Understanding rating behavior based on moral foundations: The case of Yelp reviews

Moral inclinations expressed in user-generated content such as online reviews can provide useful insight to understand and predict people's rating behavior. In this work, we extracted a corpus of over 7,000 online reviews on Yelp that express moral concerns, and associated them to five moral factors defined in Moral Foundations Theory using the Doc2Vec natural language processing technique. We compared the rating distributions between the regular reviewers and the moral-concerned reviewers, and found that their rating patterns significantly differ from each other. Our findings also indicate that people with moral concerns tend to rate lower if a moral foundation is violated. Moreover, among the five moral factors, purity is the most distinctive moral foundation.

[1]  J. Haidt,et al.  Intuitive ethics: how innately prepared intuitions generate culturally variable virtues , 2004, Daedalus.

[2]  Jeffrey Dean,et al.  Efficient Estimation of Word Representations in Vector Space , 2013, ICLR.

[3]  Victor R. Prybutok,et al.  Latent Semantic Analysis: five methodological recommendations , 2012, Eur. J. Inf. Syst..

[4]  Jeffrey Pennington,et al.  GloVe: Global Vectors for Word Representation , 2014, EMNLP.

[5]  Jesse Graham,et al.  Title: into the Wild: Big Data Analytics in Moral Psychology , 2016 .

[6]  Eyal Sagi,et al.  Moral Rhetoric in Twitter: A Case Study of the U.S. Federal Shutdown of 2013 , 2014, CogSci.

[7]  Arthur C. Graesser,et al.  The Right Threshold Value: What Is the Right Threshold of Cosine Measure When Using Latent Semantic Analysis for Evaluating Student Answers? , 2003, Int. J. Artif. Intell. Tools.

[8]  Kate M. Johnson,et al.  Morality Between the Lines : Detecting Moral Sentiment In Text , 2016 .

[9]  D. Murphey,et al.  The Righteous Mind: Why Good People Are Divided by Politics and Religion , 2013 .

[10]  T. Landauer,et al.  Indexing by Latent Semantic Analysis , 1990 .

[11]  Omer Levy,et al.  word2vec Explained: deriving Mikolov et al.'s negative-sampling word-embedding method , 2014, ArXiv.

[12]  Jeffrey Dean,et al.  Distributed Representations of Words and Phrases and their Compositionality , 2013, NIPS.

[13]  P. Rozin,et al.  The CAD triad hypothesis: a mapping between three moral emotions (contempt, anger, disgust) and three moral codes (community, autonomy, divinity). , 1999, Journal of personality and social psychology.

[14]  Kate M. Johnson,et al.  Purity homophily in social networks. , 2016, Journal of experimental psychology. General.

[15]  Scott Counts,et al.  Modeling Ideology and Predicting Policy Change with Social Media: Case of Same-Sex Marriage , 2015, CHI.

[16]  Quoc V. Le,et al.  Distributed Representations of Sentences and Documents , 2014, ICML.

[17]  Jonathan Haidt,et al.  Morality , 2008, Perspectives on psychological science : a journal of the Association for Psychological Science.

[18]  Eyal Sagi,et al.  Measuring Moral Rhetoric in Text , 2014 .

[19]  Brian A. Nosek,et al.  Liberals and conservatives rely on different sets of moral foundations. , 2009, Journal of personality and social psychology.

[20]  이주연,et al.  Latent Dirichlet Allocation (LDA) 모델 기반의 인공지능(A.I.) 기술 관련 연구 활동 및 동향 분석 , 2018 .

[21]  Christian A. Smith,et al.  Moral, believing animals : human personhood and culture , 2003 .

[22]  Ursula Hess,et al.  Testing moral foundation theory: Are specific moral emotions elicited by specific moral transgressions? , 2018 .

[23]  Sasahara Kazutoshi,et al.  Quantifying moral foundations from various topics on Twitter conversations , 2016 .

[24]  Rada Mihalcea,et al.  Values in Words: Using Language to Evaluate and Understand Personal Values , 2015, ICWSM.

[25]  T. Landauer,et al.  A Solution to Plato's Problem: The Latent Semantic Analysis Theory of Acquisition, Induction, and Representation of Knowledge. , 1997 .

[26]  Petr Sojka,et al.  Software Framework for Topic Modelling with Large Corpora , 2010 .