Made to measure?

The complexities faced by statisticians attempting to talk about immigration and population are dealt with in a special report in the latest issue of Prospect by Michael Blastland and Andrew Dilnot. And it seems that their warnings couldn't be timelier, as the ONS have just announced that, if recent trends continue, the UK population is expected to grow to 71m by 2031, something that has made the headlines in most newspapers.

The key phrase here, of course, is "if recent trends continue." As Blastland and Dilnot point out, British population projections for 1995 made in 1955 and 1965 respectively differed by more than 20m because fertility rates changed so much during that decade. Similarly, the current ONS figures are the first to take into account new data on immigration, and the chance of immigration remaining predictable between now and 2031 seems about the same as 1965 fertility estimates being valid in 2000 (as it turned out, these estimates were well over double the actual figure).

It's also interesting to note the way in which almost every newspaper report I've seen makes no distinction between "predictions" and "projections"—only one of which the ONS is actually producing. Let us know what you think here.