Confidance interval on contingency table

Xterminator

Hi,

I'm not sure hoe to get started on this assigment question.

I have a contingency table,, where a sample of 330 from unknown total were asked a question, with possible replies yes/no/undecided . The total for no is 46. But no st deviance is given

I need to calcualte a 90% confidance interval, which ithe formula is
x bar +or- z value * (sigma/root n)

Putting in the variables, 46+or-1.645 (z at 90%) * st dev/ root 330 .
But as i dont know st dev, so i cant finish the first part of question.

but as i dont have st deviation ... what do i do? any ideas please.

denachoman

Bit rusty so this may be way off the mark

If you assume binomial where p = yes
and q = no or undecided then
std dev = root of npq = root of [330*(46/330)*(284/330)] = 39.59

CI = 46 +/- 1.645[39.59/root(330)]

CI = [42.42, 49.58]

Xterminator

thks

I hope ur right!

X

Bradser

denachoman wrote:

Bit rusty so this may be way off the mark

If you assume binomial where p = yes
and q = no or undecided then
std dev = root of npq = root of [330*(46/330)*(284/330)] = 39.59

CI = 46 +/- 1.645[39.59/root(330)]

CI = [42.42, 49.58]

thats incorrect.
its a simple population porportion question.
the formula given gives all the data that is required to do the question.
i will do up a demonstration of how to tackle this problem and post it ASAP.