visual-modeling

# Visual Modeling for Grounded Theory Visual models externalize relationships among categories so you can **spot gaps, contradictions, and over-clutter**. They are thinking tools first; publication figures second. ## Types of visual models - **Concept maps:** nodes (concepts) and labeled links (e.g., “leads to,” “is a type of”). - **Process diagrams:** stages, phases, turning points—useful for temporal theories. - **Conditional matrices:** rows/columns represent conditions and outcomes; cells hold exemplar incidents. - **Network diagrams:** emphasize multiple interacting factors; good for complex ecologies of action. ## When to create them After you have **repeatable categories** and tentative relationships—mid selective/theoretical coding. Early visuals risk crystallizing premature closure; late visuals may only decorate finished text without aiding thought. ## Constructing from sorted memos 1. List major categories on sticky notes or index cards. 2. Move them until spatial proximity reflects **hypothesized relations**. 3. Draw links; label each link with a **verb phrase** (“triggers,” “buffers,” “masks”). 4. For each link, attach **2+ incident IDs** as evidence anchors. ## Conventions for representing relationships - Use consistent arrow direction (cause → effect; condition → strategy). - Distinguish **strong** vs **tentative** links (dashed vs solid) and keep a legend. - Avoid decorating with colors that have no coded meaning. ## Iterating on visual models Each iteration should respond to **disconfirming incidents** or **new theoretical sampling**. Version files (`model_v3.png`) and note what changed in the audit trail. ## Using visuals in publications Provide a clean figure with minimal jargon; define terms in the caption; ensure the figure matches claims in text. Some journals want black-and-white friendly palettes. ## Conditional matrix tips Start small (3×3). Expand only when cells remain meaningful. If a cell is empty, ask whether the absence is **empirical** or a **modeling error**. ## Software options Whiteboards, Miro, Obsidian canvas, PowerPoint, OmniGraffle, Graphviz, CAQDAS maps. Pick what keeps you **fast**; avoid tool obsession. ## Grounded theory alignment Models express **substantive theory**, not literature diagrams. If your figure recreates a framework from a prior study, scrutinize for **forcing**. ## Checklist - [ ] Model type matches the phenomenon (process vs network). - [ ] Links labeled; legend clarifies line styles. - [ ] Evidence references attached for major relationships. - [ ] Iterations versioned; changes explained. - [ ] Publication figure readable at print size. ## References (starting points) - Strauss, A., & Corbin, J. *Basics of Qualitative Research* — conditional matrix tradition (read critically with Glaserian integration goals). - Konecki, K. *Visualizing Grounded Theory*. - Northcutt, N., & McCoy, D. *Interactive Qualitative Analysis* — diagramming practices across approaches.