A data processing method is proposed for estimating the measurement error of scaled test scores; this involves scaling half-test raw scores to effective full-test raw scores and converting them to scaled test scores. It is shown that the method allows to estimate the measurement error with high accuracy. The proposed approach to data visualization involves the removal of the part of the data corresponding to the rarest values. In the case of very noisy data, this method helps to eliminate the contribution of atypical values and provides a significant increase in the visibility of scatter plots.