Use of multivariate and nonparametric statistical analyses such as similarity percentages analysis has increased in the past decade within restoration studies. While very useful to compare community composition of restored habitats to reference areas, the ease in which these analyses can be applied, coupled with their power, can result in interpretation errors that could have negative ramifications upon restoration projects. Primarily, similarity measures are often used without stipulating similarity or dissimilarity targets. Despite these drawbacks, the benefits of these methods far outweigh the risks, especially if practitioners take care to specify their community similarity targets a priori and base these targets upon a representative range of reference conditions. However, restoration practitioners should remain focused on fulfilling the objectives of ecological restoration, by ensuring the development of functional habitat that is informed, but not dictated by, statistical analyses. As such, practitioners should take steps to ensure that meaningful univariate trends (e.g. an important species at risk or invasive species) and the overall functionality of the habitat should not be neglected, in favor of these methods of data analysis.

中文翻译：

---

恢复生态学的相似性分析以及如何提高其效用

在过去的十年中，恢复研究中多变量和非参数统计分析（例如相似性百分比分析）的使用有所增加。虽然将恢复栖息地的群落组成与参考区域进行比较非常有用，但这些分析的易用性及其强大的功能可能会导致解释错误，从而对恢复项目产生负面影响。首先，相似性度量通常在没有规定相似性或不相似性目标的情况下使用。尽管有这些缺点，但这些方法的好处远远大于风险，特别是如果从业者小心地预先指定他们的社区相似性目标并将这些目标建立在具有代表性的参考条件范围内。然而，恢复从业人员应继续专注于实现生态恢复的目标，确保功能性栖息地的发展以统计分析为依据，但不受统计分析的支配。因此，从业者应采取措施确保有意义的单变量趋势（例如处于危险中的重要物种或入侵物种）和栖息地的整体功能不应被忽视，以支持这些数据分析方法。