The results of an inter-laboratory comparison of silica analysis of seven water samples (coded as IIE19-IIE25) are presented. The first six samples (IIE19-IIE24) were prepared by diluting a certain amount of silicon (Si) standard 1000 +/- 0.1% ppm, HYCEL, Mexico, in distilled water, while the seventh sample IIE25 is natural geothermal water. The conventional true values for the silica (SiO2) concentration of IIE19-IIE24 are 61.1, 213.9, 366.7, 520, 825, and 978 ppm, respectively. The present analysis database consists of 27 datasets, analyzed by colorimetry (13 sets), atomic absorption spectrometry (7 sets) and inductively coupled plasma (7 sets) techniques. The analyses were performed with two procedures: (i) individual laboratory procedure (or conventional procedure) and (ii) dilution of samples in the range 50-200 ppm SiO2. The mean ((x) over bar) and standard deviation (s) calculation after eliminating extreme values (outliers) with (x) over bar +/- 2s method was used for the statistical data treatment. Presently, the overall error in the analysis is at least +/- 10%. A comparison of analytical error and uncertainty for individual analytical technique suggests that the inductively coupled plasma (ICP) and colorimetry are relatively better techniques for the analysis of SiO2 in geothermal waters; however, there are some statistical artifacts on the colorimetric results because their number of measurements was double that of atomic absorption spectrometry (AAS) and ICP. The analysis quality (i.e., precision and accuracy) may be improved by using the data control procedure with some high quality standard (e.g., NIST standards). One of the factors of systematic error is associated with the preparation of laboratory standards. (C) 2012 Elsevier Ltd. All rights reserved.