Tip:
Highlight text to annotate it
X
We are in Cambridge, England, 1920.
The weather could not have been more perfect
for high tea on the terrace.
We gathered at the table as a friendly group of colleagues.
The gathering had been progressing in a lively fashion
and the teacups were being refreshed,
when Lady Ottoline abruptly stopped the server
and pointed out with disdain that he had poured milk first
and then tea, rather than abiding by her widely known preference
for tea first, and then milk.
Sidelong glances were exchanged by numerous members
of our assembled group as they questioned
what difference could it possibly make whether
milk or tea were poured first in the cup.
It made, according to Lady Ottoline, all the difference.
A difference she could easily taste.
It was at this point that I as a scientist
(and amateur detective) decided to step in,
and proposed a little experiment.
Safely away from Lady Ottoline's line of vision,
eight cups of tea were prepared:
4 cups with tea poured first and 4 with milk poured first,
always in equal proportions.
Happily, Lady Ottoline sampled each of the 8 cups
and provided the crowd with her judgement
of "tea or milk first".
Remarkably, Lady Ottoline identified 8 out of 8 correctly.
Could such a feat be accomplished
by sheer guess work??
Now, as I mentioned, I am a scientist above all.
As such, I left this little party with more than just some favours.
Indeed, I took with me much food for thought...
Try my experiment yourself. Consider all the possible results.
In what order should your cups be presented?
How many cups must be correctly identified to conclude
that your subject can truly tell the difference?
You will undoubtedly come away with a greater understanding
of inferential statistics.