Big Data for Policy Analysis: The Good,The Bad,and The Ugly

Big data holds tremendous potential for public policy analysis. At the same time, its use prompts a number of issues related to statistical bias, privacy, equity, and governance, among others. Accordingly, there is a need to formulate, evaluate, and implement policies that not only mitigate the risks, but also maximize the benefits of using big data for policy analysis. This poses a number of challenges, which are highlighted in this essay.     

台湾地区科技统计体系及特征分析

科技统计数据是认识一个地区科技发展特征的重要载体.20世纪80年代初期,台湾地区建立了科技统计工作制度.如今,台湾的科技统计工作既符合OECD制定的统计规范,又具有自身特点,为公众了解台湾科技发展和当局制定科技政策提供了翔实的数据.本文在分析科技统计基本职能和世界科技统计发展态势基础上,对台湾的科技统计工作体系和工作流程进行了详细介绍,总结了台湾科技统计的四个典型特征,提出了对大陆科技统计工作的若干启示,并对开展两岸科技统计合作进行了展望.     

关于旅游统计的几点思考

旅游统计对各级政府的作用,主要是描述旅游发展态势、旅游业与其它产业发展的关系、旅游发展对旅游公共服务的需求程度等。旅游统计需要先解决两个问题:其一是旅游统计的口径问题,其二是旅游统计数据的取得问题。笔者建议国家旅游局和国家统计局将年末就业人口数指标列入旅游统计体系,并使之成为核心指标。     

Using the likelihood ratio in bloodstain pattern analysis

There is an apparent paradox that the likelihood ratio (LR) approach is an appropriate measure of the weight of evidence when forensic findings have to be evaluated in court, while it is typically not used by bloodstain pattern analysis (BPA) experts. This commentary evaluates how the scope and methods of BPA relate to several types of evaluative propositions and methods to which LRs are applicable. As a result of this evaluation, we show how specificities in scope (BPA being about activities rather than source identification), gaps in the underlying science base, and the reliance on a wide range of methods render the use of LRs in BPA more complex than in some other forensic disciplines. Three directions are identified for BPA research and training, which would facilitate and widen the use of LRs: research in the underlying physics; the development of a culture of data sharing; and the development of training material on the required statistical background. An example of how recent fluid dynamics research in BPA can lead to the use of LR is provided. We conclude that an LR framework is fully applicable to BPA, provided methodic efforts and significant developments occur along the three outlined directions.     

A Bayesian Approach to Age Estimation in Modern Americans from the Clavicle*

Abstract: Clavicles from 1289 individuals from cohorts spanning the 20th century were scored with two scoring systems. Transition analysis and Bayesian statistics were used to obtain robust age ranges that are less sensitive to the effects of age mimicry and developmental outliers than age ranges obtained using a percentile approach. Observer error tests showed that a simple three‐phase scoring system proved the least subjective, while retaining accuracy levels. Additionally, significant sexual dimorphism was detected in the onset of fusion, with women commencing fusion at least a year earlier than men (women transition to fusion at approximately 15 years of age and men at 16 years). Significant secular trends were apparent in the onset of skeletal maturation, with modern Americans transitioning to fusion approximately 4 years earlier than early 20th century Americans and 3.5 years earlier than Korean War era Americans. These results underscore the importance of using modern standards to estimate age in modern individuals.     

Comparison of Frequentist Methods for Estimating the Total Weight of Consignments of Drugs

Abstract: A topic in forensic statistics is the estimation of the total weight of consignments of drugs based on subsamples of which a certain fraction may not contain drugs at all. The frequentist approach to this concentrates on obtaining confidence intervals for the total weight, based on estimation of the fraction of drugs and the mean and variance of the weights of drug units. The current study shows that the resulting confidence intervals are basically unreliable, since they are based on an underestimation of the variation of the underlying statistical process. Two alternatives are given that yield asymptotically correct results. These are not reliable for small subsamples either, though, because of the inherent multimodal behaviour of the sample mean. In cases where a relatively large fraction of the consignment contains no drugs, the confidence intervals reported in the literature should not be used in practice.     

Benford's Law for Quality Assurance of Manner of Death Counts in Small and Large Databases

To assess if Benford's law, a mathematical law used for quality assurance in accounting, can be applied as a quality assurance measure for the manner of death determination. We examined a regional forensic pathology service's monthly manner of death counts (N = 2352) from 2011 to 2013, and provincial monthly and weekly death counts from 2009 to 2013 (N = 81,831). We tested whether each dataset's leading digit followed Benford's law via the chi‐square test. For each database, we assessed whether number 1 was the most common leading digit. The manner of death counts first digit followed Benford's law in all the three datasets. Two of the three datasets had 1 as the most frequent leading digit. The manner of death data in this study showed qualities consistent with Benford's law. The law has potential as a quality assurance metric in the manner of death determination for both small and large databases.     

Z‐Transform Method for Pairwise Osteometric Pair‐matching,

A new pairwise osteometric pair‐matching approach based on the Z‐transform method is presented. In contrast to previous methods that perform a global t‐test on the summed skeletal element pair measurement distances, this approach performs t‐tests on each individual distance, facilitating the capture of measurement‐specific variation. This new approach is compared to published pairwise sorting methods using a standard reference dataset of postcranial remains maintained by the Defense POW/MIA Accounting Agency. Receiver operating characteristic curve analysis indicates significantly improved performance for the clavicle and radius over all previous methods (p < 0.01). The z‐transform method weighted by the effect size outperformed the t‐test (Byrd and Adams) and the mean t‐test (Lynch) for all elements (p < 0.01). The method performed better than the absolute value t‐test (Lynch) for five elements (p < 0.01) and performed at least as well for the remainder. To facilitate usability all methods are available at: https://github.com/spawaskar-cora/z-transform-method .     

关于做好公安院校综合统计工作的思考

文章结合贯彻《中华人民共和国统计法》和公安院校的教育改革及管理工作的实际情况,就如何做好公安院校综合统计工作的前提、措施、主要工作方面及其努力方向作了有益的探讨。