The Times, 17 October 2008
The Times claimed that the decision by ONS to publish a measure of sampling variability in its retail sales figures was an admission of inaccuracy. The National Statistician wrote to set the record straight.
ONS retail statistics are the best available and are not inaccurate. When we publish a quality measure of our statistics it does not imply that "ONS admits inaccuracy in its retailing statistics" (Oct 17). All survey estimates are, by definition, estimates. The larger the sample, the more likely it is to give an accurate value. The ONS retail sales estimates are derived from a large survey of 5,000 businesses - considerably more than those used by the CBI and the BRC, which cover fewer than 100 retailers. The ONS estimates are a comprehensive and representative reflection of retail activity.