The purpose of this study was to assess the psychometric properties of the Mental Toughness Questionnaire 48 (MTQ48) and assess the measurement invariance across elite, amateur and non-athletes. In total, 1096 participants aged between 18 and 58 years with a range of athletic experience - elite (n = 181), amateur (n = 577) and non-athletes (n = 338) - from various sports completed the MTQ48. The internal consistency of the scale was gauged through Omega for the total and relevant subscales. Factorial validity was assessed using exploratory structural equation modeling in order to provide a comprehensive estimation of the scales dimensionality. Overall, results offered support for the scales reliability with acceptable internal consistency reported at the total and subscale level. However, the validity of the MTQ48 for the use with athletes of different levels may be questioned. The MTQ48’s hypothesised four-factor model did not fit the data well, whereas the six-factor model produced acceptable levels of fit with large degrees of misspecification in the factor structures across elite, amateur and non-athletes. The results caution the use of the scale with elite athletes and call for refinement of the measure at the subscale level.