Your piece last Sunday by Neil Callanan headlined "Tesco's price cuts fail to deliver" states in the second paragraph that "Superquinn's share has dropped by more than eight percentage points."

According to TNS Worldpanel, Superquinn's share moved from 7.5% to 6.9%. This is a move of 8%, not eight percentage points. "Points" is only one word but it results in a very different meaning.

The article also mentions "Nielsen Ireland showing that 250,000 households in the Republic are now shopping in the North". Again, this is not true. Nielsen does not use households as a source of data collection on grocery sales – we use data directly from retail stores. I assume this statement should be attributed to TNS and not to Nielsen.

The article states that "SuperValu's share of the grocery market is down from 20.3% to 20.0%". All the TNS figures are drawn from a sample of 3,000 households in the Republic as stated in the article. Therefore, all share estimates from this sample are subject to margins of error in the same way that opinion-poll voting intentions are subject to margins of error. On a sample of 3,000, a movement from 20.3 to 20 is well within the margin of error and therefore not significant. Rather than being "down", it would be more accurate to say that SuperValu is "holding its position".

Lastly, the article states that "Symbol groups continue to lose market share... from 5.8% to 5.4%". Symbol groups are stores such as Spar, Centra and Londis. These are convenience stores. At the end of Neil's article he states that "They [TNS] do not include the convenience market" [in their service]. It is well established in the market-research industry that household-based surveys like TNS are not good at measuring impulse purchases made at convenience stores, hence reference to the Symbol group's share is misleading.

Tom Harper,

Managing director,

The Nielsen Company,

14 Riverwalk,

Citywest,

Dublin 24