In this study, cocaine, heroin, and MDMA were analyzed using near-infrared (785 nm excitation) micro-Raman spectroscopy. Drugs were dispersed in solid dilutants of different concentrations by weight. The dilutants investigated were foodstuffs (flour and baby milk formula), sugars (glucose, lactose, maltose, and mannitol), and inorganic materials such as powder. In most cases, it was possible to detect the presence of drugs at levels down to about 10 percent by weight. The detection sensitivity of the Raman technique depended on a number of factors, such as scattering cross-sections of drug and dilutant, fluorescence of matrix or drug, complexity of dilutant Raman spectrum, and spectrometer resolution. Raman spectra from a series of 20 mixtures of cocaine and glucose were collected and analyzed using multivariate analysis methods. An accurate prediction model was generated using a partial least squares algorithm that could predicted the concentration of cocaine in solid glucose from a single Raman spectrum with a root mean standard error of prediction of 2.3 percent. 29 references, 4 tables, and 9 figures