This paper uses data from five countries (Argentina, Chile, Italy, Portugal and Spain) to propose and test a new comparative measure of affective polarisation departing from Wagner’s recent contributions. Most of the theoretical arguments that support the concept of affective polarisation rest on social identity theory and its consequences on partisan in- and out-groups. However, current comparative measures of affective polarisation (Spread –WAPS- and Distance –WAPD-) neglect the fact that individuals’ likes and dislikes towards different parties might result from instrumental individual evaluations as much as partisan group party expressive identities (Huddy et al. 2015; Huddy et al. 2018), measured as the level to which each individual is attached to their respective group of partisans. In this paper we propose to overcome this problem with a new measure based on the following: A) a battery of feeling thermometers (like/dislike) toward different groups of party supporters (voters); B) an Identification with a Psychological Group (IDPG) scale to capture self-identified partisan social identity, designed to capture the perception of shared experiences and characteristics of group members; C) substituting the existing weights included in the WAPS and WAPD formulas that are based on the level of electoral support for each party, for the individual’s score on the IDPG latent factor.