# 98% confidence interval. For this question from a large amount of data I have calculated that the mean is 44.22, the sample size is 100 and the standard deviation is 22.0773.

98% confidence interval
For this question from a large amount of data I have calculated that the mean is 44.22, the sample size is 100 and the standard deviation is 22.0773.
From this I am asked to , make the 98% confidence intervals for the (1) true mean µ of the module mark (2) true variance of the module mark
And for each: (a) Determine what quantity to look at, and which distribution table to use, justifying your choice. (b) Determine the number of degrees of freedom, justifying your answer. (c) Calculate the actual intervals.
So far for 1, I have used the z table to look for 99% as I need 1% to the right of 2.33 and 1% to the left of −2.33, so 98% is between $±2.33$. Giving me
$\overline{x}±2.33\frac{\sigma }{\sqrt{n}}$
Which provides me with a 39.08 to 49.36 confidence interval, is this correct? And how would I determine degrees of freedom and go about answering part 2?
You can still ask an expert for help

• Questions are typically answered in as fast as 30 minutes

Solve your problem for the price of one coffee

• Math expert for every subject
• Pay only if we can solve it

Cassie Moody
Step 1
Confidence intereval for population mean. A 98% confidence interval for $\mu$ is of the form $\overline{X}±{t}^{\ast }s/\sqrt{n},$, where t∗ cuts off 1% from the upper tail of Student's t distribution with $df=n-1.$. So you are almost correct for that part. Here ${t}^{\ast }=2.365.$. I get the CI (39.00,49.44) from the following brief R session.
$a=44.22;s=22.0773;n=100\phantom{\rule{0ex}{0ex}}t.c=qt\left(.99,n-1\right);pm=c\left(-1,1\right)\phantom{\rule{0ex}{0ex}}a+pm\ast t.c\ast s/sqrt\left(n\right)\phantom{\rule{0ex}{0ex}}38.9995949.44041$
Confidence interval for population variance. To get a 98% CI for ${\sigma }^{2},$, one can use the fact that $\left(n-1\right){s}^{2}/{\sigma }^{2}\sim Chisq\left(df=n-1\right).$. So if L and U cut 1% from the lower and upper tails of Chisq(99), we have
$P\left(L\le \frac{\left(n-1\right){s}^{2}}{{\sigma }^{2}}\le U\right)=.98$
from which we get
$\left(\left(n-1\right){s}^{2}/U,\phantom{\rule{thickmathspace}{0ex}}\left(n-1\right){s}^{2}/L\right)$
as a 98% CI for ${\sigma }^{2}.$. In particular, for your data the CI is (358.4,697.0). (You could get L and U from a suitable printed table of the chi-squared distribution.) Notice that ${s}^{2}=487.4$ is contained in this interval, but is not at the very center of the interval---because the chi-squared distribution is not symmetrical. Take the square root of both endpoints if you want a CI for $\sigma .$.
Step 2
$\left(n-1\right)\ast {s}^{2}/qchisq\left(c\left(.99,.01\right),n-1\right)\phantom{\rule{0ex}{0ex}}358.3833697.0011$