households (m = 1.38). This difference is observed almost to thesame extent (m = 1.63 vs. m = 1.41) when the structures of the twosamples are weight-adjusted to correct for the difference in struc-ture by main activity types (Table 3). This confirms the existenceof the WA-MPA household vs. REM-MPA household effect. Thesame data also show that 56.0% of REM-MPA households carryon two or more types of fishing activity, whereas only 32.8% ofWA-MPA households have two or more types of fishing activity(Z-test significant difference: p < 0.02).