The article discusses how to estimate the mean of a random process when data samples are correlated. When samples are independent, the mean estimate quality can be approximated by estimating the process variance. But when samples are dependent, the mean estimate variance depends on the process covariance. To determine the stability of the mean, information on the covariance or spectrum is needed, which is unknown and must be estimated from the data. Three procedures for estimating this information are discussed in the article.