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.