The aim of this study was to carry out a gender-differentiated evaluation of the Swedish version of the Rheumatology Attitudes Index (RAI). A translation of the RAI into Swedish was performed. This translated version was applied to a study sample consisting of 37 men and 84 women suffering from non-specific neck and back pain. The RAI was administered three times during a six months period. Reliability was tested by calculating alpha coefficients and test-retest correlations. Construct validity was investigated with variables based on the theoretical concept from which the RAI was developed. Important differences between men and women were found. Internal consistency proved to be higher in the female sample, while the validity was more congruent in the male sample. Overall, the results are ambiguous, thus indicating that the RAI requires further investigation using a gender-differentiated approach before being applied in clinical and research practice. This study demonstrates that in order to obtain meaningful, quantitative and evaluated measures, gender differentiation needs to be included in the development and use of assessment instruments.