The International Program in Survey and Data Science (IPSDS): A modern study program for working professionals

1.Introduction

Data on the characteristics and behaviors of individuals often serve as a basis for decision-making in the public sector. These data can be divided into data obtained with a particular survey design and those that arise organically as a byproduct of administrative and digital processes. While the use of data from administrative procedures has been well-established in Europe, the use of new digital trace data from, for example, websites collected through web scraping or through cooperations with mobile network operators, is still in an early phase in official statistics [1, 2]. These digital trace data can be roughly defined as “records of activity (trace data) undertaken through an online information system (thus, digital)” [3]. Recent years have shown that these new digital data sources will not replace traditional surveys completely, as these data often cannot cover all aspects: Important socio-demographic variables in the social sciences like household size, education, income, or employment status are in general not part of digital trace data [4]. Still, the increased interest in the use of new and big data sources in official statistics [5] and the linking and matching of data [6, 7] has led to an increased demand for well-trained data analysts. These data analysts need both the skills to collect and analyze data from surveys and work with large semi-structured or even unstructured data sets. These skills are often not taught or are taught only on a basic level in social science or economic programs. At the same time, there is a rapidly growing number of data science programs (see [8] for a recent overview of Data Science programs in German-speaking countries). Nevertheless, the training in Data Science programs is often very tech-oriented, and the link to methodological issues in official statistics is usually neglected. This is where the International Program in Survey and Data Science (IPSDS) comes into play.

IPSDS meets the demand for experts in the diverse fields of new and old forms of data collection, data analysis, data storage, and data output through a part-time master’s degree program. The starting point for the program was the two-year Joint Program in Survey Methodology (JPSM), which has been offered at the University of Maryland since 1993 together with the University of Michigan and the Westat cooperation. IPSDS was anchored in Germany at the University of Mannheim in 2016, facilitated through strong ties between the faculty in Maryland and Mannheim and based on the experiences from the JPSM program. IPSDS provides a flexible online learning environment and courses are held by instructors from both universities (as well as other universities and organizations). After several years of constant development and testing, the curriculum developed by IPSDS is now offered as the Mannheim Master in Applied Data Science and Measurement (MDM) by the Mannheim Business School. Students can take IPSDS courses through two full Master’s programs, JPSM at the University of Maryland and MDM at the University of Mannheim and the Mannheim Business School.

Unlike many continuing education programs, the program draws heavily on flipped classroom format where students learn asynchronously in advance of the class with prerecorded videos and other materials. The synchronous class time is reserved for problem-based, collaborative learning [9, 10]. This design allows the students to arrange their time resources more freely. IPSDS implemented this flipped classroom design into an online program accessible from anywhere in the world and students can adapt their studies to their professional and personal responsibilities and combine profession, studies and family.

The target group for the Master’s program are working professionals from all over the world who are professionally involved or interested in data collection and data analysis. They are employed in various fields, such as applied demographic, sociological, and economic research. Former and current students include professionals working in official statistics in different counties, data analysts at research institutes, and lecturers in various substantive fields at universities. They use data or are in charge of data collection from surveys or new data sources. Most students expect IPSDS to provide them with the necessary contents and skills to further develop in their careers in survey and data science [11].

IPSDS courses also cover the set of learning outcomes defined by the European Master in Official Statistics (EMOS) network. EMOS is a label of European Master’s programs providing post-graduate education in official statistics, which is awarded by the European Statistical System Committee (ESSC). The MDM program at the University of Mannheim was awarded the EMOS label under conditions and recommendations in May 2021. Several IPSDS faculty members are also working for official statistics institutions like Statistics Netherlands, Destatis, or the Institute for Employment Research (IAB), the Research Institute of the Federal Employment Agency in Germany.

We will start this article with an overview of the skills needed from data science, statistics, and survey methodology in official statistics and how these skills are covered in the IPSDS curriculum. We will also present the “inverted classroom” learning format used in the IPSDS program and demonstrate that it is especially suited for working professionals. We will also present a sample course from the curriculum and possibilities to receive so-called micro-credentials, that is, certificates for single courses or course bundles through IPSDS.

2.Methodological and statistical competencies needed in official statistics today

As Radermacher [2, p. 22] noted: “The decisive factor for the quality of statistics is the staff of the statistical institution.” The fast-changing world of data constitutes high requirements to the skill set of people working with it. The demand for data literacy is accelerated by the large amount of data that is amassed in the digital age and the expectations of organizations to make use of it [8]. Based on the American Association for Public Opinion Research (AAPOR) task force report on big data in survey research [12, 13], IPSDS identified five different areas of competence for working successfully with data from different sources under current developments in the world of data (Keusch and Kreuter 2020). These five areas will be presented in the following.

2.3Data curation and storage

Once the relevant data for answering a research question have been identified and collected, they must be processed, organized, and stored before analysis. In larger organizations such as statistical offices, these tasks are often solved with the help of I.T. specialists [22]. For analysts, basic SQL knowledge is vital to load data from a database into the analysis software. The further needed data wrangling and munching skills will differ by type of data. Survey data might require consistency checks [23] or the imputation of missing data [24], non-tabular data might require restructuring, and text data often need several pre-processing steps before analysis [25].

Figure 1.

MDM curriculum.

3.Curriculum

The IPSDS program has been offered as the accredited Mannheim Master of Applied Data Science & Measurement (MDM) at the Mannheim Business School since 2020. MDM is a 30 months, 90 ECTS22 credits program taught in English and leads to a Master of Science Degree (see Fig. 1).

The 90 ECTS credits include 60 ECTS credits to be acquired through courses, from which 54 ECTS credits are mandatory courses that cover the essential competencies from the five focus areas. The students can select courses worth additional 6 ECTS credits as electives according to their personal preferences and professional needs. Every course belongs to one or several of the focus areas presented in the previous section. Following the constructive alignment principle [36], the syllabus of each course states the targeted learning outcome, teaching and learning methods used, and type of assignments and grading. The students also have to complete a project report worth 10 ECTS credits, using the skills and knowledge they gained from the program within their field of work. The program concludes with final written and oral exams and the Master’s project of 20 ECTS credits. The elements of the program will be briefly presented in the following.

3.3Project report

IPSDS encourages students to transfer their newly acquired knowledge from the courses to their workplace and apply those acquired competencies under real-world circumstances through internships or work projects. Since IPSDS students complete the study program while being employed, they can use their new skills at their own workplace. These work or external projects are rewarded with credits. Students reflect in their project reports on how they have been able to apply the skills and methods learned in their place of work. For students who cannot use their own place of work for creating the project report (due to data protection or other reasons) or who want to use the opportunity to work in a different environment, the program provides the students with a project, e.g. through traineeships. This is secured due to established cooperation with practice partners, mainly from official statistics, like the German Federal Bank and the Institute for Employment Research. They will introduce students to current research projects on which students can work as well as provide support and counseling during the project. This close collaboration between students and program partners is also generating valuable results for the cooperating institutions.

Figure 2.

Screenshot from a live meeting supported by the whiteboard function [44].

4.Learning format

Traditionally, education in survey methodology and social research has been the domain of established universities, and these have also responded to the increasing demand for expertise in data science and big data [8]. However, many non-university providers have started to offer online education and training in this area as well. Online learning platforms such as Coursera (https://www.coursera.org/), edX (https://www.edx.org/), Khan Academy (https://www.khanacademy.org/) or Udacity (https://eu.udacity.com/) offer a wide range of courses from different disciplines, including courses on working with data from new but also traditional sources. DataCamp (https://www.datacamp.com/) allows students to interactively learn how to program with R, Python, and SQL. Courses offered by these organizations are so-called massive open online courses (MOOCs) and direct interaction with the lecturers is usually not part of the course. Engagement with the learning content is typically limited to watching prerecorded videos and solving standardized, small tasks. This self-directed learning requires high motivation on the part of students. Thus, many MOOCs struggle with extremely high dropout rates [37, 38]. Some MOOC providers have tried to move towards paid certificates and offered more direct interaction with lecturers in recent years. In contrast, universities have moved towards recognizing the importance of life-long learning and offering certificates for singular courses without requiring the participants to complete a full bachelor’s or master’s degree [8].

The IPSDS program is already situated between traditional study programs and MOOC courses. It heavily relies on the flipped classroom design (also known as inverted classroom). The flipped classroom has two main characteristics: All course materials including prerecorded video lectures and readings are put online. This allows students to learn asynchronously at their own pace and to coordinate one’s employer (or family) demands and studies. In addition to those asynchronous lectures, students use discussion forums and have live meetings in regular intervals with their lecturers. This allows students to work on specific content together with lecturers and other students [10]. Each week, students participate in a 50-minute online session hosted by the instructor (see Fig. 2). IPSDS courses usually include only eight to fifteen students and the online sessions enable the discussion of students’ questions as well as difficulties in assignments and projects. They also motivate students to continue to participate in the course actively. Screen-capture software, desktop sharing, and virtual meeting rooms form a technical foundation with which instructors can effectively implement these forms of teaching. Instructors can track the progress and understanding of all students at each stage of learning.

IPSDS thus uses a hybrid format between traditional university programs and completely self-paced learning. Due to the shortened synchronous in-class time, it is also especially suitable for online classes. At the same time, the (short) meetings still allow guidance and counsel for the students, contrary to MOOCs where students study primarily on their own.

IPSDS also allows students to access a wide network of peers in the private and public sectors; a significant part of the student body works in official statistics, for example. The kick-off event Connect@IPSDS, which takes place once a year in Mannheim, plays a crucial role in that regard. This event with courses and lectures offers the opportunity to get to know each other in person before the first classes take place online or catch up with other students later.

5.An example from the IPSDS curriculum: The Data Confidentiality and Statistical Disclosure Control course

To provide a better insight into the study program, we will present one of the courses and the skills taught in more detail. The course Data Confidentiality and Statistical Disclosure Control was developed and is currently taught by Jörg Drechsler from the Institute for Employment Research (IAB), the Research Institute of the Federal Employment Agency in Germany. Most data collecting and providing institutions in official statistics like the IAB face the dilemma between giving broad data access to researchers and the general public and at the same time ensuring data confidentiality. While statistical disclosure control is an integral part of data dissemination in official statistics, few courses or workshops are available as part of regular statistics study programs. Thus, the course has been an important part of the curriculum, especially for those interested in official statistics, and is currently offered every spring.

The prerecorded lectures of the course offer a broad overview of traditional statistical disclosure control methods like local suppression, rounding, microaggregation, swapping or sampling [39]. The instructor also introduces a relatively new approach to statistical disclosure control: synthetic data sets [40]. In the online meetings, the instructor and the students typically discuss open questions from the discussion forums, the assignments, and programming examples. The discussions in the online forum benefit from the sharing of experiences of those employed at institutions that need to find the right balance between data confidentiality, analytical validity of the data, and data access. By the end of the course, students know which measures are typically taken by statistical agencies in their own and other countries to guarantee confidentiality and are aware of potential limitations. They are informed about the concept of synthetic data and can also generate synthetic data sets themselves and evaluate the disclosure risk of the generated data. Grading in the course is based on participation as well as quizzes on the more theoretical aspects of data confidentiality and assignments, which in part require practical R programming.

6.Micro-credentials

One of the key concerns in post-graduate programs for working professionals is that students have a limited time budget for further education. While the IPSDS learning format was specifically developed to accommodate the busy schedule of working professionals, part of the targeted group is still unable to complete an entire master’s program. Nevertheless, professionals employed in official statistics, research institutes, universities, or other organizations, are often interested in broadening their skill set and knowledge. And while traditional university degrees remain very important, ‘there is an increasing need for alternative credentials that can better document the informal, online, and lifelong development of skills and knowledge both inside and outside the university’ [41, p. 593]. So-called “micro-credentials” have thus become much more important and common in recent years. Those smaller learning units represent the mastery of a limited set of skills or competencies rather than broader and interrelated set of skills, like full bachelor degrees” [42, p. 2].

The increasing interest in micro-credentials can also be observed in official statistics: The topic was discussed in the January 2021 board meeting for the European Master in Official Statistics (EMOS).33 While the board stated some caution regarding the minimum requirements and the quality of the material, which is also discussed controversially in educational literature [43], the EMOS-Board acknowledged micro-credentials in their January discussion as “one way forward for personalizing teaching and providing certificates for smaller amounts of learning, especially for already employed professionals.”

Figure 3.

MDM course bundles.

IPSDS has registered many requests from people who are very interested in one or a few IPSDS courses but are not able to commit to the whole Master’s program. Therefore, since summer 2021, IPSDS offers micro-credentials through the Mannheim Business School as so-called open courses. External participants can book all courses from the IPSDS curriculum individually. The open course participants join the students of the Master’s program for their selected course(s) and receive the same qualified learning experience in a proven online environment. After passing the course, they receive a certification of completion. For people interested in developing their skill set in a specific area, the program also offers course bundles (see Fig. 3), such as the course bundle on working with missing survey data. This bundle includes the two courses on item nonresponse and multiple imputation and an introductory R course since the other courses require a certain degree of R programming skills.

7.Summary

New sources of data and analysis methods provide opportunities to deliver more efficient and effective official statistical services. However, extracting relevant and reliable information from new and traditional sources and incorporating new and traditional data analysis methods into the statistical production process is a challenging task. These developments require those in charge of data generating processes, data storage, data analysis, and data output to periodically update their skills and knowledge in the area of survey and data science.

For working professionals in official statistics, IPSDS offers a genuine online learning environment accessible from anywhere in the world since 2016. Both a master’s program, as well as certificates for single courses and bundles of courses, are available. One of the most important fields the students can apply new skills to is the field of official statistics. When the Master’s program was formally institutionalized as Master of Applied Data Science and Measurement at the Mannheim Business School and the University of Mannheim, program management applied to become certified as European Master in Official Statistics (EMOS)44 and was able to receive the label in May 2021 under conditions and recommendations.

The program emphasizes interaction between faculty and students as well as networking and working in (international) groups. A significant part of the faculty and the student body is currently working or has worked in statistical agencies in Europe and the US. The coordination of the IPSDS training with work and leisure time is possible through the flipped classroom design. Classes are a mix of synchronous and asynchronous learning in small groups of no more than fifteen students. Additional courses like the project consulting course, internships, work projects, and the Master’s project complement the Master’s program focused on applications in real-world data settings. The program covers data collection, data analysis, dissemination, and data quality, taking into account various data sources and their combination. Elective courses offer students insights into new developments like big data or geospatial data in official statistics complementing the skills they can gain in mandatory courses.

In the evaluations, Students have mentioned several aspects of the program that they appreciate. The supporting network of students who are all professionals in the same or adjacent fields is often stressed as an asset, as well as the courses that are divided into small segments that allow flexible learning. The variety of classes and cutting-edge topics that are covered are another strength that is frequently emphasized.

For education and training, there is now a wide range of offerings from universities and providers from the field of online learning. Formats range from single, free-of-charge online courses that can be taken with thousands of other students from around the world (MOOCs) to traditional face-to-face academic bachelor’s or master’s programs. In between, however, we believe there is also demand for hybrid programs such as IPSDS, that combine the flexibility of online learning with interaction in an online classroom setting, especially regarding training in official statistics.