Appendix 2 - Comparisons with Previous Mid-year Estimates

Provisionally revised 2000 mid-year estimates

A set of provisionally revised 2000 mid-year estimates, Table 5, has been produced to provide a realistic indicator of population change over the year 2000 to 2001. These were calculated by 'rolling back' the 2001 mid-year estimates, i.e., the 2001 population was reduced by a single year of age at each age and sex, deaths in the previous 12 months were added, births were subtracted and an estimate of migration was subtracted to arrive at a provisional revised set of 2000 mid-year estimates. These are considered provisional revisions as further revisions may occur once the revised series for 1982-2000 is produced, go to Appendix 3.

This provisional revised 2000 mid-year estimate is some 56,400 less than the previous 2000 mid-year estimate published by the Registrar General in April 2001. The difference between the previous mid-2000 estimate and the provisionally revised 2000 estimate is most apparent in males between the ages of 20-39. At the older ages (60 and over), the provisionally revised estimates tend to be higher than the previous 2000 estimate.

Table A2 shows the differences between the two population estimates by council area. The councils with the largest overestimate in the previously published 2000 figures are Glasgow (5.9%) 36,000 high, Renfrewshire (2.1%) 3,700 high and North Lanarkshire (2.1%) 6,800 high. Councils with the greatest underestimate in the previously published 2000 figures are Dundee City (3.0%) 4,300 low, Argyll & Bute (2.9%) 2,600 low and Moray (2.3%) 1,900 low.

Reasons for the difference

It is expected that the difference of some 56,000 between the previously published and the provisionally revised 2000 mid-year estimates is largely attributable to errors in previous estimates of migration in the eighties and nineties. It has been well documented that migration is the most difficult component of population change to estimate because of the lack of a wholly reliable source of data on migration. Estimates of net migration are the difference between large flows of migrants into and out of Scotland based on survey data (International Passenger Survey), or, for movements within the UK, from available administrative data (National Health Service Central Register recording the moves of patients throughout the UK). These estimates of net migration will, therefore, be very sensitive to measurement error in the flows in each direction and even more sensitive for particular age/sex groups.

For Scotland there are two key flows of migrants: migrants to and from the rest of the UK; and migrants to and from the rest of the world. Annually these two estimates have flows of about 50,000 each way with the rest of the UK and about 20,000 each way with the rest of the world. The estimates of international flows will be subject to greater error given the small sample sizes and the inherent non-sampling errors that exist in most surveys. However, both flows will contribute to the difference and the work to revise the population estimates back to 1982 will attempt to identify the interaction of these different flows. Some preliminary investigations suggest that the error is more likely to be the result of underestimating out-migration rather than overestimating in-migration, i.e. more people leaving Scotland than originally estimated.

The errors in estimating migration in the eighties will also have affected the adjustments made to 1991 Census counts to take account of underenumeration. The adjustments to 1991 Census counts were incorrect and added some people, particularly young men, to the population whom we now believe had actually migrated out of Scotland rather than having been missed by the 1991 Census.