Matthias Mazenauer and Andrea Schnell; photo by Mirjam Kluka

The resident population, the gross domestic product (GDP) and the number of available hospital beds during the coronavirus pandemic are all three indicators. But what are indicators? How are they created and why are they needed? Answers from Andrea Schnell and Matthias Mazenauer, who work daily with indicators at the Cantonal Statistical Office Zurich.

Andrea and Matthias, you are co-directors of the Cantonal Statistical Office Zurich. Why are you and the Statistical Office participating in the Public Data Lab (PDL)?

Matthias: The Statistical Office is responsible for the public statistics, and statistics are based on data, which explains public data. We find the lab an exciting form of collaboration because it fits in with one of our guiding principles, experimentation. It opens up space for innovation, both in the development and communication of indicators. Collaboration with the University of Zurich (UZH) and the Zurich University of the Arts (ZHdK) was therefore an obvious choice.

Andrea: The PDL allows us to exchange ideas outside our social bubble and question the status quo. The world of data is very dynamic. More and more organisations and individuals are measuring things and publishing their data. As public statisticians, we have to hold our own against these providers and prove that we are measuring and publishing ‘the right’ data. For example, how do you measure poverty? Behind this lies a definition and a concept that are determined by the purpose and the frame of reference. It helps if we can exchange ideas about which definition is the right one for which purpose.

What is your role within the PDL?

Andrea: We are the societal partner within the PDL and bring interesting questions from our work into the project to discuss them with the other project partners and/or students. These can be questions where we, from our administrative perspective, have reached a limit and are looking for external input. Depending on the scope and complexity, this can lead to a research cooperation or a Master's thesis. Furthermore, we act as door-openers for researchers to other administrative units and other offices. At the Statistical Office, we work with people from all areas of the administration. We are happy to share these contacts.

What is your goal?

Matthias: Together with the UZH and the ZHdK, we want to work on relevant questions. As a team, we can tackle issues for which we lack the resources on our own. We can step out of our comfort zone and reflect on our work thanks to new perspectives. We work for the public, which means our indicators should be understandable and relevant. The PDL supports us with new input and broadens our horizon.

Andrea: Our language should be comprehensible. We might assume certain terms are known, which isn’t always the case. The PDL can support us in rephrasing issues differently so that we are better understood. Thanks to the public events planned by the PDL, we hope to draw attention to our work: What does it take to transform data generated by administrative processes into something that meets the quality standards of a public statistic? At the same time, we hope to get closer to our data users and communicate with them directly. This would be exciting for us.

Where do you see challenges?

Matthias: Conflicting goals could be a challenge, such as speed versus quality. Take a Master's thesis, for example: Students might want to finish their thesis as soon as possible, while we would prefer to reconsult and review it again. Artistic interpretations could also result in a conflict of objectives if they are very provocative or critical, whereas we, as an Office, emphasise politically neutral and unbiased communication.

Andrea: In this context, it is also important for us that the PDL is not politically biased. We are independent of parties and not normative. Of course, we address politically relevant issues, but we strive for thematic balance. Furthermore, I am curious how we will align different deadlines. For example, do the processes at the Statistical Office fit with a Master's Thesis that has a clear timeline? Are we expecting too much too early, or are we too slow? At the same time, I expect us to manage this together. Another challenge within the PDL is the language: Since we are from different disciplines, do we mean the same thing when we use the same terms?

“We try to make behaviour or development tangible through data and indicators.”

Andrea Schnell

The PDL wants to develop indicators. What are indicators?

Andrea: Indicators help us reduce the complexity of the world. We try to make behaviour or development tangible through data and indicators. The reduction in complexity is based on many definitions to make clear what is included and what is not. This means we simplify. Sometimes we also determine something that is vague by defining it.

Matthias: Indicators are tools. They represent something that cannot be directly measured. We can't just put out a cup and see how much rainwater collects in an hour. We develop concepts for how to measure a development and compare it across time and space.

Could you give an example of an indicator?

Andrea: The starting point is a question, for example, whether women and men are equal in the labour market and how their status has evolved over time. This can't be measured directly, but various indicators can provide clues: The employment rate tells us how actively men and women participate in the labour market. Wages show whether they receive equal pay for equivalent work. And their representation in company leadership gives us insight into the professional status of both genders. If these indicators are internationally coordinated and consistently calculated over time, we get reliable insights into how gender equality in the labour market is developing in an international context.

Matthias: The unemployment rate is an indicator that many people intuitively understand because it has become established. Many can interpret the current rate, but few know how it is actually defined. For this indicator to become established, various definitions were needed—such as what “unemployed” means, but also what “employed” means. Since it's a rate, it needs a reference point. And all parties involved had to agree on these definitions. Another example of an indicator that seems simple at first glance but can be defined in different ways is the population of the Canton of Zurich: Does it include only people registered in the canton? Where do we count weekly residents or people with a second residence? How do we handle asylum seekers who are assigned or reassigned? What is the reference date?

Andrea: The City of Zurich, the Canton of Zurich, and the Federal Statistical Office (FSO) each report a different population figure for the City of Zurich, because each institution defines the population differently. What’s important here is the intended use: Is it about infrastructure load, housing needs, taxes, etc.?

Do you have an example of a more complex indicator?

Andrea: Gross Domestic Product (GDP). It’s an internationally recognised concept, but many of the data points used to calculate this indicator are estimates. It takes time for the relevant input data to become available, which means GDP for a given quarter can only be determined in the following quarter. Another requirement for an indicator is that it should be available in a timely manner—though “timely” is relative. For elections and referendums, we know the results just hours after the polls close. During the COVID-19 pandemic, we calculated many indicators daily.

Matthias: During the pandemic, the public learned a lot about indicators because they were so present. But we also learned. For example, there was the indicator “available beds” in hospital intensive care units. Through public discussion, we learned as a society that this indicator doesn’t tell us how many patients a hospital can still admit. The bottleneck wasn’t the beds—it was the respirators and the nursing staff. At the same time, we couldn’t measure how much available nursing staff there was.

How is an indicator created?

Andrea: It’s very rare that we define an indicator on our own. Usually, an indicator is developed in response to a need and is based on professional exchange. Ideally, the indicator is then coordinated between the cantons and at the federal level so that everyone measures the phenomenon in the same way. For an indicator, data availability is a prerequisite: The data used in the calculation must be collected and available in all cantons.

Matthias: When the cantons agree on an indicator and what it measures, it gains weight and value. We can then compare figures across cantons and get a sense of what is considered low or high. We can also compare over time and see whether there’s a trend. The way indicators are created has also changed: From 1850 to 2000, there was a census every ten years, where the population was surveyed in writing on specific topics. Today, much more is measured indirectly, and indicators are derived from that.  

“We try to reflect long-term developments. What will be relevant in 10 years? And what data can we collect today for that?”

Matthias Mazenauer

How is comparability ensured?

Matthias: By agreeing on the definition of the indicator, the data characteristics, and how we measure—but also by regulating the operationalisation of the definition. For example, in the building and housing register, a directive precisely defines what constitutes a building, how it is recorded, etc.

Andrea: If the calculation basis changes, this is indicated in the statistics as a methodological break as of a specific reference date. We’ve seen this with various indicators that have been collected based on registers since the abolition of the census in 2000. But even switching from a telephone survey to an online survey leads to slightly different responses. Such methodological changes are important and are noted in the statistics.

Matthias: In general, we pay attention to having long time series. We try to look ahead to reflect long-term developments. What will be relevant in 10 years? And what data can we collect today for that?

Andrea: One example is the survey by the Federal Statistical Office (FSO) on how many households in Switzerland have internet access. Today, that question seems odd, but in the first year of the survey, 2003, only 61% had internet access. The survey was discontinued in 2023 because nearly 100% of households now have an internet connection. There’s also a statistic on working from home—called “telehomeworking” by the FSO—since 2014. What was once a niche phenomenon is now widespread.
Having these long-time series—or more generally, having a public, independent statistics system—is no longer something we can take for granted. You only become aware of this when there’s a change of power somewhere, and certain data, like abortion statistics in the U.S., are no longer published or even collected.

Last question: Which data about the Canton of Zurich is most frequently accessed?

Andrea: Tax rates, population data, and land and real estate prices. After a voting or election Sunday, it’s the results.

The Cantonal Statistical Office Zurich provides the public, businesses, policymakers, and the media with independent data and analyses about the canton. During elections and referendums, it operates the cantonal voting and ballot centre. The office is part of the Directorate of Justice and Home Affairs.
It is a practice partner of the Public Data Lab and contributes practical questions to the project that could be explored through academic research.

Use of Indicators
The indicators of the Canton of Zurich are public and freely accessible. If the information cannot be found online or if individual data is needed for research purposes, enquiries can be made via a form or by telephone (043 259 75 00).
The Cantonal Statistical Office also advises interested parties on suitable data and where it can be found.