Re: Confidence Band question
- From: Jack Tomsky <jtomsky@xxxxxxxxxxxxx>
- Date: Tue, 23 Aug 2005 11:48:27 EDT
> I have a entry-level question about confidence band
> of standard normal
> distribution.
>
> According to the following webpage, for a standard
> normal distribution,
> "...The critical values at the 5% significance level
> are -0.087 and
> 0.087..."
> http://www.itl.nist.gov/div898/handbook/eda/section4/e
> da4283.htm
>
> While on another webpage of the same site:
> http://www.itl.nist.gov/div898/handbook/eda/section3/a
> utocopl.htm
> there is a formula to calculate the confidence band.
>
> +/- z(1-alpha/2)/sqrt(N)
>
> Say, alpha=0.05 ->z(1-alpha/2)=1.96
> given the sample size N=195 (quoted from the first
> webpage)
> I should have a confidence band (-0.14,0.14).
> What did I do wrong? Where is the band
> (-0.087,0.087)from?
>
> Thanks for your help!
>
That's a good question. There's an inconsistency between the formula that was given and the calculated value that was reported. You ought to contact NIST and point this out to them.
Jack
.
- References:
- Confidence Band question
- From: Seeker
- Confidence Band question
- Prev by Date: Re: ecommerce Conversion Rate Analysis
- Next by Date: Re: We can make stats say whatever we want
- Previous by thread: Confidence Band question
- Next by thread: Mediation testing in a logistic regression setup
- Index(es):
Relevant Pages
|
Loading
