“It’s amazing,” Harry Lode comments as a chart labeled “Frequency of Sex” appears on his computer screen. “From a technical point, it’s really amazing.”
Almost before the two-word sentence is out of Lode’s mouth, a CD-ROM holding 14 million records from the 1998 census in Cambodia has spit out a table with the number of men and women tabulated for a remote rural village. After more than 15 years of working on censuses in various countries, the lightning speed of processing is only one aspect of the Cambodian census that the Dutch demographer finds amazing.
The government of Cambodia also must find it amazing. King Norodom Sihanouk has awarded the Royal Order of Sahametrei, the highest honor the government can bestow on a foreigner, to Hedi Jemai, the Tunisian-born demographer who was the UN Population Fund representative when the UN-funded census was conducted. Prime Minister Hun Sen is scheduled to confer the award during a ceremony at the Council of Ministers today.
In the US, the census is both social science and a commercial enterprise—big business. In Cambodia, where this census was the first in 36 years, it is a social science in its infancy, but in many respects this particular census is superior to others in both developing and overdeveloped countries. How—in Cambodia—is that possible?
Nott Rama Rao, chief technical adviser for the Cambodian census, is the former deputy registrar general of the census for India, which now has a population of more than one billion. (Rao is quick to note that it was only 846 million when he last did a census.)
Ten minutes in his office illustrates his fanaticism for following his own rules and procedures, including obtaining signatures in various registration books for the array of documents that the census has produced.
However, the underlying factors that may have contributed most to the quality of the census are the transparency and lack of politicization of the process, from the design of the questionnaire to the dissemination of results.
“This is the most transparent census I’ve worked on,” said Lode, who has been involved with censuses in the Solomon Islands, the Yemen Arab Republic, Nigeria and Papua New Guinea.
Rao, who worked on four censuses in India as well as one in the Gambia, doesn’t like the word transparent. “In many countries,” he says, presumably speaking of India among other nations, “they do not want to part with data, especially down to the village level. It can create tensions if you say that there are so many of this type of people and so many of that type of people.”
This may be an indirect reference to the one piece of data from the Cambodian census that has not yet been released: mother tongue. Ou Orhat, secretary of state for the Ministry of Planning, admits this relates to the sensitive issue of how many Vietnamese are living where in Cambodia.
Prior to actually conducting the census, a draft questionnaire was presented to representatives from the ministries and the donor and NGO communities.
Many suggestions, such as including commercial sex workers as an employment category, were integrated into the questionnaire.
In other countries, the entire process might have been closed, as the private domain of a census and statistics department, which was itself subject to political interference about what would be included in a
questionnaire and what would be revealed to the general public.
Ou Orhat described why the government had decided a census would be useful to Cambodia. The citation for the award to Jemai
makes the key point: The result of the census is the basis for
encouraging donors and stakeholders to invest in the development of
Cambodia. The government is trying to develop a village gazetteer for each commune, including a map with roads, the pagoda, the schools, and each household on it.
When one does go down to the community level, accuracy becomes very critical. With aggregate data, errors can be invisible. At the lowest data levels, errors can become conspicuous. The post enumeration survey, a standard method of evaluating the accuracy of census data, conducted about two weeks after the completion of the census, found that census coverage was accurate to within 1.78 percent. By way of comparison, the coverage error of a US
census is usually within two or three percent.
Of course the Cambodian census had to face unique issues. “Before we had the problem of access,” said Ou Orhat. “Now we are at full peace.”
Preliminary preparations, such as mapping and training, began in 1995 before the on-going civil war had subsided. When a preliminary demographic survey was conducted in 1996, 20 percent of the country was not accessible because of the Khmer Rouge and related security problems. That survey helped to train staff and provided data on 20,000 households that could be used to develop the framework for the census.
In 1998, when the actual census was conducted over a 10-day period, only three out of 183 districts were not accessible. “In the dangerous areas, like mined areas, sometimes our people couldn’t go in,” said Lode, “but the villagers came out.”
In addition to security and access, those who implemented the census were faced with the extreme lack of trained workers in Cambodia.
Jemai, according to the citation for his award, insisted that individual and institutional skill building should be developed through technical assistance and training.
The citation comments on the quality of the technical assistance, and notes that 43,000 Cambodians have been trained under the UN-funded project.
Individual and institutional development has continued in the form of both on-going analysis—the first two of eight supplementary studies are now complete—and workshops for users.
The project is marketing its product, not to big business, as would happen in many countries, but to the public sector, the NGO sector, and the donors—which, comparatively speaking, are a big business in Cambodia.
The project segments its potential market, presenting its
software with a slant designed to pique the interest of potential
users. In a recent meeting of Medicam, an umbrella NGO with roughly 100 member NGOs from the health sector, the focus was on the potential use of data relevant to the health sector.
The census data are contained on four CD-ROMs. The first includes more than12,000 tables made up of approximately 50 “priority tables” for 250 small areas. The second includes the same data aggregated for villages, and it is used primarily by local staff from some ministries and several UN or donor agencies. The third CD is the mapping and graphing database that goes right down to the commune level. Lode says the fourth CD is the most powerful because it goes right down to the census micro data, allowing users to produce any tabulation down to the village level.
With the first two CDs, users can manipulate the data to suit their own needs, even without training, mixing and matching across administrative divisions. With the third CD, a few key strokes will produce bar charts, pie charts, and color-coded small and large area maps that can be superimposed over one another.
“We have priced the CD-ROM as cheap, cheap, cheap as possible,” said Lode.
A single CD is $15; the same software and data would cost tens of thousands of dollars in other countries. Attendance at most training courses is usually about 20 percent expatriate and 80 percent Khmer.
Asian Strategies, an Australian-based company, was one of the first private-sector firms to use the census, and has since done market surveys in Cambodia for clients in Asia, Europe and the
US. “It is the first piece of the jigsaw we need to do our job,” said Trevor Harrison, managing director.
“All from a nondescript building on a dusty or muddy—choose your season—pot-holed road in the Phnom Penh suburbs.
“We don’t even need to bribe them.”