What does integration imply in choosing a unit of enumeration: enterprise, holding or individual? Does it matter? Perspectives from Africa
MetadataShow full item record
Statistical integration is advanced as a means of making data meet particular needs and ad value to the whole statistical data collection and management system. At the data collection level it can produce significant benefits by reducing the cost of statistical collection and the burden placed on respondents, whilst also increasing the value of outputs in terms of achieving consistency and accuracy. At the data processing level, integration enables the benefits of common technology, analytical methods, tools and processes to be fully exploited. Data from different sources and different times can be consolidated to allow for richer databases to be developed and meaningful comparative analysis and interpretation of results to be achieved. Integration implies that common statistical frames, definitions and classification are promoted and used in all statistical surveys to achieve harmony in enumeration units such as enterprise, holdings or individual. However, flexibility which may be necessary at times is lost. For instance agricultural data may require zonal based sample frames than household agricultural holdings. The periods for data collection may be agricultural seasons rather than calendar months. Thus integrating general household survey data with agricultural data may be complicated. This paper examines the issues of statistical integration from the African perspective and discusses the challenges, opportunities and the desirability of pursuing and achieving such integrated systems.