Table 2.

Bias associated with point extraction from discrete rasters. Results of one-sample t-tests addressing the significance of bias associated with raster cell extraction. Point extraction is not recommended since it yielded significantly biased effect size estimates throughout most scenarios tested.

Misclassification rate	t-statistic	p-value	Mean bias	Lower 95% CI	Upper 95% CI
≤ 11%	4.69	< 0.0001	0.019	0.011	0.026
11%–20%	4.53	< 0.0001	0.018	0.010	0.026
≥ 20%	5.38	< 0.0001	0.018	0.012	0.025