Chair: Cristina Neves
Room: S3A Barbakan
Date: 28 June
Speed Talk Session 9 - papers & presentations
|Antonio J. Rueda Clausell|
|Title: <<< Calendar effects in a monthly-based CATI survey >>>
CATI is the acronym of Computer Assisted Telephone Interviewing, a collecting method used in many surveys since it is cheaper than face-to-face methods, but it has the disadvantage that we do not know anything about the household or person we call until they are contacted. In any CATI survey, at the beginning of the collecting period, the sampling units have a probability of being contacted. This probability depends on many different factors, but in any case, it decreases as the period goes on and the most difficult units remain without contact. Since it is more difficult to contact these units, interviewers use less time to conduct interviews and have more time to make new telephone calls. As a consequence, the more difficult the contact is, the more calls can be done. This gives them the possibility to make more contacts attempts. We studied the evolution of the probability of contact in a monthly-based survey, and how it acts on the number of telephone calls. On the one hand, the number of respondents increases because they are contacted at some point, although some of them refuse when they notice the large number of missed calls they receive. On the other hand, interviewers are discouraged when they make a lot of unproductive calls. Knowing how the number of calls impacts on the response rate, we can size the call center to avoid the discouragement of the interviewers as well as increase the final effective sample. It was not the aim of this study to get the “optimal” number of interviewers throughout the collecting period, but only to do a descriptive analysis of the influence of this on the results of the survey.
|Title: <<< The economy has changed, the official statistics also change: the new business statistics system >>>
Changes that happened in the last decades to economic level have changed the economies of the single countries, included in a single globalized economic system. The globalization and the role of multinationals highlighted the fragility of the measuring systems of the main economic variables about official statistics. The system of legal units as analysis unit shows its limits. The economic process splitting of business groups, based on the structure of the legal units, highlights conceptual inconsistencies, measurement problems, linked to the duplications of flows and to the ancillary and vertically integrated units. The new system, implemented since 2019, requires the aggregation of the legal units of the same group, when they don’t have decision autonomy. This will impact significantly on the production system measurement. There will be a reduction in the number of companies, an increase in the size of the same and the reduction of some economic aggregates, because of the consolidation of the intra-flows. There is also a reshaping of value added and productivity per activity economic. The new plant will have a lot of implications and will invest all the areas that deal with economic statistics. In this perspective, the process foreseen a strong integration among administrative sources, business survey, business register and EGR. In this frame is strongly used a profiling approach. These changes will impact on territorial statistics. Starting from the national data, micro-level information must be obtained. The income approach is used to calculate value added. With this work we want to explain the whole process undertaken to define new estimation in the area of business statistics in Italy to implement new BS system.
|Biljana Ristevska Karajovanovikj|
|Title: <<< Improved quality in statistical production by using standard processes in phases: design, build and collect >>>
Metadata describe comprehensively statistical data and processes, assist in their interpretation and they are fully integrated in statistical processes and represent a prerequisite for developing advanced tools for statistical production, enabling high quality statistics. Standardization and harmonization of statistical variables and code lists are essential for improvement and standardisation of processes and at the same time, they represent a base for a single standardized production line for all surveys. Electronic methods offer new challenges and opportunities to improve the efficiency of statistical processes and get high quality incoming data, reducing costs at the same time. As a tool for electronic data collection, SSO is using eSTAT system. eSTAT is an IT system for automatic generation of web forms, on-line data collection, communication with respondents, data validation and loading collected data into final observation register. All logical expressions that are frequently used in the process of data validation have been covered by an incorporated engine for generation of the validation rules in eSTAT. The statistical task is responsible for creation of web form instance for the specified survey and its reference period, conducting the collection process according to defined dates and running mass e-mailer for confirmation and reminders for respondents. All activities related to the questionnaire generation and data collection have strictly determined order and they represent highly standardized processes in the eSTAT system.
|Title: <<< Towards a standard production process in the Spanish Industrial Turnover Indices >>>
One of the five key areas identified by the ESS Vision 2020 is to “promote efficiency in production processes” and one of the objectives within this area is to “further identify and implement standards for statistical production”. In the Industry Turnover Indices, INE Spain has already implemented several standards in its monthly statistical production tasks aiming at a streamlined standardized production process. These standards comprise the use of SDMX to transmit data to Eurostat following the corresponding international standard and JDemetra+ to conduct seasonal adjustment on a monthly basis. It also encompasses the description and documentation of the production tasks is carried out under the GSBPM, and the adoption of recently proposed data and process architectures at our office. The whole production process is described and documented using the GSBPM at the third level according to a national adaptation of this model. For every single production activity, input, output, throughput, tools, documentation and unit(s) responsible are detailed and specified. The metadata reference report, which aims at documenting methodologies, quality and the statistical production processes in general, is based on the Euro SDMX Metadata Structure. For part of the production of indices, in particular, for the editing phase, we are adopting a data and process architectures recently proposed internally at INE Spain. The data architecture is based upon a key-value pair data model in which keys are composed using the system of statistical metadata. The process architecture is based on a modular organization of different processes which are being easily reused in different surveys. We describe from a bottom-up approach how these different production standards and normalized tools in the European Statistical System are specifically orchestrated to reach a standard production process in this monthly short-term business statistics.
|Title: <<< Working time, underemployment and overemployment: two different data sources with contradictory results >>>
According to German FSO’s microcensus results for 2015, well over 2.7 million employed people aged 15 to 74 years wanted to work more hours, while 1 million persons in employment wanted to work less. This means that the number of people willing to work more hours was significantly higher than that of people wanting to work fewer hours. The underemployed, on average, would like to work an extra 11.3 hours per week, whereas the overemployed would like to reduce their working time by an average of 11.0 hours. For the same reference period the German Institute for Economic Research DIW Berlin, using a different measurement approach reported contradictory results: According to the so called SOEP, the German Socio-Economic Panel study at DIW Berlin, just under 5.3 million employed people aged 18 to 64 years wanted to increase their working hours whereas 18.0 million wanted to reduce their hours of work. On an average there are wishes for more work of 9.0 hours and for less work of 7.8 hours. As the conflicting results repeatedly led to critical press coverage, in late 2016, the DIW and the German FSO launched a joint analysis project to investigate the underlying reasons for the obvious discrepancies. The paper will present results from the analysis, focussing on the different elements that may impact on the measurement: The questionnaires of both surveys point out that an increase or reduction in working hours would involve correspondingly higher or lower earnings. What are the reasons for the completely different results? Can relevant key factors be identified for measuring working time and working time preferences?
|Title: <<< Electronic data collection in the Hungarian agricultural statistics >>>
In year 2016 the Hungarian Central Statistical Office (HCSO) focused on two main challenges: in summer Farm Structure Survey (FSS) was carried out, while the sample-based population enumeration, called Microcensus was carried in autumn. In the beginning of 2015 it was decided that these key surveys have to be carried out without using paper questionnaires. Two kinds of applications were developed: one for the online period, and one for the interview period. In case of FSS more than 15 percent of the data provider private holdings completed the questionnaire online out of those who had opportunity for this. The enumerators used almost the same application during the interview period. Only a small part of them used HCSO-owned tablets, the others were using their own devices. In case of the FSS the application included more than two thousand controls within and between tables: one part was running during the data entry and the others before the finalisation of the questionnaire. The questionnaire could be sent after all errors corrected. The system contains a monitoring which is a web-based application that allows to maintain the pre-loaded address list, monitoring the field work and questionnaire processing process, perform payment and reporting. Instead of wasting time on data entry after the survey, more time could be spent on quality control. Using electronic devices and e-questionnaire allowed that the data processing was faster and the results could be published two months earlier than three years before. The budget were less by 20 percent compared to the FSS 2013 since there were no cost for the data entry, and 93 percent less paper was used. Beside the mentioned advantages of this innovation, the environment was also winner since cca. 160 trees were saved with using only a small amount of paper.