READING THE VISUAL GRAMMAR

Two views. One coherent reading.

Technical manual
01

Define the quantities

For each visual component Xⱼ, use the direct effect c′ⱼ and the specific indirect effect aⱼᵢbᵢ through each mediator Mᵢ.

Iⱼ = Σᵢ(aⱼᵢbᵢ)cⱼ = c′ⱼ + Iⱼ

Sum pathways within each component. Each component has its own Iⱼ; effects are never pooled across components.

02

Read the Mediation Map

Horizontal position represents the direct effect c′. Vertical position represents the total indirect effect I.

When their signs match, the pathways reinforce one another. Opposite signs indicate opposing pathways. The dashed diagonal is I = −c′, so c = 0. Above it, the total effect is positive; below it, it is negative.

03

Read the Fingerprint

Every row represents one component. Each fixed-colour segment is a specific indirect effect. Positive segments extend right of zero; negative segments extend left.

Their signed sum equals that component’s height on the map. A small net indirect effect can conceal larger opposing contributions. The fingerprint exposes that composition.

Compare coefficients on a defensible scale

Cross-component comparisons require effects on a consistent, justified scale and compatible model specifications. Visualscapy does not standardise coefficients, estimate a model, infer significance, or establish causality. Consult the original estimates, bootstrap confidence intervals and model diagnostics.

Input schema, precision & exports

Required long-format columns are component, mediator, specific_indirect and direct_effect. Names are case-sensitive. Every component must have the same configured mediator set, with no duplicate pairs. Numeric values must be finite; explicit zero is supported.

Displayed values are rounded to three decimal places. Calculations and CSV exports retain JavaScript numeric precision. A tolerance of 10⁻¹² identifies numerical boundaries and checks repeated direct effects; it is not a significance threshold.

SVG exports preserve editable vector text and shapes. PNG exports use a 300 ppi density tag and a high-resolution canvas. Very large figures may require SVG instead. Exported figures show every component and omit the interactive selection highlight. Simulated data remain labelled.

SEPARATE COMPANION MODULE

Visualscapy Survey

A companion module for standardised image presentation and collection of perceptual ratings. It is not the Visualscapy visualisation platform and is not required to use Visualscapy.