You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Choice in wine indicates personality?
- Thread starter Brad Kane
- Start date
originally posted by Brad Kane:
originally posted by Jonathan Loesberg:
My point about Brad's original connection is still the same: it's silly.
Wait, you can't make generalizations on six billion people by studying forty-five of them? Say it ain't so!
Sure you can.
The issue has to do with implying causation from a correlational study. With random assignment and the right experimental conditions, you can broadly generalize from a sample of 45.
Jonathan Loesberg
Jonathan Loesberg
originally posted by VLM:
originally posted by Brad Kane:
originally posted by Jonathan Loesberg:
My point about Brad's original connection is still the same: it's silly.
Wait, you can't make generalizations on six billion people by studying forty-five of them? Say it ain't so!
Sure you can.
The issue has to do with implying causation from a correlational study. With random assignment and the right experimental conditions, you can broadly generalize from a sample of 45.
From 45 people to 6 billion people reliably? Can 45 people really encompass all the relevant variables?
Jeff Grossman
Jeff Grossman
The usual way to say it is: What is the margin of error and confidence level offered by a sample size of 45 drawn randomly and independently from a population of 6,000,000,000 ?originally posted by Jonathan Loesberg:
From 45 people to 6 billion people reliably? Can 45 people really encompass all the relevant variables?
Per these lovely little calculators, the answer is: the 45-person sample represents the population with confidence level +/- 15%.
If you want to get that confidence down to something about which you can feel pretty confident, like say 5%, then you need 384 people.
Ain't statistics grand?
Jonathan Loesberg
Jonathan Loesberg
originally posted by Jeff Grossman:
The usual way to say it is: What is the margin of error and confidence level offered by a sample size of 45 drawn randomly and independently from a population of 6,000,000,000 ?originally posted by Jonathan Loesberg:
From 45 people to 6 billion people reliably? Can 45 people really encompass all the relevant variables?
Per these lovely little calculators, the answer is: the 45-person sample represents the population with confidence level +/- 15%.
If you want to get that confidence down to something about which you can feel pretty confident, like say 5%, then you need 384 people.
Ain't statistics grand?
Again, with how many variables? This makes sense if the claim to be tested is say that people who eat scrambled eggs three times a week are more likely to vote Republican. But if the claim is a correlation between wine preference and personality type where one assumes that there are more than say six personality types and probably more than 45 permutations of wine preferences depending on how specific one was going to be, would you still get the same representative numbers? I'm asking from ignorance, and I'm not skeptical of the ability to make reasonably sound statistical predictions, but it would seem to depend on what one meant to predict.
Jeff Grossman
Jeff Grossman
The way I said it is for measuring one attribute. If there are many, that will take more samples, of course.
I am going to leave the rest of the statistics to Senor Monkey, as IANAS.
My epistemologic problem with this sort of thing is that there is no simple small-ish set of things like "wine preferences" and "personality types". (Sure, I can make them up but obtaining would be nice.)
I am going to leave the rest of the statistics to Senor Monkey, as IANAS.
My epistemologic problem with this sort of thing is that there is no simple small-ish set of things like "wine preferences" and "personality types". (Sure, I can make them up but obtaining would be nice.)
Lou Kessler
Lou Kessler
No! No! Think about it, they like all wine at a demi-sec level.originally posted by Brad Kane:
originally posted by SFJoe:
What about those with no personality at all?
They drink Pabst Blue Ribbon?
Levi Dalton
Levi Dalton
If you add diode attachments, I believe the margin of error dips below 4%.
Lou Kessler
Lou Kessler
Is it true that Troiano uses only Yquem on his pancakes daily?originally posted by Thor:
Let's get Rahsaan to debate Tom Troiano, re: the borders of Boston. I bet it would surpass the limerick thread!You must be great at parties, Rahsaan.
originally posted by Lou Kessler: Troiano uses only Yquem on his pancakes daily?
Someone eats pancakes every day?!
Ian Fitzsimmons
Ian Fitzsimmons
Can the data in this study be transposed into Meyers-Briggs input? Then you could incorporate wine tasting sessions into management retreats and strategic planning sessions. Talk about an executive wine seminar!
Jay Miller
Jay Miller
Charles Weiss
Charles Weiss
originally posted by Thor:
I strongly suspect, but will never test, that we could do a fair bisection of wine preference and (U.S.) political affiliation with easily predictable results. And yes, I know outliers.
And what type of wine do you think would be preferred by the bi section?
Ian Fitzsimmons
Ian Fitzsimmons
originally posted by Thor:
I strongly suspect, but will never test, that we could do a fair bisection of wine preference and (U.S.) political affiliation with easily predictable results. And yes, I know outliers.
Then there would have to be Fox Wine. Oh, wait, there already is Fox Wine.