This study analyzes the quality of perinatal mortality and retrospective contraceptive prevalence rates
calculated from various instruments used in the Demographic and Health Surveys.
Perinatal mortality: In this report we compared methods for estimating perinatal mortality in The DHS
Program. None of the methods appear to adequately capture perinatal mortality by the standard that we
selected. However, we found that the pregnancy history and the birth history supplemented by special
questions performed better than the birth history supplemented by the reproductive calendar.
Contraceptive prevalence tabulated from the reproductive calendar: We assessed the consistency of
contraceptive use reporting in the calendar by comparing retrospective contraceptive prevalence rates
tabulated from the calendar with independently estimated current status contraceptive prevalence rates
from a prior survey. We compared estimates from the two data sources for the same point in time among
women in the same age groups. We found evidence of substantial underreporting of retrospective
contraceptive use in the majority of calendars analyzed relative to current status estimates.
Results suggest that both stillbirths and contraceptive use are underestimated in data collected using the
reproductive calendar. We recommend experiments in future DHS surveys: random assignment of some
households to receive a birth history plus calendar and others a pregnancy history, or a forward pregnancy
history versus a backward pregnancy history to assess the impact on reporting of stillbirths; and
experiments with shorter calendars and potentially alternative methods of electronic data collection to
assess the impact of these changes on reporting of contraceptive use and discontinuation.