Approaches to the problem of respondent attrition in a longitudinal panel study of nurses' careers

Longitudinal panel studies have several distinct methodological and analytical advantages over cross-sectional studies for gaining an understanding of nurses' careers and developing strategies to facilitate retention. Panel studies do, however, present the researcher with a number of substantial challenges if maximum potential is to be achieved from the investment of time and resources that such studies entail. Respondent attrition at successive phases of data collection, for example, is a major problem, often with the consequence that respondents are not representative of the study group as a whole. This paper discusses reasons for choosing a longitudinal panel design for a programme of research into careers of midwives and nurses. The way in which the problem of respondent attrition has been approached in one of the projects in the programme, that focusing on the careers of registered general nurses, is then described.