FORUM (Spotlight Weekly)

Census 2001: Quality And Credibility Of Data

By Dr. Shyam Thapa

Data from censuses (or surveys) are known to be susceptible to errors of various kinds and degrees. The conventional and uneven manners of recording dates of birth, death or similar events of life in Nepali sociocultural milieu make it extremely difficult to collect accurate and reliable information. Many respondents simply do not maintain any record of, or cannot recall, the dates of events like birth or death. This presents a major challenge to the enumerators, of census or survey, in performing their job, while weakening the accuracy level of census data.

Not surprisingly, the census data of Nepal are found to suffer from various levels of underreporting or misreporting, most of which is probably unintentional. Some families or individuals might be completely left out due to logistical or other problems or by their own choice. Even among those counted and included in the census, many may underreport or misreport, or fail to report, the number of children ever born or that of dead persons, particularly infants or young children. Because of this, the Nepalese census-based aggregate estimates of population growth and fertility and mortality often need adjustments, and such adjustments are made as necessary in estimating levels and trends using the census data. However, no adjustments are made to data pertaining to self-reported individual characteristics, such as caste or ethnicity, religion or educational attainment.

The inherent weaknesses or limitations of censuses or surveys, however, do not mean that census should not be conducted, that they are not useful or that the resulting data should not be utilized. It simply means that we have to be cautious in interpreting the data. Data need to be supplemented and verified, whenever feasible, through other sources. In this context, it will serve us well to remember that data should be presumed guilty until found innocent.

The current debate on the 2001 census of Nepal, in my view, is an essentially healthy exercise. It is likely to promote awareness of some of the census issues among the populace, while helping people to see the importance of being counted correctly and accurately. Although the ongoing debate in the media centers on data pertaining to ethnicity, religion and language, other types of information including age, births, deaths, migration, educational attainment are no less important. Quality of data pertaining to these and other areas depends primarily on accurate reporting and recording of information and the level of participation by the population.

I believe that most census respondents do not deliberately misreport their caste, ethnicity, language or religion. However, if a respondent chooses to misrepresent or incorrectly identify himself or herself, for whatever reason, then this cannot be helped; nor can the enumerator be held responsible or accountable for it. On the enumerator's part, there is hardly any reason or motivation for deliberately misinterpreting or misrecording respondents' self-reported information. Moreover, it would be highly unethical to do so. Finally, it is extremely important that enumerators are adequately trained on how to conduct the interview and record information in a consistent and accurate manner. At the same time, it is equally important to adopt effective ways and means to increase the rate at which the population participates in the census.

(Dr. Thapa is a senior scientist with Family Health International, Kathmandu. He specializes in demographic and reproductive health research and evaluation).