Interactive SEM Explorer

From model diagram to statistical story.

Explore the locked final SEM one relationship at a time. The diagram follows conventional SEM notation: ellipses are latent variables, rectangles are observed indicators, one-headed arrows are directed paths, and double-headed curved arrows are covariances.

N
689
CFI
0.950
RMSEA
0.055
SRMR
0.038
Interpretive layer · v0.5

What does the final model say?

The final MLR SEM does not treat “fatigue” as a single undifferentiated exposure. Three fatigue-related paths can be read differently: some show statistically significant positive associations with Unsafe Riding, while others do not.

N = 689
MLR · final SEM
Reading rule:β shows the standardized direction and magnitude; p indicates the strength of statistical evidence against a zero direct association. Neither establishes causality in this cross-sectional model.
Coefficient
Locked final model · conventional SEM notation
Click any ellipse, rectangle, directed path or covariance. Dashed purple curve = residual covariance.
p < .05 p ≥ .05 covariance
0.2950.174-0.0700.3950.316-0.0070.8580.000-0.0300.7180.238-0.0350.5120.087-0.1620.7540.8800.8540.9000.8410.8030.7710.5870.8200.6320.8380.8150.3940.8140.5140.4260.000cross-loading0.4200.5160.5080.172~~ .469M-CFQ4M-CFQ5M-CFQ6M-CFQ7M-CFQ8M-CFQ9M-CFQ10M-CFQ11M-CFQ1M-CFQ2M-CFQ3Rode SleepyRiding AsleepFatg. Last MonthPurpose of ridingAge groupRunning RLSpeedingBeh. ChangeCrash Hist.MotivationR² 1.000Prol. Fatg.R² 0.974Low EnergyR² 0.745Phys. Fatg.R² 0.254Mental Fatg.R² 0.273Unsafe ridingR² 0.503LATENT VARIABLESellipsesN = 689 · MLR · 57 parameters
Model fit
χ²(153) = 470.97
CFI 0.950 · TLI 0.938 · RMSEA 0.055 [0.049, 0.061]
Sequential indirect effect
0.180
Prolonged Fatigue → Low Energy → Mental Fatigue → Unsafe Riding; p = .184
Interpretation rule
Association ≠ causation
The SEM is cross-sectional. Paths describe modelled associations, not proof of causal mechanisms.

Same-cohort context

In the same cohort, occupational/service riders reported substantially greater riding exposure (about 6.63 vs 2.99 h/day), while fatigue prevalence was similar (33.4% vs 30.9%) and crash involvement was higher (32.7% vs 22.5%). These are separate analyses—not coefficients from this SEM—and help motivate examining exposure, fatigue and unsafe riding as related but distinct phenomena.