Tools for Evidence-Based Policy Making: A Guide for Civil Servants and Policy Teams

Evidence-based policy making has been a stated commitment of successive UK governments for nearly three decades. The What Works Centres have produced substantial evidence syntheses. The Magenta Book provides methodological guidance on policy evaluation. The National Audit Office consistently emphasises the importance of evidence in public spending decisions.

And yet the gap between the stated commitment and the operational reality remains wide. Policy decisions are made under time pressure, informed by incomplete evidence, and rarely reviewed against their intended outcomes in a structured way. The problem is not the absence of evidence or the absence of good intentions. It is the absence of infrastructure — the tools and processes that translate the evidence-based policy making mandate into a consistent operational practice.

The Three Infrastructure Gaps

Gap 1: No structured decision record at the time of the decision

Policy decisions in UK government are documented in ministerial submissions, briefing papers, and submissions to Cabinet committees. These documents are designed to inform and record a decision, but they are not structured in a way that makes the reasoning behind the decision systematically searchable or reviewable. When a policy is reviewed two years later, the original reasoning must be reconstructed from the submission documents — a time-consuming and imperfect process that rarely produces a clean picture of what was known, what was considered, and why the chosen option was selected.

Gap 2: No systematic outcome review mechanism

Policy evaluations are conducted, but they are typically large, expensive, and infrequent. The systematic practice of reviewing individual policy decisions against their intended outcomes at structured intervals — 30, 90, 180 days — is almost entirely absent from government practice. As a result, the learning from policy decisions accumulates slowly and inconsistently, and the same types of errors recur across policy areas and administrations.

Gap 3: No uncertainty quantification

The best available evidence rarely supports decisions with certainty. Policy decisions are made under uncertainty, and the degree and nature of that uncertainty should be captured at the time of the decision. Without a structured confidence field, the uncertainty that existed when a decision was made is invisible in the record — and the learning from how that uncertainty resolved cannot be captured systematically.

The Tool Stack for Evidence-Based Policy Making

Evidence synthesis tools

The What Works Centres, the EPPI Centre, and systematic review databases provide the evidence synthesis infrastructure. The practical issue for policy teams is access speed — how quickly can relevant evidence be surfaced during the policy development process? Tools like Nesta’s Evidence Library and the Alliance for Useful Evidence’s resources help, but the evidence synthesis process remains slower than the typical policy decision timeline.

Analysis and modelling tools

HM Treasury’s Green Book and Magenta Book provide the analytical framework. The modelling tools vary by policy area: economic models, impact assessments, regulatory impact tools. The challenge is not the absence of analytical tools but ensuring that the analysis is captured in a format that can be traced to the final decision record.

Decision logging tools

The most significant infrastructure gap is in decision logging. A decision logging tool that captures the policy decision, the evidence basis, the alternatives considered, the confidence level, and the outcome review dates would close the primary gap between evidence-based policy making as a principle and as a practice. This is the tool category that currently has the least development in the public sector context.

Making Evidence-Based Policy Making Operational

The practical implementation of evidence-based policy making requires three operational changes, none of which requires significant cultural change: first, a standardised decision log template that is completed for all significant policy decisions as part of the existing submission process; second, a scheduled outcome review requirement that is built into the policy decision record at the point the decision is made; and third, a searchable central repository for decision records that enables learning to accumulate across policy areas and administrations.

The COVID-19 pandemic response, reviewed in hindsight, shows what is possible when these three elements are in place versus what the cost is when they are not. The inquiry process is essentially a retrospective attempt to reconstruct the decision records that should have existed at the time. Building that infrastructure prospectively — as a routine part of policy decision-making — is the practical implementation of the evidence-based policy making commitment that successive governments have made.