Is this a good questionnaire? Dimensionality and category functioning of questionnaires used in nursing research.

Abstract

Background: Questionnaires are perhaps the most widely used measuring tools in nursing research, as many studies conducted by nurses focus on understanding the underlying complex factors that are amenable to questionnaires. However, most questionnaires used in nursing research continue to display inadequate evidence of validity under the traditional methods while ignoring the modern Rasch techniques with better proofs of objective measurement.

Aim: To draw researchers' attention to the recurrent limitations of the classical approach to questionnaire design and to suggest advanced psychometric analysis exemplified in Rasch methodology as a more appropriate alternative.

Discussion: For questionnaire data to be suitable for statistical analysis, transparent demonstration of mathematical assumptions embodied in the questionnaire is compulsory. The failure to engage contemporary measurement models in designing good questionnaires raises concerns about researchers' awareness of the application and usefulness of the evidence generated by the modern approach. This paper illustrates with examples the problems inherent in the traditional or classical test theory and advanced dimensionality and category functioning as requisite psychometric properties of a questionnaire. It also outlines several diagnostic parameters that proponents of Rasch techniques recommend for testing.

Conclusion: Traditional methods of assessing and analysing a questionnaire's psychometric properties are no longer tenable because the modern Rasch approach offers exemplary proofs of questionnaire validity rooted in objective measurement theories.

Implications for practice: Nurse researchers using questionnaires for clinical decisions and education purposes should apply the fundamental principles of objective measurements demonstrated in Rasch theory.