Objective The slip resistance of footwear is a critical performance indicator for ensuring the walking safety of wearers and preventing slip-related accidents. Most domestic footwear product standards adopt GB/T 3903.6 whole shoe slip resistance performance test method standard as a reference for specifying slip resistance requirements. However, the absence of a precision clause in this method standard compromises the inter-laboratory comparability of test results and weakens the basis for conformity assessment in product standards. This study aims to evaluate the repeatability and reproducibility of GB/T 3903.6, thereby providing reliable data to support the development of a precision clause for the standard and to facilitate laboratory quality control.

Methods An interlaboratory collaborative study was conducted involving 14 laboratories from across China in accordance with GB/T 6379.2. The test specimens included finished footwear, sheet materials, and S96 standard rubber. A total of 14 test levels were established, comprising 8 levels for finished footwear, 4 levels for sheet materials, and 2 levels for S96 standard rubber, with 5 replicate measurements at each level. Mandel h and k statistics were adopted for consistency evaluation, while Cochran and Grubbs tests were used for outlier identification. Based on the one-way analysis of variance model, the repeatability standard deviation (S_r), reproducibility standard deviation (S_R), as well as their corresponding coefficients of variation (CV_r and CV_R) were calculated.

Results (1) Consistency analysis revealed that one laboratory (Lab 12) exhibited systematic positive bias across all 12 test levels, with Mandel h values exceeding the 5% critical value (1.85) in 8 levels. This laboratory was excluded after confirmation by the Grubbs test, leaving 13 valid laboratories for subsequent analysis. (2) The S_r values of all test levels ranged from 0.007 to 0.010, with CV_r ranging from 0.8% to 3.2%, indicating consistently low and stable within-laboratory variability and demonstrating good repeatability. (3) The reproducibility coefficient of variation CV_R differed markedly between dry and wet conditions. Under dry conditions, the CV_R values for finished footwear, sheet materials, and S96 standard rubber were 9.7%-15.4%, 6.8%-7.6%, and 3.8%, respectively; under wet conditions, these values increased to 23.9%-34.6%, 11.9%-12.5%, and 10.9%, respectively. Notably, finished footwear exhibited lower absolute S_R values under wet conditions (0.064-0.092) than under dry conditions (0.087-0.132); the elevated CV_R was mainly attributed to the substantially lower mean friction coefficients under wet conditions, which amplified the relative variability. (4) Under identical test surface conditions, CV_R increased in the order of standard rubber ＜ sheet materials ＜ finished footwear, demonstrating that differences in specimen homogeneity and shoe last assembly were important sources of inter-laboratory variability. (5) Compared with the international precision study by Jung and Fischer (1993), the wet-condition CV_R values in this study (23.9%-34.6%) were consistent with the historical data (25%-36%), indicating that the reproducibility performance of this test method has remained relatively stable over the past three decades.

Conclusions The GB/T 3903.6 method demonstrates good repeatability, but its reproducibility differs significantly between dry and wet conditions, with markedly elevated CV_R observed for wet-condition tests and finished-footwear specimens. The systematic precision data reported in this study, covering three specimen types across 14 test levels, can be applied to routine laboratory quality control and measurement uncertainty evaluation. Furthermore, the identified influences of surface condition and specimen assembly complexity on reproducibility provide a useful reference for unifying relevant test parameters and for developing a precision clause in the future revision of GB/T 3903.6.