Probability: The Menu Problem

Re: Menu Problem minus the menu

I'm sorry, but my confusion persists. You say the mean is zero. That means 10% of your mean is -- also zero. The probability that one out of any number of samples falls between zero plus zero and zero minus zero is -- zero. You will need an infinite number of samples to reach the confidence you desire.

This answers the question you literally asked. But I suspect you really wanted to ask a different question. Which question did you mean to ask?

I've stayed away from this thread 'cuz the question is so bizarre but...

to clarify what I'm seeing so far... the definition of the standard normal curve is any normal distribution that has been standardized to a mean of zero and a standard deviation of 1 which then can be directly interpreted from the normal distribution probability tables.

1 SD on one side of the mean on a standard normal distribution represents 34.13% of the area under the curve (population distribution) so the mean +/- 1SD represents 68.26% of the population (mean +/- 2SD = 95%, which most people are familiar with as representing an alpha of 0.05).

But, I agree with you Thomas, I don't understand the question either...

It never entered my head that anybody would understand the question.

The disconnect between subjectivity and numerical measurement is profound because the former shifts about with factors outside the terms of the question.

Suppose you were grumpy, say, and your husband was taking you wining and dining to better explain about him being in bed with the maid.

Apologies to all
I want to thank all of you for your considerations, analysis, introspections and comments. Life is short and that you would spend any of it pondering my question marvels me.

But apologies are in order. I thought this a simple math problem. Ten minutes at most! It is similar to one I missed on a freshman statistics exam, even with a corresponding example in the text.

Forgetting the menu story, which is only one of many applications, this is a simple area under a curve problem. 20% is +/- .253 standard deviations. How many samples are required for an 80% confidence that one of those samples falls in that area under the curve.

I had hoped that someone would provide the elegant mathematical solution that I could plug and play any number of criteria. Instead I believe I shall now pursue the Monte Carlo simulation approach in Excel. Anyone suggestions in that arena would be much appreciated!

Thank you again everyone for the kindness of your time!
Randomly,
t