Statistical Journal of the IAOS

The purpose of the 2020 Census is to conduct a census of population and housing and disseminate the results to the President, the states, and the American people. The goal of the 2020 Census is to count everyone once, only once, and in the right place, and the challenge is to do this at a lower cost per household than the 2010 Census, while maintaining high quality results.

The Census Bureau addressed this challenge first by identifying the major cost drivers of the decennial census. From 2013 through 2015, the Census Bureau conducted research and testing related to major innovations that showed promise of significant cost savings. Across four key innovation areas, the Census Bureau believes it can avoid over $5 billion in costs relative to the cost of repeating the 2010 Census design and operations in 2020. The Census Bureau used the results of the research and testing, other key information, and input from a wide variety of stakeholders to design the 2020 Census. The Census Bureau documented and published its 2020 Census Operational Plan on October 6, 2015, accompanied by revised lifecycle costs estimates and the overall 2020 Census lifecycle budget.

The U.S. Census Bureau is researching and testing new methods to reduce the cost of the 2020 Census while maintaining data quality. One of the most costly components of the 2010 Census was Nonresponse Followup. In this operation, enumerators conducted in-person interviews at housing units that did not return a census questionnaire by mail. For the 2010 Census, enumerators were instructed to visit each of these units up to three times until the case was resolved. Additionally, enumerators were to make up to three contact attempts by telephone. In this paper, we present an overview of current research on determining the number of contact attempts that should be made to nonresponding units with an emphasis on cost containment and improved overall productivity. Rather than the fixed contact strategy employed in the 2010 Census, we consider adaptive approaches that maintain the quality of the data. We present initial results of possible approaches using data from the 2010 Census and discuss the implications of the methods. We also discuss modeling contact probabilities for each hour of the day to support the case management system.

For the 2020 Decennial Census, the U.S. Census Bureau is researching how to use administrative record information from government and other sources in place of field visits during the Nonresponse Followup (NRFU) operation. This paper describes an approach for identifying vacant and occupied housing units to be enumerated using administrative records, removing them from the NRFU workload. While the approach allows flexibility in balancing cost and quality, we evaluate one possible scenario via a retrospective study of the 2010 Census.

For the 2010 Census, the count imputation procedure filled in housing unit status and size for the small proportion of addresses (less than one-half percent) where this information was unknown. The small proportion was due in part to an extensive nonresponse followup (NRFU) field operation geared towards resolving addresses so that a status and count were known.

For 2020, the Census Bureau is researching two changes to the NRFU field operation to reduce cost. The first is the possible use of administrative records (AR) to provide a status and count for some nonresponding addresses. The second is potentially reducing the number of visits made to nonresponding addresses. Although using AR will help resolve some of the remaining unresolved cases, the proportion of addresses in need of count imputation may be higher in 2020 due to the reduction in NRFU fieldwork.

The 2010 count imputation model was developed assuming a small amount of missing data. This research looks at potential count imputation models to handle increased missingness. The paper also articulates the downstream characteristic imputation ramifications from the same missing data challenge.

Presently, many countries are discussing the future of official statistical data production. As a contribution to this discussion, we shall examine in this article a number of methodological aspects of a ``political economy of statistics'', focussing on ``statistical operationalization'', which we see as a central challenge for data production in the field of economic and social activities. In a ``political economy of statistics'' it is assumed that the producers and users of statistical data behave self-interested and will try to shape the statistical infrastructure to meet their personal needs, which do not necessarily coincide with the socially optimal form of data provision. As a result, individual rationality and collective rationality may fall apart and create welfare losses for society. In this contribution we therefore ask how the process of statistical data production can be organized to benefit society in total and not only specific interest groups. To that end we shall combine insights from political economy with insights from statistical operationalization.

This paper expounds the need for collating knowledge based economy (KBE) indicators in a holistic manner under the official statistical system. The paper argues, the official statistical user needs and demands have changed drastically in recent years. Rapid information communications technological (ICT) upheavals have made profound impacts on all spheres of life, as such the way one work, learn, interact, transact and communicate have embraced changes unprecedented in human history. The changes challenge not only public policy formulation activities but also raise questions on validity of certain statistical concepts and definitions as well as data aptness. The current official statistical system is based on agriculture and industrial settings and lack adequacy and comprehensiveness in depicting the information age developments. Beginning with the advent of online connectivity and real time interactive Internet features in early nineties, the KBE is incrementally shaped by broadband communication; borderless networking irrespective of geography and time; intensification of digitization processes and contents; miniaturization of computing devices enhancing business mobility and agility; new transactional and analytical models like outsourcing, cloud computing, cognizance computing, big data analytics and pay as you use e-services increasingly typify contemporary businesses. As such, changes in policy institutions, organizational structures, governance processes, people's lifestyles and millennial workforce new demands and behaviours have emerged, thus exerting unduly pressures not only on public policy strategies but also on national statistical system. Reckoning the imperatives, the paper is making a clarion call to scrutinize, review and realign the current official statistical system, including incepting totally new statistical surveys on gauging KBE. In support of the claim, the paper highlights the Malaysian KBE statistical experiences, wherever deemed relevant.

The Māori Statistics Framework (MSF) is a tool developed by Statistics New Zealand. This tool is known as He Arotahi Tatauranga. While it is a general resource for those working in the area of statistics for and about New Zealand's indigenous population (Māori) it is intended to be to be used primarily by Māori to organise and use their information in a way that supports their development and well-being consistent with their aspirations as a people.

The MSF has been derived from a conceptual paper that the project team then used to create a unique tool. Because the aims and concepts of the MSF come from the Māori view of well-being it can be used to define statistics from this viewpoint. The MSF also helps users understand the difference between traditional statistics that measure Māori from a standard approach and how to measure from a Māori perspective.

The MSF tool is an Excel spreadsheet with accompanying guidance and supplementary information in attached Word documents. The users of the MSF are guided in a transparent manner through:- How they can build their own information management system from a Māori perspective; - How to think statistically about the concepts and topics relevant to Māori development from a Māori viewpoint; - How to identify what type of measurement(s) would suit the user;- Whether or not an indicator already exists that could be used for such measurement.

National statistical institutes (NSIs) fulfil an important role as providers of objective and undisputed statistical information on many different aspects of society. To this end NSIs try to construct data sets that are rich in information content and that can be used to estimate a large variety of population figures. At the same time NSIs aim to construct these rich data sets as efficiently and cost effectively as possible. This can be achieved by utilizing already available administrative data as much as possible, and supplementing these administrative data with survey data collected by the NSI. In this paper we focus on one of the challenges when using a mix of administrative data sets and surveys, namely obtaining numerically consistent population estimates. We will sketch general approaches based on weighting, imputation and macro-integration for solving this problem, and discuss their advantages and drawbacks.

Statistical institutes are focusing on variety of data sources from traditional surveys to big-data. Many of these data and concepts can be expressed as crisp values. But many other data cannot be expressed by precise values. In order to collect, store and manage the fuzziness in data we have adapted the fuzzy meta model as an extension of traditional relational database. Furthermore, experts' knowledge often contains vagueness and subjectivity. If we store this knowledge in a fuzzy database we can build knowledge management systems capable to cope with fuzziness. Statistical institutes cooperate in the data exchange. We have briefly discussed a simple way of extending the SDMX standard to cope with the fuzzy data in a way that does not influence exchanging precise values. Our research was focused on examining promising ways for managing fuzziness of real world because statistical institutes have been starting to analyze variety of promising data sources where not all data are always precise.

Data editing is essential to check the survey data for possible data problems. Outlying data values are frequently encountered in sample surveys. Consequently, in working with data, the correctness of the reported values must be verified, and if a reported value constitutes an outlier, its appropriate treatment needs to be considered. In this paper, the Iterative method for the reducing the impact of outlying data points is proposed. The novelty of the Iterative method for the reducing the impact of outliers is the following: an iterative approach for determining the outlying data points is proposed; outliers are determined considering the impact of conjoined factors; estimation of weight coefficients of the outliers and estimation of the total measurement error of the non-linear regression model is carried out.