On Replication and the Implications of Findings of Nonlinearity in Crime Research