In my last post I showed how one can easily summarize the outcome of a logistic regression. Here I want to show how this really depends on the data-points that are used to estimate the model. Taking a cue from the evolution of a correlation I have plotted the estimated Odds Ratios (ORs) depending on