# Thread: Why am I having trouble with this t test!

1. ## Why am I having trouble with this t test!

In a study of the production function of the United Kingdom, bricks, pottery and glass

industry for the period 1961 - 81, researchers obtained the following results

Log Q = -8.57 + 0.272t + 0.460 log K + 1.285log H

Se = (1.40) (0.087) (0.137) (0.324) R^2=0.878

verify that the coefficients of t and log K are individually insignificant at the 5% level.

My Attempt:

t, t.s = (0.272-0)/0.087 = 3.1264 K, t.s = (0.460-0)/0.137 = 3.3577

C.V = t0.05/2, 20-1

= 2.093

is |3.1264| & |3.3577| > C.V, and therefore insignificant.

Is my degrees of freedom wrong. Gah, what am I doing wrong!

Thanks for the help

2. ## Re: Why am I having trouble with this t test!

Hey cellae.

Is the SE the standard error of the distribution or the standard error of the mean?

If the SE is the latter then the results you obtained where the co-efficients for t and log K should be significant since the confidence intervals don't contain zero.

3. ## Re: Why am I having trouble with this t test!

Hmm yeah, the answer is meant to be insignificant.

But, I now count 3 people who have found the coefficients to be significant.

Maybe there must be an error in the question ...