<p>Force-feedback or kinesthetic haptic devices are usually robotic arms with a handle at the end, which allows the user to feel both the roughness and the hardness of a virtual surface. That handle is called the end-effector, and it can be found in different shapes for different kinds of grasps and applications. While some is known about the effect of distinct end-effectors on roughness perception using this kind of devices, less is known on hardness. Actually, hardness is typically simulated using stiffness, as in a spring defined by Hooke’s law, which implies that hard surfaces that should have infinite stiffness are just approximated by a higher value of stiffness, and thus the user can still penetrate a little in that surface. The aim of this work is to compare several end-effectors in a stiffness discrimination task and assess whether there is a significant influence or not. In particular, four distinct end-effectors—Thimble, Stylus, Knob and Handle—have been compared on the same haptic device—a Phantom Premium—. The results showed no significant differences in absolute accuracy depending on the type of end-effector used, which are in line with what previous authors found on roughness, but this time confirmed on stiffness too. This suggests that people can adapt their exploration strategies to different end-effectors, and that using one or another is a matter of ergonomic preference or application requirement.</p>