Quality Guidance for Contribution Analysis in Practice – IEG
Quality Guidance for Contribution Analysis in Practice: A Critical Review
In March 2026, the World Bank's Independent Evaluation Group (IEG) released a landmark guidance document on Contribution Analysis (CA). Written by leading evaluation scholars Giel Ton, Thomas Delahais, Andrew Koleros, and Marina Apgar, the guidance aims to support high-quality application of contemporary CA — an approach that critically assesses how an intervention contributes to changes of interest in complex settings where multiple factors and actors play a role.
Overview of the Guidance
Contribution Analysis (CA) was initially developed by John Mayne in 2001 as a response to the limitations of traditional attribution-focused evaluation methods in complex settings. Over the past 25 years, CA has substantially evolved as evaluators have adapted it to emerging evaluation contexts and challenges. However, its growing popularity has led to diversity in practice, resulting in confusion about what CA actually entails and how to assess quality.
The purpose of this guidance:
To support high-quality application of contemporary CA, promote greater consistency in application, strengthen collective use of CA, and inform commissioners about the features and conditions required for high-quality CA.
The guidance is explicitly not an introduction to CA but rather a resource for experienced evaluators. A companion methodology guide with practical examples is promised for later in 2026. The document is organized around six key steps encompassing two main phases: theory building (steps 1-4) and theory testing (steps 5-6).
The Six Steps of Contribution Analysis
Step 1: Set out the cause-effect issue to be addressed
Identify changes and outcomes of interest, build shared understanding of the systems, and formulate appropriate cause-effect evaluation questions. This step determines whether CA is suitable for answering these questions.
Step 2: Develop a robust theory of change for the intervention and its pathways
Theorize how change processes are expected to unfold, how the intervention might contribute, and what other conditions and events might interact. This includes articulation of causal pathways and explanation of intervention contributions.
Step 3: Gather the existing evidence for the theory of change and contribution claims
Use secondary information and light-touch empirical investigations to assess overall plausibility and identify causal hotspots requiring deeper investigation.
Step 4: Assemble and assess the resulting contribution claims and the challenges to them
Formulate, select, and assemble contribution claims in a provisional contribution story. This concludes the theory-building phase.
Step 5: Seek out additional evidence
Conduct additional data collection to fill gaps, using a broad range of methods. This step emphasizes intentional assessment of evidence strength and active consideration of disconfirmatory evidence.
Step 6: Revise and strengthen the contribution story
Develop final revised statements about contribution claims within the overall contribution story, explicitly recognizing remaining uncertainties and avoiding false certainty.
Key insight: The steps are not applied sequentially but iteratively (repeated in a loop) and recursively (a step can be repeated several times) until reaching credible and meaningful results. This iterative, abductive process is what makes CA distinctive.
Strengths of the Guidance
✅ Clear articulation of the iterative process
The guidance excellently captures the non-linear, abductive nature of CA. The recognition that steps are iterative and recursive — not sequential — is a major contribution to the field.
✅ Emphasis on contribution stories
The contribution story — a narrative framework that structures analysis and presents findings — is a distinctive and powerful feature of CA that this guidance develops thoroughly.
✅ Stakeholder engagement as quality criterion
The guidance explicitly identifies active inclusion of diverse perspectives — especially from those directly involved with or affected by the intervention — as essential for quality.
✅ Acknowledgment of uncertainty
CA operates along a continuum of plausibility. The guidance emphasizes explicitly recognizing remaining uncertainties and avoiding conveying a false sense of certainty.
✅ Integration with process tracing and other methods
The guidance explicitly links CA to process tracing (Beach and Raimondo 2025) and other methods, supporting method bricolage in real-world evaluation contexts.
✅ Recognition of both confirmatory and disconfirmatory evidence
Quality in CA requires actively seeking and considering both confirmatory and disconfirmatory evidence — a rigorous approach often overlooked in practice.
EvalCommunity Critique & Missing Elements
While the guidance represents a significant advancement for the evaluation field, several elements are missing or underdeveloped. The following critique is offered constructively to support the ongoing evolution of CA practice.
⚠️ Missing: Practical examples and case studies
The guidance explicitly states it is "not intended to serve as an introduction to CA" and that a companion methodology guide with examples will be available in 2026. However, the absence of concrete examples in this guidance limits its accessibility. Even experienced evaluators would benefit from illustrative cases showing how the iterative process works in practice across different contexts (humanitarian, governance, health, etc.).
⚠️ Missing: AI and digital tools integration
Published in 2026, the guidance makes no mention of how AI and machine learning tools could support CA — for example, using NLP for evidence synthesis, predictive analytics for causal hotspot identification, or digital collaboration tools for stakeholder engagement. This is a notable gap given the rapid advancement of AI in evaluation.
⚠️ Underdeveloped: Guidance on selecting which claims to test
Step 4 acknowledges that "only a few claims may be further refined and tested" due to real-world constraints, but the guidance provides limited practical criteria for claim selection beyond significance, uncertainty reduction, and learning opportunity. A more systematic decision framework would strengthen this crucial step.
⚠️ Missing: Integration with quantitative methods
The guidance mentions "a broad range of data collection methods" but leans heavily toward qualitative and case-based methods. More explicit guidance on integrating quantitative methods — including quasi-experimental designs, statistical modeling, and mixed-methods approaches — would strengthen CA's credibility with quantitatively-oriented stakeholders.
⚠️ Underdeveloped: Cost and resource implications
While the guidance acknowledges real-world budget, political, and time constraints, it provides little practical advice on estimating the resources required for high-quality CA. This is a critical gap for evaluation commissioners who need to budget appropriately.
⚠️ Missing: Addressing power dynamics in stakeholder engagement
The guidance emphasizes stakeholder engagement but does not address power dynamics that can shape whose perspectives are included, whose claims are prioritized, and how contribution stories are framed. Feminist and decolonial evaluation approaches have much to offer here.
Comparison with Other Evaluation Methods
Key Takeaways for M&E Professionals
CA is for complex situations
CA is most useful when change extends beyond an intervention's direct influence and multiple actors and factors play a role. It explicitly embraces uncertainty about causal contribution.
Iteration is not optional
The iterative, abductive process is what makes CA distinctive. Quality depends on multiple rounds of reflection, data collection, and refinement.
Stakeholder engagement is essential
Quality is enhanced through active inclusion of diverse perspectives, especially from those directly involved with or affected by the intervention.
Embrace uncertainty transparently
CA operates along a continuum of plausibility. Explicitly recognize remaining uncertainties — avoid conveying false certainty.
Frequently Asked Questions
What is the difference between contribution analysis and traditional impact evaluation?
Traditional impact evaluation seeks to establish causal attribution — proving that an intervention caused observed changes. Contribution analysis acknowledges that multiple factors and actors typically play a role, especially in complex settings, and seeks to assess the intervention's contribution relative to other influencing factors.
Is this guidance suitable for beginners in CA?
The guidance explicitly states it is "not intended to serve as an introduction to CA" but rather to support experienced evaluators. Beginners should start with foundational CA literature (Mayne 2001, 2019) before using this guidance.
How does CA relate to theory of change?
Theory of change is central to CA. Step 2 requires developing a robust theory of change that includes both articulation of how change is expected to happen and explanation of how the intervention contributes to that change process.
What methods can be used within a CA framework?
CA accommodates a broad range of methods, including process tracing, qualitative comparative analysis, case studies, participatory methods, and increasingly AI-powered analytics. Method selection should be driven by the specific claims, budget, time, and data availability.
A Valuable Contribution with Room for Growth
The IEG's Quality Guidance for Contribution Analysis in Practice is a valuable contribution to the evaluation field. It successfully captures the iterative, abductive nature of contemporary CA and provides a coherent framework for experienced evaluators. The emphasis on contribution stories, stakeholder engagement, and explicit acknowledgment of uncertainty are particularly strong features.
However, the guidance would benefit from practical examples, integration of AI and digital tools, more systematic guidance on claim selection, attention to power dynamics in stakeholder engagement, and practical advice on resourcing. We look forward to the companion methodology guide promised for later in 2026.
For the M&E community, this guidance represents an important step toward greater consistency and quality in contribution analysis — an approach that is increasingly essential for evaluating complex interventions in international development, humanitarian action, and social programming.
