Abstract

This paper presents a comprehensive diagnostic of the training and working conditions of interviewers engaged in the Continuous National Household Sample Survey (PNAD Contínua) – a brazilian sample survey on labor market and sociodemographic data – conducted by IBGE, the national statistical office, in the state of Minas Gerais, Brazil. The study focuses on four dimensions: interviewers’ profile and workload, the current state of training, the assessment of work materials and institutional support, and the expressed demand for additional training. The analysis is based on a targeted survey applied to the universe of interviewers who had uploaded at least one interview to the Integrated Management and Control System (SIGC) in the seven months preceding the study and were active in one of the IBGE local agencies in Minas Gerais. A structured questionnaire, developed by the project team, was administered via LimeSurvey, with individual tokens, automated invitations, reminders and acknowledgements, and explicit assurances of confidentiality to encourage candid responses. Eleven consistency checks were implemented for date variables, comparing internal and external reference dates; inconsistent records were excluded prior to tabulation, thus strengthening data quality. Out of 334 eligible interviewers across 75 agencies, 323 responded, yielding a very high response rate. Temporary staff in the position of Survey and Mapping Agent represent 60.7% of respondents, indicating the central role of non-permanent personnel in field operations. The distributions of time in the IBGE and in PNAD Contínua reveal a high share of relatively inexperienced interviewers and suggest substantial turnover, reinforcing the strategic importance of effective and ongoing training. Regarding training, 89.2% reported having read the core Interviewer Manual in full before conducting their first interview, and 78% stated they had received a “full training”, defined as a structured, face-to-face course covering all sections of the core questionnaire. Among those trained, 93.7% at least partially agreed that the training was satisfactory; however, 58.3% had not received any refresher training in over a year. An “ideal training path” was defined as having read the manual in full, received a satisfactory full training, observed an interview conducted by an experienced interviewer, and been accompanied in the first interview. Only 44.7% met all these criteria, indicating significant scope for improvement. The results also show that interviewers frequently work on multiple surveys and often need to re-contact respondents to complete missing or uncertain information, highlighting the operational complexity of PNAD Contínua and the need for strong conceptual and procedural support. The evaluation of materials and tools is predominantly positive, with high agreement on the adequacy of the manual, the mobile data collection device and the PNAD Contínua application, although speed is perceived as a relative weakness. Finally, the study identifies clear priorities in training demand, with stronger interest in modules on labour, other forms of work and education, as well as transversal topics such as respondent approach and dwelling identification, providing concrete inputs for redesigning interviewer training strategies in official household surveys.