Previous studies using particle image velocimetry (PIV) and laser Doppler anemometry (LDA) have raised the question of how these measurements should be compared. This study reports on the difference between Favre-averaged and Reynolds-averaged velocity statistics for a turbulent burner using PIV and LDA for unconditional and conditional velocity measurements. The experimental characterization of flow fields of premixed and stratified methane/air flames is carried out under globally turbulent lean conditions (global equivalence ratio at 0.75), over a range of stratifications and swirl numbers. Unconditioned velocity data was acquired using aluminium oxide to seed the flow field. Conditioned measurements were performed using vegetable oil aerosol as seed, which burns through the flame front, thus allowing only the non-reacting flow velocities to be obtained. A critical comparison of unconditioned velocity profiles measured using both PIV and LDA, including axial, radial, and tangential components is made against conditioned and reconstructed mean velocities at different cross-sections of the flame. The comparison reveals how the differences between the Favre-averaged (unconditioned) and the Reynolds-averaged (conditioned) velocity measurements in the flame brush region can be accounted for using the mean progress of reaction, and highlights the limits of the accuracy and agreement between PIV and LDA measurements. © 2014 The Combustion Institute. Published by Elsevier Inc. All rights reserved.