A method used to adjust correlation coefficients upward because of errors of measurement when two measured variables are correlated; the errors always serve to lower the correlation coefficient as compared with what it would have been if the measurement of the two variables had been perfectly reliable.