Log or not and a couple of other Qs.

[topcat]

Hello,

I am looking at the change in a company's share price on the day that it first goes public.

The variables that I am looking at include age, assets, net_income and IPO size. I also am using a dummy for a specific industry to see if that industry has higher returns.

I have just a few (hopefully) easy questions:

For the return I am using:
((price close/price open)-1)*100

1) Is this the best way to use return in my regression?
2) Should I use the log of assets/age/ipo size? (net income is often negative so I cannot take logs of that)
3) My resulting R^2 is 0.03 but I have read and been told that it does not matter because I am using a large sample (500+). The dummy for the specific industry is the only variable that is actually significant.

Also another suggested reason for the R^2 being so low was that there may be an underlying trend given that my sample is on firms between 2005-2012?

Thanks for the help.