Thread: Statistics question interpretation

Hello,

I'm going through some past exam papers and would like it if someone could offer some help with some suggested solutions/ suggested formulas to calculate etc, I find it a bit difficult in just interpreting the questions and don't fully understand what kind of formulas they are asking for....

I need to test the hypothesis that B(x) = 1 and then interpret this result

What test should I be using... the F statistic?

and then they add a new variable into a new regression output and ask the following;

Comment on the new variable, analyse its magnitude and stat sig @ 5%

Here (I have two regression outputs) one with the old variables and now the second with the new variable added - do I just look at the new variables P-Value... (the P-value is 0.08; so its not statistically significant) and regarding the magnitude, what am I supposed to say here, that if we used the 0.1 statistical significance level then it would be accepted...

Then the next question it asks;

Rsquared of the two variables = 0.9677, calculate the VIF of variable x and variable y, explain your results...

here do we just compute [1/1-0.9677) = 0.323... to give the VIF value which suggest no multicolinearity right... as its much less than < 5...

Next question...

Looking at the regression output and your answer to the VIF - would you set an alternative specification for the model...
write this specification and justify yourself...

So here... I get abit more lost...



Any advice in the right direction would be great,

thanks

Originally Posted by entrepreneurforum.co.uk

Hello,

I'm going through some past exam papers and would like it if someone could offer some help with some suggested solutions/ suggested formulas to calculate etc, I find it a bit difficult in just interpreting the questions and don't fully understand what kind of formulas they are asking for....

I need to test the hypothesis that B(x) = 1 and then interpret this result

Surely, you have information than this? What does B(x) mean? You can't just pull a random variable out of the air!

What test should I be using... the F statistic?

and then they add a new variable into a new regression output and ask the following;

Comment on the new variable, analyse its magnitude and stat sig @ 5%

Here (I have two regression outputs) one with the old variables and now the second with the new variable added - do I just look at the new variables P-Value... (the P-value is 0.08; so its not statistically significant) and regarding the magnitude, what am I supposed to say here, that if we used the 0.1 statistical significance level then it would be accepted...

Then the next question it asks;

Rsquared of the two variables = 0.9677, calculate the VIF of variable x and variable y, explain your results...

here do we just compute [1/1-0.9677) = 0.323... to give the VIF value which suggest no multicolinearity right... as its much less than < 5...

Next question...

Looking at the regression output and your answer to the VIF - would you set an alternative specification for the model...
write this specification and justify yourself...

So here... I get abit more lost...



Any advice in the right direction would be great,

thanks

With regard to B(x) = 1, I should of made myself clearer, I mean Beta(x) where x is just a number i) 1,2...i

I can send the exam in a PM if you want...

I am not looking for the answers just some guidance on if you were asked similar questions what sort of calculations would you do...