Community Norms for the Symptom Questionnaire (SQ-48): Normalised T-Scores and Percentile Rank Order Scores

Abstract

Use of standardised scores, such as T-scores and percentile rank order scores, enhances measurement-based care. They facilitate communication between therapists and clients about test results, particularly for multidimensional measures such as the Symptoms Questionnaire (SQ-48). By transforming raw scores into a common metric, clinicians can more easily interpret and discuss patient profiles of scores on the various scales of the measure.

This study explored the advantages and disadvantages of standardised scores and percentile ranks, with a specific focus on T-scores, utilising cross-sectional data from a general population sample (N = 516) and a clinical sample (N = 242). We outline various approaches for establishing T-scores and provide illustrative examples. The analysis of the SQ-48 revealed the necessity of first normalising raw scores to obtain accurate T-scores. Normalisation based on an IRT model is deemed superior, but formulas converting summed scale scores provide a good approximation. Regarding percentile rank order scores, we demonstrated that clinical percentiles offer more meaningful interpretations than population-based percentiles, due to restriction of range for the latter among clinical subjects. Gender and age group differences were identified, with significantly higher scores for women and individuals aged 55 and older. Benefits of normalised T-scores and the need for gender- and age-specific norms for the SQ-48 are discussed.