Page 66 - IJPS-2-2

Transition in risks of higher order births in Nepal: a life table analysis

have better knowledge of health and contraception and better access to contraceptives, which might
lead to longer birth intervals among urban women. On the other hand, the lifestyle in urban areas
differs from that of rural areas. Urban women may breastfeed less frequently and for shorter duration,
which may support a shorter birth interval as compared to rural women. A longer second birth inter-
val among urban women is observed as compared to those of rural women in Egypt (Baschieri,
2004), however, place of residence had little influence on transitions to higher order births after the
second birth in Ramesh’s (2006) study in India.
Evidence shows a mixed effect of women’s education on birth intervals. Trussell et al. (1985)
found women’s education to be an insignificant predictor of risk of pregnancy in Malaysia, the Phil-
ippines, and Indonesia. In contrast, studies of Nepal (Suwal, 2001), Tamil Nadu, India (Singh, Trip-
athi, Kalaivani et al., 2012), and Banglandesh (Saha and van Soest, 2013) found women’s education
was an important factor for extending the birth interval. The longer birth interval among educated
women might be due to their enhanced capacity to adopt new contraceptive technology (Kim, 2010).
In societies where there is a stronger preference for a son than a daughter, birth intervals after a
son are found to be longer when compared to those after a daughter (Rajbhandary, 1999; Ramesh,
2006; Sahoo, 2011; Singh, Tripathi, Kalaivani et al., 2012). Birth spacing also increases with mater-
nal age (Saha and van Soest, 2013). Survival status of the previous child has strong effect on subse-
quent birth interval; shorter intervals are associated with the death of the previous child (Rajbhan-
dary, 1999; Ramesh, 2006; Sahoo, 2011). Women’s work status, their exposure to mass media, and
household economic status also influence birth intervals (Kamal and Pervaiz, 2013; Ramesh, 2006;
Singh, Tripathi, Kalaivani et al., 2012).

2. Data and Methods

2.1 Data Sources
This study primarily uses data from the Nepal Demographic and Health Survey (hereafter NDHS) in
2011. However, data from the NDHS in 2001 and the NDHS in 2006 are also used. These surveys
were based on a two-stage stratified national representative sample of households. Wards of a Vil-
lage Development Committee (VDC) and sub-wards of a Municipality were considered as Primary
Sampling Units (PSUs) in all the three rounds of surveys. In the first stage, PSUs were selected
from both rural PSUs and urban PSUs following systematic sampling procedures with a probability
of selection proportional to size. Complete house-listing was done in the selected PSUs. In the
second stage, households were selected from the selected PSUs by applying a systematic sampling
technique in 2001 and 2006 and simple random sampling in 2011. In all three surveys, oversampling
of urban PSUs was made to obtain statistically reliable estimates for urban areas. All women of age
15–49 from selected households were interviewed in 2006 and 2011; only ever married women of
the same age group were interviewed in 2001. In 2001, a total of 8,726 ever married women of age
15–49 from 8,602 households were interviewed; in 2006, a total of 10,793 women of age 15–49
from 8,707 households were interviewed; in 2011, a total of 12,674 women of age 15–49 from
10,826 households were interviewed.

2.2 Dependent and Independent Variables
The median age at first birth and the cumulative percentage of women who had a first birth by a cer-
tain age are used to examine transitions in the age at first birth. The cumulative proportions of cur-
rently married women transiting to the next higher order birth and the median length of the birth in-
terval are used to analyze dynamics of birth intervals. The determinants of higher order birth are
analyzed by using the hazard ratio of a higher order birth. Variations in birth intervals are ex-
amined by taking place of residence, women’s education, women’s work status, exposure to media,
household wealth index, age at previous childbirth, birth year of previous child, sex of previous child,
and survival status of previous child during infancy as independent variables. The same set of va-

60 International Journal of Population Studies | 2016, Volume 2, Issue 2