Notes: Regression Problems from Repeated Measurements
Some consequences of using ordinary least squares estimation procedures with remeasurement data are explained. Of particular concern is the bias of variance estimates and alteration of the confidence coefficient for confidence intervals which result from correlations in the data. Results are presented allowing consideration of the problem's severity for data with two repeated measurements. Finally, alternative estimation procedures including generalized least squares and maximum likelihood are discussed. Forest Sci. 22:382-385.
No Supplementary Data
No Article Media