Refinement and Validation of the Social Participation Restrictions Questionnaire: An Application of Rasch Analysis and Traditional Psychometric Analysis Techniques

Abstract

Objectives: The primary aim of this research was to refine and validate the Social Participation Restrictions Questionnaire (SPaRQ). The SPaRQ is a hearing-specific, patient-reported outcome measure that was originally developed through consultation with adults with hearing loss, clinicians, and researchers. This research comprised two studies. Study 1 aimed to assess the psychometric properties of the SPaRQ and to improve these properties by amending the questionnaire (e.g., removing items) as required. Study 2 aimed to validate the refined SPaRQ.

Design: In study 1, 279 adults with hearing loss completed a long-form, 53-item SPaRQ. Rasch analysis, a modern psychometric analysis technique, was used to assess a range of psychometric properties for the questionnaire (e.g., unidimensionality, fit to the Rasch model). The properties of the individual items were also assessed (e.g., response dependency, differential item functioning). In study 2, 102 adults with hearing loss completed the refined SPaRQ. In addition, they completed three questionnaires that had been designed to measure related constructs. These were a hearing-specific questionnaire (Hearing Handicap Inventory for the Elderly), a generic health and disability questionnaire (shortened World Health Organization Disability Assessment Schedule 2.0), and a brief depression and anxiety screening questionnaire (Patient Health Questionnaire-4). Traditional psychometric analysis techniques (e.g., Cronbach's alpha) were used to assess the construct validity and internal consistency of the refined SPaRQ.

Results: Rasch analysis was used to refine the SPaRQ. The result was a 19-item measure divided into two subscales. The 9-item Social Behaviors subscale measured difficulties with performing actions in a social context due to hearing loss. The 10-item Social Perceptions subscale measured negative thoughts and feelings experienced in a social context due to hearing loss. Both Rasch analysis and the traditional psychometric analysis techniques demonstrated that each subscale had strong psychometric properties. In particular, each subscale passed the test of unidimensionality, displayed good fit to the Rasch model, and had high internal consistency. In addition, it was found that, as predicted, each subscale had strong, positive correlations with the hearing-specific questionnaire and moderate, positive correlations with the generic health and disability questionnaire and the depression and anxiety screening questionnaire. Taken together, these findings support the construct validity of the 19-item SPaRQ.

Conclusions: This was one of the first studies to devise a new hearing-specific outcome measure using Rasch analysis. Rasch analysis proved to be a powerful technique for supporting decisions regarding which items to retain in order to achieve a psychometrically robust questionnaire. Additional support for the robustness of this questionnaire came from the utilization of traditional psychometric analysis techniques. Therefore, this questionnaire has the potential to be used in research and clinical practice to evaluate whether auditory rehabilitation interventions improve social participation in adults with hearing loss. The next stage of this research will be to further validate this questionnaire by assessing its responsiveness in a clinical population. The combined use of modern and traditional psychometric analysis techniques should be considered in future questionnaire development and validation research.

Figure 1

Person-Item Threshold Distribution graphs for the Social Behaviors and Social Perceptions subscales showing the distribution of persons (upper half) and items thresholds (lower half). The x-axes display the level of severity on a logit scale, y-axes display the frequencies.