Measurement Error Models (MEM) regression method to Harmonize Friction Values from Different Skid Testing Devices

Skid measurement errors are unavoidable for each kind of skid testing device. The Simple Linear Regression (SLR), used worldwide to harmonize friction measuring devices, does not consider that measurement errors affect both devices. For this reason its use provides biased and not unique estimate of the relationship between devices. The Measurement Error Models (MEM) regression method is proposed as a better method to harmonize any two skid testing devices. SLR and MEM regressions between repeated measurements (from the same device) and between measurements obtained from two different skid testing devices are performed. A comparison of the results is shown and MEM regression appears to be a more appropriate tool to harmonize friction measuring devices instead of SLR.