Psychometric testing of the Italian and French versions of the Care Dependency Scale.

AIM: The aim of this study was to test psychometrically the Italian and French versions of the Care Dependency Scale. BACKGROUND: The Care Dependency Scale assesses changes in patients' level of care dependency including important functional and mental dimensions. Evaluation of the psychometric properties of the Italian version is still ongoing. The French version has to date not been validated. DESIGN: Nationwide cross-sectional point prevalence study. METHOD: Data were extracted from the national, annual prevalence survey of hospital-acquired pressure ulcers and inpatient falls in Swiss acute care hospitals in 2011. A total of 799 Italian and 1068 French-speaking patients were included in the analysis. For the evaluation, the psychometric properties were tested for each language both separately and conjointly. RESULTS: The scales revealed high internal consistency. Factor analysis presented a one-factor solution for both versions separately as well as combined. Comparison of internal structure revealed an excellent degree of equivalence between the versions. Highly significant Spearman correlations between the Care Dependency Scale and the Braden Scale sum scores indicated satisfactory criterion validity. CONCLUSION: Both the Italian and the French versions of the Care Dependency Scale showed satisfactory psychometric properties and a high level of equivalence. Further psychometric testing, using modern test theory approaches, is required. However, the scale is recommended as a valid instrument for further use in Italian and French.

The effects of anonymity on student ratings of teaching and course quality in a bachelor degree programme.

RESEARCH QUESTION: Are there any clear differences between the outcomes of anonymous and personalised student evaluations of teaching quality? METHODS: During a two-year period students were randomly divided into two separate groups, "anonymous" and "personalised", for end-of-module evaluations. The quality of the module was assessed using a standardised questionnaire. Additionally, students were given the option to add "further comments" if they wanted to highlight specifics. These optional comments were independently assessed by three people, using a five-dimensional rating instrument: positive/negative; differentiated/absolute; naming a person/general; containing an order/neutral; visually accentuated/blank. The database consisted of 615 evaluation forms, of which 306 were completed anonymously. In order to identify whether there were any differences between the anonymous and personalised data, a multivariate variance analysis was performed. Based on the scale, the answers to the questions and the quality of the comments were evaluated. Furthermore, an assessment was made to determine if there were any differences in the number of optional comments between the two groups. RESULTS: No significant differences were identified in the informative quality of data between the anonymous and personalised student evaluations. However, students in the personalised group had a tendency to include more details in their written answers. CONCLUSION: Personalised evaluations do not generate more biased results in terms of social desirability, as long as the evaluation concept is characterised by a closed-circle process and is transparent. In other words, it is imperative that the outcomes of the evaluation are reported back to the students. Moreover, there has to be an opportunity for students to discuss any further suggestions and/or future desires in an open environment. In this way the students respect and understand that their feedback is being taken seriously; consequently, they feel able to provide a constructive and honest evaluation.