Ideas & Insights / All Perspectives / Ideas & Insights

Rewiring How We Measure Impact in a Post-Covid-19 World

Michael Bamberger — Development Evaluation Advisor
Peter York — Principal and Chief Data Scientist, BCT Partners

The rapid spread of a pandemic like COVID-19 stresses the importance of the availability of real-time and reliable data and calls for us to rewire our systems of how we predict and assess impact. One step toward achieving this paradigm shift is to bring together the capabilities of people who research and measure the impacts of policies and programs on people and the planet.

The Nexus of the Data and Social Sciences

You might have noticed that our world is increasingly dependent on big data and data science in every aspect of our personal lives and in our economic, political, and social systems. Troves of real-time information on many issues, such as the cost of food, availability of jobs, our health status, jumps in the incidence of illnesses, our mobility patterns, etc. are produced every time a purchase is made, an Uber is called, a meal is ordered, a doctor is visited, a tweet is posted.

The exponential growth in this type of transactional, human-generated data is paralleled by an increase in information and analytical capacity unimaginable even a few years ago—making it possible to predict, assess and research changes on people and planet as it is happening. Using data science, it is possible for social science researchers and evaluators to collect a vastly increased range and volume of data more easily, quickly, and economically. The ability of big data to allow us to study everyone in an entire population for example, rather than just a relatively small sample, makes it possible to avoid many kinds of selection bias inherent in research and enables disaggregation of the sample to cover many different sub-samples and categories, critical for more sophisticated problem-solving. The technologies and now-affordable infrastructure of data science also means that social research and evaluation studies can be conducted more rapidly and cheaply while advancing our understanding of the complexity of social problems–allowing for learning about larger-scale issues.

What does social research and evaluation look like post-COVID-19?

In a new, post-COVID-19 world, we imagine that traditional in-person data collection will be hindered, and real-time analysis of data and on-demand reporting will become a requirement for all types of research and evaluation efforts. As such,  leveraging and combining administrative, transactional and big datasets, like satellite images, household survey data, program administrative data, social media analytics, phone call-center data, the information generated through mobile phones, and internet searches (to name just a few), are going to become key sources of data for research and evaluation specialists.

Data science also makes it possible to collect information over a much wider geographical area, to integrate many different kinds of health, economic, socio-cultural and demographic data (to name just a few), and to track changes over time – making it possible to understand and model the “big picture”, in a way that was not previously possible.

Additionally, the computing speed of cloud-based big data architectures and data science techniques like machine learning algorithms will need to become a part of the social science toolkit to meet the rapid need for up-to-date findings. A major challenge for evaluators and other social science practitioners is to ensure the efficient use of currently available data collection and analysis tools and techniques and to learn lessons from the current emergency to ensure that the full potential of these powerful tools will become available to address future crises.

How can we rewire?

The process of transformation for adopting and adapting these new technologies will be disruptive. To date, many, but certainly not all, research and evaluators have been slower than other practitioners to adopt the tools and techniques of data science. There are many methodological, economic, organizational, and even political reasons for the slower uptake. For example, the methodologies of training machine learning algorithms to build probabilistic predictive and prescriptive models are not well understood by evaluators and other social science researchers, often making it difficult for them to analyze big data.

There are also important ethical issues that researchers and evaluators raise for using data science techniques, particularly black box techniques that hide experimental and/or social biases. Another area of concern relates to the ability of agencies or policy-makers to use big data techniques to collect detailed information on communities and to use this information to make important decisions affecting the lives of these communities—without their knowledge or the possibility of dialog. Challenges to the uptake of big data technologies and analytic techniques are real, but if we build more educational and experiential bridges between researchers, evaluators and data scientists, they are not insurmountable.

Steps toward intentionally integrating the data sciences and social sciences for more rapid, cost-effective and time-sensitive evaluation findings is paramount to the promotion of social good. For example, earlier this year The Rockefeller Foundation, in partnership with the MasterCard Center for Inclusive Growth, relaunched as a new partnership platform to build the field of data science for social impact. Looking ahead, if this potential is to be realized, many more actors must be involved in promoting and facilitating the changes required to achieve this integration, including the institutions that train data and social scientists, the organizations that plan and fund social and development programs, decision-makers in the public, business and non-profit sectors, and –of course—data scientists, researchers, evaluators and other social science practitioners themselves.

This post builds on Measuring results and impact in the age of big data: The nexus of evaluation, analytics and digital technology that takes an in-depth look at how new information technologies will dramatically impact how we track and evaluate programs and policies. Download the report below.

Interested in learning more?

The U.S. isn’t using data that could save people from getting coronavirus | Quartz

Read the op-ed

The Rockefeller Foundation Commits $20 million in COVID-19 Assistance to Strengthen Global Pandemic Preparedness and Support Vulnerable Communities

Read the press release



  • Report

    Measuring Results and Impact in the Age of Big Data: the Nexus of Evaluation, Analytics, and Digital Technology

    Download PDF