PURPOSE: This article discusses the current efforts to measure disability in a comparable manner internationally, the effects of using different types of wording in questions, and the implications of the approach of asking about ‘difficulties’ rather than ‘disability’ on the use of disability statistics. METHOD: The study design was qualitative. Twenty-one focus groups were run with adults responding for themselves. Nine groups were classified a priori by the author as ‘disabled’, six as ‘unsure’, and the last six as ‘non-disabled’. The participants completed a questionnaire using the Washington Group on Disability Statistics (WG) Short Set, the South African Census 2001 question, and the question ‘Are you disabled?’. This was followed by group discussion on these questions and on how the concept of disability is understood by group participants. RESULTS: Participants understand disability as being a permanent, unchangeable state, mostly physical, and where a person is unable to do anything. The participants in the three groups of allocated disability status (disabled, unsure and non-disabled) provided quite different responses on the three questions. All participants in the ‘disabled’ and ‘unsure’ groups reported having ‘difficulty’ on the WG questions, but the ‘unsure’ groups did not identify as being ‘disabled’ on either of the two other questions. CONCLUSIONS: Using questions that ask about ‘difficulty’ rather than ‘disability’ provides a more comprehensive and inclusive measure of disability with a clearer understanding of what is being measured. Asking about ‘difficulty’ provides an improved measure of disability status for effective data collection and analysis to promote development, implementation and monitoring of disability-inclusive policies.