suggesting an ecological influence as important as, or more, important than individual
characteristics in generating the differential fertility we observed. The ecological terms
also moderate the impact of the other antecedent variables. We interpret this as reflecting
the influence of poor labour market prospects in precipitating early motherhood. We
attempted to allow for regional variations, after flagging those cases that had moved
home since the first birth, and found no significant difference except for a tendency to
earlier births in Wales and Northern Ireland. Those families who had moved since a first
birth are unlikely to have moved region, but they are strongly over-represented among the
younger first-time mothers. Their mobility is more likely to be an outcome than an
explanation of their childbearing. However omitting the mobility indicator does not
affect the pattern of the results. The cut of points for each set of results show a similar
pattern. The distance between the cut of points are larger for those in the youngest
groups and the oldest groups and relatively constant across those who enter motherhood
in their mid twenties/early thirties.
It is also debatable whether leaving school at the minimum age is a precursor of early
motherhood for those who first child was born before or at that age. As a preliminary
investigation we excluded 528 such cases in panel 1 (all birth orders) and 169 in panels 2
& 3 (first births only), but the resulting estimates were little changed. When data on
immigration history becomes available we will be able to investigate also how many of
these very early first births occurred overseas. In addition we shall be able to adjust the
compulsory school variable for those educated outside UK for those educated outside the
UK.
15