No country or continent is untouched by COVID-19; but developing markets will see more short- and long-term negative social and economic effects than the developed world. Africa, South Asia and Latin America are particularly at risk due to their health system capacity. In these countries, people have less access to piped water and soap, and they are more exposed to the world trade cycle, as more advanced economies go into recession and focus inward1 . This is amplified in low-income communities – who are struggling to cope with the economic fallout of the crisis as it is.

Governments and health departments are reporting on the pandemic daily, and analyses of the economic consequences are being shared in abundance. However, particularly in the developing world, there is a gap in the data that is currently available. Data-driven analysis of what is happening to people and communities as they are forced to stay at home and fear for the health and safety of their families is lacking.

The lack of this information leaves governments, the private sector and civil society with a profound challenge. On the one hand, the urgency of the situation requires action. On the other hand, inadequate information makes it difficult to understand whether these actions are effective, or what knock-on effects they may have. This data gap forces governments to choose between speed of response on one side, and precision and effectiveness of response on the other.

The data gap in Africa is not new or unique to the COVID-19 crisis. Survey research targeting the full spectrum of the African population has traditionally been conducted using in-person interviews. These have formed the backbone of supporting policymakers in developing evidence-based interactions. In-person interviews ensure that even the most vulnerable and excluded people, who live in remote areas and often don’t have access to a wide range of technology, are not forgotten in evidence-driven decision-making.

In-person interviews have ground to a halt. Interviewers themselves are not able to work during the crucial lockdown periods brought about by the response to COVID-19. Data collection must be mindful of local laws. There is also an overriding ethical concern. It’s not in the best interest of saving lives to create a project that hinges on social interaction, and a team of people moving constantly around countries and communities. This may be a catalyst to the spread of COVID-19. This need not mean that Africa needs to fly blind during this crisis. In the developed world, digital surveys – particularly those conducted online – are cheap to administer and reach a wide enough spread of the population to provide a plethora of data to policymakers as they confront challenges like the COVID-19. Unfortunately, internet penetration in Africa is too low for this to be a viable option to reach those who are most vulnerable to this crisis.

However, there are alternatives, with the most obvious being to leverage the reach of mobile phone technology. Mobile phone access in planned survey markets

Mobile phone access in planned survey markets

South Africa, Nigeria and Kenya source: Pew Research Center (2018)
Rwanda source: Tasamba (2019) and Nsehe (2019)

 

Development researchers have traditionally shied away from using telephonic interviews for a few important reasons. Many households have access to a mobile phone nowadays, but there is still a significant portion of the population who do not have a personal mobile phone. This group of people are also typically those that development researchers are most interested in understanding – they are often excluded and vulnerable beyond just having a lack of access to a mobile phone.

 

Further, even when probability sampling is applied to telephonic surveys, response rates vary widely and are typically lower than 50% . This challenges the idea that they are truly representative, as such a significant portion of people are unwilling to participate.

With the data gap and the challenge to the sustainability of data collection in Africa in mind, the insight2impact facility (hosted by Cenfri and FinMark Trust) began a journey in 2018 to develop a mobile-phone-based method that can provide significant cost savings for research while providing a view of a good cross-section of the population. We developed this methodology with the ideal solution being a mixed method that includes a small in-person sample. We then tested the method with both SMS, self-completion surveys and interviewer-administered telephonic interviews. Unsurprisingly, telephonic interviews reach more people in what is considered the most vulnerable segment of the population – poor, rural women living below the poverty line. SMS interviews also reached these people, although they were heavily under-represented but not excluded. To achieve representation of all population groups of interest in a mobile sample, we’ve deployed Bayesian statistical modelling. At the heart of Bayesian modelling is using advanced statistical techniques to increase the weight assigned to under-represented groups, in line with the sample frame from the national statistics agency. This technique goes beyond traditional survey weighting to more accurately predict responses for under-represented populations – and in doing so, it ensures their voices are appropriately heard. The usefulness of this technique has been demonstrated in the lessons for more accurate political polling demonstrated by Ghitza and Gelman (2013).

Following this methodology, the insight2impact facility, with support from the Bill & Melinda Gates Foundation, is running a rapid response mobile tracker in several African markets. The survey will be ongoing for a minimum of the next nine weeks. We’ll be reporting on the results once every two weeks. The survey is currently live in South Africa, Kenya, Nigeria and Rwanda, and it’s set to launch in two more markets in the coming week. The core survey themes include health and risk behaviours, food security, work and job security, personal safety concerns, and access to government and community support. The first wave of results is set for release on 18 April 2020. Visit our website to review top-line findings or download the available dataset.

It’s our intention that the insights gained from this information can be used by interested parties, be they government policymakers or civil society organisations, to better plan their response to the COVID-19 pandemic, to assist them in ensuring that actions taken can reduce the speed-precision trade-off, resulting in increased evidence-based decisions and ensuring effectiveness of the subsequent actions. We’re looking to empower governments to address the challenge of balancing the need to protect lives with the need to protect livelihoods – both of which are key to the wellbeing of the poor.

Should you have any interest in being involved in the project – or wish to use the data to support important decision-making, or are simply interested in receiving updates as and when the data goes live – please sign up here to get updates on this work as it unfolds.

1 Surico, P and Galeotti, A (2020)

References

Ghitza, Y. and Gelman, A. (2013) Deep interactions with MRP: Election Turnout and Voting Patterns Among Small Electoral Subgroups in American Journal of Political Science Vol. 57. No. 3, July 2013, pp. 762-776 Midwest Political Science Association

GSMA (2019) The Mobile Economy Sub-Saharan Africa 2019 (Available at: www.gsma.com/mobileeconomy/sub-saharan-africa/)

Himelein, K., Eckman, S., Lau, C. and McKenzie, D. (2020) Mobile Phone Surveys for Understanding COVID-19 Impacts: Part II Response, Quality and Questions. World Bank Blogs: Development Impact

Neshe, M (2019) Ashish J. Thakkar Launches Smartphone Factory In Rwanda (Available at: ##1ac0f533481a)

Pew Research Center (2018) Internet Connectivity Seen as Having Positive Impact on Life in Sub-Saharan (Available at: www.pewresearch.org/global/wp-content/uploads/sites/2/2018/10/Pew-Research-Center_Technology-use-in-Sub-Saharan-Africa_2018-10-09.pdf )

Statistica (2018) Number of smartphones registered in Uganda from 1st quarter 2018 to 4th quarter 2018 (Available at: https://www.statista.com/statistics/1082286/number-of-smartphones-registered-in-uganda/)

Statistica (2018) Number of mobile cellular subscriptions per 100 inhabitants in Uganda from 2000 to 2018 (Available at: #)

Surico, P and Galeotti, A (2020) The economics of a pandemic: the case of COVID-19. Wheeler Institute for Business and Development: London Business School

Tambasa, J (2019) Africa's first smartphone launched in Rwanda (Available at: #)