Journal of Official Statistics最新文献

{"title":"Preface","authors":"F. Di Iorio, E. Baldacci, Dario Buono, Luca di Gennaro Splendore, D. Elliott, Rebecca Killick, Tiziana Laureti, M. Pratesi, N. Shlomo","doi":"10.2478/jos-2021-0011","DOIUrl":"https://doi.org/10.2478/jos-2021-0011","url":null,"abstract":"","PeriodicalId":51092,"journal":{"name":"Journal of Official Statistics","volume":"37 1","pages":"257 - 260"},"PeriodicalIF":1.1,"publicationDate":"2021-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43113548","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Economic Nowcasting with Long Short-Term Memory Artificial Neural Networks (LSTM)","authors":"D. Hopp","doi":"10.2139/ssrn.3855402","DOIUrl":"https://doi.org/10.2139/ssrn.3855402","url":null,"abstract":"Abstract Artificial neural networks (ANNs) have been the catalyst to numerous advances in a variety of fields and disciplines in recent years. Their impact on economics, however, has been comparatively muted. One type of ANN, the long short-term memory network (LSTM), is particularly well-suited to deal with economic time-series. Here, the architecture’s performance and characteristics are evaluated in comparison with the dynamic factor model (DFM), currently a popular choice in the field of economic nowcasting. LSTMs are found to produce superior results to DFMs in the nowcasting of three separate variables; global merchandise export values and volumes, and global services exports. Further advantages include their ability to handle large numbers of input features in a variety of time frequencies. A disadvantage is the stochastic nature of outputs, common to all ANNs. In order to facilitate continued applied research of the methodology by avoiding the need for any knowledge of deep-learning libraries, an accompanying Python (Hopp 2021a) library was developed using PyTorch. The library is also available in R, MATLAB, and Julia.","PeriodicalId":51092,"journal":{"name":"Journal of Official Statistics","volume":"38 1","pages":"847 - 873"},"PeriodicalIF":1.1,"publicationDate":"2021-03-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46757061","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Improving the Output Quality of Official Statistics Based on Machine Learning Algorithms","authors":"Q. Meertens, C. Diks, J. Herik, Frank W. Takes","doi":"10.2478/jos-2022-0023","DOIUrl":"https://doi.org/10.2478/jos-2022-0023","url":null,"abstract":"Abstract National statistical institutes currently investigate how to improve the output quality of official statistics based on machine learning algorithms. A key issue is concept drift, that is, when the joint distribution of independent variables and a dependent (categorical) variable changes over time. Under concept drift, a statistical model requires regular updating to prevent it from becoming biased. However, updating a model asks for additional data, which are not always available. An alternative is to reduce the bias by means of bias correction methods. In the article, we focus on estimating the proportion (base rate) of a category of interest and we compare two popular bias correction methods: the misclassification estimator and the calibration estimator. For prior probability shift (a specific type of concept drift), we investigate the two methods analytically as well as numerically. Our analytical results are expressions for the bias and variance of both methods. As numerical result, we present a decision boundary for the relative performance of the two methods. Our results provide a better understanding of the effect of prior probability shift on output quality. Consequently, we may recommend a novel approach on how to use machine learning algorithms in the context of official statistics.","PeriodicalId":51092,"journal":{"name":"Journal of Official Statistics","volume":"38 1","pages":"485 - 508"},"PeriodicalIF":1.1,"publicationDate":"2021-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45294034","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Weighted Dirichlet Process Mixture Models to Accommodate Complex Sample Designs for Linear and Quantile Regression.","authors":"Michael R Elliott,&nbsp;Xi Xia","doi":"10.2478/jos-2021-0004","DOIUrl":"https://doi.org/10.2478/jos-2021-0004","url":null,"abstract":"<p><p>Standard randomization-based inference conditions on the data in the population and makes inference with respect to the repeating sampling properties of the sampling indicators. In some settings these estimators can be quite unstable; Bayesian model-based approaches focus on the posterior predictive distribution of population quantities, potentially providing a better balance between bias correction and efficiency. Previous work in this area has focused on estimation of means and linear and generalized linear regression parameters; these methods do not allow for a general estimation of distributional functions such as quantile or quantile regression parameters. Here we adapt an extended Dirichlet Process Mixture model that allows the DP prior to be a mixture of DP random basis measures that are a function of covariates. These models allow many mixture components when necessary to accommodate the sample design, but can shrink to few components for more efficient estimation when the data allow. We provide an application to the estimation of relationships between serum dioxin levels and age in the US population, either at the mean level (via linear regression) or across the dioxin distribution (via quantile regression) using the National Health and Nutrition Examination Survey.</p>","PeriodicalId":51092,"journal":{"name":"Journal of Official Statistics","volume":"37 1","pages":"71-95"},"PeriodicalIF":1.1,"publicationDate":"2021-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8415180/pdf/nihms-1720820.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"39384607","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Measuring and Modeling Food Losses","authors":"Marco Mingione, Carola Fabi, G. J. Lasinio","doi":"10.2478/jos-2021-0008","DOIUrl":"https://doi.org/10.2478/jos-2021-0008","url":null,"abstract":"Abstract Within the context of Sustainable Development Goals, progress towards Target 12.3 can be measured and monitored with the Food Loss Index. A major challenge is the lack of data, which dictated many methodology decisions. Therefore, the objective of this work is to present a possible improvement to the modeling approach used by the Food and Agricultural Organization in estimating the annual percentage of food losses by country and commodity. Our proposal combines robust statistical techniques with the strict adherence to the rules of the official statistics. In particular, the case study focuses on cereal crops, which currently have the highest (yet incomplete) data coverage and allow for more ambitious modeling choices. Cereal data is available in 66 countries and 14 different cereal commodities from 1991 to 2014. We use the annual food loss as response variable, expressed as percentage over production, by country and cereal commodity. The estimation work is twofold: it aims at selecting the most important factors explaining losses worldwide, comparing two Bayesian model selection approaches, and then at predicting losses with a Beta regression model in a fully Bayesian framework.","PeriodicalId":51092,"journal":{"name":"Journal of Official Statistics","volume":"37 1","pages":"171 - 211"},"PeriodicalIF":1.1,"publicationDate":"2021-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44622863","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Editorial Collaborators","authors":"","doi":"10.2478/jos-2020-0047","DOIUrl":"https://doi.org/10.2478/jos-2020-0047","url":null,"abstract":"","PeriodicalId":51092,"journal":{"name":"Journal of Official Statistics","volume":"1 1","pages":""},"PeriodicalIF":1.1,"publicationDate":"2020-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"47013299","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Book Review: Paul J. Lavrakes, Michael W. Traugott, Courtney Kennedy, Allyson L. Holbrook, Edith D. de Leeuw, and Brady West. Experimental Methods in Survey Research: Techniques that Combine Random Sampling with Random Assignment. 2019, Wiley, ISBN: 978-1-119-08374-0, 544 pages","authors":"K. J. Thompson","doi":"10.2478/jos-2020-0046","DOIUrl":"https://doi.org/10.2478/jos-2020-0046","url":null,"abstract":"Qualitative and quantitative research are entrenched in survey methodology. Certainly, qualitative research tools that use small datasets such as focus groups and onsite visits of businesses provide invaluable information in assessing construct validity in survey instrument design, as does usability testing. However, these tools are limited, in the sense that they rely on observational data, are purposively gathered, and should not be used for any form of causal inference. At the opposite end of the spectrum, large amounts of data can be easily gathered via the internet using probability or nonprobability samples such as online convenience panels; the resulting statistics are time-dependent measures, generally considered to be externally valid. Again, however, these vehicles do not allow for causeand-effect analysis. Experimental Methods in Survey Research is an extensive collection of papers gathered together for one express purpose: to convince survey researchers to utilize experimental designs combining random assignment of subjects to treatment within random (probability) samples. Internal validity is attained via the random assignment of treatments, using blocking in the assignment process to control for “known” factors. Using a probability sample combined with successful recruitment methods should achieve external validity, assuming that the sampling frame covers the target population (i.e., has low coverage error) and that the response mechanism is ignorable (i.e., respondents are a random subsample). In assembling and organizing these chapters, the editors “embraced both the Campbell and Stanley validity framework” for research studies (statistical conclusion validity, construct validity, external validity, and internal validity) as well as the key components of the total survey error (TSE) framework, specifically coverage error, sampling error, nonresponse error, and measurement error. Thanks to the extensive usage of computer-assisted technologies in survey collection and the prevalence of internet access in the majority of target populations, the editors argue that there is little or no cost increase in embedding true experiments within ongoing programs, and therefore such tests should be carefully planned and executed on a wide scale in common practice. The book is divided into nine topical areas, comprising twenty-three chapters (the first chapter is an introduction). Each chapter covers one experimental research topic, providing a literature review and presenting at least one case study. Not perhaps","PeriodicalId":51092,"journal":{"name":"Journal of Official Statistics","volume":"36 1","pages":"941 - 943"},"PeriodicalIF":1.1,"publicationDate":"2020-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42254519","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Double Barreled Questions: An Analysis of the Similarity of Elements and Effects on Measurement Quality","authors":"Natalja Menold","doi":"10.2478/jos-2020-0041","DOIUrl":"https://doi.org/10.2478/jos-2020-0041","url":null,"abstract":"Abstract In double barreled questions (DBQs) respondents provide one answer to two questions. Assumptions how respondents treat DBQs and how DBQs impact measurement quality are tested in two randomized experiments. DBQs are compared with revisions in which one stimulus was retained while the other stimulus was skipped. The observed means and parameters when modeling latent variables differed among the versions. Metric and scalar measurement invariance was not given among the versions, and at least one single stimulus version was found to be associated with a higher validity. Response latencies did not differ among versions or respondents needed less time to respond to DBQs. The author concludes that respondents may understand the stimuli in a DBQ differently, and access one of them while disregarding the other, which can have an adverse effect on validity.","PeriodicalId":51092,"journal":{"name":"Journal of Official Statistics","volume":"36 1","pages":"855 - 886"},"PeriodicalIF":1.1,"publicationDate":"2020-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43824654","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The Representativeness of Online Time Use Surveys. Effects of Individual Time Use Patterns and Survey Design on the Timing of Survey Dropout","authors":"Petrus te Braak, Joeri Minnen, I. Glorieux","doi":"10.2478/jos-2020-0042","DOIUrl":"https://doi.org/10.2478/jos-2020-0042","url":null,"abstract":"Abstract Like other surveys, time use surveys are facing declining response rates. At the same time paper-and-pencil surveys are increasingly replaced by online surveys. Both the declining response rates and the shift to online research are expected to have an impact on the representativeness of survey data questioning whether they are still the most suitable instrument to obtain a reliable view on the organization of daily life. This contribution examines the representativeness of a self-administered online time use survey using Belgian data collected in 2013 and 2014. The design of the study was deliberately chosen to test the automated processes that replace interviewer support and its cost-efficiency. We use weighting coefficients, a life table and discrete-time survival analyses to better understand the timing and selectivity of dropout, with a focus on the effects of individual time use patterns and the survey design. The results show that there are three major hurdles that cause large groups of respondents to drop out. This dropout is selective, and this selectivity differs according to the dropout moment. The contribution aims to provide a better insight in dropout during the fieldwork and tries to contribute to the further improvement of survey methodology of online time use surveys.","PeriodicalId":51092,"journal":{"name":"Journal of Official Statistics","volume":"36 1","pages":"887 - 906"},"PeriodicalIF":1.1,"publicationDate":"2020-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41790318","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparing the Ability of Regression Modeling and Bayesian Additive Regression Trees to Predict Costs in a Responsive Survey Design Context.","authors":"James Wagner, Brady T West, Michael R Elliott, Stephanie Coffey","doi":"10.2478/jos-2020-0043","DOIUrl":"10.2478/jos-2020-0043","url":null,"abstract":"<p><p>Responsive survey designs rely upon incoming data from the field data collection to optimize cost and quality tradeoffs. In order to make these decisions in real-time, survey managers rely upon monitoring tools that generate proxy indicators for cost and quality. There is a developing literature on proxy indicators for the risk of nonresponse bias. However, there is very little research on proxy indicators for costs and almost none aimed at predicting costs under alternative design strategies. Predictions of survey costs and proxy error indicators can be used to optimize survey designs in real time. Using data from the National Survey of Family Growth, we evaluate alternative modeling strategies aimed at predicting survey costs (specifically, interviewer hours). The models include multilevel regression (with random interviewer effects) and Bayesian Additive Regression Trees (BART).</p>","PeriodicalId":51092,"journal":{"name":"Journal of Official Statistics","volume":"36 4","pages":"907-931"},"PeriodicalIF":1.1,"publicationDate":"2020-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8174792/pdf/nihms-1694536.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"39066310","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}