This week The Economist published an article (“This spreadsheet is different”, 21 September 2013) apologizing for a earlier article that was based on erroneous analysis. The problem was caused by a spreadsheet model that
…pulled data […] from the wrong place in the spreadsheet, changing the results.
We’ve written about the problems with spreadsheets before. They are opaque, brittle and difficult to test. We’ll continue to see these kinds of embarrassing errors until analysts move to more robust tools.