The main purpose for the expansion of supported community care for persons with serious mental illness in Sweden was to ensure the right for these persons to live as citizens in the community. However, earlier research shows that negative attitudes towards mental illness present an obstacle for social integration of persons with serious mental illness. The aim of this study, conducted in Sweden, was to evaluate an existing instrument's (Community Attitudes towards Mental Illness, CAMI), validity and reliability. An additional aim was to adapt and develop the questionnaire to Swedish circumstances. After translation and modification of the original CAMI, the Swedish version of the questionnaire (CAMI-S) was distributed to all student nurses at three different universities in Sweden. The overall Cronbach's alpha coefficient was 0.90 of the original CAMI-S. A corrected inter-item total correlation excluded 20 items because they showed loading <0.43. The overall Cronbach's alpha coefficient on the 20 items (new CAMI-S) that showed loading, >0.43, was 0.903. A factor analysis of these items revealed that the data could be extracted in three factors labelled as: open-minded and pro-integration, fear and avoidance and community mental health ideology. Finally, in order to reach reliable results in attitude research, it is important to measure the respondent's attitude towards the object in common as well as the respondent's attitude to interact with the object. Accordingly, it is important to add behavioural intention items to the 'new CAMI-S'. Statements exemplifying how something 'ought to be' in an impersonal way have a good degree of stability over time and place.