Survey this
- PDF / 690,122 Bytes
- 2 Pages / 585 x 783 pts Page_size
- 99 Downloads / 203 Views
T
he other day when I encountered yet another customer satisfaction survey, something seemed vaguely familiar. As this is election season in the United States (when is it not?), the results of political polls are reported ad nauseam on every news outlet. Those too rang a familiar bell. So I began to think about what a survey does, or tries to do. Something more fundamental than just collecting opinions seemed at play. Watching the news reports, the first thing to notice, if one is a skeptic who always looks for the fine print—which I am, is the quoted margin of error in the footnotes. Typically, an error such as ~4.5% is quoted, or ~3% for the better surveys. It becomes obvious that those surveys polled 500 and 1000 respondents, respectively. As a graduate student pulling an all-nighter in the lab watching the counts from the decay of a radioactive source accumulate ever so slowly in my multichannel analyzer, I was painfully aware of the statistical accuracy required before I could head home. This was worse than “a watched pot never boils.” It was the revenge of Siméon Denis Poisson and his damnable statistics. I thought root N by N would be the end of me.
But enough about the character-building experiences of my youth. Apparently the answers provided by respondents to a survey are random independent countable events, just like radioactive decay. That might have satisfied my curiosity, but the analogy seemed somewhat flawed. It did not fully explain the survey déjà vu I had experienced. Therefore, I decided that a Gedanken experiment was in order. Let’s say that you abstain from sweets altogether. You are subjected to a survey of eating habits and one of the questions is, “Do you prefer chocolate or vanilla ice cream?” No room for nuance is offered and an answer is mandatory. Based on no logical criteria at all, but perhaps based on the dessert’s color, its alphabetical preeminence, or just the flip of a coin, you choose one of the alternatives. This measurement of your fully undetermined preference has forced you into one of two ice cream camps. In the language of quantum mechanics, you were just subjected to the calamitous, but thankfully painless, process of wave-function collapse. And thus came the epiphany. These polls remind me of physics in general and quantum principles in particular—how appropriate, since both pursuits contend with an unforgiving uncertainly principle! We can stretch my Gedanken a bit further. You and your identical twin literally live poles apart. She has a sweet tooth, and at dinner on a recent visit between the entrée and the coffee, she orders chocolate ice cream for dessert. You have the fruit and cheese, just to maintain your superiority in such matters, but her likely addiction to tryptophan does not escape your notice. In your younger days, you and your sister used to change places in this circumstance to see if the waiter noticed, but the exchange energy MRS BULLETIN
•
VOLUME 41 • MARCH 2016
•
www.mrs.org/bulletin
263
FEATURES POSTERMINARIES after a full meal was too grea
Data Loading...
