CeLSIUS logo


Census logo

Study subjects

The sample for this study is taken from all women aged 15-19 who were present at the 1981 census. The sample is then further constrained by the need to know whether each woman gave birth during the next 5 years.

Note also that subjects could die or emigrate during the 5 years of follow-up after 1981, in which case they would not be at risk of giving birth throughout the whole period. We will have to exclude these people in this example, although a more sophisticated treatment would be to calculate total person-time at risk.

Furthermore, in order to test that these women are nulliparous, we will have to check that there have been no births to sample women in the 6 years before the 1981 census.

To link a birth to an LS record requires that the LS member is traced at the NHSCR at the time of birth. All of our sample must be traced so that we can tell whether or not that person has given birth. Therefore, our study subjects need to be traced before 1981.

The table below shows the number of women in the required age band at the 1981 census, and how the sample size is successively reduced by the various constraints imposed by the research question.

Sample description No. of subjects
Females aged 15-19 in 1981 20355
Females aged 15-19 in 1981, traced in 1971 18147
Nulliparous females aged 15-19 in 1981 17489
Nulliparous females aged 15-19 in 1981,
not dying or embarking in following 5 years
17411

Data extracted from LS 05/05/03.

Note that this study will be particularly sensitive to differences in the age distribution between the social housing group and the population as a whole. We would need to account for such differences before drawing any conclusions.