Fig. 4From: Investigating sources of non-response bias in a population-based seroprevalence study of vaccine-preventable diseases in the NetherlandsOrdered variable importance for predicting a QO from a FP in P3Back to article page