IMANI’S Response to Ghana Statistical Service’ Position on the 2010 Census

Tuesday, January 11, 2011

We are responding to comments made on the Joy FM channel by the Public Relations Officer (PRO) of the Ghana Statistical Service (GSS) on the 6th of January 2011 seeking to dismiss our legitimate concerns about the institution’s conduct of the 2010 National Census.

As far as the GSS’ senior staffer was concerned, nothing untoward had characterised his agency’s management of the census. If a few people hadn’t been counted, it was because, and we should all know this, people don’t spend 24 hours in their homes, and it is more than likely that on occasion enumerators would call on a particular household only to meet the absence of the members. Indeed, he reckoned, many of the complaints about the census owe to the work of saboteurs.

We are unimpressed.

Our goal in raising the concerns we did in the media was not to nitpick on random events that may or may not have happened during the census. Our goal was to question the overall credibility of the exercise, which we judged to have departed lamentably from best practice in census management.

We will elaborate. But first let us explain clearly our seeming obsession with the census.

A national population and housing census is not limited to enumeration (i.e. head counting). It goes far beyond that. The United Nations Organisation (UN) has developed a rather rigorous theory about the principles of “simultaneity, universality and periodicity” in order to guide this crucial process of collecting vital socio-economic and demographic data for the purposes of policymaking. The overriding principle is to better understand trends in the society in order to ensure that decision-making is responsive to new realities.
All too often, developing countries hold on to the hallowed realities of yesteryear, failing miserably to adapt to fast-changing situations on the ground.

For instance, even as the face of poverty becomes increasingly urban, so-called “poverty alleviation” programs continue to cling to a rural construct of poverty, thereby not only misdirecting social interventions but even more sadly missing out on the opportunities offered by urbanisation in a services-dominated economy.

The absence of good quality data about our evolving circumstances as a nation interferes so frequently with strategic planning at the socio-economic and political level that the least civil society organisations can do is protest when good money is misspent in poorly executed attempts to redress the situation.

Think about it: this is a country where the official unemployment rate hovers around 11%; where economic indicators are routinely dismissed as unreflective of the quality of general welfare; and where “official” job creation statistics are left to the vagaries of the political rumour mill. This is a country, where making an objective statement about improvements or deteriorations in any particular sector is impossible without descending into the partisan horrors of name-calling.

Think about it.

In some other places, no one has any delusions about the importance of censuses. Such is the sensitivity that attends censuses in our neighbouring country of Nigeria that four censuses have been sidelined or annulled (’62, ’63, ’73 and ’91). We may baulk at such indelicate behaviour on the part of some members of the Nigerian elite in this matter, but we cannot fail to notice the general point, which is that censuses are, and should be, grave affairs. Especially in developing countries.

Census data are a placeholder for us in IMANI for objective policy-related data generally. The lack of such data accounts for the useless squabbles in our media. It accounts for the “too much heat and no light” nature of our political conversations. It locks out independent-minded commentators and make partisan wrangling the reference point for all debate. It prevents serious comparisons between different managers of our economy and our society. It makes it impossible to hold public officers to account for the management of our affairs, except in the case of the most flagrant acts of corruption. Every comment on performance risks partisan coloration because hard-hitting observers are unable to take refuge in objective facts and figures.

Think about it.

So when we spend $50 million or close to that amount to redress some of these problems, it is just and fair that we insist on getting value for money.

When the “if you criticise us, you are out for sabotage” Statistical Service spends $2 dollars thereabouts to administer a questionnaire to each Ghanaian, we expect that they shall do that and nothing less.

When the “don’t blame us” GSS sets up more than 23,000 so-called enumeration centers , and hires 45,000 people to undertake a vital national exercise over a period of 2 weeks, we expect that they will enhance supervision and coordination. We don’t expect to hear from people living in some of our most well-laid out neighbourhoods to complain about having been neglected.

At any rate, the basic arithmetic works out to each census staff administering 38 census forms a day, far from an insurmountable logistical burden. As it turned out the period was lengthened and the burden per enumerator thus significantly reduced. So only a lack of supervision and coordination could have prevented optimal outcomes.

Which is why our concerns with the GSS relate generally to their not having sufficiently embraced “best practice”.

By establishing a parallel structure rather than integrating into the local government system (as was the approach in Kenya in 2009, where $2.5 per head was spent on an identical exercise with near-spectacular results) and then falling short of supervision and coordination, which decentralisation would have augmented, GSS was setting itself up for serious challenges.

By failing to conduct rigorous piloting in order to identify hard-to-reach spots, and thus to develop comprehensive “risk analysis frameworks” for these specific geographies, GSS was setting itself up for serious challenges.

By failing to create the framework for external monitoring and evaluation of the entire process, GSS was missing the opportunity to improve upon its internal review systems.

By ignoring the opportunity offered by the census process to create GIS and other mapping systems to give proper meaning to the household survey element of the overall exercise, GSS was falling short of excellence.

Kenya, which does not even suffer to the same extent from the kind of chaotic addressing and identification systems we have here, and operating on a similar per capita budget, made it a point to incorporate essential elements of this technological approach into their 1999 exercise.

By ignoring the need to develop a credible “non-response follow-up” mechanism, as the PRO’s comments appear to suggest (he expects people to randomly call their offices to alert them), the GSS now risks creating a census register with a margin of error too wide/loose for comfort.

Non-response follow-up mechanisms were clearly not properly advertised, nor were the avenues made available to the general public convenient enough. From what we have gleaned, “re-interview procedures” were underemphasised in the training of enumerators, many of whom were, in the first place, alienated by the haphazard approach to their remuneration. The labour resource strategy was thus strained not only by poor hiring practices but also by payroll management systems. Here is just one more area where the lack of efficient decentralisation undermined the outcomes of this important exercise.

Universality, simultaneity and periodicity very simply mean “rigour and comprehensiveness within a strictly defined timeframe”. The UN recommendations are quite thorough in this regard, and since a number of senior staffers in the GSS today have been exposed to the highest levels of the UN statistical system, we fail to understand their lack of care in these matters.

All is however not lost.

We are in the post-enumeration survey and review phase, as we understand it. The GSS can begin to focus on developing robust statistical and data management techniques to identify lags and laps in the collected data and start to make amends.

The data analysis and presentation can be approached with seriousness to take into account ease of retrieval, privacy protections, cross-referencing, and incorporation into other datasets in Ghana, particular within the local government system.

We of course shall go ahead with our petition to relevant bodies, but our hope is that the petition shall be made redundant by a renewed effort on the part of the GSS to address the concerns many have expressed about the census, prior to releasing the preliminary report. Furthermore, the GSS shall have one whole year to make amends before they release the official report in 2012.

We have never argued that the GSS is staffed by incompetents. What we have said is that they have not over the period that we and many other independent commentators have been observing them shown full dedication to their mandate of supplying credible, quality, data. And that’s a tragedy.

That can change. And, for all our sakes, we hope it will.

Courtesy of IMANI ( &