Thread: Statistics Question

I have an assignment in one of my second-year statistics courses.

My professor gave us information on the following test statistic which we haven't gone over:
Likelihood Ratio Test Statistic
Here's the website because I'm going to have trouble re-writing what he wrote:
http://www.math.mcgill.ca/~dstephens...th204-Ass3.pdf

It's on page 1, part (d).

Should it be -2 in front of the sum?

My problem is that I don't see how I would get a positive number out of this since many of the values under the ln function are between 0 and 1 (leading to negative numbers). Do you see what I'm saying? To reject the null hypothesis, I need a positive test statistic and I'm getting a large negative one. For example,
one of the expected values is 388 for an actual value of 266
266*ln(266/388)= -100.417

Thanks for any clarification. There seems to be very little on the Internet about this statistic.

Originally Posted by Five Star

I have an assignment in one of my second-year statistics courses.

My professor gave us information on the following test statistic which we haven't gone over:
Likelihood Ratio Test Statistic
Here's the website because I'm going to have trouble re-writing what he wrote:
http://www.math.mcgill.ca/~dstephens...th204-Ass3.pdf

It's on page 1, part (d).

Should it be -2 in front of the sum?

My problem is that I don't see how I would get a positive number out of this since many of the values under the ln function are between 0 and 1 (leading to negative numbers). Do you see what I'm saying? To reject the null hypothesis, I need a positive test statistic and I'm getting a large negative one. For example,
one of the expected values is 388 for an actual value of 266
266*ln(266/388)= -100.417

Thanks for any clarification. There seems to be very little on the Internet about this statistic.

You're right: it is -2 ln LR that has the chi-square distribution.