Summary: Memories of teaching a singular student: Victor Rothschild
P.M.E.Altham, Statistical Laboratory, University of Cambridge, CB3 0WB.
September 23, 2003
The excellent book by Kenneth Rose `Elusive Rothschild' has prompted me to record my own
memories of `Lord R' (as I always thought of him) with whom I had professional contact as
`statistics supervisor', mainly in the early 1980's.
In fact Lord R rst contacted my husband Jimmy, who is a philosopher, early in 1972, very soon
after the birth of our elder son, as it happened.
The problem concerned was connected with Think Tank work. Then, in 1978 Lord R came to
see me at home to ask me a small but important technical question related to his preparation
for the 1978 BBC Dimbleby lecture on `Risk'. This question was essentially the following: if one
presents an estimate of a quantity, for example p, a risk, then one should always attach some kind
of measure of uncertainty, eg a standard error, to this estimate, ^
p. Lord R was of course quite
correct on this point, and I think we then discussed how one might reasonably construct se(^p).
(Not at all easy, as we were talking about p as the probability of a very rare event.)
A note on the practical arrangements:
Lord R came once or twice to see me at home, rstly when we were in Canterbury Close, and then
when we had moved, in 1979, to our (now ridiculously valuable, as is the way with Cambridge
property) house near the railway station. Our two small boys had to be vigorously shushed, or