New births into the Longitudinal Study
In England and Wales (E&W) all live births must be registered within six weeks of the birth occurrence by law. The quality of births data for the whole of England and Wales is known to be extremely high.
All individuals born in England and Wales on one of four dates of birth - the so-called "LS dates" - enter the study as "new births" into the LS. New births are entered into the study directly from birth registration data and entry rates are expected to be very close to 100 per cent.
Methods
The LS includes linked census and events data for people born on one of four selected dates in a calendar year. Assuming an even distribution of births across the days of the year, the LS basically samples 4/365 births in non-leap years and 4/366 births in leap years. The expected numbers of new births into the LS are given by FIGURE 1.
In census years, normally only one LS date falls in the period from the start of the year up to the day before census night, with three LS dates falling afterwards, in the period from the day after census night until the end of the year. The expected numbers of new births into the LS for these two periods are given by FIGURE 2.
Results
New births occurring after 2001 Census are shown in Table 1 together with sampling fractions and entry rates.
Table 1. New births of LS members by year of birth, 2001-2005
| Year of birth | LS |
E&W |
Sampling
fraction |
Exp. in LS | Linkage rate |
| Part 2001 | 5,241 | 400,767 | 1.31 | 4,887 | 107.23 |
| 2002 | 7,217 | 596,122 | 1.21 | 6,533 | 110.47 |
| 2003 | 6,787 | 621,469 | 1.09 | 6,811 | 99.65 |
| 2004 | 6,320 | 639,721 | 0.99 | 6,991 | 90.40 |
| 2005 | 7,026 | 645,835 | 1.09 | 7,078 | 99.27 |
| Total | 32,591 | 2,903,914 | 1.12 | 32,300 | 100.90 |
Between 2001 Census day and the end of 2005, the overall achieved sampling fraction for the LS was very close to the 1.1 target. Annual sampling fractions are more variable.
This happens because births are actually not equally distributed throughout the year or by day of the week. Due to differences in the days of the week on which the fixed LS sample dates fall, the number of new births of LS members and sampling fractions may vary from year to year. This impacts on the observed variability of the annual entry rates around the 100 target, given that the expected number of LS events used as denominator of the rate assumes an even distribution of births. Over a period of years, differences tend to even out, resulting in a total entry rate between 2001 and 2007 very close to 100.
While all new birth entrants to the LS must be resident in England and Wales, the estimates for England and Wales include a small number of births to mothers who are not usually resident in England or Wales. This would appear as an undersampling in the LS.
Sampling fractions and entry rates by sex for births into the LS since 1971 are available from the Downloads box.
