Why Amazon's Ratings Might Mislead You: The Story of Herding Effects

@article{Wang2014WhyAR,
  title={Why Amazon's Ratings Might Mislead You: The Story of Herding Effects},
  author={Ting Wang and Dashun Wang},
  journal={Big data},
  year={2014},
  volume={2 4},
  pages={
          196-204
        }
}
Our society is increasingly relying on digitalized, aggregated opinions of individuals to make decisions (e.g., product recommendation based on collective ratings). One key requirement of harnessing this "wisdom of crowd" is the independency of individuals' opinions; yet, in real settings, collective opinions are rarely simple aggregations of independent minds. Recent experimental studies document that disclosing prior collective ratings distorts individuals' decision making as well as their… 

Figures and Tables from this paper

When Sheep Shop: Measuring Herding Effects in Product Ratings with Natural Experiments
As online shopping becomes ever more prevalent, customers rely increasingly on product rating websites for making purchase decisions. The reliability of online ratings, however, is potentially
Mitigating Sentiment Bias for Recommender Systems
TLDR
This paper reveals an unexplored type of bias, i.e., sentiment bias, and proposes a general de-biasing framework with three strategies manifesting in different regularizers that can be easily plugged into RS models without changing model architectures.
Eliciting and Aggregating Forecasts When Information is Shared
TLDR
An aggregation method is developed, called pivoting, which separates individual forecasts into shared and private information and then recombines these results in the optimal manner and suggests that the pivoting method provides an effective forecast aggregation procedure that can significantly outperform the simple crowd average.
Bias and Debias in Recommender System: A Survey and Future Directions
TLDR
This paper provides a taxonomy to position and organize the existing work on recommendation debiasing, and identifies some open challenges and envision some future directions on this important yet less investigated topic.
THE PHYSICS OF IDEAS: INFERRING THE MECHANICS OF OPINION FORMATION FROM MACROSCOPIC STATISTICAL PATTERNS
TLDR
A simple model that can explain why it takes so long for large groups to reach consensus and an important feature of the model that facilitates agreement with data is that individuals become more stubborn (unwilling to change their opinion) over time is introduced.
Analyzing the impact of missing values and selection bias on fairness
TLDR
An extensive empirical evaluation is conducted on both real-world and synthetic data using various fairness metrics, and it is demonstrated how different missing values generated from different mechanisms and selection bias impact prediction fairness, even when prediction accuracy remains fairly constant.
A Bayesian model of information cascades
TLDR
A detailed Bayesian model of the information gained by agents when observing the choices of other agents and their own private information is presented and an alternative model of choice to that considered in prior work is investigated: weighted random choice.
How to Collect Private Signals in Information Cascade: An Empirical Study
TLDR
This study investigates a Bayesian Inference method, which collects private signals in the information cascade, based on the choice behavior of the subjects, and demonstrates that the Bayesian algorithm works and one can know the correct choice even if the majority’s choice is wrong.
Analysing Recommender Systems Impact on Users' Choices
In this paper we introduce a novel model for simulating the choice making procedure of users under the influence of a Recommender System (RS). Our model leverages the knowledge of users’ preferences
...
1
2
...

References

SHOWING 1-10 OF 24 REFERENCES
Quantifying herding effects in crowd wisdom
TLDR
This paper develops a mechanistic framework to model social influence of prior collective opinions (e.g., online product ratings) on subsequent individual decision making and finds its method successfully captures the dynamics of rating growth, helping to separate social influence bias from inherent values.
How social influence can undermine the wisdom of crowd effect
TLDR
This work demonstrates by experimental evidence that even mild social influence can undermine the wisdom of crowd effect in simple estimation tasks.
Debiasing social wisdom
TLDR
This work uses a natural model of opinion formation to analyze the effect of social influences on an individual's opinion and estimate her propensity to conform and proposes efficient sampling algorithms incorporating these conformity values to arrive at a debiased estimate of the wisdom of a crowd.
Hidden factors and hidden topics: understanding rating dimensions with review text
TLDR
This paper aims to combine latent rating dimensions (such as those of latent-factor recommender systems) with latent review topics ( such as those learned by topic models like LDA), which more accurately predicts product ratings by harnessing the information present in review text.
Beyond the Stars: Improving Rating Predictions using Review Text Content
TLDR
This paper proposes new ad-hoc and regression-based recommendation measures, that both take into account the textual component of user reviews, and shows that using textual information results in better general or personalized review score predictions than those derived from the numerical star ratings given by the users.
Microscopic Social Influence
Social influences, the phenomena that one individual’s actions can induce similar behaviors among his/her friends via their social ties, have been observed prevailingly in socially networked systems.
Collaborative filtering with temporal dynamics
TLDR
Two leading collaborative filtering recommendation approaches are revamp and a more sensitive approach is required, which can make better distinctions between transient effects and long term patterns.
Experimental Study of Inequality and Unpredictability in an Artificial Cultural Market
TLDR
An artificial “music market” was created in which 14,341 participants downloaded previously unknown songs either with or without knowledge of previous participants' choices, and increasing the strength of social influence increased both inequality and unpredictability of success.
Behavioral dynamics and influence in networked coloring and consensus
TLDR
It is found that within a parameterized family of social networks, network structure elicits opposing behavioral effects in the two problems, with increased long-distance connectivity making consensus easier for subjects and coloring harder.
...
1
2
3
...