Imputation Procedures for Cognitive Variables in the Mexican Health and Aging Study: Evaluating the Bias from Excluding Participants with Missing Data.

Non-response of cognitive data in cohort studies is a barrier to cognitive aging research. We describe the procedures for the imputation of non-responses for cognitive data in the Mexican Health and Aging Study (MHAS). Data came from the 2001-2015 MHAS waves. We also describe the association of cognition with education, age, and other variables in 2015 with and without the imputed values. Between 12.3% and 37.9% of participants were missing data for at least one cognition variable. When we conducted the analysis with and without the imputed values, the relationships between education, age, and cognition were similar in direction and significance, but different in magnitude. Non-response of cognitive data is common and non-random in the MHAS. Investigators should use the data sets that include the imputed values, which are publicly available.