The availability of data tends to be taken for granted by the vast majority of people. The COVID-19 pandemic illustrates this vividly: estimates of case numbers and deaths have been widely quoted throughout and assumed by most to be available on demand.
However, those responsible for compiling official statistics know all too well that, even at the best of times, providing high-quality data to meet even just a small part of user needs is incredibly challenging and, on the whole, very resource-intensive. That said, the world has, in general, been steadily moving in the right direction, with more and better data being produced over time.
At the end of 2019, most users and producers of statistics would have predicted, with good reason, that the trend of increasing data availability would continue in the new decade, not least in the field of labour statistics. What no one could foresee then is that one of the cornerstones of data collection for surveys, namely the ability to visit and interview respondents, could be undermined so rapidly and drastically as was the case in 2020 owing to the COVID-19 pandemic.
Various organizations and specialized agencies in the United Nations system, including the ILO and collectively through the Intersecretariat Working Group on Household Surveys, have sought to track the impact of COVID-19 on data collection. In March 2021, the ILO launched a global survey to understand better the extent to which the crisis had affected the compilation of official labour market statistics. Information was received from 110 countries, of which 97 had planned to complete a labour force survey (LFS) in 2020. The findings point to both the tremendous challenges faced and the remarkable efforts undertaken to provide information on the world of work during the pandemic.
Nearly half of countries had to suspend interviewing at some point in 2020
Close to half (46.4 per cent) of the countries with plans to conduct a LFS in 2020 had to suspend interviews at some point in the year.
The highest levels of suspensions were reported by countries in Africa and the Arab States (70.6 per cent) and in the Americas (66.7 per cent). While some countries were able to attempt to recover those interviews later on, the majority were not, which means they completely lost data that had been expected to be available, creating a risk of gaps in data series for key labour market indicators, among others.
In contrast, suspensions were less common in Asia and the Pacific (35.2 per cent) and in Europe and Central Asia (30.9 per cent), and it was more typical for countries in these regions to be able to finish off the suspended interviews later on. Consequently, relatively few of these countries reported a complete loss of planned interviews.
This disparity across regions was clearly linked closely to the frequency and mode of data collection in individual countries. Unsurprisingly, countries with regular data collection and those using remote techniques, such as telephone interviewing, were far less likely to experience suspensions in interviewing. For example, approximately 90 per cent of countries that had remote data collection in place for some or all of their LFS sample when the pandemic started were able to conduct their originally planned interviews. This share dropped to about 65 per cent for countries that were completely reliant on face-to-face data collection.
Countries turned to telephone interviewing, but still faced many challenges
An increase in the use of remote data collection was a common response to COVID-19‑related restrictions across countries. Close to half (46 out of 97) of the countries with plans for an LFS during 2020 already intended at the start of the year to use remote methods. Among these, 40 countries reported greater use of telephone interviewing in the course of the year, with this modality replacing planned face-to-face interviews.
An additional 27, or more than half, of the other countries introduced remote data collection during the year in response to pandemic-related impacts, heralding a seismic shift in the use of remote methods. Countries achieved this in different ways, such as by allowing existing interviewers to complete telephone interviews from home instead of going out to visit households.
This shift built on the longer-term trend of growing use of remote modes over the past few decades – a trend offering the prospect of cost savings and increases in data volume and quality. However, while the adoption of remote techniques undoubtedly played a major role in maintaining data collection during the pandemic, it came with significant challenges and will not necessarily be sustainable for all countries.
The biggest challenge that countries typically faced was not having the contact details of enough households. Various approaches were tried out to deal with this. For example, a sizeable group of countries (close to one quarter) reused old samples in view of the lack of telephone numbers for newly sampled households. This was quite an effective stopgap for many countries, but it is not really sustainable in the medium to long term, since households eventually tire of repeat interviews. In order to keep response rates as high as possible, countries also fell back on other means of obtaining contact details, such as sending out advance notifications to households or changing sampling frames, with varying degrees of success.
The scale of the challenges involved is reflected in the fact that not all of the countries that turned to remote data collection during 2020 planned to keep using it in 2021 and beyond. Those that did not generally pointed to the lack of sustainable sources of contact details. Indeed, about half of the 27 countries that adopted remote interviewing in 2020 reported either being unsure whether that practice would continue or being determined to return to face-to-face interviewing as soon as possible.
Multiple lessons can be drawn from these experiences. While remote data collection did indeed prove to be an effective solution, it cannot be assumed that all countries are in a position to introduce it sustainably. International organizations and development partners need to be ready to provide countries with support and guidance to overcome the many challenges they face and fully harness the potential of remote collection. At the same time, one should not expect face-to-face interviewing to cease entirely, as it remains the most effective way to obtain high‑quality data in many parts of the world. Indeed, many countries have found success by combining in various ways face-to-face and remote interviewing for their surveys and this may be a viable path for others too.
Looking forward, among the other benefits it offers, the increased use of remote modes can make data collection more resilient to future shocks, such as new pandemics or natural disasters, which tend to be situations where access to reliable and timely information is critical.
Rising above the challenges to provide labour market data
Considering the scale of the disruptions suffered, one takeaway from the ILO survey is that countries showed remarkable flexibility and resilience to maintain, and in some cases even increase, the range of data they were publishing for users.
Inevitably, some of the responding countries had to cancel their LFS at some point in the year (21 countries in total), but over half of these undertook an alternative survey, such as a rapid telephone survey, to provide information on labour market impacts. A few countries also reported cancellations or delays in the publication of data: this occurred more frequently in some regions – such as the Americas – than in others.
However, it was more common for countries to publish additional information – by, for example, adding questions to their LFS, reporting on indicators not previously highlighted (such as changes in total hours actually worked or temporary absences) or launching additional publications devoted to the impacts of the COVID-19 crisis. Moreover, most countries reported that, by and large, they did not feel that there had been a significant negative effect on data quality during the year.
Although the ramifications of the pandemic are far from over, we should take stock of the achievements to date of the statistical community in meeting user needs for labour market data in such extraordinary circumstances. At the same time, we must also acknowledge that major data gaps have emerged and that we cannot ever take data availability for granted. Much work is required to ensure that the production of official statistics can rebound following the crisis, and countries will need significant support to achieve this.