Standard error: a poem

Hyndsight

davidgoodard

This poem was written by David Goddard from the Monash University Department of Epidemiology and Preventive Medicine. It is reproduced here with his permission. The poem won the inaugural Monash University poetry competition and will soon be published in an anthology of contemporary poetry.

For those who like this sort of thing (as I do), there is a nice collection of statistical poetry here.

Standard error

David Gordon Goddard

An inference that’s very often made –
‘a population tends to closely share
some feature that a smaller group displayed’ –
is why my students need to be aware
of standard error. Yet, it’s hard to learn,
and students’ drive to listen fades with each
attempt of mine to strive for ways to earn
attention till my message has its reach.

The teaching cycle brings around today
my yearly chance to make this topic clear.
I’ll aid my students, shorten what I say,
and they can later go and persevere …
… ask how it differs from the things it’s like,
and why we need it, what we couldn’t do
if it weren’t here. But can my students strike
the hours in busy lives to see this through …
… to think and tell themselves of things they know …
related things … then try to make the link
with standard error, find their gaps and so
seek remedy where knowledge meets its brink?

It’s four months on. A student whom I taught
then calls to see me and in measured way
explains she liked the insight I had brought
to ‘estimation’, then goes on to say:
“My research gathered data from a group,
defined a mob of which the group’s just part;
a feature in the group at six per cent
would be, I thought, like echoed in the mob.
I viewed the group as sample of the mob;
but samples vary some from whence they’re drawn.
Your standard error helped me calculate
how far from six my estimate might stray.”

She smiles and says she’d thought I’d like to know
that standard error served to underlie
her thinking. Yes, I’m happy that is so –
but more, she’d thought it worth enough to try.

comments powered by Disqus