When recruiting respondents for cognitive interviews testing translated survey questionnaires, researchers often recommend interviewing monolingual non-English speakers because they are the likely users of the translations. However, these individuals are hard to recruit, and there is no standard definition of monolingual. Using cognitive interview data collected from pretesting of Chinese and Korean translations of the American Community Survey Language Assistance Guide, we investigated whether there were differences in respondents' understanding of survey questions according to their level of English proficiency and if such differences remained after considering demographic characteristics. We found that the types of issues reported by monolingual speakers and partially bilingual speakers were similar and that differences seemed to be driven by different demographic characteristics and not necessarily by language proficiency. Our findings suggest the value of evaluating translated questionnaires with individuals having diverse demographic characteristics and recruiting both monolingual speakers and partially bilingual speakers as research participants.