Absence of evidence, bagels, discrimination of

Therefore, there is no statistically significant relationship, no causal effect, between water composition and bagel quality in any zip code in New York City.

Some people may find this uncompromising conclusion much to their liking. Others will doubtless damn it as lies, possibly damned lies, and definitely “a multiple-regression experiment”. I just wonder: why bother?

So often, there is a rush to explain things that in truth require no explanation. Pseudoscience is littered with valiant attempts to conjure mechanisms — morphic resonance, N-rays, even, in the wrong fields, gravity — to explain the unexplained, but real life is not immune. So it is with New York, bagels, and water.

Michael Muroff decided that the observation that “Residents of New York City have long touted the superiority of their bagels over those of other parts of the United States” required explanation rather than confirmation. So he conducted the “multiple-regression experiment” and wrote it up for Art of Eating 112. Muroff used complaints about water quality in an effort to explain differences among one person’s ratings of bagel quality for 202 bagel stores in New York City.¹

The thing is, and those who have been reading me for a while will know this, there is absolutely no evidence that anyone can reliably tell the difference between a good New York bagel and a good bagel made some other place. Same goes for pizza. And note, the point is difference, not superiority. I don’t care whether people prefer New York bagels, I’m much more interested in whether they can reliably distinguish between New York and Other Place. In It must be something in the water I offered one lengthy and overcomplicated scenario. It could be done much more easily with a triangle test.

The tester/taster is offered three samples, two of which are the same while one is different. Can they identify the odd one out? With suitable precautions — which would include the physical presentation of the bagel samples as well as overall similarity — it is relatively easy to discover both whether some people are expert bagel judges and whether a population in New York or anywhere else can indeed tell the difference. Extending further, are self-selected bagel mavens better than average? Maybe Montreal (or Brick Lane) bagels are actually more distinctive than New York bagels?

Once you have that evidence, we can worry about why.²

I’m available to supervise the research.