> Dear all
>
> It has been a while since I've had to calculate a sample proportion and the resulting confidence interval for the population proportion, so I am not up to date with the current thinking on what to do. My understanding is that through the central limit theorem we have that the sample proportion p is normally distributed with mean pi (population proportion) and variance pi(1-pi)/n. We can also use p instead of pi in the variance and have a t-distribution with n-1 degrees of freedom.
>
> How do we interpret it if the confidence interval has limits below 0% or above 100%? Is this just because there is not enough data in the sample, or should I be using a different method for small p and small n?
>
> (Apologies to any of my former tutors who are appalled by any of the above!)
>
> Thanks in advance for your help, I will summarise any responses if people are interested.
>
> David
>
> ---
> David Smallbone
> Strategic Analysis Unit
> 13th Floor
> 40 Melton Street
> London NW1 2EE
>
> Tel: 020 7557 8340
> Internal: 085 78340
>
>
The content of this email (and any attachment) is confidential. It may also be legally privileged or otherwise protected from disclosure.
This email should not be used by anyone who is not an original intended recipient, nor may it be copied or disclosed to anyone who is not an original intended recipient. If you have received this email by mistake please notify us by emailing the sender, and then delete the email and any copies from your system.
Liability cannot be accepted for statements made which are clearly the senders own and not made on behalf of Network Rail.
|