A semantic measure of online review helpfulness and the importance of message entropy

Jorge E. Fresneda, David Gefen

Research output: Contribution to journalArticle

Abstract

The helpfulness of online reviews and their impact on purchase decisions is well established. Much previous research measured that helpfulness by analyzing vote assessments. This study examines an alternative semantic measure based on a text analysis of the term “helpful” in those reviews. Analyzing over 20,000 reviews shows that the semantic measure has a considerably higher R2 than vote assessments. Moreover, the new measure, as opposed to those based on votes, is not affected by posting order, avoiding a known source of bias in vote measures, and is conceptually unrelated to the number of previous helpfulness evaluations. The study also examines the role of the incremental entropy of each review's content as a new determinant of both the existing measures and the new semantic measure of online review helpfulness. The potential of the semantic measure, including that it can be automatically calculated even before human review users read the review, is discussed.

Original languageEnglish (US)
Article number113117
JournalDecision Support Systems
Volume125
DOIs
StatePublished - Oct 2019

Fingerprint

Entropy
Semantics
Online reviews
Research
Vote

All Science Journal Classification (ASJC) codes

  • Information Systems and Management
  • Information Systems
  • Developmental and Educational Psychology
  • Arts and Humanities (miscellaneous)
  • Management Information Systems

Keywords

  • Ecommerce
  • Information entropy increment
  • Latent semantic analysis
  • Online consumer reviews
  • Review helpfulness

Cite this

@article{1683a81693194da09b791ba70530572c,
title = "A semantic measure of online review helpfulness and the importance of message entropy",
abstract = "The helpfulness of online reviews and their impact on purchase decisions is well established. Much previous research measured that helpfulness by analyzing vote assessments. This study examines an alternative semantic measure based on a text analysis of the term “helpful” in those reviews. Analyzing over 20,000 reviews shows that the semantic measure has a considerably higher R2 than vote assessments. Moreover, the new measure, as opposed to those based on votes, is not affected by posting order, avoiding a known source of bias in vote measures, and is conceptually unrelated to the number of previous helpfulness evaluations. The study also examines the role of the incremental entropy of each review's content as a new determinant of both the existing measures and the new semantic measure of online review helpfulness. The potential of the semantic measure, including that it can be automatically calculated even before human review users read the review, is discussed.",
keywords = "Ecommerce, Information entropy increment, Latent semantic analysis, Online consumer reviews, Review helpfulness",
author = "Fresneda, {Jorge E.} and David Gefen",
year = "2019",
month = "10",
doi = "https://doi.org/10.1016/j.dss.2019.113117",
language = "English (US)",
volume = "125",
journal = "Decision Support Systems",
issn = "0167-9236",
publisher = "Elsevier",

}

A semantic measure of online review helpfulness and the importance of message entropy. / Fresneda, Jorge E.; Gefen, David.

In: Decision Support Systems, Vol. 125, 113117, 10.2019.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A semantic measure of online review helpfulness and the importance of message entropy

AU - Fresneda, Jorge E.

AU - Gefen, David

PY - 2019/10

Y1 - 2019/10

N2 - The helpfulness of online reviews and their impact on purchase decisions is well established. Much previous research measured that helpfulness by analyzing vote assessments. This study examines an alternative semantic measure based on a text analysis of the term “helpful” in those reviews. Analyzing over 20,000 reviews shows that the semantic measure has a considerably higher R2 than vote assessments. Moreover, the new measure, as opposed to those based on votes, is not affected by posting order, avoiding a known source of bias in vote measures, and is conceptually unrelated to the number of previous helpfulness evaluations. The study also examines the role of the incremental entropy of each review's content as a new determinant of both the existing measures and the new semantic measure of online review helpfulness. The potential of the semantic measure, including that it can be automatically calculated even before human review users read the review, is discussed.

AB - The helpfulness of online reviews and their impact on purchase decisions is well established. Much previous research measured that helpfulness by analyzing vote assessments. This study examines an alternative semantic measure based on a text analysis of the term “helpful” in those reviews. Analyzing over 20,000 reviews shows that the semantic measure has a considerably higher R2 than vote assessments. Moreover, the new measure, as opposed to those based on votes, is not affected by posting order, avoiding a known source of bias in vote measures, and is conceptually unrelated to the number of previous helpfulness evaluations. The study also examines the role of the incremental entropy of each review's content as a new determinant of both the existing measures and the new semantic measure of online review helpfulness. The potential of the semantic measure, including that it can be automatically calculated even before human review users read the review, is discussed.

KW - Ecommerce

KW - Information entropy increment

KW - Latent semantic analysis

KW - Online consumer reviews

KW - Review helpfulness

UR - http://www.scopus.com/inward/record.url?scp=85070081932&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85070081932&partnerID=8YFLogxK

U2 - https://doi.org/10.1016/j.dss.2019.113117

DO - https://doi.org/10.1016/j.dss.2019.113117

M3 - Article

VL - 125

JO - Decision Support Systems

JF - Decision Support Systems

SN - 0167-9236

M1 - 113117

ER -