Page 70 - IJPS-7-2

International Journal of
Population Studies Age-adjusted measures for fertility transition

The equation [3] was used to compute the point the successive GDHS surveys were very small. When
estimates of TMFRs which were used in the comparative successive surveys have comparable levels of data quality,
assessment with the average CEB 45-49 to show trends in the impact on fertility measures is minimal. Therefore, the
fertility levels. potentially spurious trends in demographic indicators due
to improvements in data quality could not have occurred
2.5. Multivariable analysis of fertility in Ghana. These spurious demographic trends due to

We demonstrated the effect of a fertility measure on the improvements in data quality occur when there is sudden
results of multivariable regression analysis by exploring a positive change in the quality of the data.
two-fold Oaxaca-Blinder (OB) decomposition method on The sample distribution shows that for all the countries,
both CEB and ASMFRs. The OB decomposition technique the age distribution of the married women has been shifting
partitions a change in the mean estimate of an outcome toward predominance of older age groups (Figure 2). The
measure into a part explained by changes in compositional proportions of women in unions aged 15 – 24 progressively
characteristics of the sample and a part that is attributed decreased in the four countries. The pattern of education
to the behavioral changes of the sample. This regression- status was markedly different among the countries. In
based decomposition method was used to comparatively Ghana and Zimbabwe, an overwhelming majority of the
examine how the choice of a fertility measure can affect women have secondary education while, in Kenya and
the findings from the analysis of drivers of marital fertility Rwanda, they have primary education only.
change in situations, where the CEB and ASMFRs have
followed different trends. The OB technique is a counter 3.2. Trends in fertility levels
factual decomposition which estimates conditional The results from the comparative analysis of TMFRs and
contributions of characteristics and coefficients associated average CEB are presented in Figure 3 below. The
45-49
with the independent variables in relation to the dependent trends of marital fertility levels obtained from TMFRs were
variable. Applied to this study, it, therefore, means that it notably different from those constructed from average
estimates the expected magnitude of change in the mean CEB . A notable observation is the inability of CEB to
45-49
of CEB and ASMFRs based on the observed change in the capture stalls in the marital fertility transitions in all the
distribution of the sample by age and education status, four countries. The marital fertility transition of Ghana has
and differences in reproductive behaviors associated with stalled post 1998 with inter-survey increases in TMFRs
identified age groups and education status. 1998 – 2003 and 2008 – 2014 periods. In these periods, the

When the change in the dependent variable is not average number of CEB to women in the 45 – 49 age group
consistent with the change in the independent variables, decreased, suggesting a continuous decline in marital
the OB decomposition produces illogical results. fertility. We find the same contradictions between CEB and
Furthermore, when the difference in the mean outcome TMFRs in Kenya between the KDHS1998 and KDHS2003.
is underestimated, the OB decomposition tends to This 1998 – 2003 period is shown by CEB to have been
overestimate the role of the independent variables. The marked by decreasing marital fertility in Kenya. The
overestimation of the effect of independent variables can trends of CEB 45-49 and TMFRs for Rwanda also followed
potentially lead to erroneous determination of the depth different trajectories reflecting differences in their ability
and focus of investment in fertility management programs. to capture short- and medium-term changes in birth rates.
The two features of the OB decomposition relating to The pre-2005 era in Rwanda, which was characterized by
illogical results and overestimation of factors’ impact can, sociopolitical instabilities and the 1990s genocide, saw the
thus, be used to determine the comparative suitability of disruption of family planning and health infrastructures
CEB and ASMFRs in the multivariable analysis of drivers which resulted in crisis fertility-driven upsurge in birth
of marital fertility trends. rates. The CEB 45-49 is unable to capture this increase in
fertility as effectively as TMFR. If one argues that the rise
3. Results in crisis fertility in Rwanda was potentially concentrated
among young women, this increase would be reflected
3.1. Data quality and sample distribution when considering the mean CEB for all women 15 –

The quality of data was generally good for Kenya, Rwanda, 49 years. However, this too is unable to effectively capture
and Zimbabwe, where the proportions of women whose the increases in fertility in Rwanda pre-2005. The trends
ages were potentially misreported were mostly below of CEB 45-49 and TMFRs for Zimbabwe were more different
10%. It was only Ghana which had higher indices of digit compared to the other countries. The rapid marital fertility
preference with the GDHS1988 showing the highest score. transition from 1988 to 1999 reflected by TMFRs is shown
However, the differences in the Myer’s Indices between to be mild by CEB. The difference between the CEB 45-49

Volume 7 Issue 2 (2021) 64 https://doi.org/10.36922/ijps.v7i2.354