Cross-cultural research: trying to do it better. 2. Enhancing data quality.

OBJECTIVE: To describe the dilemmas for cross-cultural research in translating study instruments and implementing quality assurance methods, drawing on strategies utilised in the Mothers in a New Country (MINC) Study.
METHOD: To translation of study instruments in the MINC Study included: forward and back translations, a bilingual group review process, consultation with bilingual content experts, piloting of different versions of translations, a process for exploring unresolved difficulties and caution in interpreting unusual study findings. Interview quality was assessed by: 1. An ongoing review of interviewer-prepared English coding schedules to ensure completeness of data and identify problems with interview administration. 2. Analysis of fully translated transcripts of six randomly selected early interviews to assess the accuracy and consistency with which questions were asked. 3. A comparison of data sources for 45 randomly selected interviews (original language interview schedules, English coding schedules and translated interview transcripts) to determine the rate and nature of discrepancies.
RESULTS: Translation strategies that went beyond simple forward and back translations achieved more reliable and appropriate translations. The complexity of language and cultural differences sometimes still meant less than satisfactory results. Interview tapes played an important quality assurance role, enabling feedback to the interviewers and providing a basis of comparison for identification of data discrepancies. IMPLICATIONS: Ensuring good data quality in cross-cultural research is both critically important and difficult. Open discussion of the problems and concerted efforts to deal with them would benefit future research.