Linear Regression while comparing results?

[Modal]

Hi all

I am working on a research project for a basic stats course. I'm in the planning phase and could use some pointers to make sure I'm going in the right direction.

Question: How do I do a linear regression while assuring that the dependent variable is unique for each set of distinct input?

I am trying to fit a model for Y= rank of a song today, among the singer's discography.

I know I have independent variables such as
A=number of downloads (accumulative)
B= number of 5 star ratings
C = number of 4 star ratings
...

and I plan out to start my model with the Y~(b0)A+(b1)B+(b2)C+e

but I assume the eventual outcome would be Y(songA) = 3, Y(songB) = 4, Y(songC) =1,...
where songX will contain the array of information that I need for the independent variable.

But wouldn't the rank also have dependency with each other? For example, the fact that songC had rank 1 was because it comparative did better than songA.

Sorry if I sound confusing. I am really confused and am a little lost at the planning stage.

Thanks.

 

[noetsi]

I don't understand what you mean by rank in the above comment. If your independent variables are related to each other (which I am not sure you are saying) then you will have multicolinearity which is a problem. If independent variables are similar conceptually, then they may influence the DV in the same fashion. This does not create problems for the regression (other than multicolinearity problems).

But I am not really sure what you are asking.

 

[md azzainuri]

try to use logistic regression,