Desperate plea for help on panel data using EViews7

[stats321]

forum.xls

(234 KiB) Downloaded 232 times

forum.xls

(234 KiB) Downloaded 232 times

Hello everyone!

I need to make a model for my dissertation, but the only contact i've had with econometrics was making a simple time series model 2 years ago. I have also never used EViews before.

I am trying to find out which infrastructure indicator change has had the most significant effect on GDP using a cross section of 57 countries between 1990-2009.

I am testing for random and fixed effects, but whenever I try to change the option for time, it says I have a near singular matrix or something about positive and non-negative numbers comes up.

Also should I be doing the normal diagnostic tests that I would for time series data?

I really don't have a clue what I am doing to be honest, so any help/opinions would be greatly appreciated as I am sure I will get stuck later on also. I have attached my dataset and even a step by step guide would be so useful! I am just not very confident that what I am doing is right.

Many many thanks,

Nikki