When the number of events per variable is low, type II errors are more likely than type I, although there is decreased concern for plausible and highly significant associations. 40 Exact logistic regression was used in univariable analysis to help compensate for the low number of events, but it was not possible to do this for the multivariable model, as it was not possible to evaluate the interaction term suggested by model validation using exact logistic regression.
← all excerpts
Ultrasonographic patterns, clinical findings, and prognostic variables in dogs from Asia with gallbladder mucocele.
1
—
—