Triumphs and trials of the risk paradigm

A paper by Caraco, Martindale and Whittam (1980, Animal Behaviour, 28, 820-830) transformed foraging behaviour research by the fundamental realization that actions in nature only rarely have precisely predictable outcomes. They showed that the consequences of risk, understood as the unpredictability and variance of consequences, must play a central role in functional and mechanistic analysis of animal decision processes. Here we examine the article's theoretical and experimental contribution with the benefit of over three decades of hindsight, review developments in the intervening time and discuss the future. Our review confirms previous ones showing that although the earliest results supported the predictions of the original models, a full analysis of the literature does not. In spite of a likely positive publication bias, the majority of studies fail to support a shift from risk proneness when in negative energetic budget to risk aversion when in positive budget, as embodied in the budget rule, and the proportion of failures increases with time. Furthermore, re-examining the earlier experiments we conclude that support should not have been expected, mostly because they implausibly assumed that all probabilities were known to the subjects. We argue that such assumptions should be abandoned, being replaced by realistic, evidence-based models of learning and information processing, such as associative learning and Weber's Law. Our overall message is that models are research tools that can be fecund even if they are transformed and even refuted with the passage of time and the associated accumulation of knowledge. © 2013 The Association for the Study of Animal Behaviour.