This paper analyzes the net migration rate of the school age population in substate levels (minor civil divisions) between 1960 and 1970. The investigation includes a series of estimation procedures in order to carry out the analyses. The estimation procedures are unbiased techniques (ordinary least squares and maximum R2 improvement), biased techniques (ridge regression, generalized ridge regression and principal components regression), as well as robust techniques (least absolute deviation and Hill-Holland). For each of these estimation procedures, a model was developed, based on a sample of 50 minor civil divisions in the State of New Jersey. A cross validation was made on two subsequent samples of 50 minor civil divisions each in order to measure the amount of reduced shrinkage of the square of the multiple correlation. © 1984.