Longitudinal Convergent Validity
Also known as: longitudinal construct validity, repeated-measure convergent validity, cross-time convergent validity, temporal convergent validity
Longitudinal convergent validity evaluates whether a scale's indicators correlate with theoretically related constructs not just at a single time point but consistently across repeated measurement occasions. It extends standard convergent validity testing into longitudinal designs to ensure that the scale measures the intended construct in the same meaningful way over time.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use longitudinal convergent validity assessment whenever a scale is administered at two or more time points and the research question requires comparing scores, modelling change, or making causal claims across waves. It is especially important in intervention studies, developmental research, and clinical trials where pre-post or multi-wave comparisons are central. Do not rely on a single-wave convergent validity check alone when data are longitudinal — the single-wave result cannot tell you whether the construct-criterion alignment is stable. Avoid this approach when only one measurement wave is available, when the criterion construct is not assessed at multiple waves, or when sample sizes per wave are too small (typically fewer than 100–200 per wave) to support latent variable modelling with adequate precision.
Strengths & limitations
- Provides a more rigorous validity argument than single-wave convergent validity by demonstrating that construct-criterion alignment is not a one-time artefact.
- Detects construct drift — gradual changes in what a scale actually measures over time — that single-occasion analyses cannot reveal.
- Integrates naturally with measurement invariance testing, yielding a comprehensive picture of scale behavior across waves.
- Supports causal and change-score interpretations that depend on the scale measuring the same thing at each occasion.
- Can distinguish true score change from instability in the measurement model itself.
- Requires a longitudinal design with the same scale and a relevant criterion measure administered at multiple waves, which is resource-intensive.
- Depends on at least metric measurement invariance being achieved; partial invariance complicates the interpretation of convergent correlations across waves.
- Large sample sizes are needed to fit latent variable models at each wave with sufficient statistical power to detect meaningful differences in convergent correlations.
- Does not provide a single summary statistic; evidence must be synthesised across multiple indices and waves, which requires careful judgment.
Frequently asked
Do I need full metric invariance to assess longitudinal convergent validity?
Full metric invariance is preferred because it ensures that loadings are equal across waves, making latent-variable convergent correlations directly comparable. Partial metric invariance — where most but not all loadings are constrained equal — is sometimes acceptable if the free loadings involve items not central to the core construct, but the interpretation of cross-wave comparisons must be qualified accordingly.
How large should the convergent correlations be?
A common benchmark is that a latent correlation with a theoretically related construct should be at least 0.50 in absolute value, combined with average variance extracted (AVE) exceeding 0.50 for each factor. However, these are guidelines, not rigid rules — constructs that are expected to be only moderately related may show smaller but still meaningful convergent correlations.
Can I assess longitudinal convergent validity with only two waves?
Yes. Two waves are the minimum; you can test metric invariance and compare convergent correlations at waves one and two. More waves give a clearer picture of whether convergent evidence is stable or drifts over time, but two-wave studies are a legitimate and common starting point.
What distinguishes longitudinal convergent validity from longitudinal predictive validity?
Convergent validity compares the focal scale with a theoretically related criterion measured at the same wave. Predictive (or prospective) validity correlates the scale at an earlier wave with a criterion at a later wave. Both are valuable; they answer different questions about how a construct relates to others in time.
How is longitudinal convergent validity different from longitudinal measurement invariance?
Measurement invariance testing asks whether the measurement model — loadings, intercepts, residuals — is equivalent across waves. Longitudinal convergent validity asks whether the construct relates to external criteria in a consistent and theoretically meaningful way over time. Invariance is a precondition for the validity assessment, not a substitute for it.
Sources
- Cole, D. A. & Maxwell, S. E. (2003). Testing mediational models with longitudinal data: Questions and tips in the use of structural equation modeling. Journal of Abnormal Psychology, 112(4), 558–577. DOI: 10.1037/0021-843X.112.4.558 ↗
- Widaman, K. F. & Reise, S. P. (1997). Exploring the measurement invariance of psychological instruments: Applications in the substance use domain. In K. J. Bryant, M. Windle & S. G. West (Eds.), The science of prevention: Methodological advances from alcohol and substance abuse research (pp. 281–324). American Psychological Association. link ↗
How to cite this page
ScholarGate. (2026, June 3). Longitudinal Convergent Validity. ScholarGate. https://scholargate.app/en/psychometrics/longitudinal-convergent-validity
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Construct ValidityPsychometrics↔ compare
- Convergent ValidityPsychometrics↔ compare
- Discriminant ValidityPsychometrics↔ compare
- Longitudinal CFAPsychometrics↔ compare
- Longitudinal Discriminant ValidityPsychometrics↔ compare
- Longitudinal Measurement InvariancePsychometrics↔ compare