Résumé : We conducted an international round-robin experiment to determine the consistency of nanoindentation hardness measurements of thin films among six different laboratories, using three different samples. These samples were chosen to present a challenge of indenting at small loads (μN range). They were: 250-nm-thick TiNx, 700-nm-thick TiC, and 500-nm-thick TiB2/TiC multilayer coatings (each layer being 3-nm thick), prepared at Northwestern University using magnetron sputtering on silicon (0 0 1) substrates. Each research team was free to use whatever nanoindentor and analysis methods at its disposal. This round-robin experiment demonstrates that for the hardness range of interest (15-35 GPa) and using well-documented procedures and analysis methods, the reported results from all laboratories are essentially the same, allowing for a statistical spread of approximately ± 14%. For consistent hardness measurements, four precautions must be observed: (i) proper tip-area function calibration, (ii) using sharp indenters, (iii) performing nanoindentation measurements with minimal thermal drift and with drift correction, (iv) using smooth samples, and (v) measuring the full hardness-maximum penetration curve. © 2002 Elsevier Science B.V. All rights reserved.