Chemical Forums

Chemistry Forums for Students => Analytical Chemistry Forum => Topic started by: abcc on November 12, 2006, 09:17:52 AM

Title: confidence interval
Post by: abcc on November 12, 2006, 09:17:52 AM
A volumetric calcium ananlysis on 3 samples : 3.15, 3.25 , 3.26
what is the 95%confidence limit for the mean of data,assuming no prior information about the precision of the analysis?
i would like to know what 's the meaning of no prior information about the precision of the analysis
what's the consequent of the standard deviation ???
thx a lot  :)
Title: Re: confidence interval
Post by: Borek on November 12, 2006, 10:04:37 AM
No prior information means just that - you are using method that you have to evaluate. Sometimes you use an analytical instrument with known accuracy (known from previous measurements, or known as it is a factory data) - then you don't have to calculate it for yourself from data.
Title: Re: confidence interval
Post by: billnotgatez on November 12, 2006, 11:48:34 AM
It is interesting that a confidence interval is requested with just 3 measurements.

Would the below be the mean and standard deviation.
3.22
0.06083

Title: Re: confidence interval
Post by: abcc on November 12, 2006, 05:07:20 PM
It is interesting that a confidence interval is requested with just 3 measurements.

Would the below be the mean and standard deviation.
3.22
0.06083


i have checked the answer ,
the mean is 3.22 and the S.D. is 0.15
Title: Re: confidence interval
Post by: billnotgatez on November 12, 2006, 09:38:19 PM
What formula did you use to derive the standard deviation?
Title: Re: confidence interval
Post by: Dude on November 13, 2006, 07:06:35 AM
The basis of the question is to use probability theory to obtain the confidence interval.  Obtain the standard deviation and mean, as Bill did.  A Gaussian distribution means that 68 % or so of the measurements will fall within the mean  plus or minus the standard deviation.  The 95 % confidence interval will contain 95 % of the measurements and is usually a factor multiplied by the standard deviation.  It is dependent on the number of measurements (3 is exceeding low).  There should be tables in the appendix of your book or a statistics book.
Title: Re: confidence interval
Post by: abcc on November 13, 2006, 09:15:29 AM
i have learnt a lot from everyone nice repliy, ;)
but i still can't get the answer correct ,
answer is 3.22+/_0.15
why is 0.15???
Title: Re: confidence interval
Post by: abcc on November 13, 2006, 09:17:57 AM
What formula did you use to derive the standard deviation?

OOOO
U ARE CORRECT
i aslo find the mean and standard deviation
mean: 3.22
S.d: 0.06
BUT do u know why the CI is 3.22+/_015 :'(
Title: Re: confidence interval
Post by: billnotgatez on November 13, 2006, 09:34:01 AM
If you look at the normal statistical bell curve about 95.4 % of the area lies within 2 standard deviations. How confident would I be that with 3 measurements?

http://en.wikipedia.org/wiki/Confidence_interval
Title: Re: confidence interval
Post by: chiralic on November 13, 2006, 04:47:00 PM
Hello abcc:

If you check the book: Instrumental Analysis by Skoog and Leary Chap. 28 (I have 4th Edtion) you'll find an example EXACTLY SIMILAR to your post...and also you'll can find an explanation about what's the meaning of no prior information about the precision of the analysis .

Regards,

Chiralic
Title: Re: confidence interval
Post by: abcc on November 13, 2006, 04:55:31 PM
Hello abcc:

If you check the book: Instrumental Analysis by Skoog and Leary Chap. 28 (I have 4th Edtion) you'll find an example EXACTLY SIMILAR to your post...and also you'll can find an explanation about what's the meaning of no prior information about the precision of the analysis .

Regards,

Chiralic
THXTHXTHXTHX
big help ;)
Title: Re: confidence interval
Post by: billnotgatez on November 15, 2006, 04:53:49 AM
If the confidence interval is 3.22±0.15 that appears to be about 2.5 standard deviations.

Comments anyone.