Combines the per-construct-pair HTMT gradients with the asymptotic
variance-covariance matrix of the indicator correlations (calculateCorVCV())
into the delta-method standard error of each HTMT, sqrt(t(g) %*% Sigma %*% g).
The variance-covariance matrix is formed once and reused for every pair.
Gradient vectors and variance-covariance matrix share the lower-triangular
ordering of the indicator correlation matrix, so they align by position.
calculateHTMTasymptoticSE(.gradients, .X)A named numeric vector of standard errors, one per construct pair.
A named list of gradient vectors, one per construct pair, each aligned with the lower triangular of the indicator correlation matrix.
A matrix of processed data (scaled, cleaned and ordered).
No lower bound is imposed on the variance: since Sigma is positive
semi-definite, t(g) %*% Sigma %*% g is non-negative by construction, so a
negative value would signal a genuine problem and is deliberately allowed to
surface as NaN (with a warning) rather than being silently clamped to zero.