Age misstatement for young children in rural Bangladesh

Age data for 3,393 children, six years of age and under, in rural Bangladesh are analyzed for the level and pattern of age misstatement. Random error, age heaping at whole years, and preferences for particular ages are found in the data. Variation in age reporting is discovered to increase monotonically with age. Systematic errors in age misstatement display modest overstatement for the first four years of life and more pronounced understatement for ages 4, 5, and 6. Age misstatement is examined for its effect on one indicator often used in nutritional surveillance--weight-for-age of children. The impact of the various types of age misstatement (a) increases the difficulty of interpreting weight-for-age and (b) obscures accurate understanding of malnutrition in Bangladeshi children.