Summary#
Path tracing is for answering questions like:
- “How does A lead to B (through what intermediate steps)?”
- “What are the main routes from an intervention to an outcome?”
- “If we start from this driver, what downstream consequences show up within K steps?”
It’s best thought of as:
- a filter (keep only links that participate in the traced pathways), plus
- an interpretation rule (what counts as evidence for a pathway).
What you get (in plain terms)#
You choose:
- a From factor (or leave it blank)
- a To factor (or leave it blank)
- a maximum number of steps (how long a “chain” you want to allow)
Then the result is a sub-map containing only the links that sit on at least one allowed pathway.
When to use it (practitioner-friendly)#
- Explaining mechanisms: not just “what matters”, but “how it connects”.
- Finding mediators: what sits between A and B?
- Generating hypotheses: “these are the plausible routes people describe”, then you go back to quotes to check.
Source tracing (recommended when you care about coherent stories)#
Without source tracing, a pathway can be a composite: one source says A→X and another says X→B, so the map can show A→…→B even if no single respondent told the full chain.
Source tracing is the stricter version: it keeps only pathways that can be realised within at least one single source (a coherent within-source narrative).
Use it when you want to avoid the “stitched together across respondents” problem.
Practical tips#
- Keep K small: in interviews, chains longer than ~4 steps are uncommon and get hard to interpret.
- Be careful with transforms: if you change labels (Zoom, Collapse, Combine Opposites, clustering), you can change which pathways exist as labels. That’s often fine for presentation, but it changes what you are “counting as the same thing”.
Order matters (a conservative workflow)#
If you care about coherent pathways and you want a cleaner, summarised map:
- first do source tracing (to keep within-source chains)
- then apply label-rewrite transforms (e.g. Zoom / Collapse / Combine Opposites) for presentation
Formal notes (optional)#
If you want the precise definition, here it is.
Given one or more start factors \(S\), one or more end factors \(T\), and a maximum path length \(K\):
- keep a link \(x \rightarrow y\) iff it lies on at least one directed path of length \(\le K\) that starts in \(S\) and ends in \(T\)
- if \(S\) is empty, interpret this as “paths that end in \(T\)”
- if \(T\) is empty, interpret this as “paths that start in \(S\)”
The key is: path tracing is link-based. It should not “fill in” extra links between surviving factors.