# Interpreting regression coefficients for linear regression

#### noetsi

##### Fortran must die
I was going to title this, dumbest question ever.

I am so used to saying for a given slope a value of .07 for a predictor means for a unit increase in X y would go up .07 units. But then, maybe getting old here, I wondered what that meant in a specific case.

My dependent variable is rehab rate which is a percent (so .01 to 1). My predictor is turnover which is also a percent. So the way I interpret this is if X goes up one percent rehab rate goes up .07 percent or 7 one hundreds of a percent. Which is obviously very small.

Is that right? What I am really confused here is what the regression considers a unit for a percent. Is a one unit increase for a percent, one percent?

#### hlsmith

##### Less is more. Stay pure. Stay poor.
It does 1.0 unit, so it depends on how you inputted your data. With the percentage, the thing you would look out for is that your means aren't close to 100%, since then you could could get predictions over a hundred percent, which may or may not be non-sensical given your context