OBJECTIVES: While requirements for identifying cognitive interviewing respondents are not specifically outlined in the FDA Guidance for Industry Patient-Reported Outcome Measures, best practices dictate that respondents should be as closely representative of the study patient population as possible. For a questionnaire intended for a single patient population or a study with a single patient population, this concept is sufficient for determining criteria. On the other hand, some cases lend themselves to the use of multiple patient populations and more varied respondent criteria. METHODS: Past translation and linguistic validation projects that included cognitive interviewing of multiple patient populations were reviewed along with the process for determining their respective respondent criteria. RESULTS: For each project, the goal of cognitive interviewing was to determine patient understanding of concepts in the intended study population. As such, respondent inclusion criteria were developed to represent the intended study population as closely as possible. Criteria ranged from multiple cancer types, to multiple age groups, as well as caregiver and patient mixtures. In each case, more than one respondent of each sub-category was required in order to ensure sufficient feedback. CONCLUSIONS: Cognitive interviewing of a single patient population typically includes around 5 respondents. With multiple populations, an increased total number of respondents are recommended. Balancing this with study budget and timeline restrictions presents a challenge. While exact representation of multiple patient populations may not be possible, development of a more broadly representative criteria that still includes multiple populations proves beneficial in ensuring that a questionnaire is fit for purpose.