Understanding Attrition and Bolstering Retention in a Longitudinal Panel of Older Adults: ORANJ BOWL

Attrition from longitudinal studies can affect the generalizability of findings especially when studying developmental constructs such as successful aging.

Using data from a 12-year (6-wave) panel of 5,688 older people (aged 50-74 at baseline), we compared people retained in the panel with people lost to follow-up on demographic characteristics and measures of successful aging. After instituting expanded retention strategies at Wave 6 (i.e., a team-based approach, social media, and paid web search engines), we compared different groups of people lost to follow-up (i.e., deceased and withdrawn due to lack of interest) and different types of completers (i.e., full completers vs. lost and reengaged completers).

At baseline, Wave 6 completers were significantly younger, less likely to be African American, more likely to be married, reported higher levels of income and education, were more likely to be working full-time, had less pain and fewer chronic illnesses, and reported higher levels of subjective successful aging and functional ability than those lost to follow-up. Analyses demonstrated differences across groups based on the reason for loss (i.e., deceased, impaired, and not interested). Participants who missed an interview but returned to the panel were significantly different from those who participated in all waves of data collection. Expanded retention efforts improved generalizability, as people returning to the panel reported lower levels of education, lower levels of income, and were more likely to be African American.

Biased attrition within longitudinal research affects the interpretation of study findings, especially when studying developmental outcomes. However, expanded retention strategies can reduce bias and loss and should be used to enhance retention efforts in longitudinal work.