Figure 5.6 shows that property crime decreases consistently as guardianship intensity increases at each stage although the decrease between the monitoring and intervening stages is not significant. The positive results of this test of predictive validity lend further support to the 4-stage GIA model. Property crime drops significantly between the invisible and available stages, suggesting that visible occupancy alone makes quite a difference in terms of the amount of crime experienced at the property level. Property crime is even slightly lower at the monitoring stage and lower still at the stage of intervention. Post-hoc multiple comparisons revealed that the most significant differences occurred between the invisible-available, invisible-monitoring and invisible-intervening stages. The difference between the monitoring-intervening stages, however, was not significant. This may be attributable to the relatively low number of intervening cases that were observed. Altogether, these results suggest that the stages of guardianship depicted by the GIA model are likely to be instrumental in maintaining low property crime levels at residential properties3.