The Lady Tasting Tea - Miffy or Tiffy? Food defines science!

Do you milk-in-first or tea-in-first ?

Milk and tea, actually dairy in tea started long before the British did it, even if we think they invented it. This is one of the great origin stories in science — the moment that gave us the modern logic of the experiment, and it started as a tea-table squabble.

The location is Rothamsted Experimental Station, an agricultural research institute in Hertfordshire, England, sometime in the early-to-mid 1920s. Among the staff was Blanche Muriel Bristol, a British phycologist whose research focused on how algae acquire nutrients — an algae scientist, crucially, not a statistician. Also there was Ronald A. Fisher, a young statistician who would go on to become one of the founders of the entire field.

The anecdote, as recorded by Fisher's daughter and others: Fisher offered Bristol a cup of tea he'd just drawn from the urn. She declined, saying she preferred the flavour when the milk was poured in before the tea. Fisher scoffed that the order of pouring couldn't possibly affect the taste. Bristol insisted it did, and that she could tell the difference. Rather than let it drop, a third scientist standing nearby spoke up: "Let's test her." It was William Roach — who not long afterward married Muriel.

What makes this more than a cute story is how Fisher turned an idle boast into a rigorous test, and in doing so laid out principles still used everywhere today. His design was elegant. Muriel would be given eight cups — four prepared milk-first, four tea-first — presented in random order. She knew there were four of each, and her task was to sort them into the two groups. 

The randomization mattered enormously and was presented in random order. Fisher protected against any pattern she could exploit and against unconscious bias in how they were presented.

Here's the beautiful part — the math. With eight cups and the requirement to pick which four were milk-first, there are exactly 70 possible ways to choose four cups out of eight. 

Only one of those choices is entirely correct. So if Bristol were simply guessing with no real ability, her chance of getting every cup right by luck alone is 1 in 70, or about 1.4%.

That number is the whole point. 

Fisher framed it around what he called the null hypothesis — the assumption that she had no discriminating power at all — and the question became whether her result was surprising enough, under that assumption, to cast doubt on it. 

This was the original exposition of Fisher's notion of a null hypothesis, which is "never proved or established, but is possibly disproved, in the course of experimentation." 

As for what actually happened: the most repeated telling is that she nailed it. According to Salsburg, the lady identified every single one of the cups correctly, which would put her result at roughly P=1/70 — comfortably past the conventional threshold for "this probably isn't chance." 

The lasting significance is out of all proportion to the teacups. Fisher used this exact example to open a chapter of his 1935 book The Design of Experiments, and from it flow ideas that now underpin essentially all empirical science: the null hypothesis, randomization as a defense against bias, careful experimental design, statistical significance, and the reasoning that later evolved into what is now called Fisher's Exact Test.

His description runs less than ten pages and is notable for its simplicity and completeness regarding terminology, calculations, and design. 

Nearly every modern clinical trial, A/B test, agricultural field trial, and countless other scientific experiments trace part of their statistical logic back to Fisher's work. 

There's also a radical idea buried in it. A junior algae researcher made a claim that a table of senior scientists found laughable — and instead of letting status settle the matter, Fisher built a test that didn't care about anyone's title. The data would decide. That shift, from authority to method, is arguably the experiment's real legacy.

The story lent its name to a popular 2001 book by statistician David Salsburg, The Lady Tasting Tea: How Statistics Revolutionized Science in the Twentieth Century, which is a great read if you want the broader history of the people who built modern statistics.

Given the milk-protein chemistry — the way denaturation differs depending on whether hot tea hits cold milk or cold milk hits hot tea — there's a plausible physical basis for Murial tasting a real difference. She may not have been bluffing at all. 

And proof that people can taste a difference and food matters.

Milk and tea, actually dairy in tea started long before the British did it, even if we think they invented it.  Read more about this here