We study the principal components of covariance estimators in multivariate mixed-effects linear models. We show that, in high dimensions, the principal eigenvalues and eigenvectors may exhibit bias and aliasing effects that are not present in low-dimensional settings. We derive the first-order limits of the principal eigenvalue locations and eigenvector projections in a high-dimensional asymptotic framework, allowing for general population spectral distributions for the random effects and extending previous results from a more restrictive spiked model. Our analysis uses free probability techniques, and we develop two general tools of independent interest—strong asymptotic freeness of GOE and deterministic matrices and a free deterministic equivalent approximation for bilinear forms of resolvents.
Y. S. was supported by a Junior Fellow award from the Simons Foundation and NSF Grant DMS-1701654.
Z. F. was supported in part by NSF Grant DMS-1916198.
We thank Camille Male and Roland Speicher for helpful pointers to the strong asymptotic freeness literature.
"Principal components in linear mixed models with general bulk." Ann. Statist. 49 (3) 1489 - 1513, June 2021. https://doi.org/10.1214/20-AOS2010