Does a previous outcome shift later risk preference in the same direction across species, and what models explain outcome-dependent choice?

Does a past outcome reliably push you one way?

The intuitive story is that winning makes people bolder and losing makes them cautious. The human evidence does not support a single fixed direction. In experiments on preference malleability, prior outcomes shifted risk preferences in opposite directions within the same participants, which the authors describe as evidence that preferences are extremely malleable rather than anchored to one rule [4]. That matters for the question: if the same person can move both ways depending on circumstances, a universal "same direction across species" rule is unlikely to hold.

A field study points the same way. Following a postelection crisis in Kenya, young people became more risk averse—a shift toward lower-variance choices—rather than showing a uniform change in risk appetite [5]. The scope here is important: this is one population responding to one violent shock, so it tells us that a salient negative experience can raise caution, not that every prior outcome does so everywhere.

Why the same outcome can lead to different choices

One explanation is that what people attend to, and how an outcome is presented, changes the choice more than the outcome itself. Experiments on framing and salience found that selectively attending to the positive outcome of a risky decision led to more risk-taking [1]. In plain terms, the same gamble can feel bolder or safer depending on which part of it is brought to the front of attention.

This helps explain why prior outcomes do not map neatly onto one direction. If a past result changes which features of the next decision seem salient—the chance of gain versus the chance of loss—then the resulting shift depends on that attentional pull, not on a simple win/lose tally. The framing work is experimental and human, so it supports a mechanism for how outcomes shape choice, not a claim about species-wide patterns [1].

What animal studies and models add—and what they don't

The rodent work gets closest to the cross-species question. In rodents choosing under uncertainty, prior outcomes influenced subsequent choices, and the study tracked outcome-dependent strategy shifts alongside cortical processing; psilocybin modulated these effects asymmetrically [2]. That is evidence that outcome history shapes later choice in a non-human mammal, but the study does not compare risk-preference shifts across species, so it cannot establish that the direction matches humans or any other species [2].

Models of outcome-dependent choice in these studies are about how prior results update later decisions, not about a shared cross-species rule. The ecological and evolutionary review on diversity in fluctuating environments deals with how temporal variation maintains genetic and species diversity, and does not examine how prior outcomes shift risk preference [3]. It is relevant background for why organisms might be sensitive to fluctuating conditions, but it does not answer the direction question.

The practical answer for a curious reader

If you are asking whether a win or a loss reliably makes a person—or an animal—more or less risk-taking in a predictable direction, the honest answer is: not reliably, and not yet established across species. The human experiments show shifts in opposite directions within the same people [4], a real-world shock produced increased caution [5], and attentional framing can push choices toward more risk-taking [1]. The rodent study shows outcome history matters for later choice but does not test cross-species direction [2].

The most defensible summary is that prior outcomes change later risk choice, but the direction depends on context—what is salient, how the outcome is framed, and what kind of shock or experience occurred—rather than following one universal rule. Claims of a same-direction cross-species pattern would need direct comparisons that these studies do not provide.

Sources used in this answer

Related research questions