The object of the present study was to determine whether the non-genetic variance of a Type A scale composed of Framingham and Bortner items had gender specific psychosocial components. The study was performed on a group of Swedish twins so that variance explained by heritability for Type A could first be removed from the equation. The overall Type A score had been found to relate to self-reported CHD in this population. The dependent variable was the standardized score residual remaining after removing the genetic variance (i.e., that explained by co-twin score and zygosity). Multiple regression analyses revealed that there were differences in the psychosocial components of the Type A residual in men and women. These results are discussed in terms of culturally accepted gender roles and their possible implications for health endpoints.