environment and can empower students to develop their identities as researchers.     

Demographic and Social Factors Associated with Homophobia and Fear of AIDS in a Community Sample

Examinations of demographic and social factors associated with homophobia and fear of AIDS are limited by the frequent use of homogeneous, college student samples and limited examination of interrelationships among variables. The present study examined community attitudes toward homosexuality and fear of HIV/AIDS as a function of age, education, race/ethnicity, religious affiliation, political party affiliation, and personal contact with homosexual individuals and persons living with HIV/AIDS. A community sample of 463 adults completed standardized measures of homophobia and fear of AIDS as well as demographic and social background items. When examined separately, each demographic and social factor assessed, with the exception of race/ethnicity, was associated with homophobia and all but race/ethnicity and political party affiliation were associated with fear of AIDS. However, when entered into multiple regression analyses, 24% of the variance in homophobia was predicted by a single variable, including only personal contact with homosexual individuals, while 18% of the variance in fear of AIDS was accounted for by five variables, including personal contact with homosexual individuals, religious affiliation, political affiliation, education, and personal contact with someone living with HIV/AIDS. Findings suggest that it is important to consider intercorrelations among social and demographic factors, particularly when considering homophobia.     

Modeling censored data using modified lambda family

In the present article we propose the modified lambda family (MLF) which is the Freimer, Mudholkar, Kollia, and Lin (FMKL) parametrization of generalized lambda distribution (GLD) as a model for censored data. The expressions for probability weighted moments of MLF are derived and used to estimate the parameters of the distribution. We modified the estimation technique using probability weighted moments. It is shown that the distribution provides reasonable fit to a real censored data.     

Fitting the generalized lambda distribution to pre-binned data

Density estimation for pre-binned data is challenging due to the loss of exact position information of the original observations. Traditional kernel density estimation methods cannot be applied when data are pre-binned in unequally spaced bins or when one or more bins are semi-infinite intervals. We propose a novel density estimation approach using the generalized lambda distribution (GLD) for data that have been pre-binned over a sequence of consecutive bins. This method enjoys the high power of the parametric model and the great shape flexibility of the GLD. The performances of the proposed estimators are benchmarked via simulation studies. Both simulation results and a real data application show that the proposed density estimators work well for data of moderate or large sizes.     

Estimation of Benchmark Dose of Lifetime Cadmium Intake for Adverse Renal Effects Using Hybrid Approach in Inhabitants of an Environmentally Exposed River Basin in Japan

The aim of this study is to estimate the reference level of lifetime cadmium intake (LCd) as the benchmark doses (BMDs) and their 95% lower confidence limits (BMDLs) for various renal effects by applying a hybrid approach. The participants comprised 3,013 (1,362 men and 1,651 women) and 278 (129 men and 149 women) inhabitants of the Cd‐polluted and nonpolluted areas, respectively, in the environmentally exposed Kakehashi River basin. Glucose, protein, aminonitrogen, metallothionein, and β₂‐microglobulin in urine were measured as indicators of renal dysfunction. The BMD and BMDL that corresponded to an additional risk of 5% were calculated with background risk at zero exposure set at 5%. The obtained BMDLs of LCd were 3.7 g (glucose), 3.2 g (protein), 3.7 g (aminonitrogen), 1.7 g (metallothionein), and 1.8 g (β₂‐microglobulin) in men and 2.9 g (glucose), 2.5 g (protein), 2.0 g (aminonitrogen), 1.6 g (metallothionein), and 1.3 g (β₂‐microglobulin) in women. The lowest BMDL was 1.7 g (metallothionein) and 1.3 g (β₂‐microglobulin) in men and women, respectively. The lowest BMDL of LCd (1.3 g) was somewhat lower than the representative threshold LCd (2.0 g) calculated in the previous studies. The obtained BMDLs may contribute to further discussion on the health risk assessment of cadmium exposure.