Why Evidence-Based Policymaking Became the Gold Standard
Evidence-based policymaking (EBPM) rests on a simple idea: governments should fund and scale programmes that can demonstrate causal impact. In practice, this has meant privileging rigorous evaluation methods, particularly randomised controlled trials (RCTs), as a way of identifying “what works” rather than relying on intuition or tradition (Baron, 2018).
The approach draws heavily from postwar medicine. As Baron (2018) shows, much of early medical practice relied on anecdote and professional authority, until landmark trials such as the streptomycin study and the Salk polio vaccine demonstrated how controlled experiments could produce reliable knowledge and dramatically improve outcomes. These successes did more than improve treatment—they helped establish a new standard for decision-making that would later travel into other policy domains.
From the 1980s onwards, similar methods began to move into social policy. Experimental evaluations became more common in welfare and employment programmes, and by the 2000s had expanded into education and international development. As Haskins (2018) notes, this was partly driven by advances in the social sciences, which made it possible to run experiments in real-world settings and assess whether programmes actually caused improvements in outcomes.
At the same time, EBPM aligned closely with political and institutional incentives. Policymakers increasingly turned to evidence not just for better decisions, but for justification. Labelling a programme as “evidence-based” signalled effectiveness and accountability, helping to legitimise spending and reform. Over time, this was reinforced through mechanisms such as evidence clearinghouses and funding models that prioritised interventions backed by rigorous evaluation (Haskins, 2018).
By the 2010s, EBPM had become more than a method. It was widely treated as a gold standard—both a way of producing knowledge and a way of governing.
From Evidence-Based to Evidence-Informed Policymaking
However, this “gold standard” framing is increasingly being qualified. A growing body of work suggests that while evidence is essential, it does not translate directly into policy decisions.
Part of this shift is already visible in the language of the field. Rather than “evidence-based” policymaking, many scholars and practitioners now refer to evidence-informed policymaking, reflecting a more realistic view of how decisions are made in practice (Head, 2015).
Policy processes are shaped not only by evidence, but also by political values, institutional constraints, and organisational variation. As Head (2015) argues, the use of evidence across government is uneven, and the earlier ambition of grounding policy primarily in objective evidence has proven difficult to sustain. Even within the evidence-based movement, there is recognition that complex social problems are hard to solve consistently, and that successful programmes do not always replicate across contexts (Haskins, 2018).
The question, then, is not whether evidence matters, but how it should be used. This opens up a shift from treating EBPM as a universal model towards a more flexible, context-sensitive approach to policymaking.
The Myth vs Reality of Evidence-Based Policymaking
The appeal of EBPM is rooted in a simple analogy: just as evidence-based medicine identifies the most effective treatments through controlled trials, policymaking can identify “what works” and scale it. RCTs and systematic reviews sit at the top of this hierarchy, offering seemingly objective answers to complex problems.
The difficulty is that policymaking is not medicine.
Clinical decisions are relatively insulated. They operate within clearer goals, controlled environments, and professional norms that prioritise evidence over politics. Policymaking, by contrast, is exposed to competing interests, electoral pressures, budget constraints, and institutional fragmentation. Evidence does not enter a neutral space—it enters a contested one.
This helps explain why “what works” rarely translates cleanly into policy.
Take deworming programmes, often cited as a landmark success of evidence-based development policy. Early experimental results suggested large improvements in school attendance and informed major funding decisions. Yet later re-analyses and replications showed more modest and context-dependent effects. What looked like a universal solution became far less predictable once implemented across different systems. The issue was not flawed evidence, but the assumption that results would generalise without friction.
A similar pattern appears in global health policy. Evidence is often designed to produce generalisable recommendations, but implementation depends on local systems—staff capacity, governance, infrastructure, and social norms. As Béhague et al. (2009) show, standardised evidence can end up prioritising uniform solutions that do not align with context-specific needs, particularly in lower-income settings.
There is also a more political reality. Evidence is frequently used to support decisions rather than determine them. Policies must be feasible, fundable, and acceptable to stakeholders. In this environment, evidence competes with other pressures and is often selectively interpreted. The ideal of a linear process—from evidence to policy—gives way to a more iterative and negotiated one.
The myth, then, is not that evidence matters, but that it can deliver clear, transferable answers in a political system that is neither neutral nor controlled.
An Alternative: Evidence-Informed Policymaking in Practice
If EBPM overstates what evidence can do, the practical alternative is not to abandon it, but to use it differently. Evidence-informed policymaking (EIpM) starts from the premise that evidence guides decisions but does not settle them.
In practice, this shifts attention from replication to adaptation.
For example, interventions such as cash transfers or public health programmes often show strong results in trials, but governments rarely adopt them unchanged. Instead, eligibility rules, delivery systems, and conditionalities are adjusted to fit administrative capacity and political priorities. Evidence provides a direction of travel, but the final design reflects local constraints.
A second shift is towards testing in real policy environments, not just controlled settings. Rather than relying solely on external evidence, governments increasingly run pilots, phased rollouts, or A/B tests within their own systems. The UK Behavioural Insights Team’s work on tax compliance is a good example: small changes to messaging were tested in live administrative contexts, allowing policies to be refined based on actual behaviour rather than assumed responses.
Third, EIpM recognises the role of non-experimental knowledge. In many areas, especially where RCTs are not feasible, policymakers rely on administrative data, practitioner feedback, and user experience to adjust programmes. For instance, reducing administrative burden in welfare systems often comes not from new experimental evidence, but from analysing drop-off points, simplifying forms, and learning from frontline staff about where people struggle.
Finally, EIpM treats policymaking as iterative rather than one-off. Instead of asking for definitive proof before action, it builds in mechanisms for continuous adjustment. This is particularly important in complex areas such as housing, health systems, or labour markets, where outcomes depend on interactions between multiple institutions and behaviours.
The practical implication is a shift in mindset. The question is no longer “what works?” in the abstract, but “what is likely to work here, and how do we adapt as we learn?” Evidence remains central, but as an input into judgement rather than a substitute for it.
Conclusion
The appeal of evidence-based policymaking is unlikely to disappear. It offers clarity, legitimacy, and a sense of control in environments that are often uncertain and politically contested. But the growing body of research suggests that its strongest claims—around neutrality, transferability, and linear decision-making—are difficult to sustain in practice.
What is emerging instead is less a rejection of evidence than a recalibration of its role. Evidence-informed approaches recognise that policymaking is shaped by context, constraints, and competing priorities, and that effective use of evidence depends as much on interpretation, adaptation, and timing as on methodological rigour.
The shift from EBPM to EIpM is therefore not a wholesale replacement, but an evolution. The challenge for public servants is not simply to find better evidence, but to develop better ways of using it in systems where uncertainty, politics, and complexity are unavoidable.
References
Log in or sign up to continue the conversation