The bias is an indication of the systematic error that occurs when a plant species is predicted without being in the calibration set [32]. So, the t test (95% probability) was used to determine if the validation estimates show a statistically significant bias. Except for the TS model (Table 3), all other chemical properties presented values lower than the tcritical value, indicating that the analyses based on multivariate models are expected to give essentially the same average result as the measurements conducted by the reference method. For TS model, there is a 95% probability that the values estimated by the model will not give the same average results as the reference methods, indicating that the validation estimates show a statistically significant bias.