Alexandros Gelastopoulos,Pantelis P. Analytis,Gaël Le Mens,Arnout van de Rijt
Abstract:People are influenced by the choices of others, a phenomenon observed across contexts in the social and behavioral sciences. Social influence can lock in an initial popularity advantage of an option over a higher quality alternative. Yet several experiments designed to enable social influence have found that social systems self-correct rather than lock-in. Here we identify a behavioral phenomenon that makes inferior lock-in possible, which we call the 'marginal majority effect': A discontinuous increase in the choice probability of an option as its popularity exceeds that of a competing option. We demonstrate the existence of marginal majority effects in several recent experiments and show that lock-in always occurs when the effect is large enough to offset the quality effect on choice, but rarely otherwise. Our results reconcile conflicting past empirical evidence and connect a behavioral phenomenon to the possibility of social lock-in.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is, under social influence, why in some cases low - quality options will be locked (i.e., the "lock - in" phenomenon), while in other cases the social system can self - correct. Specifically, the authors explore the phenomenon that when the popularity of one option slightly exceeds that of another option, the probability that people choose this option will suddenly increase, which they call the "marginal majority effect". Through theoretical analysis and re - analysis of experimental data, the authors attempt to determine to what extent this effect can lead to the locking of low - quality options and explain why the lock - in phenomenon was observed in some experiments but not in others.
### Background of the Paper and Problem Definition
The paper first reviews several key concepts in social influence research, such as network externalities, information cascades, and herd behavior. These phenomena all indicate that when social influence is strong enough, it may trigger a positive feedback loop, making the popular option more popular, ultimately leading to path dependence and potential lock - in on low - quality results. However, this hypothesis has been supported in some experiments but refuted in others. For example, the classic social - conformity paradigm in psychology shows that social influence is not sufficient to maintain the majority's choice of wrong answers; platform - choice experiments in economics show that people will switch to a better platform, overcoming network effects; experiments in sociology and political science also fail to prove that social influence is sufficient to promote mediocre songs, less - suitable presidential candidates, or bad works of art.
### Research Questions
The core question of the paper is to explore under what conditions low - quality options will be locked. Although the quality difference between competing options may be an important factor in determining possible lock - in, it is still unclear whether the way people respond to social influence is also related to this. To this end, the authors extend a theoretical framework of social influence under binary choice and identify a condition, that is, when the probability of an individual choosing an option significantly increases when its popularity slightly exceeds that of its competitor, if the magnitude of this "marginal majority effect" exceeds the quality difference between the two options, then the lock - in of low - quality options is always possible.
### Main Findings
1. **Marginal Majority Effect**: The authors define the "marginal majority effect", that is, when the popularity of one option slightly exceeds that of another option, the probability that people choose this option will suddenly increase. Theoretically, when the magnitude of this effect exceeds the quality difference between the two options, the lock - in of low - quality options is always possible.
2. **Empirical Evidence**: Through re - analysis of data from three multi - world experiments, the authors provide extensive evidence that the marginal majority effect does exist, and when this effect is large enough, the system does experience lock - in. Conversely, when the marginal majority effect does not exist or is smaller than the quality difference between the two options, the system exhibits self - correcting characteristics.
### Conclusion
Through theoretical analysis and empirical research, this paper reveals the important influence of the marginal majority effect on the lock - in phenomenon of low - quality options. This not only helps to understand how social influence produces different results in different situations, but also provides a new perspective for predicting and intervening in the lock - in phenomenon of social systems.