Wang, W. and McMechan, G.A., 2019. 3D isotropic elastic reverse time migration using signed magnitudes of elastic data components. Journal of Seismic Exploration, 28: 221-244. Elastic reverse time migration (ERTM) is capable of characterizing subsurface properties more completely than its acoustic counterpart. P- and S-waves coexist in elastic wavefields, and their separation is required before, or as part of, applying the image conditions. Traditional P- and S-wave separation methods based on divergence and curl operators don't preserve the elastic vector information, and the associated polarity reversals of S-wave images are difficult to handle. Thus a preferable workflow for isotropic ERTM should include a vector decomposition of the elastic wavefields and a vector-based image condition that directly uses the signed magnitudes of the decomposed vector wavefields to produce PP and PS images. We propose a new 3D elastic image condition which is a source-normalized crosscorrelation of the signed magnitudes of the decomposed wavefields. The image condition is robust and stable for generating 3D PP and PS images and their corresponding angle domain common-image gathers (ADCIGs) with incident angles calculated from Poynting vectors. Comparisons between the proposed image condition and a vector-based dot-product image condition and show that the proposed image condition generates PP ADCIGs with a wider range of incident angles than existing dot-product image conditions.