1. ## Confidence Intervals

how do you go about constructing a 95% confidence interval. you guys have been great with the help links so far.

2. Is there a way to do them on a calculator?

3. Originally Posted by GoneFishing
Is there a way to do them on a calculator?
There is a way to compute that on your calculator.
I do not know the exact details, maybe CaptainBlank or JakeD can explain better.

Those numbers on the table of the normal curve appear as the integration answer to,
$\displaystyle \int_0^t e^{-x^2}dx$
So for example you want to calculate,
from $\displaystyle 0<z<.5$
Then you substitute that into the integral for $\displaystyle t=.5$ and you need to find,
$\displaystyle \int_0^{.5} e^{-x^2}dx$
Problem is that you cannot do that the usual way because there is no closed form anti-derivative. Yet you can approximate by Simson's rule (that is what the calculator does) under the "calculate integral function".

Of course it is not exactly like I said it, the means and standard deviation affect the function you are integration. But that is the basic idea.

4. Originally Posted by GoneFishing
how do you go about constructing a 95% confidence interval. you guys have been great with the help links so far.
It depends on the purpose/problem. Only yesterday I constructed
some confidence intervals in anger using the BootStrap because
there was no theory that addressed the problem.

RonL

(Google for "statistics bootstrap" if you want to know what it is)

5. Is there a way to calculate the 95% confidence interval for a problem:

say there are 1000 people polled and 600 say they smoke. something like that.

Or.

Say the average marriage lasts 750 days with an sd of 100 (from a poll of 50). how would yyou construct a 90% confidence interval for that?

6. Why double post the same question?. PH is gonna send his goons after you . I just answered this here.

http://www.mathhelpforum.com/math-he...questions.html

The latter is a confidence interval for the mean. The first is on the other post.

90% CI; $\displaystyle {\sigma}=100$; $\displaystyle \overline{x}=750$; n=50

$\displaystyle E=z\frac{\sigma}{\sqrt{n}}=1.645\frac{100}{\sqrt{5 0}}=23.26$

$\displaystyle 726.74<{\mu}<773.26$

You can say with 90% confidence that the mean marriage time is between 726.74 and 773.26 days.