Conducting surveys in a pandemic
As the coronavirus spreads, data collection may prove essential for emergency response in low-income countries. But how can we collect data when countries lock down?
The novel coronavirus pandemic is primed to hit its stride globally over the next several weeks. It will also finally make its dreaded landing in low-income countries, which are now scrambling to enact containment measures. The faster they act, the better: social distancing is the only known remedy for COVID-19, and poor countries are particularly vulnerable to the spread of the pandemic due to their already-strained healthcare systems and food security issues.
As the virus moves south, the burden of social distancing policies will also make itself felt for any decision-maker and scientist who relies on survey data from low-income countries. The face-to-face surveys that have long been our primary source of information on the health, income, and wellbeing of the poor will likely be decommissioned until at least late summer. Data collection will grind to a halt just as it is needed the most.
An impending darkness
Developing countries already struggle to produce reliable and timely statistics, strapped by limited budgets and inaccessible populations. Long breaks between data points are the norm, making international comparisons difficult in the best of times; an extended interruption in the data stream for all poor countries at the same time would essentially plunge the developing world, and all of Sub-Saharan Africa in particular, into statistical darkness.
Such a breakdown in data availability would, in turn, place a major burden on decision-making and crisis management. Development agencies in rich countries and ministries in poorer countries rely on survey data to allocate financial and humanitarian aid. Without up-to-date numbers, the distribution of limited resources, funds, and aid workers will become a (likely politicized) guessing game, and public health officials will be unable to react to outbreaks of the disease in a timely fashion. This may lead to countries missing crucial windows of opportunity to stamp out the spread of the virus before it starts.
In normal times, decision-makers faced with missing data might extrapolate from past observations, operating under the assumption that the numbers follow stable, predictable patterns. In times of major upheavals, such as a natural disaster or a financial downturn, the indicators that measure the impacts of these events — and are thus of most interest to researchers and politicians — are precisely those that become entirely unpredictable. COVID-19 is, in this sense, the perfect storm, leaving bureaucrats without one of their most reliable tools for coordinating a response while the pandemic lays waste to both their fragile healthcare infrastructure and national and local economies.
There’s an app for that
Fortunately, the rapid spread of mobile connectivity may prove to be the lifeboat that keeps developing countries’ data collection infrastructure afloat. Mobile phone surveys have been gaining traction as a viable, low-cost alternative to face-to-face interviews, particularly when respondents are nomadic, potentially contagious, or living in conflict areas that may be dangerous for enumerators. They have made it possible to follow displaced populations during the northern Mali conflict, for instance (see study here), or isolated cases during the West African Ebola crisis (here, here, and here). The latter experience may be of particular relevance as we enter a period of widespread social distancing.
Apart from being the most feasible form of data collection in a pandemic, mobile surveys have two advantages over traditional face-to-face surveys: their low cost and high frequency. The cost per response has been estimated to be from three to ten times lower for cell-phone surveys than for traditional face-to-face surveys. These savings, which depend mostly on the number and geographic dispersion of the respondents, let researchers increase the frequency of their questionnaires and generate valuable time-series data. High-frequency times-series, in which the same respondents are tracked over time, are particularly important in rapidly developing situations. Mobile phone surveys may thus prove crucial in the fight against the exponential spread of COVID-19, helping with the early identification of outbreaks, coordinating emergency deliveries to areas hit by food or water shortages, and promoting the effective communication of new lockdown measures and other official diktats.
Despite their advantages, mobile phone surveys are no panacea for social distancing and come with their own challenges. Without the proper incentives, they tend to be beset by low response rates and high attrition — the number of people dropping out of the survey over time. It is harder to keep respondents’ attention over the phone than in person, meaning mobile phone surveys also need to be shorter and more concise. Most importantly, mobile phone surveys require phone numbers. For researchers who did not collect contact information from their respondents before the lockdown, getting access to phone numbers is difficult and raises privacy concerns.
Moving to mobile
These can represent major hurdles for academic researchers, but national governments generally have all the tools at their disposal to transition to mobile phone surveying without a hitch. They have access to citizens’ contact information through ubiquitous census data, and their credibility as a national body is a strong incentive for participation.
Around 80% of countries reporting data to the International Labour Organisation (ILO) rely on face-to-face collection techniques. Only in Europe is it already commonplace to interview respondents remotely. Now that COVID-19 has shelved pen-and-paper surveys for the foreseeable future, the right time has come for developing countries to start moving towards mobile phone and computer-assisted remote surveys. It could prove to be a vital tool for the coordination of emergency response if the virus spreads in developing countries in the coming weeks, and will improve their lagging data infrastructure in the long run.
What if you’re not a national government, and would like to collect data in a low-income country during COVID-19? At NADEL, we’ve already garnered some experience running mobile phone surveys for our research in Benin, and are in the process of launching new studies in South Africa and Burkina Faso. Find out more about running a successful mobile phone survey by following these links:
- How to improve SMS survey response quality and completion rates in emerging markets (Aimee Leidich)
- Best practices for conducting phone surveys (J-PAL)
- Surveying Young Workseekers in South Africa (Nilmini Herath, Emmanuel Bakirdjian, Wim Louw)
- Using Mobile Phones to do Research in the Time of COVID-19 Lockdowns and Beyond (Mywish Maredia)
- Mobile Phone Surveys for Understanding COVID-19 Impacts: Part I and Part II (World Bank)
Written by Bart Kudrzycki, PhD student at the Center for Development and Cooperation (NADEL) at ETH Zurich. Bart is using face-to-face and mobile phone surveys to study youth employment in Benin. He will shortly publish more tips on running a successful mobile phone survey on his own blog.
