You are here

An investigation of user-generated book reviews: Goodreads or Badreads?

Representation of a resource's subject content using descriptors from a pre-existing vocabulary raises questions regarding the relationship between controlled vocabularies and natural language. A controlled vocabulary strives to ensure consistent semantic representation of resources by normalizing the indexing vocabulary, resulting in what Lancaster (1977) has characterized as an “artificial language… in which the terms… have assumed special meaning by the way they have been used in indexing” (p. 23). In this situation, user-generated content, such as tags and book reviews, could complement and enhance the representation, organization and discoverability of resources. LIS studies have investigated both the functionality and subject representativeness of user-generated vocabularies ( Munk & Mork 2007) and the level of agreement among tagging vocabularies generated by multiple users (Kipp & Campbell 2006), often by comparing user-generated vocabularies to traditional systems of information representation and organization (Bruce 2008; Yi & Chan 2009). However, there is a lack of studies evaluating online reviews through the lenses of theories of representation and organization to find how reviews might enhance the representation, organization and findability of resources.

Besides serving publishers’ commercial interests, Goodreads’ users’ reviews make the site a valued resource for library acquisition and reference services (Thelwall & Kousha, in press). Goodreads’ ratings and reviews of academic books also provide alternative measures (i.e., altmetrics) of the impact of scholarly publication ( Zuccala, Verleysen, Cornacchia, & Engels, 2015). While most online book reviews tend to provide either negative or positive opinions about a book (Kousha, Thelwall, & Abdoli, in press), user-generated reviews also provide insights into the book’s content. However, there is lack of studies analyzing the content of user-generated book reviews through a lens of representation and organization of resources. This paper aims to investigate the following broad question: How Goodreads book reviews, engagements and average ratings can complement traditional representation of resources? Preliminary analysis of over 2000 randomly selected reviews assigned to Goodreads 2015 top rated books demonstrates multifaceted representation of the reviewed books, including genre, relation to other relevant resources, personal experiences and affections.

Presentation Type: 
Audio Size: 
Presentation Visual: