Current understanding of invasive species impacts cannot be ignored: potential publication biases do not invalidate findings

Abstract

Guerin et al. (2017) believe many nonnative species do not cause ecological harm and, therefore, underlying biases towards studying harmful species render meta-analysis unhelpful for designing effective management strategies. Invasion biologists already recognize this bias (Py?ek et al. 2008; Hulme et al. 2013). We argue that meta-analyses are indeed useful for managers for three reasons. First, most meta-analyses explicitly and honestly address bias. Second, for our meta-analysis (Kuebbing and Nuñez 2016), it is unlikely that more even sampling across types of nonnative species would lead to a different conclusion. Finally, and perhaps most importantly, the bias of studying nonnatives with suspected or known impacts focuses research on the exact subset of nonnatives most relevant to managers. It is important to clarify terminology to understand the nature and implications of bias. Ecologists classify nonnative species into three categories: (1) casual nonnatives that do not form self-sustaining populations; (2) naturalized nonnatives that do form self-sustaining populations; (3) invasive nonnatives that form self-sustaining populations and spread beyond their original introduction point (Richardson et al. 2000). There is disagreement whether the definition of invasive should include a negative impact (Young and Larson 2011), but the best available evidence suggests that impacts increase with increasing spread and abundance (Simberlof et al. 2013; Hulme et al. 2013).