Comparing sentiment expression in movie reviews from four online genres
| Pages | 317-338 |
| Date | 20 April 2010 |
| Published date | 20 April 2010 |
| DOI | https://doi.org/10.1108/14684521011037016 |
| Author | Jin‐Cheon Na,Tun Thura Thet,Christopher S.G. Khoo |
| Subject Matter | Information & knowledge management,Library & information science |
Comparing sentiment expression
in movie reviews from four online
genres
Jin-Cheon Na, Tun Thura Thet and Christopher S.G. Khoo
Wee Kim Wee School of Communication and Information,
Nanyang Technological University, Singapore
Abstract
Purpose – This paper aims to investigate the characteristics and differences in sentiment expression
in movie review documents from four online opinion genres – blog postings, discussion board threads,
user reviews, and critic reviews.
Design/methodology/approach – A collection of movie review documents was harvested from the
four types of web sources, and a sample of 520 movie reviews were analysed to compare the content
and textual characteristics across the four genres. The analysis focused on document and sentence
length, part-of-speech distribution, vocabulary, aspects of movies discussed, star ratings used and
multimedia content in the reviews. The study also identified frequently occurring positive and
negative terms in the different genres, as well as the pattern of responses in discussion threads.
Findings – Critic reviews and blog postings are longer than user reviews and discussion threads, and
contain longer sentences. Critic reviews and blogs contain more nouns and prepositions, whereas
discussion board and user reviews have more verbs and adverbs. Critic reviews have the largest
vocabulary and also the highest proportion of unique terms not found in the other genres. The most
informative sentiment words in each genre are provided in the paper. With regard to content, critic
reviews are more comprehensive in coverage, and discuss the movie director much more often than the
other genres. User reviews discuss the scene aspects (including action and visual effects) more often
than the other genres, while blogs tend to talk about the cast, and discuss the music and sound slightly
more often.
Research limitations/implications – The study only analysed movie review documents. Similar
content and text analysis studies can be carried out in other domains, such as commercial product
reviews, celebrity reviews, company reviews and political opinions to compare the results.
Originality/value – The main contribution of the study is the sentiment content analysis results
across genres, which show thesimilarities and differences in content and textualcharacteristics in the
four online opiniongenres. The insights will be useful in designingautomatic sentiment summarisation
methods for multiple online genres.
Keywords Internet, Film,Attitudes
Paper type Research paper
Introduction
With the explosive growth of Web 2.0 sites and applications, there is a tremendous
amount of user-contributed material on the internet expressing opinions on all sorts of
subjects, issues, events and products. Blogs, discussion boards and revie w sites
(containing both critic and user reviews) are channels commonly used to express
opinions on movies, products and social issues. Researchers are turning their attention
to a kind of automated text analysis method called sentiment analysis to mine opinion
information found on these sites.
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1468-4527.htm
Sentiment
expression in
movie reviews
317
Refereed article received
6 July 2009
Approved for publication
21 September 2009
Online Information Review
Vol. 34 No. 2, 2010
pp. 317-338
qEmerald Group Publishing Limited
1468-4527
DOI 10.1108/14684521011037016
One challenge in mining public opinion information on the web is that the texts
from these sites represent different genres of documents with different characteristics
which must be taken into consideration when developing automated methods to mine
sentiment across different types of web sites. When a new product is released in the
market, different user groups may publish different perspectives of the product on the
web. The sentiment analysis programme should summarise user sentiments on
various aspects of the product and present the different perspectives of the different
user groups. For example, the summary generated can highlight specific product
features appreciated by customers, but not by expert critics. Such opinion summar ies
will be useful not only to potential buyers but also to product makers.
This paper reports an analysis of the characteristics and differences of four online
genres in the way opinions are expressed, which will be useful for developing
automatic sentiment summarisation methods that take into account the characteristics
of different online genres. The study focused on movie reviews, because there are many
of them on the web, and they are written, with various levels of complexity, by
different user groups. Critic movie reviews tend to be lengthy and comprehensive,
covering most aspects of the movie (storyline, director, cast, etc.). User reviews tend to
be informal and more explicitly emotional, and not all aspects of the movie are
analysed. Discussion board users interact and respond to one another directly, and
sentiments are expressed using stronger language and tend to sound personal. In blog
postings, bloggers may discuss multiple topics and discussion of a topic may be
interspersed among other topics.
To understand how users perceive the four online genres of movie review s, we
conducted a small survey with ten final-year undergraduate journalism students. The
respondents consider user reviews, discussion boards and blogs to be useful sources of
opinions although they do not provide expert viewpoints. In fact some even find user
reviews more useful than critic reviews as user reviews reflect the opinions of ordinary
people and are more likely to carry the same perspectives. However they also find critic
reviews useful in clarifying issues relating to various aspects of the movie and in
helping them to understand the movie better, even after watching the movie. Among
the four genres, they find discussion boards least useful due to spamming and
irrelevant postings.
The following sections introduce the online genres and discuss the research on
automatic sentiment analysis. Then the results of our sentiment analysis are discussed.
Online genres
Genres are commonly used for organising and presenting information to serve
particular purposes. Genres cover very broad concepts and are sometimes hard to
define and distinguish compared to topical categories, such as education vs computer.
When numerous documents have similar socially recognised characteristics, such as
content, forms (or styles) or intended communicative purposes, they are categorised
under the same genre (Kwas
´nik and Crowston, 2005). For instance, science fiction and
horror are two different genres that are differentiated by content. Product
specifications and product reviews are different genres that are distinguished by the
intended communicative purposes of the documents. A product specification describes
the features of a product, while a product review provides opinions about the product
features. Montesi and Owen (2008) identified and analysed several article genres in the
OIR
34,2
318
To continue reading
Request your trial