Does fitting of a regression line as such contribute to mass significance?

Is the fitting of a regression line (least squares, for a given number of predictors) as such to be regarded a statistical test that can contribute to mass significance? -
I am searching for the best linear regression model and I have 10 potential predictors. I calculated the Schwarz Bayesian Criterion value for every possible combination of predictors (SAS: proc reg / selection=adjrsq best=1024 sbc). Do I have to consider these 1024 models when I adjust my results for false positives? How could that be done?

- I rewrote the question and started a new thread with it on 03-13-2012 in the forum "Statistical Research \ Statistical Research" under the heading "Mass significance - information criterion".
Thanks for your consideration.
Last edited: