regression when the dependent variable is a proportion or percent?

[proctice]

If my outcome was dichotomous, I would use logistic regression.
It my outcome was continuous, I would use linear regression.
If my outcome was a count, I would use poisson regression.
What do I use when my outcome is a percent or proportion and not centered between 20% and 80%.

Thanks.

[Dason]

You could still try to fit to a logistic function if you think that it seems appropriate. It's still possible it's just that it would be a nonlinear regression at that point. You'd have to just use an iterative least squares procedure or something similar to get estimates for your coefficients.