MIGRATION
Nicholas Soames: To ask the Minister for the Cabinet Office what level of reduction in UK net migration from 2010 to 2011 would have been regarded by the Office for National Statistics as statistically significant. [121381]
Mr Hurd, Parliamentary Secretary: The information requested falls within the responsibility of the UK Statistics Authority. I have asked the authority to reply.
Letter from Stephen Penneck, dated September 2012:
As Director General for the Office for National Statistics (ONS), I have been asked to respond to your question concerning what level of reduction in UK net migration from 2010 to 2011 would have been regarded by the Office for National Statistics as statistically significant (121381).
On August 30, ONS published provisional estimates of Long Term International Migration (LTIM) for 2011. Estimates from the International Passenger Survey (IPS) are the main component used to calculate LTIM. As with all survey estimates, they carry a degree of uncertainty due to sampling variation. The uncertainty around the IPS based migration estimates was expressed as a confidence interval in this publication. A confidence interval provides an upper and lower limit in which we would expect the true value to lie with 95% probability.
These confidence intervals were published alongside inflow, outflow and net flow estimates. The change seen in the net migration estimates from 252,000 in 2010 to 216,000 (provisional) in 2011 was described as not statistically significant. This is because the confidence intervals for the IPS component for each estimate overlapped and the difference between the figures could have been due to sampling variation.
The 2011 provisional estimate for net migration would have needed to be below 180,000, for ONS to regard the change from 2010 as statistically significant. The upper limit of the confidence interval associated with an estimate below 180,000 is less than the lower limit of the 2010 estimate, so the confidence intervals do not overlap and the difference between the two estimates is considered to be statistically significant.