EMOS reloaded: Unlock the future of education in official statistics with a new partnership with Universities

1.Introduction

EMOS is a European network of University Master programs in official statistics. EMOS is currently a unique and innovative path within various courses of study already active in universities. This is its peculiarity and novelty: there is no such path in the world, and together with the Joint Statistical Programme in Survey Methods [1] it represent an example of cooperation between Universities and National Statistical Institutes and other producers of official statistics.

The basic idea originated in Southampton in 2010, when a Group of Experts from Universities, NSIs and Eurostat started the discussion on the EMOS. For a clear description of the history of the EMOS, its actors, governance and procedures we refer to the paper by Annika Näslund [2]. The paper by Mojca Bavdaž and others, published in this special issue, gives additional hints on the EMOS network and its rationale [3].

Here we highlight that at the basis of the project there is the concept that there is a core of knowledge, skills and competences (quantifiable in ECTS and called Learning Outcomes) essential to those who aspire to be educated to work in official statistics, both in the production of data and in their analysis and dissemination and also to simply read and use and analyze data and Official Statistics. This “core” knowledge is defined after consultation, redesign, contributions and constraints by the Universities and the other Institutions involved in a process well described in the cited works [2, 3].

In this paper we try to see EMOS from the perspective of the Universities already in the network and eventually ready to join it, trying also to describe the role of the NSIs, Banks and other possible actors to join the network. Section 2 describes briefly our ideas on the emerging needs of statistical knowledge. In the last decade there has been an evolution of official statistics following the so-called data deluge. Also in University education, topics such as Data Science, Artificial Intelligence, Data Analytics have raised interest in the education of students and in their training. In fact, statistics and statistical literacy have also evolved. Section 3 describes how EMOS is implemented in the Universities and how should it be implemented to face the challenges, summarizing also what students and Universities says on their needs. Section 4 is dedicated to the description of the needs of the National Statistical Institutes and in general of statistical agencies and authorities in National Statistical Systems and in Not Governmental Organizations (e.g. Ministries, Statistical Departments, Chambers of Commerce, Banks) in terms of competences and skills of their employees comparing them with what EMOS offers. Section 5 concludes with our findings and recommendations.

2.Statistics and knowledge on data in the University programmes: An evolving scenario

Today life requires understanding many complex phenomena: globalization, with international trade, value chains, consequences on employment, real incomes; environment protection, quality of life and last but not least sustainability. Glocal is the adjective relating to the connections or relationships between global and local problems, businesses, etc [4]. Moreover Sustainable Development Goals launched by United Nations are tracing an idea of progress that requires new data and statistical analysis to monitor its development and the global and local target to reach [5].

The scenario is challenging and the need for specific competences in statistics and data analysis especially in the education of new generations is clearly urgent. In addition, an integrated and articulated information system that coordinates the data and the analysis on the phenomena and is able to allow for global and local perspectives is urgent as well. At the same time, our increasing daily dependence on technologies, social media and electronic transactions is producing huge amounts of data. The data produced refer to the most varied areas of natural, economic, and social sciences. They are labelled as “big” also because of their complexity (texts, diagnostic and satellite images, signals (and can provide information on events, behaviors and relationships that a few years ago were not easily measurable.

It is clear that Universities are following these changes and preparing adequate programmes of study to prepare the students to understand and possibly guide these complex phenomena and maybe contribute to the design of the information system. However, it is not clear according to which paradigms the Statistics and Data Science could and should analyze the phenomena in order to bring benefits to the society and to everyone’s life, as Melchiorre Gioia wished two centuries ago [6]. Especially for Official Statistics.

These are big challenges and they have been faced by the European Statistical System (ESS) finalizing the Trusted Smart Statistics strategy [7]. Anyhow, Universities in the EMOS network and any statistical organization (national, regional, and international) have a role of promoter of the understanding, development and good practice of statistics and official statistics.

Nothing new under the sun. The development of methods for statistical and socio-economic and demographic analysis has always been accompanied by the most modern storage and computing technologies available, which provided efficient and complex analyses of the data processed. Techniques like interaction of sources, data integration, data fusion, data mining, analysis of functional data, multivariate analysis, simulations, estimation algorithms, and so on, would not have been possible without a continuous dialogue and exchange with computer scientists. Already in the second half of the last century, there were discussions about data analysis and data driven analysis, learning by data, Data Science, and about the validity of the level of knowledge reached.

However nowadays statistical “thinking” is evolving fast as the landscape in which statistics operates is changing fast, as seen in previous paragraphs. The concept is evolving including not only traditional statistical literacy, but we need to give more space to concepts like data literacy and data awareness [8].

In our view we have evolved from statistical literacy to a more comprehensive concept named statistical thinking [9], that include also data literacy and data awareness.

Statistical literacy is the ability to understand and reason with statistics and data. Being statistically literate is sometimes taken to include having the abilities to both critically evaluate statistical material and appreciate the relevance of statistically-based approaches, including uncertainty in all the reasoning applied to all aspects of life in general or more specifically when evaluating, design, and/or produce scientific work (see the International Statistical Literacy Project by the IASE)

Data literacy is the set of skills to collect, manage, evaluate and apply data in a critical manner [10]. Generally speaking, the term ‘awareness’ can be defined as “the state or condition of being aware; having knowledge; consciousness.” For data awareness, we might refine this definition as “having knowledge of the data that exist on a phenomenon and an understanding of its contents.”

To think statistically is to think in distributions and probabilities (the need for statistical literacy) and also to understand the role of statistical data (data literacy and data awareness). Also data production process can generate uncertainty as it is subject to errors and misleading. Data production and data analysis should be considered within the greater machinery of generating scientific knowledge.

Official statistics agencies and statistical literacy educators have been working together in the past. See for example [17], where Statistics New Zealand, like some other NSOs, introduces a number of products designed specifically to support school statistics learning and a larger number designed for public and/or professional audiences. Statistics Education to promote statistical literacy at the University level is addressed in [18] and aspects for future research are stressed in terms of technology use, content and pedagogic approaches.

In the context of this evolving scenario, the EMOS project can play an important role in co-creating and reinforcing statistical thinking, in line with the European Data Strategy. The challenge for EMOS is to pick up this trend in its cooperation with the universities. The objective is to educate students and contribute to empower people, businesses, and organisations to make better decisions based on insights from official and non-personal data.

There are three aspects stemming out: i) the importance of the continuous exchange between different disciplines in university Master programmes; ii) the necessity of a pervasive statistical thinking in the university education; iii) the necessity that the “core” knowledge, competences, skills be adequate to the evolving scenario.

3.EMOS in the Universities

In general, we must admit that the Academy is not interested in issues related to the production of official statistical data. Already in 2015 Danny Pfeffermann noted this and cited EMOS as an example of University courses dedicated to the themes of official statistics [11].

The EMOS network builds on existing and nationally accredited Master of Science (MoS) programmes.

Very shortly: Universities can participate to a call to candidate a MoS to host the EMOS track. The 32 Master of Sciences now in the network range from MoS in Statistics, Survey methods to MoS in Economics, Finance and Data Science. The label is awarded by the European Statistical System Committee to the university Master programmes, certifying that the graduates familiarized with the system of official statistics, production models, statistical methods, and dissemination.

The title obtained by the students is not a degree in Official Statistics but a University degree labelled EMOS, certifying that in the course of study space has been given to that “core” of knowledge, able to familiarizing the student with the system of official statistics, production models, statistical methods and dissemination.

EMOS labelled Master programmes collaborate actively with the National Statistical Institutes or other producers of official statistics for relevant master thesis topics and internships in the sphere of official statistics.

This is the aim of EMOSin the scope of its cooperation with the universities. But there are new opportunities of picking up important developments where the universities are forerunners: universities are involved in European Programs, using modern online teaching tools, and are investing in laboratories. EMOS “reloaded” should make even more use of these developments coupled with eventual challenges in these type of cooperation.

There are needs and questions emerging by a pilot Focus Group conducted in few Universities already in the network and others not yet involved.11 They all knew the EMOS project. We reserve to repeat the experience enlarging it to the whole set of Universities of the network.

4.EMOS and the agencies’ needs

The EMOS network is now working in the field of education and cultural promotion of Official statistics more than in the field of specific training of future or present officers. Anyway, an essential part of the “core” knowledge is an internship in a statistical agency recognized by the European Statistical System. The topics of the internship are generally decided by tutors from the statistical agency and the University staff. It is desirable that the Master thesis is linked to the internship.

This is good as it is, as noted in other contributions [3] and it is a way to introduce the students in activities of a statistical agency. Obviously, it can be seen as a limitation and in a recent survey on the EMOS students22 the hint from their side is to open to the possibility of offering internships also to banks and other Institutions/agencies involved in the production and in the analysis of official data.

As the topic of the internship is the result of a “negotiation” between the University supervisor and the internal tutor of the hosting institution or agency it is important to gather the ideas on the needs of the institution hosting the student. In this section we try to understand and comment on these general needs underlying the “core” knowledge required and the emerging areas of interest.

However an important insight is that on the competences of the employees of a National Statistical Institute to understand if and at what extent these can be covered by the EMOS project.

In a recent experiment conducted at the Italian Statistical Institute a team of experts in human resources [12] have collected a large range of technical capabilities, knowledge, and expertise – not only statistical ones – needed to work in a national institute of statistics. The experts have defined six thematic areas: statistics, information technology, administrative, communication, organization, and foreign languages. Each area is subdivided into several subareas and each subarea provides a list of individual competencies. Overall, the classification consists of 29 subareas and 276 skills. It also provides a self-assessment scale of the mastery level structured in five modes: 0-none, 1-limited, 2-discret, 3-advanced, 4-expert.

For each of the thematic areas the skills were classified into owned by personnel, acted by personnel, expected by the Institute.

Not considering the administrative, organization and foreign language areas, the focus of EMOS is on statistics, information technology and communication.

However, in the long run, trust in official statistics can only be maintained based on a continual striving for the best quality, with leadership based on profound knowledge of the business and with customer orientation as the supreme orientation. This implies that statistical products must meet the expectations of users in their design, production, and communication [13]. In the following, we focus on four aspects that we think are currently the most important in terms of official statistic needs, shared also by statistical agencies of national statistical systems: Administrative sources, Big Data/Smart Statistics, Statistical Confidentiality, and Social Issues.

5.Conclusions

In his paper Pfeffermann identified the “core” topics of interest related to the production of official statistics as those contained in courses as Survey sampling, Seasonal adjustment (SA) and trend estimation, and National accounts. This because it is reasonable that the concepts taught in these courses are required in the background of an officer working in a statistical agency.

Today the “core” set defined by EMOS is larger involving also the system of official statistics, production models, including the enterprise architecture concepts applied to official statistics and treatment of big data, smart statistics, Statistical confidentiality, Administrative data and new social topics, but also time series analyses, seasonal adjustment, outlier treatment, index theory, multivariate statistics, econometrics, spatial statistics, knowing the concepts of metadata, paradata, data integration, critical capacity of framing analysis of statistical data.

A very large core set that requires an interdisciplinary approach and the support of the ESS to the Universities, together with a tight cooperation of researchers from statistical agencies and other data producers in teaching courses. Statistical thinking is evolved and still evolving, with a focus on data awareness and data literacy, and on the uncertainties and measurement errors emerging in the analysis of modern data sources.

But knowing this is not enough to design a university programme. Universities are asking for an Institutional support by the ESS to enlarge the sets of Master programmes eligible to host the EMOS track. This can be done using the already existing Erasmus and H2020 programmes, dedicating calls to the topics relevant to the education for official statistics and statistics in general; providing teaching materials and training for teachers and professors; also encouraging research and teaching innovation in partnership with National Statistical Institutes, Banks, and other statistical agencies, especially encouraging data laboratories and “big data sand-boxes”.

All in all EMOS should invest more in cooperation with Universities, through the involvement in European Programs. It is a matter of fact that collaboration among universities is a strength of the EMOS project. This must evolve and be reinforced also in view of the Alliances for innovation as supported by the Erasmus + programme. Europe’s innovation capacity can be boosted through stronger cooperation and knowledge flow among universities, official statistical agencies, businesses and the broader socio-economic environment, including research.

Statistics and statistical thinking is evolving and calls for designing European core curricula and training programmes on official statistics, building on the EMOS experience and reloading the EMOS project.