Abstract
Probability is a fundamental metric for assessing risk and uncertainty. However, probability estimates are rarely static; they carry "trends" that function like physical forces on the human mind. This paper explores the trend effect of probability estimation, categorizing it into two forms: dynamic changes over time and one-sided linguistic phrasing (e.g., "at least 40%"). By applying the theory of psychological momentum (PM), we propose an integrated model to explain how these trends trigger mental simulations and influence decision-making. This framework reveals how probability trajectories can lead to non-linear biases in fields ranging from climate change to medical risk.
Introduction
In traditional decision science, probability is often treated as a static "snapshot"—a fixed percentage representing the likelihood of an event. Yet, in real-world scenarios, information is constantly updated. A 40% chance of rain feels fundamentally different if the previous forecast was 30% (a rising trend) versus 50% (a falling trend). This paper addresses the critical gap in how these trajectories influence human choice, arguing that the path taken to reach a probability is often as significant as the probability itself.
Manifestations of the Trend Effect
The authors identify two primary ways the trend effect manifests in cognitive processing:
- Dynamic Probability Change: This refers to sequential updates over time. A rising trend creates a "sense of gain" or increasing certainty, while a falling trend induces a "sense of loss" or growing doubt.
- One-Sided Probability Estimation: Linguistic hedges like "more than X%" or "less than Y%" are not merely vague ranges. They "leak" information about the past. Stating a probability is "at least 40%" implies the value was lower previously and is currently trending upward, creating a directional vector in the observer's mind.
The Mechanism: Psychological Momentum
To explain why these trends have such a powerful grip on decision-making, the authors propose a model based on Psychological Momentum (PM). In this framework, probability estimates are viewed as vectors with direction and magnitude. The PM engine consists of two core components:
- Perceived "Mass": The perceived importance, severity, or resistance to change of the event.
- Perceived "Velocity": The rate at which the probability is changing.
When a trend is observed, it acts as a catalyst for mental simulation. If an individual perceives a rising trend, their brain "projects" that movement into the future, leading them to believe the event is even more likely than the current data point suggests. This predictive processing narrows the gap between current perception and future expectation.
As illustrated in the integrated model (Figure 1), the process begins with external stimuli (the trend) which triggers mental simulation. This simulation acts as a catalyst, forming a psychological momentum experience characterized by perceived mass and velocity. This experience then dictates decision behavior through perceptual inference and predictive processing.
Empirical Evidence and Impact
The trend effect has been observed across diverse domains:
- Climate Change: Studies show that when expert predictions for global warming increase (e.g., from 60% to 70%), the vast majority of people expect the trend to continue upward. Conversely, a drop from 80% to 70% leads to expectations of further decline, despite the final probability being identical.
- Medical Risk: Rising mortality rates for a disease are perceived as significantly more severe than falling rates, even when the current mortality rate is the same in both scenarios. This suggests that the "velocity" of risk change heavily weights the perceived threat.
- Sports Strategy: Coaches' perceptions of momentum—such as a team's rising probability of winning—directly influence strategic choices, such as whether to maintain a current plan or switch tactics.
Conclusion and Future Directions
Understanding the trend effect allows for a more nuanced view of human irrationality. Future research should focus on how individuals process conflicting trends from multiple sources and how long-term trends interact with sudden, high-certainty data "snapshots." Furthermore, in the realm of risk communication, framing information (such as vaccine side effects) as a "falling trend" may prove more effective in reducing hesitancy than simply providing static percentages. By accounting for the psychological momentum of probability, we can better predict and nudge decision-making in an uncertain world.
Summary updated