# Desperately Need Help with Regression Analysis..Please :)

#### jayde

##### New Member
hello

I'm new here (and in doing statistics in general).

I have to do a regression analysis on panel data (10 groups, 9 years - 90 observations).

I'm using STATA and i tried performing regression with fixed effect, which showed that basically my model sux.

But what about heteroscedascity and multicolinearity and all this?? Should I perform any tests or anything? I read the theory and I'm trying to understand, but I don't know how to apply it to my problem.

I don't even know how to start..

#### kenasto

##### New Member
You start with deciding what you want to predict and which variables to use for that.

#### jayde

##### New Member
i have a model, variables and data

data is already log transformed and put into STATA..I figured out how to run a regression analysis, but I'm not sure which one to use (fixed, random effect..) and what to do with the data

#### bugman

##### Super Moderator
How do expect help if you do not provide details about your data?

ten groups - of what?
what is your response variable (observations of what)?
What are you trying to model exactly?

#### jayde

##### New Member
well I tried to change the thread's name but I can't..besides, I was (and still am) panic stricken due to the fact that this statistics is killing me

well I am studying the influence of FDI (foreign direct investment) on GDP

I have panel data for 10 countries through 9 years. The model contains GDP as the dependent variable and six independent variables (labour, investments, fdi, openness and stuff like that). In some variables change value (delta) is required, so these are computed with first log difference, while others are just log transformed. I got those values in Excel and put them into STATA (don't ask about the effort, since at first all of them were identified as string variables, but now they are numeric - flow variables).

Now I pretty much get lost..