Table 2. The pair-wise mean difference in resistance between scenarios.1 .
| Comparison2 | Position | Mean Difference3 | 95% CI4 | Effect Size5 (d) |
| No buffer vs. Real | Inner | 8.84×10−3 | 8.28×10−3–9.40×10−3 | 1.50 |
| Outer | 6.02×10−2 | 4.17×10−2–7.91×10−2 | 0.90 | |
| Randomized vs. Real | Inner | −2.24×10−3 | −7.17×10−3–2.69×10−3 | 0.13 |
| Outer | −5.40×10−5 | −6.24×10−3–6.13×10−3 | 0.002 | |
| No buffer vs. Low | Inner | 7.18×10−3 | 4.53×10−3–9.83×10−3 | 0.75 |
| Outer | 5.28×10−2 | 7.01×10−2–3.55×10−2 | 0.84 | |
| Randomized vs. Low | Inner | −3.90×10−3 | −7.80×10−3– −3.93×10−6 | 0.28 |
| Outer | −7.46×10−3 | −1.31×10−2– −1.84×10−3 | 0.37 | |
| No buffer vs. High | Inner | 8.97×10−3 | 6.21×10−3–1.17×10−2 | 0.90 |
| Outer | 6.45×10−2 | 4.45×10−2–8.44×10−2 | 0.90 | |
| Randomized vs. High | Inner | −2.11×10−3 | −6.05×10−3–1.84×10−3 | 0.15 |
| Outer | 4.23×10−3 | −8.50×10−4–9.31×10−3 | 0.23 |
1
Across 10 random sites in each of 5 plots; n = 50.
2
The randomized buffer is composed of habitat data randomly assigned to pixels in proportion to the map itself, the real buffer is what is truly on the map, the low buffer is biased toward low-quality habitat data, and the high buffer is biased toward high quality data.
3
The difference is calculated as the first scenario minus the second scenario (e.g., for the first row, no buffer minus real buffer).
4
95% confidence interval of the mean pair-wise difference.
5
Cohen's effect size (d) for paired comparisons [21]; d = 0.2 is a small effect, d = 0.5 is a medium effect, d = 0.8 is a large effect.