In 2000 and 2004, the National Annenberg Election Study (NAES) was an exclusively telephone‐based rolling cross‐section survey. For 2008, the fieldwork strategy is more diverse. This paper outlines the reasoning behind the transition and describes the new structure. Because of declining telephone response rates and diminished coverage of households by landlines, some of the sample has been moved to the web. This also enables the study to leverage the statistical power of its sample through panel re‐interviews, thanks to the high panel retention rate of the web supplier, Knowledge Networks (KN). Each wave will be released as a rolling cross‐section of households in the KN panel, with further provision for replenishment of the sample. Altogether, five waves will be released between October 2006 and January 2009.