In the book, such methods of covariance estimation have been studied, which allow to make the measurement results more precise in case the observation data have different degrees of precision and are correlated. To compare the results of the measurements accomplished at different times, a test of the homogeneity of several correlation coefficients has been applied. It is a known fact that during the estimation of any output value it is impossible to be sure that the input values'' correlation ratios can be neglected. Therefore, in practice any function under investigation has to be estimated with the use of sampled data and such methods of statistics estimation, which will specify the covariance of dependent variables. For example, if the function under investigation is hypothetically linear, the input values'' correlation ratio is estimated with the use of a sampled correlation coefficient, applying statistical or empirical methods. Operating just a small number of sampled input values, it is not possible to be sure that the function is linear.