Measurement errors normally present in measured process variables can mask the actual process conditions and lead to wrong operational decisions. This is due not only to the imprecise measured values but also to the propagation of the measurement error through process models used to perform control and optimization tasks. One possible alternative to deal with this problem is to characterize the measurement errors online and to incorporate the obtained information in the process operation. The present work analyzes the performances of different methods used for characterization of measurement errors in different scenarios commonly found in the industrial practice. Based on the obtained results, a method is proposed for direct evaluation of measurement errors at plant site.