BACKGROUND: Mastery refers to the degree to which people perceive that they can control factors that influence their life situation, and has been found important for people's quality of life and well-being. It is thus essential to be able to measure mastery in a valid and reliable way. AIM: This study aimed at using the Rasch measurement model to investigate the psychometric properties of a Swedish version of the Pearlin Mastery Scale (Mastery-S). METHODS: A sample of 300 healthy individuals and 278 persons with mental illness responded to the Mastery-S. Item responses were Rasch analysed regarding model fit, response category functioning, differential item functioning (DIF) and targeting, using the partial credit model. RESULTS: The Mastery-S items represented a logical continuum of the measured construct but one item displayed misfit. Reliability (Person Separation Index) was 0.7. The response categories did not work as expected in three items, which could be corrected for by collapsing categories. Three items displayed DIF between the two subsamples, which caused a bias when comparing mastery levels between subsamples, suggesting the Mastery-S is not truly generic. CONCLUSIONS: The Mastery-S may be used to obtain valid and reliable data, but some precautions should be made. If used to compare groups, new analyses of DIF should first be made. Users of the scale should also consider exempting item 6 from the scale and analyse it as a separate item. Finally, rewording of response categories should be considered in order to make them more distinct and thereby improve score reliability.
BACKGROUND: Mastery refers to the degree to which people perceive that they can control factors that influence their life situation, and has been found important for people's quality of life and well-being. It is thus essential to be able to measure mastery in a valid and reliable way. AIM: This study aimed at using the Rasch measurement model to investigate the psychometric properties of a Swedish version of the Pearlin Mastery Scale (Mastery-S). METHODS: A sample of 300 healthy individuals and 278 persons with mental illness responded to the Mastery-S. Item responses were Rasch analysed regarding model fit, response category functioning, differential item functioning (DIF) and targeting, using the partial credit model. RESULTS: The Mastery-S items represented a logical continuum of the measured construct but one item displayed misfit. Reliability (Person Separation Index) was 0.7. The response categories did not work as expected in three items, which could be corrected for by collapsing categories. Three items displayed DIF between the two subsamples, which caused a bias when comparing mastery levels between subsamples, suggesting the Mastery-S is not truly generic. CONCLUSIONS: The Mastery-S may be used to obtain valid and reliable data, but some precautions should be made. If used to compare groups, new analyses of DIF should first be made. Users of the scale should also consider exempting item 6 from the scale and analyse it as a separate item. Finally, rewording of response categories should be considered in order to make them more distinct and thereby improve score reliability.