The Evidence Barometer provides a fast, evidence-based way to assess whether a policy or program is likely to be effective.
Given a policy and an outcome, our solution scans relevant academic research, extracts key findings, and combines them into a simple directional signal. The goal is to help users quickly understand whether an intervention appears promising and worth deeper investigation.
This tool is designed as a barometer, not a replacement for a full systematic review or formal meta-analysis.
Beta version
The current version of the system is an early-stage implementation designed to test how users feel about the tool.
How It Works
- Finding Relevant Research
The system identifies academic studies related to a given policy and outcome. It expands search terms to capture variations in how programs and outcomes are described, and uses citation relationships to discover additional relevant papers.
Only studies with accessible full texts are included, ensuring that results are based on directly verifiable evidence.
- Identifying Key Results
From each study, the system identifies the main quantitative finding that reflects the effect of the policy on the outcome of interest.
This step focuses on extracting the most relevant estimate from each paper, rather than attempting to capture every reported result.
- Combining Evidence
Because studies report results in different units and formats, findings are converted into a standardized measure of statistical signal.
These standardized signals are then combined to produce an overall directional indication of whether the evidence tends to support or oppose the intervention.
- Country-Specific Insights
The system also uses information about where studies were conducted, together with country-level characteristics, to generate a rough prediction of how the intervention might perform in a specific context.
What the Output Means
The output should be interpreted as a directional signal, not a precise estimate of impact.
It is intended to answer questions such as:
- Is there consistent evidence supporting this policy?
- Does the literature point in a positive or negative direction?
- Is this worth investigating further?
Limitations
- Only studies with accessible full texts are included.
- Automated extraction may not capture all nuances of complex empirical results.
- The system summarizes direction of evidence, not effect size.
- Predictions across countries are approximate and depend on available data.
Ongoing Development
This tool is actively being improved. Future versions will expand coverage, improve extraction accuracy, and increase the speed and scale of analysis.
The long-term goal is to provide near-instant, high-quality evidence summaries across a wide range of policy questions.
Access Evidence Barometer