I am working on a problem in which I need to calculate a 97% confidence interval.
This would make my alpha=.03
alpha/2=.015
However, .015 is not on my t-distribution chart! I don't know what to do!! Anyone know??
2007-12-13 12:56:23 · 4 answers · asked by Anonymous in Science & Mathematics ➔ Mathematics
the df is large
n=4000
2007-12-13 13:15:21 · update #1
There are two answers, one requires calculus, the other does not.
The method that does not require calculus is to interpolate the table. It will be close, but slightly biased.
Hopefully you know calculus because it would be impossible, due to lack of symbols, to post the answer on Yahoo Answers.
You set the cumulative density function equal to .985 for Student's t, given the sample size it should be very close to the z distribution, which is far easier to calculate and find the appropriate value plus or minus the expected value. There should be trivial bias presuming none of the assumptions of ordinary least squares were violated. You can then either use the symmetric distance or redo the problem with the cdf equal to .015.
2007-12-13 15:39:49 · answer #1 · answered by OPM 7 · 0⤊ 0⤋
It means that according to your 6-item sample, the true underlying mean of the distribution that generated the data has a 95% probability of falling between 86.14 and 97.86. I'm not sure what that means in terms of interpreting your outcome, since you haven't described the hypothesis you are testing. Statistical data aren't really useful without a context within which to apply them.
2016-05-23 11:22:47 · answer #2 · answered by Anonymous · 0⤊ 0⤋
Don't look on a chart... try using either the R software package (just google R statistics to find it) or use the site in the source. You didn't mention the degrees of freedom or else I'd a done it myself. Hope this helps!
2007-12-13 13:03:51 · answer #3 · answered by highschoolmathpreparation 3 · 0⤊ 0⤋
maybe you read it wrong. it's probably 97.5%
2007-12-13 13:01:45 · answer #4 · answered by Anonymous · 0⤊ 0⤋