On evaluation of different methods for quality control of correlated observations