Using logit with linear regression analysis

I'm trying to figure out the best way to do regression analysis for a dependent variable that is a ratio. The independent variables will include both categorical and numerical variables. I have been told that what I should do is convert the ratio to a continuous number with logit, then do linear regression. I would like to confirm that this is correct, and that I would not have to do logistic regression because the independent variables will include categorical variables. I have read that the correlation coefficient is not an accurate measure of goodness of fit of logistic regression. I'm really interested in being able to use the coefficient of determination and correlation coefficient with my research because it is what I am most familiar with. Is there any recommended resources or advice for this type of analysis?