| Abstract: | The population data of statistically under-developed countries are usually subject to many substantial errors. Techniques are required to detect and, by adjustment, to reduce these errors. In the past, much of this work has been of an ad hoc nature. This present study attempts to contribute to the building up of a general methodology for dealing with these problems. Each technique for detecting error is paralleled by techniques for reducing the error, and a secondary objective was to integrate closely each detecting technique with its associated adjustments. The criteria used are based on the need for internal consistency, and the fields studied are the ogive, the proportion of each generation surviving from one census to the next, sex ratio, digital preference and age overstatement at the older ages. |
