Monday, January 15, 2007

Nigerian Census – Gender Population Distribution Shift, a Red Flag

Many have commented on the figures of the March 2006 National census, which puts the Nigerian population at 140 million. Comparing this with the November, 1991 figure of 88.992 million, this is a growth rate of about 3.07% per annum, which approximates the estimates that have been put forward by various statistical models. What is however intriguing is the population distribution along gender lines. In 1991, males accounted for 50.04% and females 49.96% of the population (a male/female ratio of 1.001). The distribution in 2006 has males accounting for 51.21% and females, 48.79% (a male/female ratio of 1.050).

This shift in population distribution raises a red flag. Several questions come to mind as to what may be responsible for this gender distribution shift. Could it be cultural, social, biological, economic or man-made? Are more male children being born than are female ones, or are more females dying at a faster rate than are males? Or were there cultural issues that limited the accessibility of females to census enumerators 2006 more than in 1991? Could it be that the ravages of HIV/AIDS is taking a toll on the female population, especially since females have a higher susceptibility to the disease? Using the 1991 census figures from the Federal Office of Statistics (FOS) for statistical analysis to help answer these questions the following are the three most striking findings: -

  • The November 1991 census showed that 72.27% of the entire population was made up of 29 year-olds and younger
  • Of these, about 51.74% (or 36.67% of the entire Nigerian population) were females and 49.26% (or 35.6% of the entire Nigerian population) were males
  • The population growth rate for males in Nigeria between 1991 and 2006 was 3.23% versus 2.90% for females
While various political dissenters have raised allegations of fraud as regards the validity of the census figures, equally important, perhaps even more so, is an investigation into the cause of the gender distribution (sex ratio) shift. Although this shift appears not to be statistically significant at the international accepted 95% confidence interval, however, it is important to know if females have a higher death rate as a result of violence, diseases or some female-specific morbidity, or if there is a lower birthrate for females. If either of these two scenarios is not the case, then why were fewer females counted in the last census?

3 comments:

Nilla said...

Interesting analysis....

Especially since I'm still in doubt of the ratio of males to females.

imnakoya said...

It's great to see numbers popping from the census. Great and interesting insight.

One other possible cause for the observed shift is a wrong baseline estimate (1991), or even an inaccurate reference point (2006).

You mentioned 95% confidence Interval, but you didn't state clearly which point estimates or the magnitude. Is 95% C.I even relevant since we are not dealing with samples??

What could provide the most accurate answer to your questions will be national birth and death records/registry between 1991 and 2006. Extremely crucial piece to any census which we don't have in Nigeria, yet.

Veracity said...

Imnakoya,
Thank you for your comment. Sorry for the confusion. The 95% Confidence Interval refers to the male/female ratios of 1.001 in 1991 and 1.05 in 2006 compared to a sample of the same ratio from 30 randomly selected countries. I omitted to include this in the write-up, since it was a background analysis.

As you rightly mentioned, the accuracy of any statistical analysis depends on the reference points, and the 1991 and 2006 figures may well be flawed.