The use and quality of reporting of Rasch analysis in nursing research: A methodological scoping review. (August 2022)

• Record Type: Journal Article
• Authors: Stolt, Minna; Kottorp, Anders; Suhonen, Riitta
• Is Part Of: International journal of nursing studies. Volume 132(2022)
• Publication Date: 2022-08
• Subjects: Nursing -- Psychometrics -- Rasch analysis -- Reproducibility of results -- Scoping review; Nursing -- Periodicals; Nursing -- Periodicals; Soins infirmiers -- Périodiques; Nursing; Periodicals; 610.73
• DOI: 10.1016/j.ijnurstu.2022.104244
• Languages: English
• ISSNs: 0020-7489

Abstract

Abstract: Background: Rasch analysis is widely used in the life sciences. Rasch analysis is a mathematical and probabilistic model based on the assumption that the probability of passing a single item is governed by a person's ability and the difficulty of the item. However, its use in nursing science remains unclear. Aim: To (i) describe the use of Rasch analysis in nursing research and (ii) determine the quality of reporting in nursing studies using Rasch models. Methods: A methodological scoping review of literature was conducted. The systematic electronic literature search was initially conducted on 1 February 2020 and updated on 16 April 2021 from PubMed/Medline and CINAHL databases. The search was limited to covering the timeframe from the earliest literature available until 31 December 2020. The search terms used were Rasch, IRT, item response theory, and nursing. The search was limited to the English language and title/abstract level. The analysis included quantification and content analysis. Results: In total, 388 hits were identified. Following a two-phase retrieval process, 88 articles were included in the final analysis. Rasch analysis was used to test the psychometric properties of the newly developed instrument, and validate or test a short version of the existing instrument. The reporting of Rasch analysis demonstrated large variability in quality. Rating scale functioning, internal scale validity using goodness-of-fit statistics, and unidimensionality were the most frequently reported outcomes. Conclusion: The use of Rasch analysis in nursing science was found to be unsystematic. Rasch analysis could provide new possibilities for investigating measurement properties. However, robust, comprehensive, and precise reporting of the methodological choices and results of Rasch analysis is needed. Furthermore, the use of Rasch analysis in nursing science is encouraged. What is already known: Rasch analysis is a mathematical and probabilistic model based on the assumption that the probability of passing a single item is related to a person's ability and the difficulty of the item. Rasch analysis is widely used in the life sciences. However, its use and quality of reporting in nursing science have yet to be explored. What this paper adds: Rasch analysis is rarely used in nursing science, although its use is increasing. This review has identified deficiencies in the reporting of nursing validation studies using Rasch analysis. Minimum standards for the reporting of Rasch analysis in nursing research are proposed.