A regression equation with 4 independent variables is estimated using 20 data points. The R2 is 0.46. An analyst is testing to see whether all of the coefficients are equal to zero. The p-value for the test is: A) between 0.025 and 0.05. B) lower than 0.025. C) between 0.05 and 0.10. D) greater than 0.10.
B.
can you share your calc?
In this case t = r*sqrt(n-2) / sqrt (1-r^2) = sqrt(0.46)*sqrt(20-2)/sqrt(1-0.46) = 3.91578 In the normal distribution, 97.5% of the observations will be be less than or equal to 0 + 2S. 3.91578 >> 2 --> this observation should be, confidently, below 2.5%.
I think both A and B work… Here is my explanation: Since R2 is 0.46 it means that atleast one of the coefficients being greater than 0 has a good chance… which would mean that the test should say that “at least one of the coefficient(s) is/are statistically significantly greater than 0” i.e. the null would be rejected and that means p-value will be lower than a given significance level… which for me in this case would be either somewhere between 0.025 and 0.05 or lower than 0.025. Anybody else?
so it means we have to use the F test instead of t test in this scenario? forgot 80% of what I have read so far…
F-test is need when there are more than independant variable
yes the answer is A. Thank you all!
mvwt, sorry did n’t get your point…is there an omission in ya answer
What don’t you get?