In this study, a novel approach combining different techniques, including Fourier transform infrared (FT-IR) spectroscopy, photoacoustic spectroscopy (PAS) with an interferometric cantilever microphone, and principal component analysis (PCA) along with a proper data preprocessing procedure, have been used in the investigation of hair samples for cocaine abuse. Hair fibers from cocaine-overdose patients have been measured using a simple procedure involving cantilever-enhanced photoacoustic Fourier transform infrared spectroscopy. In addition, a reference group of hair samples from subjects with no cocaine abuse has been measured. We present a first approach to discriminate the cocaine users from the reference group with the photoacoustic method and PCA. With proper data preprocessing methods, the two groups were successfully separated according to their spectra. The results were confirmed with two different classification methods independent of the principal component data analysis.