Using multigroup structural equation modeling in a large sample of online-survey respondents (N = 6,744), the present study examined the reliability and dimensionality of the Male Role Norms Inventory-Short Form (MRNI-SF), a popular measurement of traditional masculinity ideology (TMI), and also tested measurement invariance between individuals that do and do not fit the White heterosexual male TMI reference group. Results indicated that (a) it is appropriate to model the MRNI-SF using either a bifactor or unidimensional model but not a second-order model, (b) the raw MRNI-SF total score is a suitable measure of the general TMI construct, (c) the raw self-reliance through mechanical skills and negativity toward sexual minorities subscale scores may be appropriate measures of their respective specific factors (akin to subscale factors), and (d) SEM or ipsatizing procedures should be used to model the 5 other specific factors, given the insufficient model-based reliability of their raw subscale scores. When comparing men to women, White men to Black and Asian men, and gay men to heterosexual men, the MRNI-SF demonstrated configural invariance and at least partial metric invariance (i.e., measured similar constructs). However, scalar and residuals invariance were only supported for Asian men compared to White men. Taken together, these findings suggest that a general TMI factor of the MRNI-SF is best represented by a bifactor model, even in individuals that do not fit the White heterosexual male TMI reference group, but the instrument may be tapping somewhat different constructs in women, Black men, and gay men.