It's possible that statisticians have an odd definition of 'interesting'

That being said, I'm always interested in stories about where the numbers come from, like this article from CNNMoney:

Two recent price comparisons of grocery and household goods revealed that Target's prices are lower than at No. 1 retailer Wal-Mart.

Craig Johnson, president of retail consulting firm Customer Growth Partners, compared 35 brand-name items sold at Wal-Mart and Target stores in New York, Indiana and North Carolina. They consisted of 22 common grocery goods such as milk, cereal and rice; 10 general merchandise products such as clothing and home furnishings; and three health and beauty items.

Target's shopping cart rang in at $269.13 (pre-tax), a hair lower than the $271.07 charged at Wal-Mart.

"For the first time in four years, our price comparisons between the two has shown that Target has a slight edge over Wal-Mart," said Johnson. A smaller study by Kantar Retail found similar results.

Though I don't have enough information to say for sure, the CGP study looks pretty solid (particularly compared to some of the research I've seen from other consultants) and it's backed by additional analysis and a separate study.

What really stands out for me, however, is the slipperiness of assigning even a well-designed metric to something like the cost of shopping at a certain store. In order to get these results you have to buy the same brands of the same items in the same quantities as those in the studies. Your own shopping list would certainly produce different results (though the difference probably wouldn't be all that large).

It's always useful to remember that metrics are, with few exceptions, arbitrary. They may be useful and well thought-out but they should be approached with that caveat in mind.