Causal Mapping 1 What’s in this guide?I Introduction 2 Manifesto: piecing together fragments of causal information3 Causal maps: a unifying idea4 Causal mapping5 The causal map app in four pictures6 We need a “soft arithmetic” for causal maps7 We need rules about how to encode causal information in a causal map (and decode it again)8 Soft arithmetic is also the answer to understanding causal maps9 Metalanguage: some words for talking about causal mapsII The rules of Soft Arithmetic for causal maps 10 The mini-map coding rule11 The mini-map coding rule – functional form12 The extras rule: adding extra information, in particular about the levels or values of the variables13 The juxtaposition rule14 Rules for joining maps15 Rules for joining maps: what counts as “the same” variable?16 The chaining rule, functional form (zooming out / black-box rule)17 The shared consequence rule, functional version18 The shared influence rule, functional version19 The shared arrow rule, functional form.20 The rule for merging arrows. Weight of evidence.21 The chaining rule with loops22 The rule for conceptual links23 Combining “extra” information like the values or levels of variables24 Context25 Maybe the nodes are not variables, maybe they are … propositions, events, schemas?26 Variable-based and propositional-based approaches to causal diagrams: chalk and cheese??27 Limitations to the proposition-based approach28 Clarifying how to ask QuIP questions in contexts when the final outcomes are not necessarily “changes”29 Coding individual causal fragments using propositions (revised)30 Combining causal fragments from different sources using propositions31 Coding causal maps with propositions32 Types of variable33 Causal thinking is essentially contrastive thinking34 Lo/hi Variables, types of variable, and contrasts35 “For each ..….” variables36 Coding specific influences – individually and in combination37 One function to rule them all?38 Coding a claim about the absence of a causal link39 Strength / importance40 Causal Inference41 Upstream (Bayesian) inference42 Contour43 Valence44 Valence and direction45 Effect: two kinds46 Interventions and differences47 Contribution48 Information about the source of our causal information49 Clusters of similar maps50 Maps of maps51 Probability density functionsIII Simplification 52 Simplifying causal maps: aggregation and filtering53 Aggregating and filtering beliefs54 Aggregation and filtering based on face value55 Aggregation and filtering based on particular research questions56 Aggregation and filtering based on metrics57 Citation intensity58 Conspicuous absence59 Not forwards60 Homogenity of pathsIV Visualising 61 Visualising and formatting causal maps62 Coding using the UI (outdated)V Summary 63 Summary of the rules for inference in causal maps, aka “Soft Arithmetic”VI The why question 64 The “why question” as a generic method in social researchVII The why question 65 The “why question” as a generic method in social research66 The “why question” – asking about changesVIII Using the causal map app 67 Using the causal map app – already partly outdated!68 Using the causal map app for real-time, collaborative theory constructionIX Case studies 69 Case Study: the Strawberry Line70 Case study: Global Young Academy. Tracing the paths of GYA’s impact71 Appendix: previous work72 ReferencesPublished with bookdown
Aggregating and filtering beliefs
How to combine multiple such fragments? (this is QuIP bread-and-butter)
e.g. if source P believes A (completely) causes B and that B (completely) causes C, do we know that they believe that also A (completely) causes C? It’s a logical consequence, but how rational can we assume our sources to be?
(Fiona’s problem) e.g. if source P believes that A (completely) causes B and source Q believes that B (completely) causes C, can we conclude anything about A and C? In other words, if one source tells me about one arrow in a chain, and another tells me about another which links with it, who told us about the whole chain?
(Generalisation problem) e.g. if source P believes A (completely) causes B, can we conclude anything about what some larger population believes?
How to summarise multiple such fragments?
In particular:
How do we combine different overlapping statements in which we have different degrees of trust? See xx