Page 35 - IJPS-2-1

Bruno Yempabou Lankoande

the time of the survey or at the time of death is used to assess the gap in mortality. In women, esti-
mates obtained by using children’s place of residence yielded an urban advantage. This advantage
was reinforced when parent’s place of residence were taken into account.
These results suggest that information on children’s place of residence is not a good proxy for
their parents’ place of residence. In addition, misclassification errors seem to operate differentially
according to the sex. Nevertheless, even if urban areas benefit from a slight advantage, this differ-
ence is not significant at the 0.05 level, because confidence intervals considerably overlap in each
period.
3.3 Sibling Survival Histories

Estimates of adult mortality based on sibling histories are presented in Figure 3 for both types of
residence. The dating procedure is more precise here (compared to orphanhood estimates) and the
availability of three surveys provides a rough idea on the trend in mortality. On the one hand, adult
mortality rates have decreased since 1990s to 2000s in urban areas. This mortality decline was more
marked among men. On the other hand, rural areas experience stalls, sometimes reversals, in adult
mortality. Rural areas tend to have an advantage in terms of adult mortality that fades over time. In
the period of 1992–1997, for men as well as women, the value of 45q 15 in urban areas was around 0.4
while this value was less than 0.3 in rural areas. An inverse situation was observed in the period of
2004–2009, with higher mortality in rural areas. However, for each period, confidence intervals
largely overlap, indicating that DHS sample sizes are too small to detect any significant difference.
Again, I should reiterate that the place of residence of interviewed women was used as a proxy for
their sibling’s place of residence.
To assess the quality of sibling survival histories in the different DHS, Figure 4 shows the mean
number of siblings reported by the five-year age group of respondents. This number is the average
parity of the respondents’ mother; it should be closed to the completed fertility in Burkina Faso, and
should increase with the respondents’ age since fertility has declined in recent decades. Two major
observations stand out from these plots. First, the average parity is globally higher in urban

Figure 3. Estimates of adult mortality with 95% confidence interval, by sex and according to place of residence based on siblings survival data from
1992 to 2009 (DHS data).

International Journal of Population Studies | 2016, Volume 2, Issue 1 29