Do social explanations work?: studying and modeling the effects of social explanations in recommender systems

By Amit Sharma, Dan Cosley. 2013

147 citations recorded by the source corpus.

Recommender systems associated with social networks often use social explanations (e.g. "X, Y and 2 friends like this") to support the recommendations. We present a study of the effects of these social explanations in a music recommendation context. We start with an experiment with 237 users, in which we show explanations with varying levels of social information and analyze their effect on users' decisions. We distinguish between two key decisions: the likelihood of checking out the recommended artist, and the actual rating of the artist based on listening to several songs. We find that while the explanations do have some influence on the likelihood, there is little correlation between the likelihood and actual (listening) rating for the same artist. Based on these insights, we present a generative probabilistic model that explains the interplay between explanations and background information on music preferences, and how that leads to a final likelihood rating for an artist. Acknowledging the impact of explanations, we discuss a general recommendation framework that models external informational elements in the recommendation interface, in addition to inherent preferences of users.

Subfields: Information Systems, Management Science and Operations Research, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Information Systems

In Management Science and Operations Research

In Signal Processing

Citations in the corpus

RippleNet: Propagating User Preferences on the Knowledge Graph for Recommender Systems (2018, background): “The explanations are usually based on community tags [30], social networks [24], aspect [2], and phrase sentiment [44] Since RippleNet explores users’ interests based on the KG, it provides a new point of view of explainability by tracking the paths from a user’s history to an item with high relevan”
Explainable Recommendation: A Survey and New Perspectives (2018, background): “Sharma and Cosley [2013] studied the effects of social explanations in music recommendation context by providing the target user with the number of friends that liked the recommended item (Figure 2.”
Who Also Likes It? Generating the Most Persuasive Social Explanations in Recommender Systems (2014, background): “The nodes are with different types, corresponding to different types of users (e.g. celebrities, experts and friends).”
Explainable Entity-based Recommendations with Knowledge Graphs (2017, unknown): “For example, [10] produced explanations that showed whether a good friend of the user has liked something, where friendship strength was computed from their interactions on Facebook.”
A Reinforcement Learning Framework for Explainable Recommendation (2018, methodology): “Post-hoc methods [3], [5], [6] generate explanations after the items have been recommended.”
Explicit factor models for explainable recommendation based on phrase-level sentiment analysis (2014, background): “Many meticulously designed strategies have been investigated to tackle the problem, ranging from the simple ‘people also viewed’ explanations in e-commerce websites [38] to the more recent social friends or social tags based explanations [31, 39].”
The Limits of Popularity-Based Recommendations, and the Role of Social Ties (2016, background): “Will unknown books become hits and vice-versa? How and how much will the habit of the typical reader change? A large body of work has investigated questions of this kind at the individual, user level [8, 25, 26, 35].”
Make it personal: A social explanation system applied to group recommendations (2017, background): “Or Sharma and Cosley (2013), where the authors reported that explanations combined with technical information about the recommended item, like presenting movie posters where leading actors can be seen, increased the acceptance of recommendations.”
Designing Explanation Interfaces for Transparency and Beyond (2019, background): “, enhancing the system explainability) can achieve different explanatory goals which help users to make a better decision or persuading them to accept the suggestions from a system [14, 16].”
Co-Attentive Multi-Task Learning for Explainable Recommendation (2019, methodology): “Post-hoc methods explain a black-box model after it is trained [Vig et al., 2009; Sharma and Cosley, 2013; Peake and Wang, 2018].”
Distinguishing between Personal Preferences and Social Influence in Online Activity Feeds (2016, background): “It may be even stronger at the individual level; Sharma and Cosley showed that trusted friends’ opinions have more influence on people’s willingness to try unknown items than item popularity [43].”
Social Collaborative Viewpoint Regression with Explainable Recommendations (2017, methodology): “Trusted social relations have been applied to enhance the performance of collaborative filtering [7, 18, 35].”
Explaining Recommendations: Design and Evaluation (2015, background): “In [67] there is a distinction between the likelihood of finding out more about a recommended artist and the actual rating given to the artist after listening to several songs.”
TEM: Tree-enhanced Embedding Model for Explainable Recommendation (2018, background): “Content-based works have considered various side information, ranging from item tags [38, 40], social relationships [31, 37], contextual reviews written by users [13, 15, 28, 31, 48] to knowledge graphs [3, 8, 47].”
Taming the Unpredictability of Cultural Markets with Social Influence (2017, result): “This in line with past experiments in online trial-offer markets for music where after-consumption ratings of songs have low correlation with social influence [22].”
DualDiv: diversifying items and explanation styles in explainable hybrid recommendation (2019, background): “Although most studies re-lated to ERs had focused on a single ES, where a single explanation is displayed for each item [6, 11, 12, 15], it has been reported that users prefer multiple ESs, where many explanations are displayed for each recommended item [8].”
Hidden factors and hidden topics: understanding rating dimensions with review text (2013, background): “to suggest new products they might enjoy [20], to summarize the important points in their reviews [13], or to identify other users who may share similar opinions [27].”