Abstract Recent innovations in portable energy-dispersive X-ray fluorescence (PXRF) spectrometry have increased its utility for the geochemical characterization of obsidian artifacts for archaeological provenance research. However, concerns over the utility of PXRF instrumental analyses have been raised, focused on the validity and reliability of the geochemical data produced. Here we adopt the framework of Richard Hughes ( On Reliability, Validity, and Scale in Obsidian Sourcing Research, 1998), whereby reliability addresses instrument stability and issues of measurement while validity pertains to an instrument’s ability to discern geochemical source provenance. This is done in order to test the utility of PXRF instruments for archaeological provenance research. k-Means cluster analysis was used to test the accuracy of PXRF through statistical comparison of data acquired via laboratory and portable energy-dispersive XRF instruments. Multivariate analysis was employed to demonstrate obsidian source representation at two Classic Maya archaeological sites in southern Belize – Uxbenká and Ek Xux – and to test the validity of data obtained from a PXRF instrument in answering archaeological research questions pertaining to regional interactions between lowland Maya polities. Results suggest that portable XRF instruments produce internally consistent results. However, data acquired from a PXRF instrument are not statistically equivalent to other XRF instruments. This is to say that while PXRF is not a reliable technique, it is valid for questions pertaining to geochemical source representation.