The Poor-Poor Divide in Northern Nigeria

Place comments on the Facebook page 'The Poor-Poor Divide'

In 1980, whilst working as a Medical Adviser to Oxfam in Tanzania, I came across an article by S.A. Meegama on Socio-Economic Determinants of Child Mortality in Sri Lanka. He had analysed a cohort of approximately 25,000 live births born in the period 1948 to 1974 to ever-married women aged less than 50 years.

The remarkable finding was that 9% of the women had suffered 52% of the child and infant deaths whilst 71% had had no deaths of children at all during that period.

Naturally, distribution of deaths would be skewed, since not all households would have one child death each, but this skew was extreme. Amongst the 9% who had the most mortality, the distribution was also skewed: 1% had had 4 or more child deaths each, 2% had had 3 deaths each, and 6% had had 2 deaths each. You can see the distribution in the following diagram:

[Figure based on S.A. Meegama, Socio-Economic Determinants of Infant and Child Mortality in Sri Lanka An Analysis of Post-War Experience, WFS Apr 1980]

Meegama was at a loss to explain the skew, and eliminated the usual suspects: including child spacing, numbers of births, and availability of health centres. He did find a difference between urban, rural and plantation-based households in that the skew was roughly similar in urban and rural areas whilst among plantation workers (all of whom were very poor) it was worse -- but even in the plantations 53% of women had had no child deaths in the 25 years studied.

I found that a few demographers had already discovered similar distributions amongst a variety of populations - including 19th Century Sweden, Senegal, Nepal and Tanzania. In every case the authors postulated a variety of potential causes of such distribution, but could not pinpoint any particular one -- let alone a combination of factors.

This situation intrigued me. I tried many times to bring this to the attention of a whole host of agencies with a view to focusing programmes on this disparity but none were interested in moving from their traditional areas of focus (poverty, lack of resources, gender inequality etc). I therefore moved from trying to persuade people to undertaking my own studies whenever an opportunity arose in the various programmes that I had links to or advised.

Sure enough, every time I studied a population, the same kind of skewed distribution occurred, but I never had the means to study why.

It was only in 2009, when working on a programme of Maternal and Child Health in Northern Nigeria, that I at last had a chance to undertake a meaningful study. In order to prepare for this we conducted numerous interviews with villagers about how they recognised women who needed more support and their attitudes to those women. We also asked women what kinds of support they needed most and what determined whether a woman felt respected or not.

We then went to 4 villages in 3 States and surveyed every woman in every household in each village. A total of 1,688 women who had had at least one live birth was interviewed. The results were very similar to those found in other demographic surveys in terms of child death. 65% had no child deaths, 15% had one child death each and 20% had multiple child deaths (having an average of 3 deaths per woman).

20% of the women (those having had multiple child deaths) had over 80% of all the child deaths.

Further clustering was seen to occur within polygynous households – where one of the wives suffered all (or nearly all) the child mortality whilst others have none. Of the 267 polygynous households, 37% had women with multiple (2 or more) child deaths, and, amongst these, 62% had a significant skew in which one woman had all the deaths whilst others with similar numbers of births had none.

Similarly the figures showed that within compounds of related families, some of the families suffered child deaths whilst others did not – i.e. an entire compound may have had no child deaths despite having tens of births, whilst in another compound only one of the families might have child deaths.

In the analysis none of the usually cited factors accounted for the skew: wealth, resources, child spacing, religion, education, employment, or distance from a health facility. This was not surprising as almost all the households surveyed were of similar levels of wealth, were the same distance from a health facility, had the same level of education (almost none had completed primary education), had similar numbers of resources, were of the same religion and were almost all subsistence farmers.

Support, respect, appearance and numbers of births

6 factors were found to be associated predominantly with child deaths.

The woman rarely or never had anyone older to look after the children (vs always)
The woman believed she had no or little respect from relatives, in-laws, husband or others (vs some or very much respect)
The woman had no one to turn to for support if her children had difficulties (vs some or lots to turn to)
The woman had no one to turn to for support if she herself had difficulties (vs some or lots to turn to)
The woman had no or little general support from own relatives and in-laws (vs some or a lot)
The extent to which the woman cared for herself, her children or her household was very poor (low versus average or high).

The levels of association varied by both age of mother and numbers of births -- but the association was not the simple one of age and numbers of births. As the woman grows older she naturally is likely to have more births. With greater numbers of births the need for support of various kinds becomes much greater. As a result, should support be witheld or not provided by the family or other members of the household, then naturally the woman is going to have greater difficulties in caring for children, will become depressed and in turn will start to neglect herself and her household.

These results led to further surveys amongst younger women aged 15-19, and it was not surprising to find the same patterns of vulnerability resulting from the same factors -- although for the younger women the impact was greater in terms of their not attending social services (such as health), their ability to communicate with others, and their care of themselves or their children.

It is therefore very clear that

Money plays a part, but by no means the only part
Respect and self-confidence have a very big role for women in caring for their children and themselves
Respect is highly dependent on social interactions, including support from those around the woman.

None of this should be in the least surprising. We have known for a very long time the importance of moral economies in the entire lives of people who are very poor or who have little access to resources or services. Another page on this site discusses the idea of Moral Economies and why they are so important to survival. The point here is to emphasise that we can measure the real impact should we choose to do so. Unfortunately most aid agencies choose not to do so.

Implications for improving support

The reasons for any particular woman lacking in support or respect from her family or others are far too numerous, complex and interactive to make any attempt at reducing their existence impossible. There are personalities, family histories, land disputes, jealousies, passions, hatreds, alliances, fears, obligations, financial worries, and beliefs amongst the many that would have to be taken into account.

The fact is the woman lacks support and respect, and this has a direct impact not only on her own health but that of her children as well. It is this aspect that needs strengthening.

There are two broad ways of achieving this:

Encourage existing groups in different types of community to encourage and welcome any woman into their number.
Encourage community leaders and others to create various support opportunities -- such as child care, counselling, dispute resolution -- which are available to any woman.

Group psychology is covered on another page on this site. Suffice it to say that whilst groups in general are exclusive and tend to self select members from existing friendships, alliances and those with similar social backgrounds, but with the right guidance can become more inclusive, supportive and beneficial. In such cases membership of such groups can lead to improved confidence by women that they can be respected, communicate with others and work with them. If this is achieved the health, capability and involvement of under-supported women improves greatly, as does their use and attendance of a variety of social services including health. As a result the health of their children also improves.

Most projects run by aid agencies depend on group formation. As discussed in the page on Group Psychology, the success of such groups results from the attention paid to the groups and the strengthening of mutually reinforcing and supportive interactions by the people in the groups. What most projects do not do is ensure that their groups are more inclusive of those who need greater respect and support.

Tony Klouda