Validation and reliability of the Cardiff Visual Ability Questionnaire for Children using Rasch analysis in a Turkish population.

ABSTRACT

AIM: To develop a Turkish version of the 25-item Cardiff Visual Ability Questionnaire for Children (CVAQC) and to evaluate the validity and reliability of the questionnaire. METHODS: The study involved two main phases. The first phase involved a cross-cultural adaptation of the CVAQC from English into Turkish. The second phase involved the completion of the Turkish version of the CVAQC by 150 partially sighted children (6-18 years old) and validity and reliability checks. Extent and construct validity were investigated using Rasch analysis and reliability by internal consistency and person separation index (PSI). RESULTS: An adequate conceptual equivalence was achieved following the linguistic adaptation process. The dataset for validation comprised 150 participants, 88 (58.7%) of whom were male. Evidence of disordered thresholds was found for one item (item 17). This item was recorded by collapsing two categories and ordered thresholds were evident. All items of the CVAQC were found to fit the Rasch model (χ² (df)=59.90 (2), p=0.159). The internal construct validity was good (mean item fit (SD) -0.054 (1.132), person fit (SD) -0.629 (2.079)) indicating a single underlying construct. The reliability was good with Cronbach's α of 0.91 and PSI of 0.94. Differential item functioning (DIF) was tested for age, sex, diagnosis, degree of visual impairment, and comorbidity. Evidence of DIF was found on age for one item (item 10, reading the board at school). CONCLUSIONS: The Turkish version of the CVAQC is a valid, reliable, and unidimensional questionnaire for partially sighted children aged 6-18 years. TRIAL REGISTRATION NUMBER Ankara University Ethics Committees registry number 06-177-12.