|The first National Identity Card issued in Pakistan|
Monday, 16 November 2015
by Azmat Budhani and Hussain Bux Mallah
Many of us field researchers have had frustrating and at times amusing experiences probing respondents’ ages. For most types of household surveys we need to fill out a family roster which includes an age column where we record the age of each household member. Recording this data is not an easy job because a respondent is often unaware of his or her own precise age, let alone the ages of other household members. This shows up as discrepancies in the data. An adult son, for instance, might have been noted as 25 years and his mother’s age will be recorded as 35, implying that the mother gave birth to her son when she was 10 years old. Clustering of ages around multiples of five years has been reported across surveys in Pakistan and other countries with similar conditions. Although researchers have many tricks to get to the year of birth within certain margins of error, accurate recording of the actual date of birth remains a challenge.