As analytical problems addressed using time-of-flight secondary ion mass spectrometry (ToF-SIMS) increase in chemical complexity, multivariate analysis (MVA) methods have become standard tools for simplifying the interpretation of ToF-SIMS spectra and images. MVA methods can significantly simplify ToF-SIMS datasets by providing a comprehensive description of the data using a small number of variables, typically in an automated fashion requiring minimal user intervention. However, successful and widespread application of MVA methods to SIMS data analysis is limited by a lack of understanding of the outputs of MVA methods and optimization of these methods for ToF-SIMS data analysis. Appropriate selection of data pre-processing and MVA tools are critical for accurate interpretation of ToF-SIMS spectra and images. As an example, an image dataset of a selectively ion-etched polymer film was analyzed to identify and characterize the chemically distinct regions in the image. Principal component analysis (PCA) and multivariate curve resolution (MCR) after pre-processing using normalization or Poisson-scaling were compared to identify the etched and non-etched regions of the image. The utility of each pre-processing and MVA method was examined, with MCR coupled with Poisson-scaling being the appropriate choice for identifying the different chemical phases present in the image. However, appropriate selection of data pre-processing and MVA methods generally depends on the specific dataset being analyzed and the goals of the analysis. (c) 2006 Elsevier B.V. All rights reserved.