HTMT Ratio

The Heterotrait-Monotrait ratio assesses discriminant validity by comparing the average correlation between items belonging to different constructs against the average correlation among items within each construct. If items measuring supposedly distinct constructs correlate about as strongly as items within a construct, those constructs are not empirically distinct.

The common thresholds are 0.85 for conceptually distinct constructs and 0.90 for conceptually similar ones. Values below the threshold support discriminant validity; values above it indicate overlapping constructs that should be merged, respecified, or stripped of redundant items.

HTMT was introduced because the older Fornell-Larcker criterion and cross-loadings examination were shown to detect discriminant-validity problems unreliably, particularly in PLS-SEM. Best practice is to report HTMT with a bootstrapped confidence interval and confirm that its upper bound does not include 1.