Path Analysis

Path Analysis tests a network of hypothesised causal relationships among observed variables, decomposing into direct effects, indirect effects (through mediators), and total effects.


Step 1 — Define your model

Specify paths (directed arrows): for each path, the from variable and to variable.

  • Variables with only outgoing arrows = exogenous
  • Variables with incoming arrows = endogenous

Supports Excel Import, Sample Data Generator, and Manual Entry.


Step 2 — Results

Path coefficients:

  • Standardised beta, SE, t-value, p-value per path

Direct, indirect, and total effects:

  • Direct = path coefficient of direct arrow
  • Indirect = product of coefficients along mediated paths
  • Total = direct + sum of indirect

Variance explained:

  • R-squared and Adjusted R-squared per endogenous variable

Path diagram — visual model with labelled coefficients

A UniversalChatBot is available for discussion.


Statistical methods used

Estimation: Standardised OLS regression per endogenous variable.

Data standardised first (mean 0, SD 1), so coefficients are standardised path coefficients.

Per-path inference: t = beta/SE, two-tailed p from t-distribution.

H0: path coefficient = 0. H1: != 0. alpha = 0.05.

Indirect effects: Product of path coefficients along mediating routes.

Total effect: Direct + sum(all indirect effects).

Limitations: No global fit indices (use SEM for those). No formal test of indirect effects. Observed variables only.

Causal interpretation requires theory and study design, not just statistics.