HTMT — Heterotrait-Monotrait Ratio for Discriminant Validity
HTMT (the heterotrait-monotrait ratio of correlations) assesses discriminant validity between reflective constructs. This guide explains what HTMT measures, when to use it, its assumptions and how to interpret the result. For choosing between the .85 and .90 criteria, see the HTMT threshold guide (.85 vs .90).
What it is
The heterotrait-monotrait ratio (HTMT) measures discriminant validity - whether two constructs are truly distinct. It compares the average correlation of items across different constructs to the average correlation of items within the same construct. If two constructs really measure different things, between-construct correlations should be clearly lower than within-construct correlations, so HTMT stays low. HTMT was introduced by Henseler et al. (2015) and is now the recommended criterion, because it detects validity problems that the older Fornell-Larcker test often misses. In short: common cutoffs are below .85 for conceptually distinct constructs and below .90 for similar ones (Henseler et al., 2015). For the full decision guide, see HTMT thresholds (.85 vs .90), and for the side-by-side comparison with the older criterion, see HTMT vs Fornell-Larcker for discriminant validity.
When to use it
- You need to show that your constructs are empirically distinct (discriminant validity).
- You are running PLS-SEM or CB-SEM and a reviewer expects modern evidence of validity.
- Your Fornell-Larcker results look fine but you want a more sensitive check.
- You have reflective constructs measured with multiple items each.
When not to use it
- Your constructs are formative — HTMT is designed for reflective measurement.
- A construct has a single indicator — there is no within-construct correlation to compare.
Assumptions
- Reflective constructs with multiple indicators per construct.
- Indicators of each construct are positively correlated (the same direction).
- Reasonable reliability of the items entering each construct.
Thresholds & cutoffs
| Criterion | Rule of thumb | Source |
|---|---|---|
| HTMT (conceptually distinct constructs) | < .85 (strict) | Henseler et al. (2015) |
| HTMT (conceptually similar constructs) | < .90 (liberal) | Henseler et al. (2015) |
| HTMT inference | Bootstrapped CI upper bound should be below 1 (some test against the .85/.90 value) | Henseler et al. (2015); Franke & Sarstedt (2019) |
How to report it in APA
Discriminant validity was assessed with the heterotrait-monotrait ratio of correlations (HTMT). All HTMT values were below the .85 threshold (max = .XX), and the upper bound of the bootstrapped 95% confidence interval did not include 1, supporting discriminant validity (Henseler et al., 2015; Franke & Sarstedt, 2019).
How Explonia runs it
Explonia computes the full HTMT matrix in the browser from the data you collect or upload. It flags any pair above .85 or .90 so you can see at a glance which constructs may overlap, and it can bootstrap a confidence interval for inference. The results feed an APA 'Article description' you can review and edit. Export to R or SPSS (.sav) is available on paid tiers, and HTMT sits alongside AVE, CR and Fornell-Larcker so you can report convergent and discriminant validity together.
Frequently asked questions
What does HTMT measure?
HTMT (the heterotrait-monotrait ratio) measures discriminant validity: whether two constructs are empirically distinct rather than capturing the same underlying concept. It compares the average correlation between indicators of different constructs to the average correlation between indicators of the same construct. A low HTMT value means the constructs behave as separate things, not one construct in disguise. For which value counts as low, see HTMT thresholds (.85 vs .90).
When should I use HTMT?
Use HTMT whenever you report discriminant validity for reflective constructs, typically alongside CFA or PLS-SEM results. It is now the standard criterion expected by reviewers in marketing, management and information systems research, because it catches validity problems that the older Fornell-Larcker test often misses. See HTMT vs Fornell-Larcker for a side-by-side comparison.
Is HTMT for reflective or formative constructs?
HTMT is designed for reflective constructs with multiple indicators each. It relies on correlations between indicators, so it does not apply to formative constructs or to a construct measured with a single item. Formative constructs are assessed differently, for example with weights and collinearity checks.
How is HTMT calculated?
HTMT divides the average correlation between indicators of two different constructs (heterotrait correlations) by the average correlation between indicators of the same construct (monotrait correlations). Run the HTMT calculator to get the value directly from your data.
HTMT or Fornell-Larcker — which should I report?
HTMT is now preferred because it is more sensitive to validity problems that Fornell-Larcker can miss. Many authors report both. See our HTMT vs Fornell-Larcker guide.
Last updated: September 2026
Run HTMT on your own data
Collect or upload your data, run the analysis in your browser, and get a draft APA paragraph you can edit. Free to start. No credit card.
Start free