Appendix I. Fit statistics for structural equation models used to study the relationship between floral neighborhood density, pollinator visitation, pollination, and reproduction of Alstroemeria aurea in grazed sites.
Full model
Reduced model
difference
Site
df
P
NFI
df
P
NFI
df
P
Arroyo Goye
5.29
5
0.3810
0.94
4.96
3
0.1748
0.96
0.33
2
0.8479
Cerro López
11.26
5
0.0465
0.91
2.68
2
0.2624
0.98
8.58
3
0.0354
Mascardi
32.04
5
<.0001
0.39
3.87
2
0.1445
0.75
28.17
3
<.0001
Quetrihué
8.72
5
0.1210
0.86
4.76
4
0.3131
0.91
3.96
1
0.0466
Notes: Fit statistics are maximum likelihood
statistic and associated degrees of freedom (df) and probability value (P), and Normed Fit Index (NFI). Full model corresponds to Fig. 5; reduced models (Fig. 14) were obtained by constraining nonsignificant paths to zero. Analysis was conducted using the CALIS procedure in SAS, using observed covariance matrices given in Appendix K.