Separation occurs in binomial response models when a combination of the predictor variables perfectly predicts the response. In such a case the estimates of the coefficients for these variables diverge to (-)infinity, and the numerical algorithms typically fail. To anticipate such a problem, the fitting functions in spaMM
check for separation by default, except for large data sets where this may take time, in which case a message notifies this to the user. The threshold size of the data is set by spaMM.options(separation_max=<.>)