The article by Svancara and colleagues (“Policy-driven versus Evidence-based Conservation: A Review of Political Targets and Biological Needs,” BioScience 55: 989–995) has numerous failings that make the authors' analysis highly questionable. Space allows me to discuss only four. First, no corrections were made for discrepancies in how minimum area was expressed. The policy targets and most of the conservation assessments state minimum area as percent of total area. However, at least seven of their assessments reported minimum area as percent of available land or of survey area. This error is most apparent in figure 5, where the data points (10, 34.25), (10, 36), (25, 49), and (50, 76) become (10, 20.5), (10, 12.9), (25, 25), and (50, 49.9) after the proper correction. Incidentally, the point at (20, 70) should not be included in the regression. The analytical target for this study (Noss et al. 2002; see Svancara et al. for citation) does not conform to the independent variable in figure 5. After fixing these mistakes, a line with a slope close to 1 fits the data reasonably well.
Second, the data set suffers from at least two forms of pseudoreplication. Seven of the references were review articles that based their minimum areas on some of the other references in the data set. In addition, at least three references were published in two forms (e.g., journal article and government report), and in all three cases both publications were used in the analysis.
Third, the authors overlooked the distinction between preservation and conservation. The policy targets specify the amount of land dedicated for nature preserves. The policy targets do not include other lands where biodiversity conservation is just one of many important goals. In contrast, some conservation assessments, such as those done by The Nature Conservancy, specify the amount of land needed for both preserves and conservation zones where extractive uses are allowed. Likewise, many of the threshold results could be accommodated by a mix of preserves and conservation zones. Hence, comparisons between the policy-driven targets and many of the evidence-based targets are not valid.
Last, and most important, the authors overlooked the significance of what they called “predefined analytical targets.” Figure 5 (whether corrected or not) suggests that these targets may be the most important factor determining minimum area requirements. A majority of the conservation assessments had predefined analytical targets, and many of these targets were based on the subjective judgment of scientists. Noss and colleagues (2002) typify how such targets are formulated. They stated that their targets were those they “felt comfortable with.” In other words, the targets were actually based on the scientists' feelings about acceptable risk. Attitudes toward risk are based more on ethical value judgments than on scientific expertise. (In research studies, subjective predefined analytical targets represent hypothetical ethical value judgments.) Svancara and colleagues rightly exposed the lack of science in the policy-driven targets, but they failed to recognize the value judgments underlying many of the “evidence-based” conservation assessments.