Rasch analysis of positive changes following adversity in cancer patients attending community support groups
Article first published online: 3 MAR 2010
Copyright © 2010 John Wiley & Sons, Ltd.
Volume 20, Issue 1, pages 98–105, January 2011
How to Cite
Chien, T.-W., Wang, W.-C., Chien, C.-C. and Hwang, W.-S. (2011), Rasch analysis of positive changes following adversity in cancer patients attending community support groups. Psycho-Oncology, 20: 98–105. doi: 10.1002/pon.1713
- Issue published online: 3 MAR 2010
- Article first published online: 3 MAR 2010
- Manuscript Accepted: 29 DEC 2009
- Manuscript Revised: 28 DEC 2009
- Manuscript Received: 18 JUN 2009
- Perceived Benefits Scale;
- Rasch rating scale model;
- Differential item functioning
Objective: This study assessed the 38-item Perceived Benefits Scale (PBS) by examining whether the items constructed a single latent trait and formed an interval scale. This would justify its use to measure the advantages of benefit-finding brought to patients with different cancers from participation in community-based cancer support groups.
Methods: A total of 300 patients were randomly recruited from a 1300-bed medical centre in Taiwan. The Rasch rating scale model was used to examine the model-data fit. Differential item functioning (DIF) analysis was conducted to verify construct equivalence across groups. Comparisons were made among demographic characteristics for various types of patient support groups.
Results: Of the 38 items on the PBS, 28 were applicable to cancer patients and were divided into two distinct unidimensional domains; both met the Rasch model's expectation to constitute a single construct. DIF was found between types of cancers patients, but not between genders. Positive changes following adversity were statistically significantly associated with and ascribed to the duration of patient attendance in community-based cancer support groups.
Conclusion: The two domains verified by Rasch analysis can be used through Rasch-transformed measures to make further statistical inference when comparing positive changes following adversity within and between cancer groups. The psychometric properties of the PBS verified by Rasch modeling fit to the unidimensionality, but need a huge sample size to support its validity and reliability in future studies. Nonetheless, we should be cautious to make comparisons among types of cancer patients due to DIF exhibited in scale. Copyright © 2010 John Wiley & Sons, Ltd.