The Oxford Happiness Questionnaire: Transformation from an Ordinal to an Interval Measure Using Rasch Analysis

Abstract

The scientific study of happiness requires accurate measurement of the construct that satisfies assumptions of parametric statistics and thus allows both researchers and clinicians to make reliable and valid comparisons with the relevant data sources. The 29-item Oxford Happiness Questionnaire (OHQ) is a widely-used scale for assessment of personal happiness. While its psychometric properties are acknowledged to be acceptable, it presents scores on an ordinal scale and may thus not discriminate precisely between individual happiness levels. The current study aimed to improve precision and item functioning of the OHQ by applying Rasch analysis to a sample of 281 participants. To correct disordered thresholds items were rescored in a uniform fashion. Four items displayed poor relationships with the latent trait of happiness and were removed. Best fit to the unidimensional Rasch model was achieved after locally dependent items were combined into subtests and adjusted for personal differences. Using the ordinal-to-interval conversion tables published here, ordinal OHQ scores can now be transformed to interval level data and thus subjected to parametric statistical analysis without violating fundamental assumptions. The precision of the instrument can be improved significantly by these minor modifications without the need to modify the original response format.