One problem of practical interest is the evaluation of catalyst activity and selectivity of commercial fluid catalytic cracking (FCC) catalysts. Although different statistical methods may be used for this purpose, these methods are usually neglected or oversimplified in catalysis. In a previous study [1], it was shown that experimental errors of independent Variables may exert a significant impact upon the discrimination of catalyst performance in FCC processes and should not be disregarded. In this work it is shown that actual experimental measurements can present strong correlation and that the existence of covariance among experimental measurements can also influence the catalyst discrimination procedure. The influence of covariance among variables on parameter estimation and on the confidence interval (CI) of predicted values is discussed. It is shown that considering the covariances may have either a positive or a negative effect on the CI for predicted values. Practical examples regarding the microactivity test (MAT) of FCC catalyst characterization are used to illustrate the discussion, although the conclusions can be extended to systems other than catalytic cracking.