The present study, aligned with the structural approach on social representations (SR), aimed at comparing dichotomous (central vs. peripheral) and continuous models of structural status of SR elements. The sample was formed by 114 undergraduate students who completed context independence tasks and resistance to change evaluations related to 30 elements from 3 SR objects. Non-nested regression models having resistance to change as the criterion variable and dichotomous and continuous operationalizations of symbolic value as predictors were compared. Statistical results complemented by follow-up t-tests challenge the dichotomous model and support a continuous quadratic one. The curve is interpreted as being related to different evaluations when elements are attributed to a majority or a minority and point out to the need to employ within subjects designs in future research assessing more properties with composite measures.
