What do you think is important in social research?
Here's what we believe, do you agree?
1 ⏹️ Social research works best when the concepts come from people closest to the subject — not from experts. Quantitative tools can't do that.
2 ⏹️ People's worlds differ. Not just in details — in main features. One size doesn't fit all. Especially in a changing world.
3 ⏹️ Social research must welcome data and narratives that don't fit: messy structure and messy contents.
We also believe social research needs a causal lens, do you?
1 ⏹️ Many of the most important questions in research and evaluation are causal: What drives X? What does Y lead to?
2 ⏹️ Asking "how does your world work?" can be a really useful question — for interviews and for analysis.
3 ⏹️ Beliefs matter. What do people think drives what? That is critical to know when we're dealing with people — even when they're wrong.
4 ⏹️ People are causation experts. Mostly, people are right about causation. We make mostly successful causal judgements thousands of times a day. We're the best causation detectors there are.
☑️ Causal mapping is not an "evaluation method" but an "evidence broker" for evaluation methods like QuIP, Outcome Harvesting and Process Tracing. It finds masses of causal information, organises it, and feeds it in to other methods for evaluative judgement.
🤖 Causal mapping is a perfect fit to make use of the power of AI — not by using it as a block box but as a low-level coding assistant. We can apply a relatively generic causal coding template to hundreds of interviews documents and be ready to ask and answer causal questions about them very quickly.
🗺️ Stories in, stories out. People tell stories. Causal mapping takes them and outputs stories — and maps.
Causal Mapping — the evaluation evidence broker▸
Marshalling causal evidence at scale for Contribution Analysis and beyond.
Causal Mapping is not primarily an evaluation method; it is a way of identifying and organising causal claims in support of evaluative judgement.
It fits naturally with Contribution Analysis (CA), while still adding value to other evidence-based frameworks...
Outcome Harvesting: strengthen contribution claims by identifying the precise causal chain from outcome back to intervention.
Realist Evaluation: identify Context-Mechanism-Outcome (CMO) linkages mentioned by sources.
QuIP: strengthen causal contribution claims through explicit, source-grounded chains.
1 Link = 100 Claims
An arrow in a map can look simple (for example, Training -> Knowledge), but that single link can represent dozens or hundreds of auditable claims from different sources.
It lets evaluators “double-click” into and examine the different sources for each link.
Every claim has verbatim text attached to it.
!
Scaling analysis with AI
When you move from 10 interviews to 200, manual coding alone is too hard. "Asking ChatGPT" surrenders human judgement to a black box.
Causal Mapping uses AI as a low-level assistant to automate the extraction of traceable evidence nuggets, leaving evaluative judgements to the evaluator.
How causal mapping can help with Contribution Analysis
John Mayne's six Steps, plus one
1 Set out the Attribution Problem
Define the evaluation questions and the level of evidence required.
2 Develop the Theory of Change
Establish the logic of how the intervention is expected to lead to results.
Often there are multiple versions of the "official" theory, or none at all.
Causal mapping helps: Assemble theories of change from official documents.
3 Gather Evidence on the ToC
Collect existing and new evidence to populate the causal links.
Causal mapping helps: Assemble "empirical theories of change" from stakeholder evidence and test if the official theory matches up.
There is even a metric for that.
4 Assemble the Performance Story
Build the contribution narrative based on synthesised evidence.
Causal mapping helps: Synthesises individual claims into verifiable chains with path tracing and source tracing.
5 Assess Alternative Explanations
Account for external influences and other drivers of observed change.
Causal mapping helps: Explicitly maps non-project influences mentioned by sources.
6 Revise and Strengthen
Refine the story based on gaps identified in the evidence base.
Causal mapping helps: Highlights weak links where evidence count is low.
7 Extend
Causal mapping helps: A single causal coding of all documents creates a causal database which can provide inputs to all the above steps, and a lot more too.
See how different stakeholder groups view the the project differently.
