Methods and tools to evaluate the quality of vital statistics
Approaches to measure the quality of reporting of birth characteristics
The quality of reporting of characteristics of the birth or persons directly involved in the birth is paramount to ensure good-quality fertility statistics. The UN’s Principles and recommendations for a vital statistics system, revision 31 recommends a set of characteristics that should be collected from the birth event. Some of which are discussed below.
The United Nations recommends that the following characteristics are collected in connection with the birth record for birth statistics purposes:
- Sex of child
- Weight of child
- Gestational age of child
- Birth order
- Age of mother
- Date and place of the occurrence of the event – that is, the place of delivery
- Place of usual residence
- Attendant at birth
- Date and place of the registration of the event
- Type of birth – for example, single, twin or triplet
- Time since last birth.
The United Nations also recommends collecting a comprehensive list of characteristics of the child’s parents which are used for different tabulations of natality and fertility statistics:
Sex of child
The percentage of births with unspecified sex should be very low and is a sign of poor quality data.
The sex ratio at birth (SRB; the number of male births per 100 female births) should be in the range of 103–107. A SRB in excess of 107 generally signifies son preference in fertility behaviour and abortion of female fetuses, unless it is a result of poor-quality data.
Birth order is an important fertility characteristic that is used to help understand fertility behaviour. It should measure all previous live births of the mother, not just births in the current marriage or relationship. Average birth order should be higher if the mother is older. A different pattern would suggest poor-quality data. The percentage of births with unspecified birth order should be low.
- Place of delivery
- Attendant at birth
- Gestational age
- Time since last birth
- Place of usual residence.
A large percentage of data missing for these characteristics could signal poor data quality and should be further evaluated.