Anthropomorphism is a key construct in social human-robot interaction, with a variety of measurements. Our experiment compares the three most used scales on construct and convergent validity as well as reliability. We used a mixed methods design where participants first viewed a video featuring either a Sota or a TIAGo robot, and then completed each scale with randomisation of the order of presentation. Results indicate that the Mind Attribution scale (Gray et al., 2007) and the Individual Differences in Anthropomorphism questionnaire (Waytz et al., 2010) experienced some order effects and appear to be founded on a different latent process than the revised Godspeed questionnaire (Ho & MacDorman, 2010; Bartneck et al., 2009), though the different scales yield mostly comparable results.