Has the World Changed? My Neighbor Might Know Effects of Social Context on Routine Deviation

Tilmann Betsch, Stefanie Lindow, Christoph Engel, Corina Ulshöfer, Janet KleberSSRN 2011 · DOI

Introduction

In a stable environment, routines are an evolutionary superpower. They allow for rapid, low-effort decision-making based on a history of past success. However, when the underlying "rules of the world" shift—such as a reliable investment losing value or a familiar shortcut being blocked by construction—these same routines can become maladaptive. This paper explores how individuals detect environmental shifts and deviate from established routines, specifically focusing on the role of social context. While it is well-documented that humans learn from others, this study investigates whether observing a peer provides a unique "cue for novelty" that purely data-driven sources lack.

Experimental Framework: The Iowa Gambling Task

To test the mechanics of routine deviation, researchers utilized a modified version of the Iowa Gambling Task across two distinct experiments involving female undergraduates from the University of Erfurt.

Phase 1: Establishing the Routine

Participants were tasked with choosing from four card decks to earn poker chips. One deck was designed to be clearly superior, offering a positive expected value. Participants repeated this selection process until a routine was established, typically after 30 to 40 choices.

Phase 2: The Environmental Shift

Unbeknownst to the participants, the payoff structure was altered during the test phase. The previously optimal "routine" deck became mediocre, while a different deck was promoted to the new optimal choice.

The Social vs. Non-Social Variable

The core of the study lay in the type of model provided to the participants to help them navigate the shift:

  • Social Condition (Experiment 1): Participants worked alongside a human confederate. In specific conditions, they could observe the confederate's choices and outcomes. The confederate was programmed to adapt rapidly to the new optimal deck.
  • Non-Social Condition (Experiment 2): The setup was identical, but the human was replaced by a computer-labeled "card-information box." Crucially, the sequence of data (draws and outcomes) was identical to the social condition, removing "sample size" as a confounding variable.

Key Insights: The "Cue for Novelty"

The study differentiates between a "routine effect" and a "routine bias." While participants were slow to change—a routine effect caused by the need to dilute a long history of success with new failures—they were not acting irrationally. Bayesian updating models confirmed that participants were integrating evidence logically rather than clinging to habits out of pure inertia.

The most significant finding was the role of the human model as a meta-signal. When a human peer deviates from a shared routine, it creates a cognitive conflict in the observer. This conflict suggests that the environment itself has changed, triggering a meta-evaluation of the world's rules. In contrast, the computer-generated data stream failed to provide this "cue for novelty," even though the raw information provided was identical.

Results and Adaptation Speed

The presence of a social model significantly accelerated adaptation. By the tenth trial of the test phase, participants in the social condition were already choosing the new optimal deck more frequently than the old routine.

Conversely, participants in the non-social condition struggled. They were far more likely to cling to their established routines, and their decision-making data was significantly "noisier," suggesting they were not effectively using the computer's information to re-evaluate the situation. Multinomial logit regression indicated that social context allowed participants to give greater weight to new evidence, effectively "unlocking" their ability to ignore irrelevant historical data in favor of current environmental realities.

Limitations and Future Directions

While the study provides strong evidence for the social "cue for novelty," several limitations remain. The experimental environment provided immediate and precise feedback; in "wicked" real-world environments where feedback is delayed or ambiguous, routines may be even harder to break. Furthermore, the routines established were relatively "weak" (30-40 repetitions). Deep-seated habits formed over years may require more than a simple social cue to disrupt. Finally, as the study only involved female undergraduates, future research should explore whether these effects hold across different genders, social hierarchies, and in-group/out-group dynamics.

Conclusion

This research highlights that humans are not just data processors; we are social evaluators. A peer's change in behavior serves as a powerful signal that the "rules of the game" have changed, providing a level of environmental awareness that raw data alone cannot replicate. In an ever-shifting world, our neighbors may indeed be our best indicators of when it is time to leave the old routine behind.

Summary updated

BibTeX
@misc{betsch2011has,
  author = {Tilmann Betsch and Stefanie Lindow and Christoph Engel and Corina Ulshöfer and Janet Kleber},
  title = {{Has the World Changed? My Neighbor Might Know Effects of Social Context on Routine Deviation}},
  year = {2011},
  doi = {10.2139/ssrn.1920296},
  howpublished = {SSRN},
}
Figure