Geography Reference
In-Depth Information
cut the data for the 1,800 1931 districts up into 2,916 fragments and then
reassembled them into the 1,472 1939 units. The table was very carefully
cross-checked by comparing the 1931 populations of 1939 units computed
by applying the boundary change information to the 1931 census figures
with the 1931 populations listed in the 1939 report. The results showed
that people brought up in Britain's depressed areas in the 1930s have sig-
nificantly worse health today, even if they later moved to other localities.
9
The third approach works with samples of people, inevitably small,
who have been surveyed repeatedly starting at birth. The first such Brit-
ish project, the National Survey of Health and Development (NSHD),
covers 5,362 people born in March 1946.
10
Similar “birth cohorts” were
started in 1958, 1970, and 2000, but the NSHD is the most relevant to ag-
ing research, as its members are now in their midsixties. Our work on the
NSHD has two distinct elements. One task was assisting in converting
postal addresses for cohort members into geographical coordinates. The
3,354 addresses from 1999 all include postal codes, and locations were
easily found for all members. All but 50 of the 3,519 addresses from the
1972 survey were automatically matched against a modern address list.
However, of the 4,856 addresses from 1950, 2,658 (54.7 percent) could not
be matched automatically. Although the number of residential addresses
today is vastly greater than in 1950, the majority of houses and flats that
existed in 1950 still exist, so many of the nonmatches were due to detailed
variations in spelling and punctuation. However, our particular focus
was the relatively small fraction of 1950 addresses that simply no longer
exist, because these were concentrated in areas in which the street pat-
tern has been completely replaced by “comprehensive redevelopment.”
These are inner-city areas with very poor housing conditions, whose
consequences the project is especially interested in. Locating these ad-
dresses involved working with georeferenced scans of mid-twentieth
century mapping at 1:10,560 scale.
The other focus of our work is including additional area-based data
in the study. There are two distinct reasons for doing this. One is that
when the NSHD was started it was more narrowly medical than today,
so while we know a lot about the babies' physiology, we know litle, for
example, about the houses they lived in. Aggregate data from the 1951
census about, for example, the percentage of homes lacking their own
