Interactions in statistical models: Three things to know

In ecological studies, the magnitude and direction of interactions among two continuous explanatory variables x1 x2 are commonly evaluated by fitting a statistical model form , where x1x2 is an interaction term that measures departure from additivity effects. Here, we highlight three issues associated with evaluating in models this appear underappreciated literature, but which have important implications for how fit correctly identify interactions. First, scale (additive or multiplicative) on outcome variable y modelled matters. Transformations change analysis alter interpretation terms can hide importance. Second, spurious arise when correlated there unmodeled nonlinear relationships, situation likely to non-experimental data. Third, be such will not capture all interest. We illustrate each these result potentially misleading outcomes using examples linked impacts multiple stressors biodiversity. provide recommendations aimed at identifying effects models.