Skip to main content
. 2018 Jun 26;13(6):e0199526. doi: 10.1371/journal.pone.0199526

Table 5. Average effect of the CFP on AGC and sensitivity analysis by forest category.

(1) Forest Category (2) No. of plots in CFP/ non-CFP (3) Average SMD of Observed confounders (before/after match) ATTC (Comparisons of Median) Hidden bias
(4) Point estimate (5) Lower confidence limit-95% (6) Upper confidence limit-95% (7) p-value (8) Critical level of bias (Γ)
Overall forest 325/70 0.40/0.11 -15.11 -26.35 -3.49 0.012 1
Lower altitude 170/60 0.37/0.21 11.21 -7.42 31.02 0.243
Higher altitude 155/28 0.24/0.08 -22.81 -37.41 -9.39 0.001 1
Lower slope 89/28 0.39/0.11 25.51 0.98 55.14 0.041 1.10
Higher slope 236/56 0.26/0.17 -17.72 -30.93 -4.22 0.010 1
Terai 120/43 0.36/0.13 5.87 -15.88 32.80 0.585
Hill 205/41 0.16/0.10 9.76 -1.48 22.04 0.089
Open canopy 149/41 0.42/0.09 25.84 12.22 41.36 0.000 1.66
Closed canopy 176/53 0.39/0.13 -2.93 -18.06 12.11 0.694

Columns 2 and 3 provide the number of CFP/non-CFP forest plots and average SMD of confounders before and after matches. Columns 4, 5, 6 and 7 present the ATTC, lower and upper confidence levels of ATTC and p-values. Column 8 provides information about the sensitivities of estimated results to unobserved confounders and the p-values.