Missing data constitute a common but widely underappreciated problem in both cross-sectional and longitudinal research. Furthermore, both the gravity of the problems associated with missing data and the availability of the applicable solutions are greatly increased by the use of multivariate analysis. The most common approaches to dealing with missing data are reviewed, such as data deletion and data imputation, and their relative merits and limitations are discussed. One particular form of data imputation based on latent variable modeling, which we call Multivariate Imputation, is highlighted as holding great promise for dealing with missing data in the context of multivariate analysis. The recent theoretical extension of latent variable modeling to growth curve analysis also permitted us to extend the same kind of solution to the problem of missing data in longitudinal studies. Data simulations are used to compare the results of multivariate imputation to other common approaches to missing data.