Table 6. Partial r-square, significance levels, and average squared canonical correlation of stepwise discriminant analysis of loco characteristics. Successful and unsuccessful fisher groups were split using median values. In the discriminant analysis, independent variables entered and remained in the model if p-value ≤ 0.15.


Step Variable Partial
R-Square
Pr>F Average squared canonical correlation Mean Values Successful Mean Values Unsuccessful
Loco
1 WEIGH (g) 0.43 <0.001 0.44 126 104
2 BORE (%) 0.17 <0.001 0.53 10 19
3 PHOT (mg/m3) 0.15 <0.001 0.65 2 2
4 ROADV (km2) 0.10 <0.001 0.68 96 49
5 BEN (#) 0.04 0.03 0.69 4 6
6 YEAR (y) 0.03 0.07 0.70 23 23
 
Congrio
1 ROADV (km2) 0.23 <0.001 0.24 47 92
2 FAC (#) 0.05 0.05 0.28 1 2
3 AREAMA (ha) 0.04 0.10 0.30 95 115
4 BOAT (#) 0.06 0.04 0.34 20 16
5 OAC (#) 0.07 0.02 0.39 1 2
6 BORE (%) 0.08 0.02 0.43 22 23
7 EDU (y) 0.05 0.06 0.47 9 8
 
Management area
1 BORE (%) 0.16 <0.001 0.16 13 25
2 AREAMA (ha) 0.07 <0.001 0.21 144 112
3 DISTHQ (km2) 0.05 <0.001 0.25 1887 1418
4 FAC (#) 0.04 <0.001 0.28 2 1
5 BOAT (#) 0.02 0.03 0.29 16 19
6 BEN (#) 0.04 0.01 0.32 6 6