Business survey error achieves complete statistical coverage
A business survey error affected every measure in the Office for National Statistics Annual Business Survey release, achieving a level of reach usually promised only by broadband adverts. At last, a problem that treated every column equally.
The ONS said a processing issue caused the error. Its early assessment suggested approximate gross value added might rise by around 0.7 percent. GDP estimates remained outside the blast radius, sitting quietly behind the safety glass.
The spreadsheet enters a period of reflection
Analysts received a request to stop using the figures while corrections were prepared. Consequently, several spreadsheets were closed gently and told none of this had anything to do with their formatting.
“The error was impressively even handed,” said Quentin Pivot Table of Numbers Adjacent to Numbers. “It did not discriminate by sector, column or person who had already built forty seven slides from it.”
The survey samples tens of thousands of firms and measures turnover, purchases and value added. Meanwhile, the correction demonstrated a newer indicator, the speed at which a caveat can travel through a consultancy deck.
The amber cell asks not to be quoted
The business survey error had no impact on GDP. Economists relaxed in the traditional way, by opening another release and debating whether 0.1 percentage points meant resilience or doom.
A corrected workbook was promised. Until then, every cell received an amber background, the recognised colour of a number asking not to be put in a press release.
The error will eventually be fixed. The forty seven slides will then need forty seven footnotes, proving that no correction comes free once somebody has booked a meeting room.
Read the real story: The ONS notice on its Annual Business Survey processing error.
