Abstrak
According to the characteristic of parameter research, there are three kinds of experiment design models, those are; fixed model, randomized model and mixed model. Estimation and testing of variance (variance component) can be performed only for the randomized model and mixed model. There are severals model that can be used to estimate the variance component, those are ANOVA method, maximum likelihood method (ML), restricted maximum likelihood method (REML) and MINQUE method. MINQUE method is variance component estimation method that to minimize the distance between two sum square, as such as founded the unbiased estimator. MINQUE method has several disadvantages also, that are; that solution does depend on the pre assigned value used in w, different values of w can give different estimators from the same data. The estimation result can be negative value, and using with manual method becomes problem, especially for R calculation. T will have IxI dimension, as such as the T will have a very big dimension and the calculation of T is not easy but we can use computer program in software SAS with statement PROCVARCOM to calculation.