Are we thinking correctly about ‘data’ and ‘data use’ in global education?

These days, international development networks, donors, and education organizations will take almost any opportunity to issue their refrain: Education in the Global South will improve only when countries possess sufficient, good data presented accessibly and used effectively by decisionmakers. As a result, educational management information systems in low- and middle-income countries (LMICs) have proliferated. “Evidence-based decisionmaking” is a ubiquitous term. And innovations like dashboards or searchable databases that organize existing data for easy use are omnipresent. We are in a data zeitgeist.

On one hand, that’s a good thing. In the absence of accurate, timely education data (of any kind), educationalists and politicians—whether they’re teachers, school heads, mid-tier officials, nongovernmental organizations, or central government policy makers—will over rely on experience, instincts, political calculus, and splintered views of education to make important decisions. That’s how the past repeats itself. And evaluating schools cannot effectively or equitably occur without careful collection and use of data. Having and using good data is a powerful component of education improvement.

Yet on the other hand, significant barriers, complexities, and exaggerations dilute both the accuracy and strength of this popular prescription. What counts as data—or whose data counts—is contested. Cultures that have historically used neither education research nor data (in the modern, Global North sense) do not always see the need to change. Critics view the technical-rational reform paradigm informed by quantitative data as a colonialist legacy. Government education decisions are frequently based on political expediency as much as (or more than) evidence. And there are pragmatic, if dispreferred, reasons those in power may work to circumvent the transparency and attention increased data use often produces.

What to do about this incongruity?

Forthcoming findings from our current ROSIE study on the role of middle-tier education governance and support (below central government but above individual schools) for scaling innovation will discuss promises, realities, and potential improvements around data use. Stay tuned as we release this research in different formats soon. In the meantime, though, here are four emerging suggestions.

1. Organizations from the Global North working with LMICs can widen how they define ‘data’

Economists have long dominated modern research for education development. They have their role, but it’s time to more fully invite qualitative, participatory, and innovative research designs that value stories, participant perspectives, activity, indigenous ways of knowing, and realia (even dreams and smells, as Elizabeth St. Pierre reminds us). When locals and outsiders work together to build research and data systems that derive from a shared, integrated foundation of local and technical-rational epistemologies, we will not only produce better, more location-sensitive data practices but also organically increase data use and therefore sustain its impact. In education, it’s always better, though more difficult, to invite local commitment than to impose control.

2. Those pushing for increased data collection and use can graft their innovations and goals onto existing decisionmaking practices and ways of knowing

External funders and education organizations can demonstrate respect for traditional ways of knowing at the same time they point out cognitive biases and weaknesses in anecdotal evidence or folk theories. If done with humility, curiosity, and ethical collaboration, outsiders can nudge mindset shifts in locations and increase demand for using facts, statistics, and replicable findings in education—and then leverage the demand into stronger support and commitment for data use. But the “collaborative” part means that external groups must themselves be open to change, too. Taking seriously the value of stories, personal knowledge, and other cultures’ epistemologies is key. Each side has something to teach the other.

In practice, data proponents might jettison some of their inflexible exhortations, unwelcome practices, and imposed attempts at behavior change on the ground and instead attach their versions of data use in smaller, softer ways to existing practices so that what results are hybrid, data-forward practices built on—and supplementing—local ways of capturing practice and making education decisions. In this way, a more systematic, less subjective culture of data use can take root and grow in “complementary and additive ways,” as Yue-Yi Hwa said to me recently—not as wholesale replacement.

3. We can combine improved data use with accompanying broader systems support.

We see at least three promising approaches here: providing effective professional development for middle-tier education actors, school principals, and local governments; increasing the decisionmaking autonomy of local and municipal education officials; and incentivizing governments and schools to be candid about their shortcomings. Our current study on the middle tier in LMICs finds it to be a system level able to yield exponentially improved functionality if education officers, instructional leaders, and support personnel receive extended and relevant professional training for their roles, have more authority to engage in real decisionmaking, and don’t feel overly burdened to simply monitor schools and ensure pro forma compliance.

Increasing both the valuing of data and a comfort with data use (and not simply focusing on more data or accessible data systems) in this crucial middle-system level can increase the ability for these actors—uniquely positioned to view local realities of communities and teachers yet see across schools and municipalities/villages—to actualize their potential to improve education.

4. We can be honest about the limits of data use to improve education

Sometimes it can feel like the donor community is more focused on processes of data use than on exactly how data use will add specific, pragmatic value. Therefore, encouraging these external education organizations to critically interrogate why and how they’re introducing data systems in countries, and to consider alternative ways to pursue this work should ferret out less-useful practices to discard and identify better ways to advance their efforts.

Also, like in so many aspects of life, too much of a good thing can go too far. Higher-income countries have already experienced negative effects of data singlemindedness. These include data-based accountability systems that identify failing schools but don’t support their improvement; data overuse that inadvertently incentivizes data cheating, overly punitive measures, and teacher mistrust; and widespread victim blaming or deficit thinking.

Additionally, it must be acknowledged that data use is just one tool in a complex repertoire of education decisionmaking, accountability, and improvement. Electoral politics will always play a role. Philosophical questions around purposes of education matter. Diverse decisionmaking coalitions to work with policymakers (including inside-outsider perspectives) and strong leadership are crucial ingredients of a healthy education decisionmaking system.

And finally, if the global education community oversells data use as a panacea and their approach doesn’t dramatically improve education in locations, then the global community has yet again lost a layer of trust and credence with those it is meant to serve.

We would like to hear your views on this topic and learn your examples of successful or promising efforts to integrate data use with existing decisionmaking histories, systems, and processes in education systems in low-and-middle income countries. Please let us know if our insights resonate with your work!

The author would like to thank Minahil Asim and Yue-Yi Hwa for helping him think through parts of this piece.