Quantitative testing of potential modifications to item wording and presentation in the EQ-HWB long form

Ethics approval was granted by the Research Ethics Committee of Corvinus University of Budapest (KRH/284/2024). All respondents provided informed consent prior to participation. All procedures involving human participants were conducted in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Declaration of Helsinki and its later amendments, or comparable ethical standards.

Abstract

Objectives This study compared the psychometric performance of alternative item wording and formats in the experimental EQ-HWB long form: (1) presentation formats for positively worded items, (2) positively worded items from EQ-HWB v1.1 versus negatively worded counterparts from EQ-HWB v1.2, (3) hopelessness item wording, and (4) vision item wording. Methods Overall, 1,200 general-population adults completed the experimental modified EQ-HWB (v1.2) itemby-item with additional item variants and were randomized to three presentation formats for positively worded items: text warning, reversed response scale, or unmodified. Positively worded items (‘accepted’, ‘feeling good’, ‘do things one wanted to’) were compared with negatively worded counterparts (‘excluded’, ‘negative feelings’, ‘difficulties with enjoyable activities’). To identify response errors from negative-to-positive wording switches, positive-item responses were benchmarked against external measures (e.g., happiness). ‘Hopelessness’ was compared with ‘nothing to look forward to’, and vision items with and without reference to aids. Psychometric performance was assessed using distributional properties, and convergent and known-group validity analysis. Results Response errors for positively worded items occurred across formats but were higher in the unmodified format (5.8–6.8%) than across alternative formats (3.3–4.5%). These items showed lower ceilings versus negatively worded counterparts and slightly improved instrument-level known-group validity. ‘Hopelessness’ showed stronger known-group validity than ‘nothing to look forward to’. Referencing aids slightly reduced reported difficulties, while both vision versions similarly distinguished those with chronic eye disease. Conclusions Our findings can inform key design decisions on item wording and presentation in the EQ-HWB long form and should be considered alongside qualitative evidence from the literature.