# Confidence intervals

1. Jun 23, 2008

### sstudent

we need to show with 90% confidence that the probability of the "treshhold number" being correct is 99.99%. "So we want P=0.9999 with C=0.90". we can consider 5 trails with out put { 4, 7, 12, 9, 11}..we can say the treshhold # is the "2nd trail = 7"... i tried doing it using t alpha but i failed to get it... i havent taken statistcs in years..thanks for help folks

2. Jun 23, 2008

### sstudent

is it that hard??

3. Jun 23, 2008

### Dick

Yes. I don't understand what you are talking about. Can you give a reference for the background to this problem?

4. Jun 24, 2008

### sstudent

5. Jun 24, 2008

### Dick

Re: confiendnce

Ok, now can you explain what your problem is? You have five trials? You want to say the probability of something is 0.9999? The probability of what exactly?

6. Jun 24, 2008

### sstudent

Re: confiendnce

the probabilty of the treshhod # being correct, which is the 2nd trail..thanks

7. Jun 24, 2008

### Dick

Re: confiendnce

This isn't helping. You are just repeating what I don't understand. What do the trials represent? Are they samples from some distribution? What does it mean for one of them to be 'correct'. What is correct? What's a 'threshold'? Can you define all of these things?

8. Jun 24, 2008

### sstudent

Re: confiendnce

the trails are just sample from some distributaion, for the treshhold # u can say its just a numer, so we are trying to prove the the output of the samples larger than 7 "treshhold" is 99.99% with 90%confience..i hope that makes it more clear

9. Jun 24, 2008

### konthelion

Re: confiendnce

I believe you are talking about the tolerance interval. In this problem, your confidence level is 95% and the threshold interval for capturing at least 99.99% i.e. % of these samples falling in the tolerance interval is:
$$\bar{x} \pm (tolerance)s$$ , where tolerance is your tolerance critical value.

Your statistics book should have a tolerance critical value table.

Last edited: Jun 24, 2008