Abstract. The assumption of reference station conditions is investigated for the first time across 60 rock stations belonging to the broadband and accelerometric networks of the National Observatory of Athens. We select the stations based on the established belief that they lie on rock, and provided their data are publicly available through EIDA/NOA and have been for long enough to yield a substantial number of recordings. No site effects studies have been conducted before for the ensemble of the stations under study. Furthermore, no ad hoc field campaigns have been performed to characterise them, save in 2 cases. The first step is to compile all existing information for these stations from all publicly available sources and past studies, including geology, topography, station installation, Vs30 estimates and any other known metadata. The second step is to compile ad-hoc information from maps combined with the operator’s first-hand experience of the sites, to better describe the geological unit and age, along with other characteristics such as station installation and morphology. The third and largest step is to compile the first Greek ground-motion dataset on rock and perform a detailed analysis of the recordings to estimate site-specific transfer functions and hence assess local site response characteristics for each station. A strong-motion dataset of 6840 recordings is developed and curated for this purpose, visually inspected and processed in the time and frequency domains. Single-station amplification functions (horizontal-to-vertical spectral ratios, HVSR) are estimated from the seismic data, and the site resonance characteristics are assessed, not only in the conventional way of combining components, but also assessing directional sensitivity. ‘True’ reference site behaviour implying low, flat amplification with no directional dependence, these elements these transfer function characteristics are combined with the compiled existing and new metadata, to evaluated the stations’ capacity as reference sites. The stations are grouped in terms of behaviour and the preferred ones are recommended, hoping to facilitate the better use of data in future hazard applications.