Human-robot interaction (HRI) is a topic of major interest in the robotics community, where robots are envisioned to be used for interacting with humans not only socially, but also physically. Therefore, safety has been one of the main concerns, mainly when physical contact may happen or is necessary. Safety standards for collaborative robots have been introduced in ISO/TS 15066 in 2016, where a set of limitations and strict rules have been imposed. In the state-of-the-art research on HRI, physical humanrobot interaction (pHRI) is often more focused on the development of controllers, while social human-robot interaction (sHRI) is more focused on the evaluation of the perception, e.g. sense of comfort, safety of the user with respect to robots, often by means of post-experiment questionnaires. In sHRI there is generally no physical contact between the human and the robot, whereas in pHRI the social aspects are usually not taken into account, and robots are often overconstrained to ensure the aforementioned safety, becoming rather “passive” and “slow”. We believe that the efficiency of physical interaction with robots could be boosted with “active” interactions. We refer to this as active physical human-robot interaction (active pHRI), i.e. an interaction during which robots may take autonomous physical actions on the human without prior notifications. To this end, robots should be able to adapt to the human counterparts during the interaction, i.e. be “active” and at the same time be “human-aware”. In this work, we attempt at addressing the identification of possible useful interaction factors during active pHRI that could be used for the future development of active controllers, by combining methodologies from both sHRI and pHRI, analyzing the users’ state via both physiological data, physical data, and questionnaires. As physiological and physical data alone are of difficult interpretation, we try to find correlations that may give interesting insights.