Ghana's population is based on assumption

The above-named article carried on ghanaweb of 1 March 2006 raised a lot of issues which I would kindly wish to have clarifications for per your medium.

Population projections the world over are usually based on certain assumptions. Very useful statistical and mathematical modelling techniques are used to reduce the level of error or sometimes smoothen the waves in such projection models.

But to argue that the population data, (and by that I understand the census data of 2000) is based on assumptions requires further clarifications from the Regional population Officer. Possibly he was wrongly reported by the GNA correspondent. Secondly population censuses are not projections they are hard facts collected and documented. And they dealt with aggregate or macro data. Intercensal population is rather estimated and postcensal population is projected based on previous census data as well as certain micro-demographic data and other vital statistics which are based on scientific assumptions.

Andrei Rogers already in 1975 published his research on projection and interpreting population data using what he termed: multistate and multiregional mathematical models. This considers migration by origin and destination. This is a break from the long established method of unistate and single region models.

In the multistate model, factors (as mentioned by the population officer) such as ?migration, family planning, malaria, HIV/AIDS, provision of potable water, disasters and infectious diseases that could claim lives? are factored into the population projection model simultaneously. But this model of Rogers requires huge data and very sophisticated matrix computational facilities which we may not be ready for in Ghana.

The assertion that the census failed to capture seasonal minors and immigrants is nothing new. What is true is that migration data is the best kept secret of almost all of sub-Saharan Africa. Ghana is no exception. But on census night (here an assumption for the population census, assuming all the people were counted on one day, viz. 26 March 2000) one of the census questions captured this group of people.

The UN already in 1980 approved for the migration data the ?last-move? question as an acceptable alternative to ?N-year ago? questions. But in Ghana?s census for 2000 (20 years later) question P08 reads: Place of Residence 5 Years Ago??. And this type of data was very suitable for Rogers?s multiregional analysis. But for the technical difficulties associated with the analysis of ?last move question? data, the last-move question was perhaps a sure way of capturing the seasonal minors of the Officer.

In the absence of much detailed data, micro-surveys such as Demographic and Health Surveys or Ghana Living Standards Survey or the Indepth Network are wonderful opportunities to update the census data of Ghana. The sets of data obtained by these surveys are used for projection based on certain time tested statistical assumptions. It would also be appreciated if the Officer could corroborate his assertion that women outnumber men in Ghana, as he put it: ?for the simple reason that women took good care of themselves than men?. This assertion is important for population studies. Scott Duncan et. al. in 2002 developed a thesis that ?malnutrition, from which the bulk of the population suffered, was the major factor that regulated demography in historical times, its controlling effect operated via the mother before, during and after pregnancy?.

Our population officer again raised a very serious issue. I quote the report of GNA on what he said: ?Also he explained the male foetuses are destroyed in the menstrual cycle than the females and as such the mortality rate of boys is greater than that of girls, especially in their early stages.?

What does this mean? Is it similar to the practice of female foeticide (aborting female foetuses) as a result of cultural practices in parts of India? Is there evidence of male foeticide in Ghana? And if male foetuses are already destroyed in the ?menstrual cycle? (?) what effect would it then have on the mortality rate of boys?

In any case we could investigate if the mortality rate of boys and girls can be explained in terms of other factors like the length of period of weaning and breast-feeding. Are boys weaned earlier than girls in Ghana, etc. And possibly we could be looking for some computer-intensive methods such as bootstrapping to model available data for more accurate projections.

Ghana is being hailed as a success story on checking the fertility rate and in the spread of AIDS. Could we reach the population transition stage earlier than projected? So I guess we would continue to rely on Population Officers to implement sound policies and as well help investors and forces of the market determine the direction of businesses.

Thank you for the attention.

Views expressed by the author(s) do not necessarily reflect those of GhanaHomePage.