Approaches to identify the amount of publishable information in business surveys through waivers

Abstract

Statistical agencies make use of waivers to increase the amount of information that they release in tables of business data. Waivers, obtained from large contributors to sensitive cells, allow the agency to publish the cell value even though doing so may lead to the disclosure of these large contributors' values. Waivers allow sensitive cells to be released and also reduce the need to suppress nonsensitive cells to prevent the derivation of these sensitive cells' values. As waivers are not easy to obtain it is desirable to identify businesses whose waivers will have the largest impact on the amount of publishable information. Two heuristic approaches are presented for the identification of priority cells for waivers that factor in properties of neighboring cells. Empirical investigation on two business surveys shows that the approaches have merit but that the structure of the tabular data plays an important role in the results obtained. © Springer-Verlag 2004.