How to interpret results of logistic regression? I am struggling with interpreting the output of logistic regression correctly. The dependent variable is leaving the university (=1) and I have 7 significant independent variables. The coefficient for the independent variable "age" is -0,057. Is my interpretation correct that:
exp(-0,057)=0,945 1-0,945=0,055
-> The increase of the age by one year reduces the possibility of leaving the university by 5,5%?
 A: No, that is not correct.  If that were true, then eventually I would become old enough that the probability I leave university would be less than 0.  The correct interpretation is that the odds of the outcome are reduced by 5.5%.
If $p(x)$ is the risk of the outcome conditioned on covariates $x$, then logistic regression makes the assumption that
$$ \log\Big( \dfrac{p}{1-p} \Big) = x^T\beta $$
The LHS of this equation is called the "log odds" because the argument to $\log$ is the odds. Hence a unit increase in covariate $x_j$ means the log odds will change by $\beta_j$ or alternatively that the odds will change by a factor of $\exp(\beta_j)$.
In order to know what the change in probability is, you need to transform the odds onto the probability scale by solving for $p$, yielding
$$ p = \dfrac{1}{1 + \exp(-x^T  \beta)}$$
and taking the difference between estimated risks.  Note, a change of $\exp(\beta_j)$ in the odds will result in bigger/smaller changes depending on what the baseline risk is.  A doubling of the odds results in a smaller change in probability when the baseline risk is small/large as compared to when the baseline risk is near 0.5.  We can plot this very easily to see the change in risk.

