To find is the confidence interval for the population correlation 'rho', assuming normality, where the sample size is less than 30. So I think we should apply the formula involving tanh and all, ie. The formula in normal case. But in the solution they've used the other formula which is generally used in case of t distribution. Please correct me if I'm wrong.
Q13(b)? that's odd - but I expect they've used the formula from page 25 of the Tables to have an additional rho on top. But it's a bit naughty...
Sir is it because of the next part of the question which requires a comment on significance of rho? If we assume that so, then is it justified to go with t-tests, given the bivariate normality?
