Psychometric limitations of the 13-item Sense of Coherence Scale assessed by Rasch analysis

Abstract Background A person’s sense of coherence (SOC) reflects their perception that the world is meaningful and predictable, and impacts their ability to deal with stressors in a health-promoting manner. A valid, reliable, and sensitive measure of SOC is needed to advance health promotion research based on this concept. The 13-item Sense of Coherence Scale (SOC-13) is widely used, but we reported in a previous evaluation its psychometric limitations when used with adults with morbid obesity. To determine whether the identified limitations were specific to that population or also generalize to other populations, we have replicated our prior study design and analysis in a new sample of adults with inflammatory bowel disease (IBD). Methods A sample of 428 adults with IBD completed the SOC-13 at a routine clinic visit in Norway between October 1, 2009 and May 31, 2011. Using a Rasch analysis approach, the SOC-13 and its three subscales were evaluated in terms of rating scale functioning, internal scale validity, person-response validity, person-separation reliability and differential item functioning. Results Collapsing categories at the low end of the 7-category rating scale improved its overall functioning. Two items demonstrated poor fit to the Rasch model, and once they were deleted from the scale, the remaining 11-item scale (SOC-11) demonstrated acceptable item fit. However, neither the SOC-13 nor the SOC-11 met the criteria for unidimensionality or person-response validity. While both the SOC-13 and SOC-11 were able to distinguish three groups of SOC, none of the subscales could distinguish any such groups. Minimal differential item functioning related to demographic characteristics was also observed. Conclusions An 11-item version of the sense of coherence scale has better psychometric properties than the original 13-item scale among adults with IBD. These findings are similar to those of our previous evaluation among adults with morbid obesity and suggest that the identified limitations may exist across populations. Further refinement of the SOC scale is therefore warranted