#
Assumptions and approximations

This is a section from my text book *An Introduction to Medical Statistics,
Fourth Edition. *I hope that the topic will be useful in its own right,
as well as giving a flavour of the book. Section references are to the
book.

##
1.4 Assumptions and approximations

Many statistical calculations give answers which are approximate
rather than exact. For example, if we carry
out a clinical trial and obtain a difference in outcome
between two treatment groups, this applies only to the
people in the trial. It is people not in the trial, those who
are yet to come and to be eligible for the trial treatments,
for whom we want to know the difference. Our trial can
only give an approximate estimate of what that might be.
As we shall see, statistical methods enable us to get an
idea of how precise our estimate could be, but this, too,
is only approximate. It depends on some assumptions
about how the data behave. We have to assume that
our data fit some kind of idealized mathematical model.
The great statistician George Box said that ‘essentially, all
models are wrong, but some are useful’. We might add
that, in medical statistics, all answers are approximate,
but some approximations are useful.
The important thing will be to have an idea of how
good our approximations are. We shall spend quite a
lot of time investigating the assumptions underlying our
statistical methods, to see how plausible they are for
our data.

Adapted from pages 2–3 of
*An Introduction to Medical Statistics* by Martin Bland, 2015,
reproduced by permission of
Oxford University Press.

Back to *An Introduction to Medical Statistics* contents

Back to Martin Bland's Home Page

This page maintained by Martin Bland.

Last updated: 7 August, 2015.

Back to top