Non-Native Language Questionnaire Usage [Letter]

Dear editor

I read with great interest the Basamih et al study concerning beta-blocker usage among medical students at Umm Al-Qura University.¹ I commend the authors for drawing attention to a staggering rate of self-prescribing among those taking beta blockers for anxiety. Nevertheless, it is pertinent I highlight a principal limitation of this study not referenced in the discussion section.

I note with concern the questionnaire was not validated formally preceding data collection. The authors’ questionnaire was written in English for students attending university in Mecca, Saudia Arabia. The native language of Saudia Arabia is Arabic,² and consequently participants answered the questionnaire in their non-native language. This introduces bias and impacts the reliability of the results, for which I will provide evidence.

A previous study has shown that when fluent non-native English speakers complete questionnaires, they culturally accommodate their answers. When replicating an identical questionnaire in their native language, their answers differ.³ This could lead to different rates of reporting of self-prescription due to cultural variations in self-prescribing between native English-speaking nations and Saudi Arabia.

Furthermore, the exclusion criteria removed incomplete responses. Another study proved that non-native speakers of English are less likely to respond to self-completion sections of questionnaires.⁴ Use of a non–native language questionnaire may have influenced the number of incomplete responses received. This could introduce bias and distort results by reducing the number of completed questionnaires received. A breakdown of the quantity of incomplete responses is not included in Basamih et al’s study, and thus it is difficult to assess the level of distortion.¹

A final criticism of the study design is to give emphasis to additional research demonstrating English questionnaires have a higher rate of middle responses from non-native speakers, whereas when you reproduce the same questionnaire in the participants’ native language, you get more extreme responses.⁵ This is particularly relevant when assessing for side effects and degree of symptom relief when using beta blockers. Non–native language questionnaire use may cause middling responses as opposed to extreme-end responses.

Concluding, to minimise cultural bias and maximise objective, reliable, and complete responses, questionnaire-based study designs benefit from being performed in the native language of the participants. Care must be taken in study design to mitigate factors affecting reliability and bias. If formal validation had been performed for the questionnaire, this issue may have been highlighted prior to data collection.