Structural Equation Modeling: Theory and Applications in Forest Management
Table 3
Estimated direct effects on the endogenous latent variables in the overstory-understory SR model, along with corresponding standard errors and values (see Figure 4).
Path Model
Direct effect
Standard error
valueb
Late-seral herb cover (LSHERB) =
Overstory cover (TREE)
−0.191
0.075
0.011
Tree density (ln(TPH) × 10)
−0.083
0.168
0.620
Stand age (AGE)
0.156
0.044
0.000
Aspect (ASPECT × 10)
0.423
0.162
0.009
Fine litter (FLITTER)
0.273
0.181
0.132
Coarse litter (CLITTER)
0.205
0.219
0.351
Understory cover (UNDER)
0.176
0.047
0.000
Fine litter (FLITTER) =
Overstory cover (TREE)
0.171
0.032
0.000
Tree size (QMD)
−0.102
0.038
0.007
Tree density (ln(TPH) × 10)
−0.400
0.127
0.002
Aspect (ASPECT × 10)
0.000
0.040
0.999
Understory cover (UNDER)
0.072
0.012
0.000
Coarse litter (CLITTER) =
Tree density (ln(TPH) × 10)
0.169
0.055
0.002
Stand age (AGE)
0.030
0.011
0.007
Understory cover (UNDER) =
Overstory cover (TREE)
−0.508
0.181
0.005
Tree density (ln(TPH) × 10)
−0.989
0.391
0.012
Aspect (ASPECT × 10)
1.543
0.419
0.000
Coarse litter (CLITTER)
−0.284
0.281
0.313
bStatistical tests were based on a t-distribution with 1180 degrees of freedom, with adjustments for the nested sampling structure on standard errors.
