A component analytic method for analyzing multivariate longitudinal data is presented that does not make strong assumptions about the structure of the data. Central to the method are the facts that components are derived as linear composites of the observed or manifest variables and that the components must provide an adequate representation of the observed variables. Specifically, the components are derived so as to minimize the sum of squared errors in the linear regression of observed component variables. Although no structural assumptions are required, the method derives components under a variety of stationary constraints. An advantage of the method for researchers studying change at the individual level is that scores on the derived components are uniquely calculable and have clear properties. The method might, therefore, be usefully applied as a precursor to the application of strong models for change in the component scores at the individual level. The method can easily be generalized to encompass data measured longitudinally in multiple groups. A program has been written in FORTRAN to perform the method's analysis, although the program is not "exportable" at this time. An example of the application of the component analysis method is presented. (TJH)