This chapter presents the findings from this study organized by method. Given the sequential, multiphase research design used (see section 3.3), this chapter begins with findings from content analysis of messages in the nine groups, then turns to the results of the survey of members of and visitors to the nine groups, and concludes with reporting analysis of the interviews with select survey takers. All three methods address the two research questions introduced in Chapters 1 and 3, repeated here for reference:
- RQ1: What roles do LibraryThing and Goodreads play, as boundary objects, in translation and coherence between the existing social and information worlds they are used within?
- RQ2: What roles do LibraryThing and Goodreads play, as boundary objects, in coherence and convergence of new social and information worlds around their use?
Each method addresses each of the two digital libraries studied, LibraryThing and Goodreads (see section 3.2); findings from each case are integrated together in this chapter, per method, for a fuller, more descriptive picture of the phenomena of interest that impact the roles the cases play, as boundary objects, in existing and emergent social and information worlds. Differences and commonalities in results between the two cases are mentioned here where appropriate. A synthesis and discussion of the findings across the three methods and in relation to the literature is provided in Chapter 5. The briefer summary of the findings to be found in section 5.1 may serve as a more useful entry point for some readers, who are encouraged to return to this chapter for further details of findings that interest them.
4.1. Content Analysis
The research design for the content analysis phase of the study was reviewed in section 3.4. After groups were sampled following the procedures in section 3.4.2, 286 messages were collected on May 24th, 2013 from the discussion boards of five LibraryThing groups, and on June 21st and July 9th, 2013 233 messages were collected from the discussion boards of four Goodreads groups (three in June, one in July), for a total of 519 messages. Five Goodreads groups were sampled at first, but data for one of these groups had to be dropped in late June, before content analysis began, when it was discovered that its sole moderator was under 18 and could not provide consent. A pilot test was conducted with two separate groups prior to the main content analysis phase beginning; see sections 3.4.3, 126.96.36.199, and 3.7.1 for further discussion of this test. Qualitative analysis of messages then proceeded as mentioned in section 3.4.4, using the codebook and associated procedures provided in section 3.7.
For the purposes of intracoder reliability testing, a subsample of 20% of the messages analyzed during the content analysis phase was coded a second time at the end of the study. This subsample consisted of threads selected at random from across the nine groups, with an eye to a sample balanced between messages posted on LibraryThing and Goodreads. 103 messages were recoded from 8 threads, 41 from LibraryThing and 62 from Goodreads. Intracoder reliability testing produced an average of 96.2% agreement on codes applied, with a Cohen’s kappa of κ = 0.7011. Such a value is considered to represent “substantial” (0.60 < κ < 0.80; Landis & Koch, 1977, pp. 165) or “fair to good” (0.50 < κ < 0.75; Fleiss, 1981) agreement in coding. Brief review of those codes and sources where agreement was not “substantial” (κ < 0.60) on Landis and Koch’s scale indicated most disagreements were on the potential meaningfulness and importance of codes, with one or more codes applied to sentences during the initial analysis that did not contribute significantly to later holistic analysis. Some variation in the codes applied and in the relative significance of codes is understandable, given the different knowledge held during the initial content analysis phase versus once all data analysis was complete.
The subsections below cover results of relevance to each of the phenomena of interest under the theoretical framework, with emphasis on those phenomena found to play significant roles in the coherence, convergence, and information behaviors of users of LibraryThing and Goodreads. Quotes were selected from those coded for these given phenomena (see section 3.7) based on their representativeness and potential insight in answering the research questions and explaining the roles played by the two digital libraries in relation to each of the phenomena and groups. Names and other information that could identify participants—including the book series of interest in two of the groups—were changed to protect the confidentiality of users. Pseudonyms were chosen with an eye to representing participants’ perceived ages, genders, and ethnic backgrounds; for example, a hypothetical user whose real name was “Michelle” and believed to be middle-aged would be given an American female pseudonym appropriate for a 40-year-old, such as “Amy,” instead of being called “Tony” (wrong gender) or “Gertrude” (wrong perceived age). The nine groups are referred to by letters: Groups A-E are the five LibraryThing groups, while Groups F-I are the four Goodreads groups. The topics of each group can be characterized as follows:
- Group A: Science fiction literature
- Group B: Audio technology in relation to books
- Group C: C. S. Lewis’s The Chronicles of Narnia series of young adult (YA) fantasy books
- Group D: Books by a given publisher (who founded the group)
- Group E: The universe of Frank Herbert’s Dune science fiction series
- Group F: Eclectic, diverse, and challenging reading across different genres
- Group G: Historical fiction
- Group H: The love of literature and reading
- Group I: Hobbyist writing of fiction
Many of the phenomena of interest were found to play one or more roles in the processes of coherence and convergence of communities within and around the groups and the two digital libraries. Phenomena that played particularly important roles included emergent information value, sites, technologies, social types, and social norms; these are discussed first below, followed by the remaining concepts.
4.1.1. Information Value
188.8.131.52. Key in two groups
In many groups information value played a role in coherence and convergence, and it played a key role in the latter around and within LibraryThing Group B and Goodreads Group F. In Group B, members spent much time establishing common values of which audiobook narrators were best and of the qualities they valued in narrators. The narrators would sometimes be socially typed in relation to their qualities or their status as narrators, as with one user who stated “I agree with Marie, some recordings are pretty good considering they are done by volunteers.” In this first example, the type of “volunteers” is raised, and given the characteristic of producing lesser quality recordings than professional audiobook narrators. In a second example a user responded to another, saying “I’m with you though, if a reader does bad voices / cannot characterize one sex well / reads with an unnatural cadence I just do not stick with it (unless the book is REALLY short).” In this case, the qualities and information valued are made explicit; this user wants a narrator who can do many voices well and read the audiobook naturally. Other users discussed their personal values in relation to audiobooks and audiobook narrators, finding common ground that became the values for an emergent information world and community formed by a significant subset of Group B, albeit implicitly in most cases.
In Group F, the welfare of a member’s cat became a common value for building community, with other users expressing shared concern in a affable and caring way over time (e.g. “I do hope your cat’s doing better now”), often with humor and sincere emotion injected (e.g. “keep my paws crossed” and “sending purry thoughts his way :)”). Group F also built common values around the thread topics they wanted to see established; when two new topical threads were suggested, many users chimed in to express their approval and agreement.
184.108.40.206. In other groups
Other groups expressed common information values, but with lower observed frequency. In these groups information values often focused on being open to contributions from other group members. For example, in Goodreads Group F a thread setting up a “challenge” where group members could score points for their reading activities included a message noting “there’s still a couple of weeks before we start too, so if there’s a really exceptionally good category you’d like to suggest, go ahead!” This sort of comment indicated a shared sense that informative contributions towards the group’s activities—in this case, types or categories of reading activities that could score points for group members—were of value to members and moderators. This sense was shared and coherent within an existing information world—that of Group F—and less so convergent for an emergent subset of that group. A similar example is this following exchange, posted in a LibraryThing Group C thread where the users were playing a forum game:
April: Brad, you may want to check out this information thread [link]. It will give you an idea of who all the players are and what our nicknames are!
Brad: April- Thanks; I forgot all about it. It is starred but I never posted to it and so it got lost in the shuffle!
Here, Brad had forgotten about a useful thread that gave information about significant members of Group C, and April valued sharing this thread with him and others, believing it to fit the cohered information values of the group. Brad implies a shared valuing of the thread in his response, although this convergence is on a small scale within a small, emergent information world of two members, April and Brad. The level of convergence should be considered low given that Brad had “never posted to” the thread in question.
Smaller sets of users established common ground over the value—or potential value—of reading given books, authors, or genres and discussing them. Some typical examples of this included the comment “I’m also a great admirer of Fieldman’s work”; a comment from a user stating
…I am going to set this aside for a soon-to-read selection based on your recommendation and then I will totally be down for some discussion about it, so I will be sure to check in when I have finished to see if I can help you get one going on it.
…and, after two users had both recommended a particular author, “ah, great minds think alike - think you’re obliged to check out Cook now :-)”
220.127.116.11. Coherence of individual information values
Individual expressions of information value were common, as users shared their opinions of books, authors, genres, products, and so on. The kinds of information behavior and information-related activities users chose to engage in implicitly indicated their information values and their thoughts on what they expected others would find of value. In some cases these values were seen to align with those of other group members, leading to coherence and—on occasion—convergence; in other cases this could not be determined from content analysis alone; and in still other cases values were divergent. When users disagreed, they were most often polite and did not cause conflict over differing opinions; “flame wars” or other disruptive behavior due to value disagreements were rare occurrences in the threads collected and analyzed.
Consider the following exchange in Group A; it began with a post by Will that was later removed, to which Brian responded “Please read [LibraryThing’s author policy]. And when your message is flagged so as to hide it from view to prevent its use as advertising, do not take it too personally.” Will, realizing the (apparent) error of his ways in violating existing, cohered LibraryThing information values and social norms around self-promotion, said “Sorry about that. I did the member giveaway and thought this was the right place to put up links for it to those who did not get it or would want it. I removed it.” Brian had no desire to start a major conflict, as shown in his subsequent response “Good show! We’re a very forgiving group. :)” This and other examples showed understanding of the coherence of information values and associated social norms shared in each of the digital libraries and in broader society and on the Internet as a whole, a role played by these phenomena in almost all of the groups.
In a few cases people beyond the group were included in the establishment of common information values, as seen in the following insightful example. One user posted a link to Group E to a knitted representation of a character from the Dune series of books, and another user responded by saying “I’m sending the link to my daughter, a Dune fan and a knitter.” In this case, we see the sharing of common information values beyond the group, with multiple information worlds present that intersect: the world of the group, the broader world of Dune fans, and the world of knitters. Each world has a different set of information values, but all three overlap in sharing a common value of this knitted representation of a Dune character. The values are existing and coherent within each world, but emergent and convergent across the worlds as they establish an overlap via this user’s comment.
The phenomenon of sites played somewhat of a lesser role than that of information value, but in many cases it was an important one in explaining the roles played by the digital libraries in users’ communities. The Goodreads groups were more apt to use the digital library as an emergent site for information behavior and activities.
18.104.22.168. Emergence of sites in Goodreads
In Goodreads groups group moderators played an active role, setting up threads and folders (grouped collections of threads) as sites for structured, purposeful discussion and behavior. For example, threads and folders were set up
- “for people to post their progress reports and score updates, ask questions and just generally chat as we go” as part of a gamified reading activity;
- as “where we will post threads for the chosen group read”;
- as a place for writing “the best [fictional, character-based] stories we can”; and
- for group reads themselves, in one case with a date given when discussion would begin and a note that “in the meantime, people can stop by this thread to chat, and I might post some bonus material about the book - but no spoilers until discussion opens please.”
In these and other similar cases, the thread had a clear, defined purpose that users were not supposed to stray too far from. Norms would sometimes be enforced or develop as a way of enforcing this purpose (see section 4.1.5 below).
In a couple of cases, threads served as emergent sites for a few users to engage in discussion that, while not on the intended thread topic, was allowed to continue. One example was a discussion of American Civil War literature in Goodreads Group G that continued for several long messages among three users (and may have continued further beyond the data collected and analyzed). Goodreads, Group G, and the thread in question served as an emergent site for this information behavior, despite the thread in question being intended for “threads for the chosen group read.” This shows that enforcement of the purpose of a given thread (or site for information-related activities, in social world terms) was not always immediate.
22.214.171.124. Emergence of sites in LibraryThing
LibraryThing users, in comparison, established common sites less often than Goodreads users. In cases when such sites were established, topic drift would often occur; for example, a thread in Group D with the initial purpose of discussing one out-of-print book began discussing other books by the same publisher that were also out of print. Group C had a special thread where “derailment” of the topic was the intended purpose, resulting in a site within the group where non-normative information behavior and activities could take place and were encouraged. Group D, with formal connections to a book publisher, served as a partial exception to this rule, with more emergent sites existing—for discussion of the publisher’s books and book series—in that group than in the other LibraryThing groups analyzed.
126.96.36.199. Existing sites: Weaker role
Invocations of existing sites for information behavior and activities were fewer, and often suggested a weaker role being served by LibraryThing and Goodreads in such behavior. Other external sites serving as information resources—online and brick-and-mortar booksellers, libraries, an audiobook web site, blogs—were mentioned and played a role, albeit not always a successful one. Some representative quotes include the following:
- “Amazing 5* and 4* reviews on Amazon and Goodreads already from those who got it on Amazon.”
- “They do not have this book in Shell’s book shop”; and later in the same thread, “it’s not at the library :(”
- “Right now most of my choices are dictated by which books are cheapest at the Kindle store.”
- “I wanted to review it [a particular book] for LibraryThing and for a new blog that a friend and I have started: [link]. But alas, Amazon (or [the publisher]) keeps changing the Kindle release date…”
Libraries were not a frequent topic of conversation in the threads analyzed; only two cases of them playing a role as external sites for information behavior and activities were seen, one of which was unsuccessful (as seen above).
188.8.131.52. Types of use
Many users of LibraryThing and Goodreads were frequent users of technology to link within the digital libraries to pages for books, authors, or series, within the flow of their information behavior. Both sites provided this functionality in their group discussion board interface, although uptake of the feature was seen to vary from group to group and from user to user. Larger Goodreads groups featured more technology use of this kind, while technology use in LibraryThing was diffuse across the five groups and users within them, in most cases use being slightly less frequent than in the Goodreads groups. The two groups with the least use, Groups C and E, focused on a specific author (Herbert) and book series (Narnia) known to members of each group; linking to books by that author or to books in the series was not as necessary an activity. In the other seven groups sampled, linking to books, authors, or series made up a majority of occurrences of technology use identified in the data, and served to cohere the group together and help converge subsets of it into new communities (as discussed further below).
Other uses of technology were nevertheless prevalent. Users linked beyond their group and used the technology provided by LibraryThing and Goodreads to support the creation and continuation of sites for common information behavior. In one example a LibraryThing user tried to encourage others in Group B who might be interested in a particular book to come and join another group’s group read, saying that “we’re having a group read of it, so anyone interested in joining, the thread is here [link].” Other examples referenced technology provided by LibraryThing or Goodreads, such as discussion board threads and the “folders” that threads can be organized in within a Goodreads group’s discussion board. A link to a page on LibraryThing relating to posting policies for authors was posted to Group A by Brian in one case (see section 184.108.40.206), resulting in Will using the technological features provided by LibraryThing to remove his offending post. Technology provided beyond the discussion boards was referred to, including polling features, user profile pages, and search functionality.
Users referenced organizational features provided by the digital libraries, such as the ability to create lists (or “shelves” in Goodreads parlance) of books one would like to read in the future. For example, one Group F member stated they “actually stumbled across that book on [an audiobook service] a couple months ago, so it’s on my library e-book wishlist but somehow it never made it to my Goodreads TBR [To Be Read shelf].” This same functionality was referenced by other users as playing a role in their information behavior, such as one—also in Group F—who said she had “been adding a LOT of books to my want to read list (thanks everybody).” In one case in Group G “the group’s bookshelf” was referenced, a shared list used by the group to catalog all of the books read together as a group. Metadata for books was referenced, such as in these two examples—one from each digital library—where the technology was providing information that could be seen as incorrect:
“I’m almost afraid to nominate this book [for a group read] since it does not have any of the Goodreads genre classifications (very loose, very subjective) attached to its record.”
“Series info is part of Common Knowledge that anyone [any members of LibraryThing] can edit, so if you find errors or missing books then feel free to change/add as appropriate.”
220.127.116.11. Technology for community convergence
In a good example of technology use to support community convergence, Mia, a Goodreads user, had expressed interest in Group G in German literature and in Western European war literature, and in learning more about books within these genres. Another user, Jared, responded with information that could help the user satisfy their information need, using the technology provided by Goodreads to support Mia’s finding of others sharing common interests, activities, and behaviors. Mia was appreciative, as seen in the exchange below:
Jared: Have you found the … German Literature group ([link]), yet?
Mia: Oh wow thanks! :) I’m sure to find more authors there.
Jared: My pleasure. I also sent you a friend request …. Also there is a “Books set in Germany” thread in this group, under [folder name], as well as many other interesting threads.
Mia: Thank you for the tips and for the friend request. I accepted it. I’ll look into that as well. Cheers.
Jared: Have fun! I’m sure you’ll enjoy [this group] (and Goodreads). It’s a nice group of folks.
A similar, albeit simpler and less convergent example of technology use is shown in the sharing of a link to another thread in Group C by April in a conversation with Brad (see section 18.104.22.168 above). The case related in section 22.214.171.124 of a user “sending the link to [his] daughter, a Dune fan and a knitter” shows technology—in the form of the link in question—playing a role in the convergence of Group E and of a broader community beyond the group of Dune fans who are knitters.
126.96.36.199. Technology for community coherence.
Uses of technology that related to cohering existing social worlds were somewhat less common, but still often played a role in users’ information behavior and in groups and communities. Occasional references to different forms of books—audiobooks, e-books, movies—were found scattered across the groups, and a couple of references to dictionaries (e.g. the Oxford English Dictionary or “OED”) and reference books were seen. More common were links to external web sites that served as information resources, such as online booksellers, libraries, audiobook web sites, blogs, publisher web sites, and author web sites. These were used by many users and in many groups to aid in the exchange and sharing of information about books. Few links to social networking or social media-type web sites—other than blogs—were found, although many of the web sites that were linked included user-generated content (most notably Amazon.com, but also deviantART, Wikipedia, Huffington Post, and other examples).
This latter kind of use was prevalent in LibraryThing Group B, who spent much time discussing audiobooks and audiobook narrators, as links to audiobook providers or particular audiobooks were of primary relevance to and played a large role in this information behavior. In another thread in the same group, users engaged in a short discussion of e-book readers after one user used their Nook to post, noting they could not link to the books they were discussing as the “Nook keyboard does not have the right brackets” necessary to insert a link. Another thread in this group was started by a user who wanted “advice and suggestions for an MP3 player that would be user-friendly, especially for someone who is not very comfortable with technology.” Members of Group B discussed MP3 player hardware that could be used to listen to audiobooks and podcasts and quirks of the software used to transfer audiobooks to these devices, making suggestions and giving advice to help satisfy the user’s information need.
188.8.131.52. Based on information needs
Translation occurred quite often in the messages analyzed, with users making frequent requests—whether they realized it or not—for the translation of meaning and understanding from the worlds of others to their world. This was most true when users would make requests based on their information needs or desires. Some examples typify this: “is this [book] a good place to start if you’ve never read any of his novels before?”; “… my question to you: does this line [an idiom] mean anything to anyone?”; and the following from LibraryThing Group B:
I’d like to get an MP3 player for my mother-in-law. She enjoys reading, but I do not think she’s ever listened to audiobooks. … I’m looking for advice and suggestions for an MP3 player that would be user-friendly, especially for someone who is not very comfortable with technology.
In most of these cases, the translation process continued in later messages, as users provided their understandings and the collective sense of a group (or subparts of it) around the request (offering e.g. the best MP3 player for a mother-in-law, the accessibility of an author’s novel for those who have not read his work before, or the meaning of a phrase used by an author). In the more elaborated examples, meaning and understanding became socially constructed as multiple users engaged in discussion and translation activities. The Group E thread that began with the question about an idiom used by an author was the strongest example of this, featuring multiple interacting and intersecting information worlds. Users with different backgrounds, from different locations, or having different information and knowledge shared their understandings, meanings, and translations of the expression and its context.
184.108.40.206. Disruptive circumstances
In another case, translation took place in the explanation of circumstances from other social and information worlds that could disrupt the boundaries of the group-as-community and their information behavior and activities. This lengthy, but insightful example included multiple members of Goodreads Group F participating in a group read:
Amelia: Well my copy [of the book] has not shown up yet. To add to my challenges, I’m getting migraines … The fact that I clearly need glasses at this point is not helping! I’m hoping to participate still, but I seriously doubt I’ll be finished [in time] :(
Amelia: [A couple of weeks later] I’m definitely not going to make this [in time]. New kitten says no. How is everyone else going?
Frank: I just finished, however it was my second read of [that novel] so it went fairly quick.
Sherri: I most likely will not be done [in time], either.
Angela: I will not be done by [then]. One of my senior cats required emergency surgery and that caused a lot of nail chewing on my part, poor little guy. He seems to be doing well but he has a 14 day recuperation ahead and he is twelve so kind of watching out for him.
Carmen: [a day or two later, after other messages about the cat’s welfare] I hope your little guy is still doing well. I probably will not be done by [then] either. … Lots of work and a short hockey trip kind of got in the way. This book has been a tough read for me so far.
Amelia: Okay, so should we go for [two weeks later] or a little longer?
Sherri: I think [that date] would work for me.
Angela: … Thanks, [that date] sounds good to me.
Amelia: OK, [two weeks later] is the new start date. I’m only 10% in, but starting to really get into it.
Nate: Hope that everyone’s reading is progressing well :)
Amelia: [nearly two weeks later] I’m only just over halfway through. And those footnotes are killing my eyes!
In this example, we see multiple reasons given by multiple users of why they can not participate in the group read discussion on the original date, with an eventual suggestion from Amelia (group moderator) to move the date two weeks later. The meaning and understanding of the deadline and of the purpose for the group read are socially constructed and translated in the back-and-forth conversation that took place, and may need to be again if Amelia’s last comment (the most recent collected and analyzed from this thread) is any indication.
In another case, unintended disruption was caused by a user not familiar with the norms and values of the community represented by LibraryThing and one of its groups, Group A, with Brian posting a link to LibraryThing’s author policy (see section 220.127.116.11 above). Translation took place of the meaning and understanding of the purpose of a thread (or site, in social world terms) in the group, with reference to LibraryThing and group norms and values.
18.104.22.168. Explaining point of view
Translation sometimes included explaining the point of view of oneself, another user, or an author. A good, prototypical example of this is seen in the following comments from Goodreads Group F in relation to an author and one of their books:
John: [The author] did a great job developing [a particular female] character. I have not read his bio, but it seems like it would have to come from real life experience parenting a teenager. I kept feeling like I was reading an autobiographical account of [his] attempt to better understand his daughter, niece, or someone close to him.
Amelia: Absolutely. As someone who has been that daughter, it rang true for me too!
A few other examples typify translation and elaboration of one’s point of view, such as a user who stated they “usually ignore the reviews and choose a book just because I like the title or the cover. (I know I should not judge a book by its cover, but I’m a photographer so I get attracted by the visual ….)” Another user of LibraryThing Group A questioned “the purpose of even having a ‘join to post’ requirement [in Group A] when it is so easy to join or to leave. It’s not much of a barrier …” A third user and group moderator shared their views on the differences between graphic novels and comic books:
The difference mainly lies in the format of the story contained. A graphic novel is usually a self-contained novel-length story, while most comics are serialized month to month. Most graphic novels tell a single, complete story and their length allows for greater depth and character development.
This last example led to some discussion of the merits of graphical storytelling, with users comparing and translating their meaning and understanding of comic book and graphic novel media.
22.214.171.124. Other examples
Other examples of translation observed included explaining the connection of a comment to the thread topic; discussion of the intended purpose and topic of an existing or new thread (when not set in stone by a moderator); giving necessary context for information found, e.g. as potential spoilers from early drafts of a book; back-and-forth on the merits of a book (often in response to or as part of a posted review), a character from a book or series, or of an author; and discussion of the understanding of a news or opinion article shared in a group. Plot points were translated and placed into context for other users, such as with one user’s explanation that “incoming war with the Lunar Alliance … in context is like Java going to war with the United States.”
One final example of translation is atypical in that it comes from a publisher’s account in LibraryThing Group D, responding to a user’s “quibble” about the style of notes (i.e. footnotes or endnotes) used in a particular book edition:
The placement and format of notes and note markers is always a tough call. In this case there was a feeling that, given the number of notes in particular sections, it would be distracting to have markers on the actual page. Clearly this would not have been true for you.
126.96.36.199. Convergence not guaranteed
Of course, translation was not always successful, with one member of LibraryThing Group A admitting as much in his response to a request for suggestions of non-American authors in a genre. After making some suggestions, he concluded his reply by saying it “rather depends on your tastes in humour of course….” This illustrates an important caveat of the translation process, that meanings and understandings may still differ between social and information worlds despite users trying to reconciling them together—discussing and explaining differences in meaning and understanding—across those worlds and converging a new world as a result. While the thread, Group A, and LibraryThing served a role as a boundary object in this discussion, there was no guarantee that it would be a successful role in every way and at every stage of the translation process, nor that a new community would converge from the process.
188.8.131.52. Note on frequency
As a final note on translation, three of the groups—Goodreads Groups F and H and LibraryThing Group A—featured a plurality (and near majority) of the examples seen in the data, showing clear signs of being more encouraging and welcoming of the process and of this form of discussion. Two of these groups—Groups A and H—were more overtly international in their make-up than others, and one—Group F—had a strong moderator who was engaged and invested in the success, coherence, and convergence of the group.
Few emergent organizations were seen; when they did occur, they were often due to the efforts of boundary-spanning individuals. For example (and as mentioned in section 184.108.40.206), when mentioning a particular book a user of LibraryThing Group B related that another group was reading it and linked to the group. The thread on the other group served as an emergent organization for this group read, crossing (or at least trying to cross) the boundaries of the established groups and organizations of the digital library (albeit remaining within its broader scope). Other group reads and a reading challenge that occurred could be considered semi-emergent organizations, as not all members of a group would participate and they might draw in other users who are not or who would not consider themselves members. The degree of emergence in most of these cases is small, and most relate to the existing organization of a given group.
References to existing organizations, besides the group a thread was part of, were a little more common, but still infrequent. Users referred to organizations such as online and brick-and-mortar booksellers, publishers, libraries, audiobook providers, blogs, and LibraryThing and Goodreads themselves; these served as sites for information behavior and activities (see section 4.1.2 on sites, above). References to organizations—in particular, a given publisher—were most common in LibraryThing Group D, associated with that publisher; such references were expected. Other, more unique examples of organizations mentioned that did not serve as sites (in social world terms) for users’ activities included the Pulitzer Prize for Fiction jury, a seller of artwork related to a book series, a set of bloggers working together, a book distributor (as distinct from a publisher), and a television network adapting a series of books.
4.1.8. Open Codes
Three additional open codes were assigned during the analysis process which had not been included in the codebook. These were other boundary objects, boundary spanner, and outsiders.
220.127.116.11. Other boundary objects
There were quite frequent invocations of other objects as boundary objects that played an important role in the social and information worlds of LibraryThing and Goodreads users. Some were existing sites, technologies, or organizations (in social world terms; see sections 4.1.2, 4.1.3, and 4.1.7). For example, in a LibraryThing Group B discussion about audiobooks two web sites for audiobook providers became the discussion topic, with their various merits discussed and related and a site (and its associated organization) becoming important to the information behavior and activities of users posting in the thread. The sites served as boundary objects between the communities of the group and thread and the community and organization of these providers, with the LibraryThing thread acting as an emergent site for discussion and translation of users’ meanings around and understandings of these objects.
In LibraryThing Group D, started by and facilitated by a publisher, the publisher themselves often acted as a boundary object or boundary spanner, since they edited and published books that users of the group, in at least one instance, discussed the layout and appearance of. In another discussion within Group D, the process of getting “Early Reviewer” books via LibraryThing (before they are available to the general public) became a boundary object for brief discussion, with a representative of the publisher stating “I think that LibraryThing gets the advance copies from our distributor….” In another thread in LibraryThing Group E, TV networks served as organizations and boundary objects in a discussion of adaptations of fiction as TV series.
Other web sites and online resources became boundary objects when they were posted as sources of information that then underwent a socially constructive process of sense- and meaning-making, helping a community understand them individually and collectively. A newsletter—which was not specified to be offline / print or online / digital—served as a boundary object in another discussion. While not quite the same kind of resource, a quiz posted in LibraryThing Group C was socially constructed (and deconstructed) in the ensuing discussion of users’ individual and collective performance on it. Technologies could serve a similar role, as in short discussions about e-book readers and MP3 players.
Books served as other boundary objects in threads where a particular novel was the focus of discussion; users would discuss and translate their understanding of the book’s plot and characters and their valuing of them. Books could serve as boundary objects between smaller groups of users when they were suggested by one user to be of interest to one or more others. In one case from Goodreads Group H, the nature and medium of a book as a boundary object changed over the course of the discussion:
Rosa: I’d love to read this one but I’m afraid I do not have access to it. If anybody is willing to lend it to me, I’d appreciate it :)
Sandi: I’m almost sure you can find it at the library, or at a thrift shop. :-?
Rosa: It’s not at the library :(
Sandi: Or if you have an e-reader, or the Amazon Kindle for PC, you can download it and read it in digital format.
This book first changed from one that Rosa would borrow from another member of Group H, to one she could borrow from the library or buy from a thrift shop. Then, in Sandi’s last response, the medium of the book changed from a printed book to a digital one.
In another intriguing case from LibraryThing Group E, the British English language became a boundary object of importance to the information behavior and activities of users, as a thread was started to determine the meaning of an idiom used by a British author (as mentioned earlier in this chapter). In this discussion other potential boundary objects were raised in users’ explanations or educated guesses.
18.104.22.168. Boundary spanners
Users or other individuals would, on occasion, act as boundary spanners between social and information worlds. While such occurrences were less common than other boundary objects serving this role, they still contributed to the role of LibraryThing and Goodreads in the social and information worlds of at least a few users. In many of these cases the boundary spanning was between two groups in LibraryThing or Goodreads; a user frequenting both groups could make comments about one group’s activities and information behavior in another group, e.g. “I believe the [horror] group’s about to tackle it [as a group read],” and “I’ll start this off with[book suggestion], which I’m reading for the International Readers theme read on 20th century Germany.” The conversation between Mia and Jared on German war literature (see section 22.214.171.124 above) showed Jared acting as a boundary spanner between Mia’s own information world, that of Goodreads Group G where they were posting, and the German Literature group found elsewhere on Goodreads.
Boundary spanning occurred when members of a group would invite new members to join, or new members had seen mention of the group elsewhere. These new members would then explain this in their introduction messages, such as one who posted “I’m very glad I found you, so I have to thank Terry for unknowingly pointing me in this direction :)”; and another who stated she “got invited by Rosa here :D” and expressed her thanks: “I’m glad she did, this looks like an active group.” A related case was when one user introduced another to a book and claimed to have introduced his significant other to a book series, with a similar introductory element present in the boundary spanning.
One final case of boundary spanning occurred when a publisher’s representative, in LibraryThing Group D that was started and facilitated by the publisher, explained some factors in how they format endnotes and footnotes in their books. The representative served as a boundary spanner and translator between the publisher’s information world and Group D’s information world.
People and animals who were not members of the group or of the community most germane to a message were mentioned on a few occasions, with the open code “outsiders” applied in these cases. The animal cases—in the majority—were all in LibraryThing Group C and all in reference to the ability of cats to disrupt individual and group activities: a “new kitten says no” to finishing a book by a certain date, and a “senior cat” requiring “emergency surgery” leading to group discussion (see sections 4.1.2 and 4.1.5 above for more on the latter). Cases of people as outsiders included a “stepfather [who] recommended we move a few of the bookshelves into the garage.” After the user, a member of Goodreads Group H, asked where the books would go, she said that the stepfather “gave [her] a really weird look, then walked away.” The stepfather is seen as someone who does not understand the book-loving culture present in Group H and in Goodreads as a whole. Another case is of a former insider. A former member of LibraryThing Group C, Melanie, was referred to as “Mel” by one user, AJ, who stated she had reminded him of something which he contributed to the thread, implicitly on her behalf. When another user asked who she was, a third user responded that “Mel (melanie123) is AJ’s daughter. She’s in college now and not around as much.” Melanie was an outsider to the current makeup of the group-as-community, despite still being a member of Group C for the purposes of LibraryThing (i.e. as determined by the system and technology).
The research design for the survey phase of the study was reviewed in section 3.5. Using the groups sampled from the content analysis phase (see sections 3.4.2 and 3.5.2), users of the five LibraryThing groups and four Goodreads groups were invited to complete the survey following the procedures given in sections 3.5.2 and 3.5.5, with further details of this process given below. The survey was pretested prior to the main survey phase beginning; see section 126.96.36.199 for discussion of this. After the survey completion period was over, quantitative analysis of the findings was conducted using SPSS using descriptive, inferential, and nonparametric statistics as appropriate (see section 3.5.5 and further details below). The survey instrument that was used is included in Appendix B.
Invitations to take the survey were posted on August 26th, 2013 to the nine groups selected—five in LibraryThing, four in Goodreads—and sent in private messages to 113 LibraryThing members (one member in the sample from the content analysis did not allow sending of private messages and could not be included). Reminders were posted on September 9th and September 23rd; the survey was closed on October 7th. During the six weeks the survey was open, 264 users started the survey, with 163 completing it to the end (a completion rate of 61.7%). Of these, 99 were users of LibraryThing, while 64 were users of Goodreads. 78.5% of surveys were started or completed during the first two weeks of the time period; all but nine (94.5%) were completed during the first four weeks. Once incomplete responses and responses from users who did not fit the age criteria for the study (between 18 and 65 years old) were removed, a total of N = 142 survey responses—94 from LibraryThing users and 48 from Goodreads users—were kept as usable for analysis. Ten $25 gift cards were distributed to seven LibraryThing users and three Goodreads users, selected at random, on November 8th; all ten users chose to receive Amazon.com gift cards.
Due to the nature of the sampling and invitation methods, response rates cannot be confirmed. While group membership numbers totaled 7,875 for the four Goodreads groups and 8,935 for the five LibraryThing groups at the time of the survey, it is possible not all of these members saw an invitation message, and visitors to the groups who are not members and not included in these counts did see the message and respond. Since usernames were not requested, calculating response rates for that sample is impossible. Since sampling was not purely random—users could choose to participate or not and not all users of the nine groups were guaranteed to see the invitation to participate—relying on traditional inferential statistics—given that they assume a representative sample obtained through simple random sampling—to infer beyond the sample ( i.e. the participating users) is not possible. Selection bias—a form of sampling bias—may have generated results that are not fully representative, but one may assume that survey respondents are at least moderately representative of the population of users of the nine LibraryThing and Goodreads groups. If such an assumption is true, conclusions can be inferred about users in these nine groups. Nonparametric statistics are included and used to strengthen these inferences and findings. These are accepted as limitations of this exploratory study (see also section 5.4).
The following subsections present analysis of the reliability of the scales for each phenomenon of interest under the theoretical framework; a summary of participants’ demographic and background characteristics; analysis of relationships between demographics, background, and the phenomena of interest under the theoretical framework; and analysis of the phenomena in relation to one another and to the two case studies. Since not all participants answered every survey question, effective sample sizes at each stage and for each result are stated.
4.2.1. Scale Reliability
The Likert scales from the survey, as constructed (see section 3.5.4 and Appendix B), were analyzed to determine the internal consistency and reliability of the scales via Cronbach’s alpha. As per the procedures related by George and Mallery (2010), individual items were dropped from a scale where their removal would increase the Cronbach’s alpha for and the reliability of the overall scale. The results of this procedure are shown in Table 4.1, which shows the numbers of items that were dropped (refer to the survey instrument in Appendix B, section B.1 for details) and the resulting scale reliability, based on common cutoffs for Cronbach’s alpha discussed by George and Mallery (2010).
All scales reached at least the level indicating “acceptable” reliability (a Cronbach’s alpha of at least 0.7; see George & Mallery, 2010), with the exception of Boundary Objects and Information Value. While the value for the latter, 0.697, fell below the 0.7 cut-off, it was so close that it was decided to keep the scale in further analysis, while keeping in mind that the scale might be of questionable reliability. Due to the “poor” resulting reliability—even when an item was dropped—shown by the Cronbach’s alpha of 0.520, the Boundary Objects scale was dropped altogether from further analysis. It may be that the concept of a boundary object is too difficult to convey in five items, as tried here. In the theoretical framework the other concepts measured relate to the concept of a boundary object, one with a strong, clear role that can be considered successful. One might argue that the concept has adequate measurement through the combination of the other scales. Other qualitative methods of determining if LibraryThing or Goodreads played roles as a boundary object for a given participant could be useful; the follow-up interviews included in this study explicitly allowed for this possibility.
|Scale||Initial alpha||Item(s) dropped||Final alpha||Scale reliability|
|Coherence / Convergence||0.749||None||0.749||Acceptable|
|Information Behavior and Activities||0.773||4, 5||0.823||Good|
Note: As measured using Cronbach’s alpha and procedures from George and Mallery (2010).
* These final scales are of only three items. While it is normal to try for scales consisting of at least four items, in these cases the improvement in Cronbach’s alpha was considered worthwhile. The four-item scales for these concepts were included in parts of the further analysis for comparison, but no substantial differences in the results and findings were observed.
The final scales are given in Appendix B section B.2, after the initial survey instrument. Averages were calculated for these scales for each participant by summing the scores on each Likert scale item and then dividing by the total number of items. These were then used in further analyses, as related in sections 4.2.3 and 4.2.4.
4.2.2. Demographic and Background Characteristics
More participants were female than male; 86 (or 59.3%) reported their gender as female, while 51 (or 35.2%) reported male (n = 139); one participant reported other gender, while one declined to respond to this question. Differences were observed in the gender distribution of participants from LibraryThing vs. Goodreads. For LibraryThing, 40 (or 43.5%) reported to be male and 51 (or 57.3%) to be female, with 1 (or 1.1%) reporting other gender; for Goodreads, 11 (or 17.0%) reported to be male and 35 (or 76.1%) to be female, with 0 reporting other gender. Chi-square analysis suggests that such a difference is meaningful (if the participant who reported other gender is ignored; χ2(1) = 5.253, p = 0.022), although this finding is limited to the sample and inference to the nine groups assumes representativeness.
Participants ranged from 18 to 64 years old, with a mean of M = 42.12 years old and a standard deviation (SD) of 12.50 (n = 141); the median age was 42. Age distribution differed significantly between participants from the two digital libraries, as determined via an independent-samples t-test and an independent-samples Mann-Whitney U test. Goodreads participants (M = 38.48, SD = 12.12, median = 38, n = 48) were found to be younger (t(137) = 2.540, p = 0.012; for Mann-Whitney, Z = –2.504, p = 0.012) by an average of 5.55 years than LibraryThing participants (M = 44.03, SD = 12.33, median = 45, n = 91).
Of the 140 participants who reported their educational level, most had completed at least a bachelor’s degree, but many educational experiences were reported (see Table 4.2). No meaningful differences were observed in educational level between participants from the two digital libraries; chi-square analysis—limited to the sample and with inference beyond that level assuming representativeness—did not find any significant differences.
|Less than high school (or equivalent)||1||0.7%|
|High school diploma (or equivalent)||16||11.3%|
|Associate’s (two-year) degree||8||5.7%|
|Bachelor’s (four-year) degree||53||37.6%|
|Professional degree (JD, MD, etc.)||11||7.8%|
|Doctoral degree (PhD, EdD, etc.)||8||5.7%|
|Prefer not to answer||1||0.7%|
Note: n = 141
188.8.131.52. Use of the Internet
Participants spent an average of 27.82 hours (or about 27 hours and 49 minutes) on the Internet each week (n = 139, SD = 16.43); the median number of hours was 21 with a mode of 20 hours. Participants ranged from 5 to 100 hours of Internet use per week. Differences were observed between LibraryThing and Goodreads participants in their Internet use, but an independent-samples Mann-Whitney U test analysis showed this was not significant. If the samples are assumed to be representative, an independent-samples t-test analysis finds this difference significant: Goodreads participants (M = 32.58, SD = 21.54, median = 25, n = 48) may use the Internet somewhat more during an average week than LibraryThing participants (M = 25.31, SD = 12.38, median = 21, n = 91; t(137) = –2.160, p = 0.035), on average 7.28 hours more in this sample. Since sampling was not purely random, inference of this result beyond the sample requires assuming the sample is representative, and inference beyond the nine groups is impossible.
184.108.40.206. Use of the digital library
Of hours spent on the Internet each week, participants reported spending an average of 7.65 of those (or about 7 hours and 39 minutes) per week using the digital library in question (n = 140, SD = 7.52); the median number of hours spent on LibraryThing or Goodreads was 5 with a mode of 2 hours. Hours spent using the digital libraries per week ranged from 0 to 40; one can assume values of 0 reflect a few minutes per week at best, given all participants had used LibraryThing or Goodreads at least long enough to see the survey invitation. Significant differences were found between LibraryThing and Goodreads participants; analysis through an independent-samples t-test and an independent-samples Mann-Whitney U test showed the latter used Goodreads more during an average week (M = 9.83, SD = 8.98, median = 7, n = 48) than the former used LibraryThing (M = 6.51, SD = 6.40, median = 5, n = 92; t(138) = –2.283, p = 0.025; Mann-Whitney Z = –2.340, p = 0.019), on average 3.33 hours more and with a median difference of 2 hours.
220.127.116.11. Use of groups
Not all of this time was spent using the groups feature; participants reported spending an average of 5.20 hours (or about 5 hours and 12 minutes) per week in groups (n = 139, SD = 6.00), with a median of 3 hours and a mode of 1 hour. As with hours spent using the digital libraries, hours spent in groups per week ranged from 0 to 40. No significant differences were found between participants from LibraryThing and Goodreads in their use of their respective digital library’s groups, based on an independent-samples t-test and an independent-samples Mann-Whitney U test.
4.2.3. Demographic and Background Relationships
Independent samples Kruskal-Wallis, median, and Mann-Whitney U test analyses were used as appropriate to examine potential relationships between each continuous demographic or background variable and the phenomena of interest under the theoretical framework introduced in Chapter 2, for those scales that were considered reliable (see section 4.2.1 above). Inference of these results may apply to the population of users of the five LibraryThing and four Goodreads groups, if the samples are assumed to be representative of those users with minimal selection bias.
When ages were grouped into categories (each category spanning five years), age was found to be significantly related to information value via the median test for independent samples (χ2(9) = 26.014; p = 0.002; n = 136) and an independent-samples Kruskal-Wallis test (χ2(9) = 18.833; p = 0.027; n = 136). Younger participants tended to give a higher rating on information value than older participants, with the distribution of these ratings differing by age group. Age was found to be significantly related to technologies via an independent-samples Kruskal-Wallis test (χ2(9) = 19.951; p = 0.018; n = 135), implying differences in the distribution of technology scores across age groups. The median test for independent samples was not significant (p = 0.067). Examination of the median rating for technology per age group shows few major differences beyond the two youngest groups (18–27 years old), who rated lower on technology, but not significantly so. No other relations were found between age and the phenomena of interest.
A significant relation was found between education level and information behavior and activities, using the median test for independent samples for analysis (χ2(6) = 13.589; p = 0.035; n = 128); the median rating given for information behavior and activities decreases as education level increases. An independent-samples Kruskal-Wallis test was not significant (p = 0.096), implying the overall spread and shape of the distribution of ratings is similar at all levels of education. No other relations were found between education level and the phenomena of interest.
18.104.22.168. Hours of use
When responses for the number of hours spent on the Internet, on LibraryThing or Goodreads, and in groups in the two digital libraries were grouped into categories (in groups of four, three, and three hours, respectively), many relationships were found with the phenomena of interest. The statistical analyses are summarized in Table 4.4. Significantly higher ratings—based on median tests for independent samples—were given by those who spent longer using the digital library each week for a sense of shared organizations (χ2(9) = 17.735; p = 0.038), and by those who spent longer in groups each week for translation (χ2(9) = 20.404; p = 0.016) and a sense of shared organizations (χ2(9) = 18.142; p = 0.034); higher ratings were most observed among heavy users of each digital library and of the groups features. Among users who spent different lengths of time on the Internet each week, different distributions of ratings were observed—based on independent-samples Kruskal-Wallis tests—for coherence and convergence (χ2(13) = 23.975; p = 0.031). Among users who spent different lengths of time using the digital library each week, different distributions of ratings were observed for translation (χ2(9) = 20.706; p = 0.014), a sense of shared organizations (χ2(9) = 22.442; p = 0.008), and shared technology use (χ2(9) = 18.303; p = 0.032). Among users who spent different lengths of time in groups each week, different distributions of ratings were observed for translation (χ2(9) = 25.583; p = 0.002), coherence and convergence (χ2(9) = 17.347; p = 0.044), and a sense of shared organizations (χ2(9) = 22.317; p = 0.008).
|Hours on Internet||Hours on LT / GR||Hours in groups|
χ2(9) = 20.706
p = 0.014 *
χ2(9) = 20.404
p = 0.016 * Distribution differs
χ2(9) = 25.583
p = 0.002 **
|Coherence / Convergence||Distribution differs|
χ2(13) = 23.975
p = 0.031 *
χ2(9) = 17.347
p = 0.044 *
|Information Behavior and Activities||NS||NS||NS|
χ2(9) = 17.735
p = 0.038 * Distribution differs
χ2(9) = 22.442
p = 0.008 **
χ2(9) = 18.142
p = 0.034 *
χ2(9) = 22.317
p = 0.008 **
χ2(9) = 18.303
p = 0.032 *
* p < 0.05; ** p < 0.01; NS = not significant (α = 0.05)
4.2.4. Roles of Phenomena of Interest
Through independent-samples Mann-Whitney U test analyses, differences in a couple of the phenomena of interest were found between the two sites. Goodreads users rated their sense of shared information value judgments significantly higher than LibraryThing users (Mann-Whitney Z = –3.021, p = 0.003); on average, Goodreads users (M = 3.24, SD = 0.660, median = 3.25, n = 48) rated 0.403 higher per scale item (indicating more agreement with each item) than LibraryThing users (M = 2.84, SD = 0.722, median = 2.88, n = 92). Goodreads users rate their sense of shared information behavior and activities significantly higher than LibraryThing users (Mann-Whitney Z = –2.293; p = 0.022); on average, Goodreads users (M = 3.78, SD = 0.730, median = 4.00, n = 48) rated 0.240 higher per scale item (indicating more agreement with each item) than LibraryThing users (M = 3.54, SD = 0.676, median = 3.75, n = 90).
Descriptive statistics were analyzed for the phenomena of interest. These are reported in Table 4.5, with results from Wilcoxon signed rank tests to determine if the median rating is significantly different from 3 (a neutral value representing neither equal proportions of agreement and disagreement with the statements). Ratings on most of the phenomena of interest were found to be significantly above this value, indicating significantly more agreement than disagreement, with two exceptions: social types (median = 3.000, p = 0.323) and information value (median = 3.000, p = 0.709). While not as appropriate given the lack of simple random sampling, parametric testing using independent-samples t-tests to test for means different from 3 produced similar results.
|Sample size (n)||Mean||Median||SD|
|Coherence / Convergence||139||3.733||3.750 ***||0.667|
|Social Norms||140||3.736||3.800 ***||0.550|
|Social Types||140||2.945||3.000 NS||0.748|
|Information Value||140||2.975||3.000 NS||0.725|
|Information Behavior and Activities||138||3.620||3.750 ***||0.702|
*** p < 0.001; NS = not significant (testing H0: median = 3; α = 0.05)
Kendall’s τ correlations were computed among the nine scale averages that were found to be valid in the earlier validity analysis. These are reported in Table 4.6. Most correlations were significant (ps < 0.01), but three were notable in their non-significance: translation vs. information value (τ = –0.003, p = 0.958), technologies vs. information value (τ = 0.082, p = 0.201), and organizations and information value (τ = 0.046, p = 0.467). While significant, the correlation between social norms and social types (τ = 0.129, p = 0.044) and information value and sites (τ = 0.139, p = 0.028) were weaker (0.01 < p < 0.05) in comparison to the other significant correlations. Values for correlations using Spearman’s ρ were higher (implying stronger correlations when using that measure), but no major significance level differences were observed when compared with the Kendall’s τ values presented in Table 4.6.
The research design for the interview phase of the study was reviewed in section 3.6. The interview instrument was pretested prior to the main interview phase beginning; see section 22.214.171.124 for discussion of this. Interviewees were purposively sampled from survey participants, following the procedures in section 3.6.3 and using an interview instrument and guide constructed as discussed in section 3.6.4. The interview instrument used to guide the semi-structured interviews is included in Appendix C. Once interviewees had agreed to an interview and a date and time had been set, the collection of interviews took place following the procedures discussed in section 3.6.5. After interviews had been transcribed, qualitative analysis of interview transcripts proceeded as mentioned in section 3.6.6, using the codebook and associated procedures provided in section 3.7.
A total of 48 potential interviewees were purposively sampled based on the survey findings and contacted, of which 36 were sent a second reminder invitation prior to saturation being reached. From the 48 contacted, a total of N = 11 participants responded and were interviewed from January 13th to March 26th, 2014. n = 7 interviews took place with LibraryThing users and n = 4 took place with Goodreads users, a similar proportion as in the survey responses. Four of these interviewees responded after receiving a second reminder invitation; one potential interviewee responded after receiving a second reminder invitation and was scheduled for an interview, but then was not contactable at the time of their interview and did not respond to requests to reschedule. Full details of sampling and interview procedures can be found in Chapter 3, section 3.6. The average interview length was 50 minutes and 33 seconds, with the median interview length being 50 minutes and 3 seconds. The shortest interview was 34 minutes and 46 seconds; the longest, 60 minutes and 47 seconds. Six interviews were conducted via phone (with phone calls made through Google Voice and Google Hangouts), four interviews via Skype, and one interview via FaceTime; no interviewees chose Google Hangouts as their medium of choice.
Intracoder reliability testing was conducted using two of the interview transcripts (about 20% of the interview data), after the remainder of the data analysis was completed. These transcripts were selected at random; both were with LibraryThing users. This testing produced an average of 96.3% agreement on codes applied, with a Cohen’s kappa of κ = 0.7374. Such a value is considered to represent “substantial” (0.60 < κ < 0.80; Landis & Koch, 1977) or “fair to good” (0.40 < κ < 0.75; Fleiss, 1981) agreement in coding. Brief review of those codes and sources where agreement was not “substantial” (κ < 0.60) on Landis and Koch’s scale indicated most disagreements were on whether a code applied to an existing world, an emergent world, or both; or the degree to which a code applied, such as whether the digital library was an influential, standard boundary object for an emergent social and information world, or a place where information behavior and activities and technology use happened, without much influence. As with the content analysis, some variation in the codes applied and in the relative significance of codes is understandable, given the different knowledge held during initial analysis of the interviews versus once all data analysis was complete. The time difference was less than in the content analysis, which may have contributed to the better agreement.
The interviews provided the richest source of data in this study. Before the results of the interviews are reported in detail, section 4.3.1 below provides a summary, based on the survey findings, of the demographics and background characteristics of the interviewees and their Likert scale ratings for the phenomena of interest. Then, the remaining sections cover detailed results of relevance to each of the phenomena of interest under the theoretical framework. (As with the content analysis results, names and other information that could identify participants were changed to protect the confidentiality of users. Refer to section 4.1 for details.)
4.3.1. Demographic and Background Characteristics of Interviewees
Comparing means and medians for interviewees vs. survey takers on the phenomena of interest, no major differences are observed (see Table 4.7). Social norms, social types, information value, information behavior and activities, organizations, and sites were all rated less (indicating less agreement) by interviewees when comparing means and medians, although differences were minor; interviewees rated their agreement with statements on coherence and convergence and on technologies a bit lower on average, but medians were the same.
|Interviewees||All survey takers|
|Translation||3.894||3.667||3.882||3.833||(n = 140)|
|Coherence / Convergence||3.455||3.750||3.733||3.750||(n = 139)|
|Social Norms||3.418||3.400||3.736||3.800||(n = 140)|
|Social Types||2.788||3.000||2.945||3.000||(n = 140)|
|Information Value||2.795||2.750||2.975||3.000||(n = 140)|
|Information Behavior and Activities||3.409||3.500||3.620||3.750||(n = 138)|
|Organizations||3.636||3.750||3.824||4.000||(n = 138)|
|Sites||3.673||3.800||3.939||4.000||(n = 137)|
|Technologies||3.455||3.667||3.659||3.666||(n = 139)|
Note: For interviewees, N = 11; for all survey takers, N = 142.
Interviewees were more likely to be female, older, and have a bachelor’s degree or better; they used the digital library in question (LibraryThing or Goodreads) and the groups feature a little more often; and they were a little less frequent in their use of the Internet (see Table 4.8). Interviewees were less likely to use Facebook, Instagram, and Twitter; they were more likely to use blogging services, Pinterest, LinkedIn, and Google Plus. Use of the other digital library (Goodreads for LibraryThing users, and vice versa), Flickr, Shelfari, and Foursquare was similar among interviewees to that of all survey takers (see Table 4.9). Due to the small sample size of the interviews, no inferential statistics were used for comparison.
(N = 11)
|All survey takers|
(N = 142)
|Digital library||n = 7 LT (63.6%)|
n = 4 GR (36.4%)
|n = 94 LT (66.2%)|
n = 48 GR (33.8%)
|Gender||77.9% female||59.3% female||(n = 139)|
|Age (mean)||48.27 years||42.12 years||(n = 141)|
|Education||81.8% bachelor’s degree or better||76.6% bachelor’s degree or better||(n = 141)|
|Internet use (mean)||27.20 hours||27.82 hours||(n = 139)|
|Use of the DL (mean)||8.80 hours||7.65 hours||(n = 140)|
|Use of groups (mean)||5.45 hours||5.20 hours||(n = 139)|
Note: For interviewees, N = 11; for all survey takers, N = 142.
|Site||Interviewees||All survey takers|
|6 (66.7%)||103 (81.7%)|
|Blogging (e.g. WordPress, Blogger)||5 (55.6%)||43 (34.1%)|
|3 (33.3%)||32 (25.4%)|
|3 (33.3%)||30 (23.8%)|
|Google Plus||3 (33.3%)||24 (19.0%)|
|1 (11.1%)||45 (35.7%)|
|Goodreads / LibraryThing||1 (11.1%)||16 (12.7%)|
|Flickr||1 (11.1%)||13 (10.3%)|
|0 (0.0%)||16 (12.7%)|
|Shelfari||0 (0.0%)||6 (4.8%)|
|Foursquare||0 (0.0%)||5 (4.0%)|
Note: For interviewees, n = 9; two interviewees did not answer the survey question on their use of other social media sites. For all survey takers, n = 126.
A moderate amount of translation was mentioned by interviewees. Examples included users trying to come to agreement, getting to know each other, and negotiating and explaining norms and rules. In some cases translation was partial at best.
126.96.36.199. Coming to agreement
Translation would often involve users trying to come to agreement on values, norms, or behaviors. In one example, Tanya had a conversation with an author in Goodreads Group G about his book and how it had been categorized in Goodreads:
An author posted to [Group G] saying that his book had been characterized as modern history and he did not understand why. And I looked at the description on Goodreads and I also did not understand why. I told him that I thought that it was urban fantasy.
Then, later in the interview, Tanya looked at the discussion thread and related further details of the ongoing translation process:
Tanya: I am looking at the thread now, he made a comment in return and, he says that he thinks he has a “Dan Brown” sort of book. And, a “Dan Brown” sort of book makes it a, a thriller of some kind. But, and it has a historical element in it, apparently. But it also looked like it had a fantasy element from the description.
Adam: Um-hmm. So it sounds like there’s a lot of things going on in this book.
Tanya: Yes, yes, it looks like it may be hard … hard to categorize.
This translation process was still going on, since as Tanya stated, “the interaction is not developed enough yet … for [a] common understanding to have been built.” Of course, “readers can classify books any way they like. And they do!” Author disagreement with readers or publishers over meanings and understandings, like this, were mentioned by multiple interviewees.
Another example of users discussing the process of translation came from LibraryThing user Ann:
We were talking about gender and characters, and someone [brought up] Rich Berlew’s Order of the Stick [comic]; there’s a character in there that is, four, five, six books in, has not been fixed to male or female. … And it’s not known, it’s absolutely not stated. And if you go hunt up on boards and the rest, everyone’s got an opinion.
In this case, the thread Ann was participating in focused the translation at the meta-level, discussing gender roles in fantasy literature with this as one example, instead of trying to reach any shared understanding of the gender of Berlew’s character. Others brought in academic perspectives, adding another layer to the translation process. Later in the interview she mentioned that, through the process of having
… five or six people all on, all who’ve read different kind of books, and some overlapping and so on, you tend to get really quite good coverage. … When you’re actually to-ing and fro-ing about examples … it [the conversation] can go quite detailed.
188.8.131.52. Getting to know each other
Translation processes were mentioned as part of users getting to know each other better through their interactions in LibraryThing and Goodreads, as in the following examples from Goodreads users Rachelle and Kevin and LibraryThing user Betty:
Rachelle: That thread lets you get to know the people, and … a lot of the people who post in the random thread also post in the book threads.
Kevin: …[I was] interacting with several new members, learning what interests them and, you know, if they were writing or if they were just interested in, in reading different things.
Betty: …as time went on and there was more interaction, while you did not know people personally, you did get a sense of who they work, and sort of felt that you knew them at least a little bit, at least at an acquaintance level, just by virtue of the way they interacted and what they said.
Rachelle: …so you get to know who likes what, who’s into Bill Lawne and who liked [Lawne’s latest book] but does not like his Japan series, and stuff like that. So you know their book taste as well as some personal stuff that people talk about.
Such translation could, given time, lead to a close-knit community where social ties and relationships were important and common. This had happened to Sam, who related the extended process of translation—and eventual convergence—that he felt had occurred to him within a private LibraryThing group:
Because rules change when you change relationships, right? So you go from a stranger to a friend, and you know, there has to be common understandings, and there are things that you know not to mention to strangers. But by the same token the same happens with friends. You learn about people and you find out that, you know there are certain things that are sensitive topics and so on. So it works that way also.
Sam agreed that a sense of community and commonality had been established through this translation process and the process of establishing relationships with other users of LibraryThing.
184.108.40.206. Norms and rules
Translation could involve group moderators, who in the process of enforcing group norms or LibraryThing and Goodreads’ rules would help members understand the rules and their intended meaning. Sometimes established members would be reminded of the rules if they broke them, as Goodreads user Rachelle related:
…if a longtime member oversteps their bounds … they do not get, like if you’re a longtime member and you’re well known, you do not get kicked out right away. You [the moderator] are probably going to be like, “that just sounded a little mean,” you know, and the longtime member will be like, “oh, sorry! I did not mean it!”
220.127.116.11. Partial translation
Translation provided for a social element, where users were not just reading and making sense of a novel or other information source themselves, but could discuss it with others to come to an understanding. It was not always the case, though, that such translation could be considered an unqualified success, in the sense of being shared by others. For example, Miriam had started a thread on LibraryThing to discuss illustrations for a popular children’s novel from the early 20th century; as part of this thread a discussion came up about what features would be best for illustrations targeted at young children. As Miriam related,
…we did have a difference of opinion there; somebody thought that the simpler the better, and my thought was that quality is the key, that you can train babies’ eyes with good illustrations.
Another perhaps more unusual perspective, from a discussion of reviews of novels, can be found in these firm words from LibraryThing user Jennifer:
You know, whatever my opinion is, it is, and you can either agree with it or not agree with it. But, you know, most likely I’m not going to change it unless you give me a really good argument, and over a book; but it’s subjective, either you like it or you do not. Like, I cannot stand romances, I could not read one to save my soul, like I cannot, I just could not do it; and yet, some other people absolutely love them.
Other comments from Jennifer make it clear that she is willing to read and discuss other people’s opinions, values, and thoughts on novels, and engage in a form of partial translation, but based on her comment above it is unlikely that a full and complete understanding, and full coherence (let alone convergence), would occur.
4.3.5. Information Value
Information value plays a huge role in users’ individual and collective information behavior, in the existing and emergent communities they are part of, and in the roles that LibraryThing and Goodreads play in. A coded mention of information value in either an existing information world or an emergent information world occurred, on average, every 28 seconds in the interview data. This section cannot present all of these, but will provide an illustrative overview focusing on cases where information value is important to the roles played by LibraryThing or Goodreads in users’ individual and collective information behavior and in user communities. This includes coverage of the impact of existing information worlds on information values, the information values users shared with others, how these shared information values led to a sense of community and commonality, and conflicts and disagreements over information values.
18.104.22.168. Existing impacts
Many users felt that the existing information worlds that other users came from would impact on their information values of given books. For example, Goodreads user Kevin said that he could often sense on the site that
…if, you know, one of the people has religious views, another person does not, they might have different opinions about a book that either has a religious slant or, you know, lampoons a certain religion, or so on.
The views and values of outsiders played a role. Goodreads user Tanya mentioned that, in her interaction with an author where she needed to determine what one of his novels was about, she “was basing [her determination] on the description, and the description may be inaccurate, even if it is author provided.” This implies that authors may not value accurate descriptions of the subject of their books to provide one, or to work with an information professional who can give them. Given Tanya’s interaction, though, it appears they may value them enough to argue about them once they are made available by Goodreads and publishers.
There were a few cases where users appeared to value something that a group could provide them, but did not take advantage of it for various, most often pre-existing reasons. Most of these cases imply users’ existing judgments of information value can complicate the processes of coherence and convergence. Kevin provides another example from Goodreads Group I, a group for hobbyist writers and authors, which allowed members to post their writing up for critique and discussion. He said, “you know, it certainly would be something that would be worth pursing, if I had the time.” Elsewhere, he made it clear that he valued the group as a whole and the opinion of its members, and already posted some of his writing to another online community. Kevin seemed to feel there would still be too much of a commitment to make to getting his writing up in Group I to make it worth his while, despite clear convergence with the information world of the group and coherence with the worlds of some of its members.
22.214.171.124. Sense of community and commonality
Many other users related shared information values to the sense of community and commonality they felt. LibraryThing user Lindsey had a similar “happy” experience to that of Rachelle (in the previous section) realizing how the digital library and online communities could connect her with others who shared similar information values and interests about books:
You know, when I was reading books when I was a child, little did I know that there … would be such a thing as the Internet, and you could talk to people you do not know in real life and, it’s—[pauses]—I mean, it’s a happy surprise.
Miriam felt a sense of community, stating that
…very few people in my real life community have a passion for the things that I have a passion for… and it’s affirming, to know so many other people [via LibraryThing] who like the same things, and, we know that since we like the same things, we like each other too.
Sam seemed to share the same sense, focusing part of his interview on what he valued from LibraryThing from a discussion and community perspective. He said that “it allows people to make a little bit more of a choice of who they’re spending time with”; one could make fine-grained choices about who would be part of a group—especially a private group like the one Sam was pat of—than in a real-life venue such as a “tavern” (mentioned elsewhere in his interview). He struggled a bit to explain why he felt a strong sense of community in the private group, but when member-checked with the idea of “everybody knows your name” from the American sitcom Cheers, non-American citizen Sam seemed to agree with the general sense that there was a certain something, helped along by the shared and common values and interests members of the group had, that helped them feel part of a common, convergent site and community emerging from their use of LibraryThing. He tried to explain this process—of what might be termed translation on steroids—himself, and did not do too bad; see the quote given earlier in section 126.96.36.199 on translation.
LibraryThing user Melissa’s comments, referenced in section 188.8.131.52, on “deal[ing] with [differences] like a real friendship would be dealt with” are indicative of common and shared information values of a strong sense of community, despite any “differences you might come across” in other values. As she stated later in the interview, “some just want the connections to people who are interested in something they like” (although this is, of course, not true for everyone).
Ann shared an amusing view of this sense of differences in values being OK in a strong community:
No one in the [LibraryThing friendly fantasy fiction] group would ever, they would not slag you off at all; not even if you went on there and said that you loved Twilight and said it was the best thing ever written. No one would, sort of, jump on you and give you a kicking for that. They’d just go, “OK, well yes a lot of people like them.” [laughs] No one’s going to chat to you about it, but…
While the Twilight series would not be of value to this group, and they might share a social type of fans of the series in private, Ann’s comments indicate they are accepting that others’ information value judgments are different, and would not ostracize anyone from the community for such activity. This can be compared with the strong norms and values in some groups against unwanted content, such as advertising (see section 184.108.40.206); the latter is against explicit rules and may get a user kicked out of a group, but differences in information value judgments are not a rule violation, but a different normative view than the majority of a group. Later in her interview, after mentioning that “this is, yeah, I guess, real, real community” (see the discussion of norms in section 220.127.116.11), she observed again that the sense of community goes beyond having common interests:
…people are making connections above and beyond, you know, oh it’s nice that I have something in common with him, we’ve read the same book; you know, this is, it’s obviously meaning quite a lot to people.
18.104.22.168. Conflicts and disagreements
This provides a contrast to the examples of conflict and disagreements over values that were seen or reported by interviewees. Goodreads user Taneesha was asked what common connections there were between her and the other users she was interacting with in a group she was part of, and she said “yes, all [valued] the YA, the young adult book group,” but that she “cannot think of anything [as a common interest], apart from that.” Over the course of the interview, it became clear that she had little else in common in other phenomena (as seen elsewhere in this chapter), and focused on the one thing she valued most in Goodreads: using it “to find new books, and to discuss books,” nothing more. Her values and interests did show strong coherence with those of the group—with a shared interest in YA novels—and with Goodreads as a whole—as mentioned above with her love of reading—but strong convergence and emergence of a new information world that included Taneesha as a member did not seem to be occurring. She still placed clear value in Goodreads as a site, and felt that if it went away
…it would kind of suck, ‘cause you would not have … I mean, it’s kind of like a one-stop shop; you get to find books as well as find people who want to read those books or have read them.
As such, Taneesha’s degree of actual conflict with other group members was not high, and she could remain cohered with the group, but did not have the same level or kind of experience as Ann or Melissa. Of course, she seemed happy with and accepting of that, given her narrow focus on book reviews, discussions, and recommendations.
Jennifer felt there was not a match between her information values and those of others she interacted with (in a loose sense, since she was a rare user of the groups and focused on reviewing). She stated the following at two separate points in the interview:
I’ve found through these sites, you know, you can think you know people and then they’ll say something that’s, like, totally off the wall, and you realize you do not know them at all.
I like that I, you know, if I wanted to I could talk to other people who also enjoy books. And you would think that that would … you would think you would have more in common with them because they do like books, but I’ve found that’s not true. You know you do have that one thing in common … but a lot of times that’s the only thing you have in common.
While Jennifer, like Taneesha, seemed happy enough with her use of the digital library, she was a more frequent user of BookCrossing, another web site and online community that allowed her to give away books to others and track their progress. Her experiences there were of greater value, overall, and showed that any convergence for her on LibraryThing was not significant. (See section 4.3.8 for further discussion of sites for information behavior and activities.)
When there was actual conflict and upheaval within a group, thread, or community’s membership was where the clearest contrast was drawn with those groups sharing an understanding about information values. Betty’s relation of a disagreement over the information value of Isaac Asimov’s writing, and how a selection of people “exited en masse” from LibraryThing Group A because of the disagreement is the best example of such conflict. This occurred despite Betty—and perhaps others—feeling that those that left “actually were one of the people … [who] had the most interesting conversations about the books.” Once the conflict occurred, the group was no longer as useful or valuable for Betty and she “did not find as much to participate with in [Group A],” although she still follows it and checks in “once a day to once a week.” Conflict like this strengthens the argument that emphasizing the boundaries and barriers between users and user communities will lead to lower levels of coherence and convergence—i.e. less consistency, common understandings, and feelings of community—over time. Encouraging common, shared value judgments of information across a group, and accepting that values may differ for some information and at some levels, will lead to a stronger community and a greater role being played by the digital library in users’ information behavior and their communities.
4.3.6. Information Behavior and Activities
Information behavior and information-related activities were frequent mentions by interviewees, albeit not quite to the same level as information value. Participants discussed their individual information behavior and activities, behavior and activities they shared with others, examples of divergences from normative behavior, and how some information behavior related to community ties and information values. As with the messages analyzed during the content analysis phase, some examples in this section can be characterized as true normative information behavior, with a clear connection to information seeking, use, sharing, or avoidance. Other examples do not have such a clear connection, but still fall under the view of “information-based occupations or pursuits” being considered under the view of activities, drawing from the social worlds perspective. Emergent distinctions between these two and difficulties encountered in coding will be returned to in Chapter 5; as with section 22.214.171.124 earlier, this section treats information behavior and activities as one phenomenon, following the coding scheme and procedures established in Chapter 3.
Interviewees related their normative information behavior and activities when characterizing how they used LibraryThing or Goodreads and the reasons for their use. Some examples of this were included in section 126.96.36.199, such as Tanya’s posting of books to Goodreads for discussion, not cataloging purposes; and Lindsey’s investment in LibraryThing over other book sites. Another example was referenced in section 188.8.131.52: Miriam’s completion of a survey of the illustrators and illustrations of an early 20th-century children’s novel. Miriam stated she would often introduce herself when she joined a group, explaining what she was interested in and why she joined.
Kevin’s opportunity to post his writing to Goodreads Group I (mentioned in section 184.108.40.206) led him to relate differences in his normative information behavior on Goodreads compared with another site he used:
I guess the difference between this site and the other site I mentioned earlier, is that on the other site I’ve actually posted some of the things I’ve written, which I have yet to do on Goodreads.
Melissa referenced going to the library as part of her normative behavior, but stated that “it’s hard to find people who are into the same books that you’re into” by doing so, resulting in her normative use of LibraryThing as an additional space for finding connections. A further comment she made later in the interview stressed the nature of this normative behavior:
…I know that I just feel like this was, just, such a great group of people that I wanted to make sure that, every once in a while, I say, “Hey, still thinking of you! [chuckle] Come on if you get a moment.”
This comment shows Melissa was encouraging people, in a gentle way, who have not been around LibraryThing Group C in a bit to come back to the group and participate, if they can. Maintaining social ties within the context of the community, Melissa saw this as normative information behavior and activities for herself and as a role—a social type—that she played within the group, to help “get the energy back into the group,” as she put it (see discussion of Melissa’s roles within the group in section 4.3.4).
220.127.116.11. Divergences, digressions, and non-normative behavior
A notable set of examples of shared, normative information behavior and activities came from Melissa, a member of LibraryThing Group C. As observed from the content analysis, this group liked to play a lot of forum games, and Melissa confirmed this by reviewing many of the games and the simple rules that they took part under:
I mean, we have a word association game where pretty much you jut say what’s the, what the first thing is that comes into your head. And there’s a Narnia hangman which is just straightforward hangman. The movie quotes thread, kind of, it’s just, whatever quote you think would fit next, you write in. And sometimes you do not know what the person before you was thinking, but it’s OK, it does not matter!
These rules governed the expected normative information behavior within the thread for each game. Melissa mentioned something else:
…sometimes things go off on a tangent, sometimes a whole thread will get completely derailed because someone started going in a different direction entirely. And, and that’s OK too; I mean, it happens, and you can start a new thread for the game or you can just keep going with whatever it turns into.
In essence, it became normative within the emergent world of Group C that there would be non-normative information behavior, and in many ways that non-normative behavior became normative though the acceptance of divergences of topics and derailment of threads as something that would—and perhaps should—happen.
A different example relating to divergences in topic came from Miriam, who was a member of two LibraryThing groups and had started a thread in one of them—on pre–1950 literature—on the illustrations of editions of an early 20th-century children’s novel. We had the following exchange:
Miriam: Some of the people, a lot, the core people in this [pre–1950 literature] group also belong to another group that I’m heavily involved with, the [gardening group]. And sometimes conversations that belong in one place, you know, slip over into the other group."
Adam: Sure. Why would you say there’s such overlap there?
Miriam: Because we’re getting to know each other, you know, in a more rounded way. … Somebody in the [illustrations thread] asked me to post pictures of a gardening project, and, that would have been more appropriate for the gardening group.
In this case, the divergence was caused by there being three overlapping worlds—the two groups and the thread within one of them—consisting of many people who shared the same interests, social ties, and sense of normative information behavior. While as Miriam said the gardening topics were not normative for the illustrations thread, such overlapping divergence appeared to be accepted, and so to an extent became normative like divergences did in Melissa’s Group C.
Mention was made by Goodreads users of new topics coming up and diverging from existing threads, with new threads created by moderators to turn what was non-normative information behavior into normative information behavior. This will be discussed further below in section 4.3.8 on sites.
18.104.22.168. Community ties, everyday life information, and values
Those who felt a sense of community and commonality due to shared information values (see section 22.214.171.124), and who had often established social ties and connections within the communities they were part of, referenced their interactions with these users as being normative. Since such interactions were almost always mutual engagements, this should be considered information behavior or an information-related activity occurring within an emergent information world, albeit sometimes a world containing only two people. Sam provides a good example in the context of the private LibraryThing group he was part of:
…by now, he’s someone I’d welcome to my home, you know, after interacting with him for, I suppose, nearly three years now, something like that.
The welcoming style of interaction is an information-related activity that Sam and this other member engage in, and have done so for some time. At the same time, Sam did not feel the group shared common information needs, wants, or desires, but agreed that “the most important aspect of it is engaging in [shared] activities.” In member-checking, he agreed with my statement that “here it’s much more about the social activity and the information is very much secondary to that.” Sam’s statement also indicates a degree of shared social norms around which activities are appropriate to engage within; both the social norms and the information behavior and activities codes apply to his circumstances.
It became clear during Sam’s interview that, in the tight-knit community of the private LibraryThing group where Sam participated, his interactions fell within a broad definition of information behavior and activities, where the interactions could be about values, interests, types, information, or building social connections. These information-related activities (in most cases) and information behavior (in those cases with a stronger connection to information use or sharing) explained much about the role that LibraryThing and the group played in his life and in the emergent information world of the private group. As stated elsewhere in this section and chapter, normative information behaviors and activities often had relations with the other phenomena of interest; Sam’s experiences provide the clearest evidence of this.
Other interviewees discussed how community ties and connections related to the normative information behavior within a thread or group. Miriam, a LibraryThing user, stated:
You see people on a regular basis and talk about things that are, interest you, and the rest of your life does creep in; you know, people know when you’re having health issues, and financial issues, or whatever. … But, there’s an awful lot of emotional support when you allow that to be there.
This indicated that, at least in her experience being part of multiple groups, that exchanges of everyday life information might not be normative (in a narrower sense) for the topic of the groups in question, but it is normative for human interaction and for LibraryThing, providing for social and emotional support based on the community ties and connections that emerge over time.
Goodreads user Rachelle provided evidence of similar experiences in that digital library. Within a “random,” off-topic thread in a group on historical fiction she was part of,
…people are like, you know, like somebody a couple [of] weeks ago was like, “oh, I’m going out and the husband and I are building a chicken coop today.” … so it’s like, that thread lets you get to know the people …. So you know their book taste as well as some personal stuff that people talk about. You know, right around Christmas everybody was taking about what they’re doing with their families …
This is illustrative of the range of everyday life information sharing, focusing on community ties that went beyond a common interest in historical fiction, that was accepted as normative in this thread. While the content was not always the same—not everyone was building chicken coops or spending time with family—the type of information behavior and activities taking place was shared across the emergent world represented by the thread and—since Rachelle said many users posted in both this thread and others within the group—the group as a whole.
LibraryThing user Ann discussed one “friendly fantasy” group where there was “loads of, kind of irrelevant chit-chatty stuff about what people are doing at the weekend and, you know, blah-blah stuff,” such as talking about their pets or family. Melissa mentioned a “rants thread” in LibraryThing Group C where people could vent about and discuss upsetting situations. Both of these served a similar purpose to the “random” thread in Rachelle’s group. According to Melissa Group C took part in a “secret Santa” gift exchange every year, a selection of members had visited a traveling exhibit related to the Narnia series, and some members had shared knitting and other “crafts that [they] had done.” Ann mentioned a couple of community efforts in the “friendly fantasy” group, one where a user’s husband had a medical problem and “40 or 100” people set up “a collection, PayPal thing” for him, and another where “in Australia, someone, either in the fires or the floods, lost all their books, and a whole load of people were sending books in for them”; this latter case of a shared, information-related activity was also mentioned by fellow LibraryThing user Miriam. These were all examples of shared normative activities—at least for a period of time—that strengthened community ties and connections, with less emphasis on the normative topic (Narnia books or fantasy fiction, respectively).
Overall, fewer mentions of or allusions to organizations were made by interviewees than most of the phenomena of interest. More were made to existing organizations than those that emerged as part of new social and information worlds.
Many references to existing organizations were to LibraryThing and Goodreads themselves, while other mentions were of other organizations that they were part of or aware of, and their relation to LibraryThing or Goodreads. In many cases these were better described by considering the sites associated with each organization, and so much of this will be discussed in section 4.3.8 below instead. Many of the examples of references to organizations that did not involve a site (i.e. where an existing organization was coded for, but an existing or emergent site was not) were due to interviewees’ answers being split up by pauses or acknowledgements from myself as the interviewer, or in some cases—since at some level almost anything can be considered an information behavior or activity—the role of such a site for information behavior and activities being minimal and implicit. Nevertheless, a few of these examples are notable and worth mentioning:
Discussion of groups as organizations, with most focusing on how well or not they are run. These mentions went outside the scope of the information behavior and activities that occurred within groups’ threads.
LibraryThing user Betty mentioned that most of the group reads she had been part of were not organized by LibraryThing themselves, but within groups by ordinary members. “Like, the group reads, science fiction group reads that I was part of. The LibraryThing staff had nothing to do with that.” They did organize a group read that she discussed at some length in her interview, this being “the first time since I joined that I’ve seen them do something like that. I think it was a trial…”
References to the organizational features of LibraryThing or Goodreads; or to the activities of other organizations such as publishers, booksellers, or libraries; that did not involve much normative information behavior or activities within an information world. For example, much of Tanya’s examination of the metadata for an author’s book or of his profile was not normative, although it could be considered information behavior. Kevin referenced his book being released on Amazon and the traffic another author friend received to her book on that site, which only implicitly included any information behavior (the writing of the books in question). Miriam ordered “different copies” of a book “from my library system,” which did involve information behavior that could be considered normative for her: checking out books from her local libraries. This was, however, a case of a minimal role being played by LibraryThing as a site for her information behavior and activities. Its role was more significant as an organizational resource furthering Miriam’s activities.
Comments on other book organization web sites that did not fall under consideration of information behavior and activities occurring within a site of an existing or emergent social world, such as the views of LibraryThing held by Goodreads users Tanya (“LibraryThing has a very weak social networking component. It’s not in any way, shape, or form competition to Goodreads”) and Rachelle (“…from what I’ve heard that one actually costs money”).
Comments on other social media or social networking web sites that, again, did not fall under consideration of information behavior and activities occurring within a site of an existing or emergent social world. There was little of this, with the most notable occurrence being Kevin’s comments on Facebook as an organization with motives that are not in line with his (see the discussion of external technologies in 126.96.36.199 for a few more examples).
Cases where the presence of LibraryThing or Goodreads as an organization was much more prominent and important than any implied role as a site. For example, Rachelle mentioned one reason she wrote a review of each book she read and posted it on Goodreads was so she could look back “ten years from now … and go, ‘OK, well ten years ago I hated that book, or ten years ago I loved it….’” This displayed faith that Goodreads as an organization would still be around in ten years; its role as a site for the activity of reviewing was much more implicit. Other examples included discussion of Amazon’s recent acquisition of Goodreads and of the roles played by LibraryThing and Goodreads in the running of the groups.
A brief mention of LibraryThing in Tanya’s interview when I asked her a question about it instead of Goodreads—the digital library she was being interviewed about—before correcting myself. She said “actually I do use LibraryThing, but … I use it differently.” No elaboration and no discussion of its potential role as a site for information behavior and activities took place.
Fewer emergent organizations were mentioned in the interviews. Those that were included groups that had changed platforms or settings, transcending a particular environment and organized themselves as social worlds to further their activities. Tanya’s “most important group” had first existed on AOL, but AOL abandoned its group platform; as a result, she “was the one who recommended that this group come to Goodreads.” It did so, in the process establishing the group as an emergent organization independent of either AOL or Goodreads. Another group she was part of was discussing potential transitions off of Goodreads, engaging in a similar emergence process as it had become an organization independent of the platform. While the private LibraryThing group Sam was part of had not switched platforms, the social world associated with it had moved from a previous public group to the current private group. He held an unofficial role in recommending people to join the group or not, and someone else had taken the lead on moving the group from public to private in the first place, making it seem like an emergent, but informal, organization had formed to further the private group’s activities.
In other cases, interviewees treated their groups as emergent organizations in the language they used to describe them. Some users from Goodreads considered the groups’ level of independence from the Goodreads administrators, with “their own structure and their own rules” (Tanya), as an indication that they were emergent organizations in and of themselves. While Rachelle did not consider most of the groups she was part of to be organizations, there was the possible hint that one larger group might be because it might have “different threads that might create sub-communities.” Kevin kept his thoughts on a potential emergent organization out of the hobbyist author and writing Group I on Goodreads to an informal level, commenting that “it almost feels a bit like, you know, a social gathering … call it a virtual social gathering.” Social gatherings may have loose organization at times, but the simile implies a degree of emergence of Group I as an informal organization looking to further its activities. LibraryThing users raised other factors that could be indicative of an emergent organization, including
- a “resource-like” approach (Lindsey), where the group focused on providing information about given topics;
- “little subgroups [existing] within” a bigger community (Betty), similar to Rachelle’s ideas about one of her Goodreads groups;
- activities organized outside the group’s online boundaries, such as the secret Santa exchange related by Melissa in Group C or the collections of money and books Ann and Miriam mentioned were set up for particular members;
- pages or threads that introduce the group, its rules, and its members (as existed for Group C, per Melissa); and
- “connections above and beyond” reading the same books (Ann).
Along with social norms, information value, and information behavior, sites were one of the most frequent phenomena mentioned or implied in interviewees comments. Sites in both existing and emergent social and information worlds were discussed. Many interviewees related their perceptions of the sites they used to engage in information behavior and activities as existing or emergent communities, and a few cases of sites fell into a middle ground that could be considered both existing and emergent.
Many users made references to LibraryThing and Goodreads serving as sites for information behavior and activities, with this taking place within the existing social worlds of users of the two digital libraries. Taneesha felt the role of Goodreads was a vital one in facilitating her interactions:
I mean, I would not be a part of the group or I would not know these people if it did not provide different groups to be a part of.
Taneesha felt Goodreads, as a site encouraging information-based interactions around particular topics, was necessary to connect her to the group she was part of, talking about YA literature. She further clarified later in the interview (and as mentioned in section 188.8.131.52) that she saw Goodreads as “the web site to find new books, and to discuss books.” Tanya commented on the classification of books provided by Goodreads members, using Goodreads as the space to complete this task in. The set of people that she said had left LibraryThing and come to Goodreads to engage in social networking could be considered to be using the site for this activity within an existing social world they had begun to cohere with.
While most of Ann’s interactions could be considered to be within emergent social and information worlds (see the next subsection below), her group memberships on LibraryThing were quite broad in their distribution, and at one point during the interview she could not remember which group a particular discussion thread was in. From this it seems clear that she valued the digital library and the role it served across all the conversations she took part in to facilitate and support them, no matter which social or information worlds were at play.
A few users discussed other existing sites where interactions could take place. Kevin mentioned a different online community, a space where artists of all kinds could post their work and portfolios for comment and distribution, that he frequented. As stated above, he had posted “some of the things [he’d] written” on the other site, but had not done so on Goodreads, despite a group he was part of—Group I—providing that ability. Both sites could provide a space for the kind of information behavior he was engaging in, but for reasons that were left somewhat unclear, Kevin chose to post his work to the other online community and not to Group I. He had made a choice—based on objective reasons or subjective preference—which helped determine the role that Goodreads and Group I played in his information behavior and his experience within the community of Group I and the broader social world of hobbyist authors.
Jennifer brought up her use of “a site called BookCrossing where you basically give books away … and then you can track where they go.” As stated in section 184.108.40.206, she valued this site more than LibraryThing because she found the latter difficult to use and navigate. There was some crossover in the people she saw on each site, but LibraryThing’s role was not central, and it did not serve as her chosen site (in both social world and Web terms) for much of her interaction-focused information behavior and activities; most of her actual interaction about books was in BookCrossing’s discussion spaces, not LibraryThing. Jennifer did say if BookCrossing was to go away LibraryThing would be her next choice, although she “would end up going over there and probably bullying people into fixing the forums so they could be a little more user-friendly.”
Rachelle discussed the importance of a site for information behavior and activities fostering interaction:
I think [interactions] could move somewhere else, as long as the other site fostered that kind of interaction. Whenever the whole Amazon [as new owners of Goodreads], you know, deleting reviews thing happened, there were a lot of people that tried to see if they could move stuff onto BookLikes, I believe, is the web site? And … the thing with BookLikes is, as great as it is, it’s blog style. And with it, the interactions are not as much among your friends as it is among everybody that, you know, just subscribes to look at your blog.
Moving from Goodreads was something many users wanted to do, but Goodreads had advantages as a site for information-related activities and behaviors within existing and emergent communities. The blog style of BookLikes did not provide for the same interactions among friends as Goodreads; it would be a different social and information world, and not one that Rachelle and others might want to be a part of. She later characterized “some blogs [as] more of talking to people instead of with people,” further elaborating on the reason Goodreads could foster interaction better than she felt BookLikes would.
Tanya brought up moving the opposite way, to Goodreads, believing many former LibraryThing members had moved over to Goodreads “because they were interested in the social networking,” as a normative information behavior—incorporating information and knowledge sharing with other users—and information-related activity shared within an emergent information world. In her view, Goodreads as a site was more supportive of and would better facilitate this practice. As learned from other interviews there are users of each digital library who use it in this way, and users who have left Goodreads for LibraryThing and vice versa (or are thinking of doing so) for varied reasons.
As mentioned above, Taneesha felt Goodreads was the only place that could facilitate and foster the kind of discussions she was looking for around books. She felt “a Facebook page” would not lead to people being “as interested … and as active as they were” on Goodreads. Further comments she made on the technologies provided by the digital library (see section 4.3.9 below) were indicative of coherence around the site, but without having considered other sites—such as LibraryThing—that could also serve as a space for her chosen information behavior.
While Rachelle’s comments implied potential conflict in how BookLikes might support information behavior, and Tanya and Taneesha made personal arguments for why Goodreads supported information behavior better than other sites, Betty had direct experience with other users deciding that a LibraryThing group was not for them, due to the conflicts in social types and information values (see sections 220.127.116.11 and 18.104.22.168) that had led them to leave “en masse.” Both the new forum they joined (which was implied to not be on LibraryThing, but its location was never made clear) and the existing group served as social and information worlds where the users interacted with each other and engaged in information behavior and activities; while the break-up of the group was mostly about values and types, it lowered the degree of convergence and reduced the level of facilitation, as a site for information behavior and activities, the LibraryThing group could provide. Betty, for one, was now using the group less as a result.
Melissa made brief mention of the physical space of her local library, saying she goes there often but finds it hard to meet people who share her interests there. In her comment is the implicit notion of a library serving as a site for information behavior and activities, with looking for books of interest to oneself being a normative behavior for library-goers. However, Melissa had not found anyone with the same information wants and desires, only those sharing the same broad information-seeking task. LibraryThing was a more successful space and site for her; as mentioned earlier she found a group that turned into a convergent and emergent social and information world, facilitated by LibraryThing and the community of Group C.
When users in LibraryThing and Goodreads groups engaged in information behavior and activities, in many cases the group could be considered part of the digital library serving as an emergent space—an influential, standard boundary object—for this information behavior. The degree of the emergence varied from case to case, but most users found they were engaging in new behavior, in a new space, that would not have been possible without the LibraryThing or Goodreads group and web site bringing them together.
Rachelle’s discovery of sharing common values with others over “the same books that I read” (see section 22.214.171.124) led her to engage in shared information behavior with those users—“I can talk to them about it”—within the space of the groups she joined and participated in. One of those groups, deemed her “favorite … to interact in,” included “a great atmosphere”; she got to know a lot about the people in the group and what they value (as seen above), adding to the sense of an emergent social world wherein Goodreads and the group served a role as boundary objects, bringing users together.
Over on LibraryThing, Sam referred to the “exchange [of] ideas about a writer you like or something like that” as a potential behavior he might engage in, and in relation to this “the one group [and] LibraryThing’s role is to—the little extra I would say is that they provide a space where you can choose who to be with, in a sense.” Here Sam felt that a new space had been made from the group for particular information behavior with specific people; it was the combination of these that made it a successful role, in his view. Melissa, as mentioned above, found common information values (like Rachelle), information behavior, and activities to converge, facilitated by LibraryThing and Group C acting as a site for these to take place in the context of an emergent social and information world. Ann’s experience on LibraryThing was similar; she found four or five people who shared interests and values and who all participated in a thread about gender roles in fantasy fiction, with LibraryThing supporting this emergent social and information world.
Miriam’s use of LibraryThing was perhaps the most insightful case of a site being created for an emergent social and information world’s information behavior. Miriam had an interest in the illustrations from editions of a early 20th-century children’s book, and decided to create a thread as a place to organize her overall project and to discuss it with others. She related the process she went though as part of her interview:
So I started [the new thread] in the group that I’m most comfortable in, and then I went and invited a couple of people, that I know might not have been in that group, but [I] know that they like [the book]… and invited them to come and participate.
A bit later in the interview, she added:
And, I’d had some interactions with a couple of people who helped to set up series and change the Common Knowledge [metadata], and we’d talked back and forth about that sort of thing before. So I already knew there was an interest there.
After I asked if anyone else joined in, Miriam responded by saying:
It’s interesting because, other people mentioned it to other people, and would give a link to the thread, and so they found it … through other people talking about it. And joined the group, and either lurk and enjoy it or actually participate.
The thread that Miriam created became part of a emergent social and information world. It crossed existing group boundaries; brought in people from a pre-existing, emergent world from the past that had already shared some values and interests; and served as a emergent and influential space and site for information-related activities and behaviors related to the illustrations in editions of the book. This latter role made the thread a clear boundary object, with the processes of translation, coherence, and convergence on display through Miriam’s actions and the contributions of others.
While not at the same level as Miriam’s actions, back on Goodreads Rachelle recollected a few cases of the moderator of the historical fiction group she was part of creating new threads when new topics came up (as mentioned in section 126.96.36.199). This first began because tangents were possible in the main threads:
…we have the book threads but we do not want to go too much off on a tangent, so … she [one of the group moderators] created the [random / off-topic] thread.
The moderator’s activity did not stop there, as Rachelle continued:
And then sometimes when people do go off on tangents, she’ll create a thread. Like we ended up with a theater and opera thread, were we were talking about theater and opera [in the random thread] and [the moderator] said, “well maybe we should start a thread devoted to theater and opera if so many people want to talk about it!”
I asked Rachelle if this happened often, and she replied that
…it happens whenever anybody goes off on a tangent that seems like it could … like, if enough people are interested—like the theater and opera discussion started as a discussion between three different members, and then when the thread was started it was like all of a sudden, you know, seven or eight people started regularly posting on the theater and opera [thread]. And the mod’s like, “wow, a lot of people here like theater and opera,” you know! … It just kept going for so long, she’s like, “well I’ll just dedicate a space to it.”
A similar occurrence happened with an on-topic discussion of author Bernard Cornwell, which the moderator split off from group readings of his work so that there was a dedicated thread for discussion of this popular historical fiction author. In both cases, the moderator had created a site (in social world terms) for what had been non-normative information behavior and activities to become normative within, and where the group could facilitate and support it, with the thread serving as an emergent information world for discussion of theater and opera or Bernard Cornwell, respectively.
188.8.131.52. Perception of existing / emergent communities
A few users explicitly characterized—without prompting—the spaces where they engaged in information-related activities and behaviors as what were or could become emergent “communities,” including Goodreads user Tanya and LibraryThing users Lindsey and Ann. (See also the discussion on a sense of community in relation to information values in section 184.108.40.206.) In contrast, Goodreads user Taneesha and LibraryThing users Jennifer and Betty perceived themselves as interacting with those from multiple other existing social and information worlds when on the respective site. Although Betty saw LibraryThing as “one big community” when it came to “the way that you behave [and] the way people interact,” and Jennifer seemed to hold that view to an extent, the rest of their interviews implied they did not see it to be as coherent as others did. Goodreads user Kevin and LibraryThing user Melissa felt like they were part of one emergent social and information world, the group that they spent most of their time in. Melissa felt that the community of Group C had “splinters” when people had different interests and engaged in different threads. Along with Tanya, Lindsey, and Ann (all LibraryThing users), fellow LibraryThing user Miriam and Goodreads user Rachelle felt they were part of multiple emergent social and information worlds, given their participation in multiple groups and threads. Of this group, Rachelle did not see the digital library as a whole as an overarching community, while Miriam and Lindsey did, and Tanya and Ann expressed strong considerations of it as such. Miriam’s view of potentially overlapping and nested communities was elaborated on through her discussion of the divergences in topic that occurred within the groups and threads she was part of (see section 220.127.116.11). LibraryThing user Sam could not decide whether the private group he was in was best perceived as one convergent community (and emergent social and information world), or multiple communities (existing social and information worlds) that intersected with each other and were nested within a broader community (an emergent social and information world); analysis of his interview implies either is possible, but that substantial convergence existed within the emergent social and information world of the private group.
For Lindsey, when the topic of community first came up she said she “would not necessarily [have] thought of it that way, if you had not described it that way.” When asked later in the interview about the nature of the group she was in, she replied “it’s definitely a community,” albeit she maintained the distinction that it was “a virtual sense. It’s a virtual community, it’s not a, you know, a real life community.” Later still, she gave what could almost amount to a textbook definition of a site emerging, within an emergent social and information world: “…people found, I guess, other people that they shared interests with and started a group.” By attracting other people like Lindsey to their groups through shared, common information behaviors, the groups—and by extension LibraryThing—became emergent and influential spaces for those behaviors within an emerging social and information world.
Tanya told me she had “met some of these people in person,” implying that led the emergent social worlds of one or more of the groups she was part of had become a community to her. Ann was willing to call it “real community, rolling over into, very much into real life”; as discussed earlier in this chapter she valued and appreciated the ability to build connections beyond the common interest of reading a particular book, with this sense of community and discussions within the community facilitated through the groups that emerged from LibraryThing as sites for valued interactions, information behavior, and related activities.
18.104.22.168. Both? Importance of context
In a few cases throughout the interviews, users referred to sites beyond LibraryThing and Goodreads, but implied they were supporting the information behavior and activities of an emergent social and information world that used a group or thread in the digital library. For example, in the context of discussing the information behavior and activities of the private LibraryThing group he was part of, Sam mentioned that “I’ve never had a blog, but some, quite a few of the people have a blog … and they’re posting their reviews [there].” The blogs could be sites for the information behavior associated with reviewing, and considered to be being used as a new, local standard for the emergent world of the private group. An alternative view would perceive the blogs as sites for an existing social and information world that overlaps with the group and LibraryThing in some way, such as the world of hobbyist authors.
In another example, Betty had been a member of another online community, which was
…not actually book-related in nature, and we actually formed a book group read, at one point, and did something fairly similar to [what is done on LibraryThing]. Just sort of on our own. … [E]ven though the group started as, for a particular game, it branched out all over the place, including a group read which lasted for about a year, until people got, you know, people get busy and whatever.
In this case, within the context of the other online community the group read was a site for an emergent social world to engage in new, convergent information behavior, as it was not part of the original intent or purpose of the group. From the context of LibraryThing, however, this was an external site, and the world of its users —including Betty—would be seen as an existing information world that cohered (or not) with LibraryThing. A political discussion she mentioned that occurred in the other community was both within an emergent world and an existing world, depending on the context used. Betty mentioned later on that one could “start something”—meaning a discussion about books or a group read—on Facebook, if one had friends with similar interests; the word “start” implies an emergent world, and it would be in the context of Facebook, but not in the context of LibraryThing.
In the cases where something like this was seen in the interview data, I reasoned that the sites in question were external to LibraryThing or Goodreads, it was not clear that they should be considered in relation to an emergent world from the context of the two digital libraries, and my given definition of a site within an emergent world included the words “digital library” (which most external sites were not perceived as, although some could fit the definition being used in this study). This led to analysis as a site within an existing world, but was a theoretical wrinkle that will be discussed further in Chapter 5.
Most interviewees mentioned using technologies, in some form, as part of information behavior and activities; overall this phenomenon appeared with moderate frequency, and occurred in the context of both existing and emergent worlds. Many of these were talked about in earlier sections of this chapter, where the technology influenced sites, organizations, information behavior, information value, social types, and social norms. (The relationship and distinction between sites and technologies, first discussed in Chapters 2 and 3, will be returned to in Chapter 5.) This section provides an overview of users’ pre-existing and emergent technology use, beginning with their use of features of LibraryThing and Goodreads, then discussing their use of existing external technology, and finishing by examining their technology use in the context of emergent social and information worlds.
22.214.171.124. Existing use of LibraryThing and Goodreads
An important point in Taneesha’s comments about why Goodreads was the best site and technology (in social world terms) for supporting her information behavior (see also section 126.96.36.199 above) was that she felt both the organizational features (cataloging, search, and other traditional digital library features) and the discussions and groups (social, online community-like features) were necessary for Goodreads to support her information behavior. As mentioned in section 188.8.131.52, when I asked her what she would think if Goodreads only provided discussion boards, she replied that she would miss the “one-stop shop” provided for the purposes she has: to find books, review them, and read and discuss others’ reviews. In a similar vein, she did not want the discussion board features to go away and leave only the organizational features; both technology aspects were vital for Goodreads to be a facilitating and supportive site for her information-related activities and behaviors. While Facebook was mentioned, other web sites that would have the same features and technology—including LibraryThing—did not come up in the discussion.
In a similar vein to Taneesha, Lindsey used a technological feature provided by LibraryThing, the “members with your books” list, which shows “who has the most books—either weighed or raw—in common with you”; she sorted it from least popular to most popular so she could identify people with similar interests in “stranger books” as opposed to “To Kill A Mockingbird.” Miriam mentioned collaborative use of the technological feature of the “Common Knowledge,” the metadata and other material posted about a book that LibraryThing users could edit and contribute to (LibraryThing, 2013). Miriam also used LibraryThing’s technology to connect some books together that belonged in a series. On Goodreads, Betty related a critical incident of interacting with an author over why his book had been classified—in an inappropriate way, in his view—as modern history, which implicated the metadata provided on the site through the technological features of Goodreads.
Kevin, a hobbyist author, mentioned that he “might use some of the Goodreads campaigns to try and drum up some publicity” for his book; this would involve using the technological features Goodreads has put in place for authors to use to promote their books in various places within the digital library and on the web site. I asked the other hobbyist author, Sam, if he used any of the features LibraryThing provided for authors, but he did not, saying “I do not know what they really offer, but, if I did know at one time I’ve forgotten”; he added that my question “may spur [him] to look into it.”
Most, but not all interviewees used the organizational and cataloging features of their respective digital library with some frequency to manage their book collections, lists, and reviews. Goodreads users Tanya and Kevin and LibraryThing user Jennifer made particular use of these features, and a few of the long-time LibraryThing users had used the site when these features were all that were available, continuing to use them through to their interview date. LibraryThing user Ann’s use of the cataloging features had fallen off some in recent weeks, but LibraryThing user Sam was the only user who explicitly said he did not use these features except on rare occasions to facilitate his group membership.
The ability to flag messages that violated the explicit norms and rules of the site was mentioned as being used—and useful—by LibraryThing users Melissa and Miriam. Melissa and Lindsey addressed the technological limits that tended to have threads cap out at around “200 or so posts,” per Lindsey, with continuation threads created when that happened (see also section 184.108.40.206 on norms). Concerns over loading times—mentioned by Melissa—led to this practice, seen in the content analysis (see section 220.127.116.11).
Explicit mentions of the ability to “friend” users on LibraryThing and Goodreads were not as frequent as one might expect. Of the Goodreads users, Rachelle and Taneesha made many mentions of the word in social network and traditional contexts; Kevin used both contexts but focused on the latter usage; and Tanya made brief use of each a couple of times. LibraryThing users almost all used the word in the traditional context; all of Sam, Ann, Melissa, Miriam, and Jennifer‘s uses were in this way. Betty used the word in the context of Facebook twice, but did not use it to describe the ability to “friend” on LibraryThing. Lindsey did not use the word “friend” or stemmed variations of it at all during her interview.
18.104.22.168. Existing external technology
Technology beyond LibraryThing and Goodreads was mentioned in interviews, including other competing sites (by Betty, Rachelle, and Jennifer), Web search (explicitly by Sam and implicitly by Taneesha), e-mail (by Ann), blogs (by Sam, Rachelle, Ann, and Taneesha), Amazon.com (by Rachelle, Kevin, Taneesha, Ann, and Jennifer), and the Kindle app and store (by Kevin). Apple’s iPad was mentioned by Kevin, who used Goodreads on that device much more than any other, and by Jennifer (as owning one, but without further references to it).
Of these, Amazon merited some short discussion by interviewees in reference to seeing it in Web searches, purchasing books, and (from the authors) making their published books available there, but only blogs were discussed at much length. Taneesha, Sam, Rachelle, and Ann saw that blogs were of potential use within existing social and information worlds as an adjunct alongside LibraryThing or Goodreads. Taneesha and Ann focused on blogs as venues where people could post their opinions and let people comment. Taneesha mentioned “book bloggers” that would review and recommend books to their audiences, such that a blogger might become “that one go to person that [someone will] trust.” Ann talked about authors and writers’ blogs, and she stated “it does not necessarily get really interactive … I mean, yeah, things can develop, but yeah…” Sam’s references to blogs were discussed at some length in section 22.214.171.124, and will not be repeated here, but he was also concerned if “they get a lot of readers.” Rachelle expressed the strongest concerns of the three in saying that the blog technology in use by BookLikes—a similar site she was discussing—might not provide the same level of facilitation for interactions as the group structure of Goodreads, the digital library she used:
… if they had the structure of the site that would facilitate groups like that, then it could happen. But if it’s blog style … you’re not going to have groups like that, if all you’re doing is posting your reviews and books blog style.
While her values and opinions were a factor—“… some blogs are more of talking to people instead of with people,” she stated—when combined with the comments of Sam and Ann this indicates that blogs may not contribute to the coherence—the consistency and common underststanding—of existing communities and to interactions between people from those communities (to say nothing of the convergence and emergence of new communities). They may in some cases (as expressed in Taneesha’s view), but care is needed in considering their role.
Many social media services and online communities were mentioned, sometimes prompted by interview questions based on participants’ survey responses. Some of these mentions were complex in their invocation of the concepts of technologies and sites from the social worlds perspective, with some of users’ mentions and activities incorporating both concepts, while others only focused on technologies. This complexity will be returned to in Chapter 5. Mentions of social media services and online communities by interviewees included the following:
Facebook: Users referenced having been sent links via the site (Rachelle), having migrated a previous community to a private Facebook group (Betty), using it to interact with friends first made via a LibraryThing group (Melissa), the idea of the “like” (Ann), that is was lacking in structure to support book discussions (Rachelle, Ann, Taneesha), feeling overwhelmed by it (Melissa), using it for family (Ann), that they were not “a Facebook person” (Lindsey), and that they stayed away from it for privacy reasons (Kevin, Jennifer).
Twitter: Ann said she dislikes Twitter “because it’s just, yeah, that’s just shouting at the void, that’s not conversation”; she “follows a few people” but does not post herself. No other interviewees mentioned Twitter.
LinkedIn: Kevin mentioned that it was his go-to professional social network, and Goodreads might become his social network “for the other side of my life,” that where he is a hobbyist author. Betty had said she used LinkedIn on the survey and I included it when summarizing her survey responses to ask about her use of social media, but she did not mention it in her interview.
Pinterest: Miriam and Jennifer said they used it, Jennifer “a lot” as “her place to gather things so that I do not have to dust them, because I’ve downsized…” Miriam mentioned using it with her daughter to “pin pictures for each other, and we include comments to each other in the little descriptor place.” She elaborated, when asked if she could use Pinterest for other purposes, that she would be unable to complete much of her illustration project on there because “you cannot post to their board [directly], you have to post to a particular pin,” terming this a “primitive setup.” She had invited another Pinterest user with interests in illustration “to join LibraryThing, and she came! And said, ‘this is how I found you!’ [laugh]” Betty had said she used Pinterest on the survey and I again included it when summarizing her survey responses to ask about her use of social media, but she did not mention it in her interview.
Flickr: Melissa mentioned that “we used to post pictures to a Flickr account that we made up for [LibraryThing Group C],” but did not mention any other use. Ann did not mark Flickr on the survey, but said she “liked” it until “they kind of destroyed it a bit ago”; she still used it to make photos of her daughter available to her family in another country. No other interviewees mentioned Flickr.
Skype: Ann mentioned using this to keep in touch with family and friends at a distance, but not for book discussions. No other interviewees mentioned Skype.
WhatsApp: Ann made brief mention of this as an example of the kind of technology she saw as having potential use for book discussions, although “it would have to be a cast of people that you already knew, that you could sort of group together and throw conversations to.” Ann was the only interviewee to mention WhatsApp.
Digg / Reddit: Ann referenced the idea of “up-voting” content without naming a particular service or web site. It is present in multiple online services, but was popularized by the Digg and Reddit web sites. No interviewees mentioned these sites by name.
Other online communities: Betty mentioned a private Web-based forum; Kevin mentioned a hobby community he had been part of and a community for artists he still participated in; Ann mentioned the discussion boards of a massively open online course (MOOC).
There were fewer mentions of use of the two digital libraries providing emergent and standard technology, as used in the information behavior and activities of emergent social and information worlds. Some referenced specific technology, and will be discussed first; then general comments will be reviewed.
Tanya, Taneesha, and Rachelle mentioned that users of Goodreads can use the “shelves” feature to create classified lists of books, an emergent behavior that they could do “any way they like,” as Tanya put it. For members of the site, this becomes a semi-emergent social and information world where the technology is used as the standard means to create classifications, although the degree of emergence depends on how many people use a particular shelf as part of their information behavior and activities. As seen elsewhere in this chapter, authors can disagree with readers over this and Goodreads has begun enforcing certain rules and deleting shelves as a result; these activities may decrease the degree of emergence and of convergent use of the Goodreads-provided technology.
The use of profiles (discussed in section 126.96.36.199) by LibraryThing users Ann, Miriam, and Lindsey, and Goodreads user Tanya, invoked a technological feature that interviewees used to learn about other users who had posted to groups and threads. Because it encouraged establishing social ties and connections with other users, this feature could lead to an emergent social and information world being created; the technology served as an emergent, standard boundary object that could be used for doing so. Rachelle’s discussion of a moderator creating new posts as new topics came up (see section 188.8.131.52) showed use of the discussion board and threading technologies provided by Goodreads (in this case) to support new, emergent social and information worlds. Ann made brief mention of the groups feature and the ability to post private messages for other LibraryThing users as ways they had enabled interaction and information behavior in both existing and emergent social and information worlds. She referenced PayPal as a means to organize a collection for a member’s husband (see section 184.108.40.206), another example of technology being used to organize an activity for an emergent social and information world.
Melissa mentioned the ability to post pictures in LibraryThing messages as something that “people forgot how to” do, since “you have to remember all the code … in the midst of the thread.” She remembered that threads would sometimes get “too heavy with pictures, [and] some people [would] have a lot of trouble loading them.” She said she had not seen many pictures posted “in a while” in Group C, but pictures were common in her thread on book illustrations, which is not too surprising given the content of interest; she considered this “a really nice feature.” The pictures are a technological feature that could support information behavior and activities within an emergent social and information world, but it appears their use may have diminished over time, at least within one LibraryThing group.
Linking—as seen in the content analysis—is more common, although not mentioned much by interviewees. Ann mentioned that pages on LibraryThing for authors and reviews were linked to and many external links were made, all supporting interaction and information behavior within groups in an emergent fashion. Miriam mentioned the links that other users had sent to each other to inform them of the thread she had started (see section 220.127.116.11). She used the author and book pages to find cover images for the editions she was posting illustrations from. Miriam found a newer feature “on each work page that now tells if there are active discussions on that book,” and provided links to them, to be “really cool.” The ability to post pictures and links, and have her discussions and those of others connected automatically to the pages for books in the digital library, supported and facilitated her information behavior and her creation and strengthening of an emergent social and information world.
General comments referring to technology supporting emergent social and information worlds were made by interviewees. Some were quite simple; for example, Sam said LibraryThing had facilitated interaction “just by being available”—that is, having the technology present to interact online—and by “put[ting] a lot into their design.” Usability concerns caused the opposite reaction by Jennifer, who used the competing site BookCrossing more than LibraryThing because she found the latter hard to use and navigate (see sections 18.104.22.168 and 22.214.171.124). While Miriam showed clear appreciation of the LibraryThing site and had done a lot to create a community there, she felt having a place for discussion was “a given,” implying she may feel that its use of technology to provide this is easy.
Others had more complexity behind their thoughts. Rachelle stated the following as the reasons she felt Goodreads had been successful:
…it’s just they had, you know, the free site with really good structure that facilitated not just reviews of books, but also groups, where you could interact with others. They gave the readers freedom to pretty much do whatever they wanted with the structure, and that is the formula; then, you know, they’ve been successful at keeping people.
The structure that Rachelle referred to was not solely technological, but the technology of Goodreads played a big role in allowing for book reviews, groups, and interactions, and the perceived flexibility allowed for users to form many different groups. While the degree to which this technology has supported the emergence of these groups as new social or information worlds varies from group to group, Rachelle’s comment could explain a lot of the success Goodreads has had in recent years.
Melissa made a shorter, but similar comment about what technological features helped make LibraryThing work as well as it did:
But I know the idea of having your, having people set up discussion boards, based on their interests and being able to draw people in that share that interest, is really a feature I cannot see working in many other ways…
While the sense of what makes LibraryThing work in this comment goes beyond the technology to sites and information values, the technology is at the root of this: some programming and system configuration went into allowing users to create discussion boards, and it is how that feature is set up that Melissa feels has contributed to LibraryThing’s success.
4.3.10. Open Codes
Four phenomena not explicitly part of the codebook at the beginning of the study were seen in the interviews, to which open codes were applied as per the qualitative analysis procedures detailed in section 3.7. Three of these were seen in the content analysis phase: other boundary objects, boundary spanners, and outsiders. A fourth phenomenon, lifecycles, was seen in a few interviews.
126.96.36.199. Other boundary objects
Many other objects, not connected with LibraryThing or Goodreads, served as boundary objects and were identified by interviewees. These included:
Books were, by far, the most popular “other boundary object.” This included their classification, as in the incident Tanya raised where there were multiple views of what an author’s book was about and the genres it fell under. Books were explicit objects of common interest and discussion in many cases, with the social act of reading that LibraryThing and Goodreads provide increasing their status as boundary objects during their reading. This was most true in group reads. A few cases were made specific to a genre (e.g. Rachelle mentioned historical romances), series (Narnia, in LibraryThing Group C, as mentioned by Melissa), titles (e.g. Kevin’s second book, which had not been released), or editions (e.g. the multiple editions of the book from which Miriam was examining illustrations). Goodreads users Taneesha and Rachelle mentioned books in boundary object roles with some frequency in their interviews.
Writing was referenced as a boundary object a few times, most often by the two hobbyist authors Sam (a LibraryThing user) and Kevin (a Goodreads user).
A web comic strip, Rich Berlew’s Order of the Stick, and one of its characters became boundary objects in a discussion of social types in relation to gender that Ann mentioned in her interview.
The illustrations that Miriam was discussing and posting reviews of became a boundary object as other users responded to her reviews.
Book reviews became a boundary object in one case where Kevin mentioned the relative praise (or lack thereof) he had received on Goodreads and Amazon.com, and in another case where Jennifer believed her reviews would be used by a publisher. Reviews serve as a negotiation between author and reader; as part of that process, Kevin explained his reaction to the reviewers’ comments and opinions in the interview.
A library was a boundary object for Betty, as she placed herself “on the waiting list at a … local library” for a book; since the list was long, she had “actually kind of forgot about the book until the library notified they had it,” explaining why she did not begin a group read until many weeks after it had started.
The weather was treated as a boundary object in one case where Rachelle and other members of one of her groups were “talking about the extreme weather that we’ve been having” in different locations around the world.
Web search results were a boundary object in one case, when Taneesha referred to them in the process of “looking up a book” and finding a review or author page from the Goodreads site near the top of the results, often right before or after Amazon.com.
188.8.131.52. Boundary spanners
Few occurrences of a potential or confirmed boundary spanner were seen in the interviews. Rachelle explained that, in some situations, users might be part of two communities, but the degree of boundary spanning they perform could be low:
…like I said [the people in Goodreads Group G] have a lot more romance in there, and, you know, there are people that like both, the romance and the non-romance, and they’ll straddle the two groups. But they, you know, their participation in the groups will be, may be different as well, especially since [Group G] also has challenges and stuff like that …. So even if there’s an overlap between people, it’s not like the same people, you know, it’s only a [small] overlap, as far as I know.
Rachelle herself was a member of at least four groups, with “three or four of the same people,” and was knowledgeable about the norms, values, and behaviors of two of those groups; as such, she—and perhaps the others—could serve as a boundary spanner if called on to do so. Her comments indicate sufficient differences between groups that boundary spanning could be difficult, and she reflected elsewhere in her interview that Goodreads was not an overarching community. Nevertheless, she agreed that overlaps existed where a boundary spanner could sit themselves.
Betty uncovered similar overlaps between LibraryThing groups, but claimed there were not many people that crossed between the groups she was part of: she said “it’s a totally different group of people that I run into” in each one. In such a case a boundary spanner might find it difficult to bridge the overlap, but it is possible, as proved by two other LibraryThing users.
Lindsey sometimes acted as a boundary spanner, saying
… I might invite somebody I’ve come to know in another [LibraryThing] group because I think they’d be good for this group … and, I’d suggest that they read a couple of the threads and get an idea of what it’s like, and see if it’s something for them.
By serving as a bridge between the group and another individual, Lindsey furthered the emergence of the social and information world of the group and facilitated translation, coherence, and convergence between it and potential new members.
The other case of a boundary spanner was Miriam. In starting a thread on illustrations from an early 20th-century children’s novel, she went beyond one group’s membership to consider others who she knew valued illustrations or this particular novel, and invited them to join the thread she had created. Other users then passed the link around, serving as limited boundary spanners. Some who joined the thread ended up joining the group, which was a greater level of coherence and convergence than Miriam herself had perhaps hoped for. Her social type as a boundary spanner was evidenced in her comment that she would introduce herself and her interests if she joined a new group. By spanning these boundaries between individuals, threads, and groups, Miriam had helped facilitate the emergence of a new social and information world around her thread and, to an extent, the strengthening of the already emergent social and information world of the group the thread existed within.
A few mentions were made of three perceived outsiders to the social and information worlds of users of LibraryThing and Goodreads. These were publishers (mentioned by Tanya and Jennifer), family members (in the form of Jennifer’s husband, who made a brief interjection during her interview), and authors (from some perspectives). The latter, of course, was a source of much tension in Goodreads, and in notes made after Tanya’s interview I wrote that I perceived authors to be both insiders and outsiders from her perspective, which—from further analysis—does not appear to be far from the truth.
As analysis concluded, it became evident that a few examples of the lifecycles of social and information worlds were present in the data. While a longitudinal view was not examined or expected in this study, a quick review of these could be useful for planning future research. These examples included:
Tanya’s “most important group” had transitioned from AOL to Goodreads, at her recommendation. She termed this a “very successful” transition, with around 75% of members having moved with the group (“233 members” on Goodreads, vs. “300 on AOL”). Activities had changed to reduce in-joke use and special terminology unique to the group, to encourage new members to join from Goodreads (with “minor success with that”).
Kevin commented on Goodreads Group I that “it seemed to be, you know, a group of a few people, but the people change from time to time.” He mentioned another online community that he had “participated [in] for about eight or nine years, and then at one point the site … as we would say, ‘went south’… and eventually the site went down” a few years ago.
Sam discussed the history of the private group he was in and how “most people in the group … were in a different group before that sort of slowed down,” and so he knew most of the members “very well.” The other group had begun as a “very very lively” group when he joined it, but then
…at some point a couple of people who were … constant posters, disappeared. Some unannounced, and some not. And, it was clear in a couple of cases that, you know, people were angry, offended by something; there were a couple of quarrels. … But that group still exists, but, I stopped writing on it because so little was happening, other than, I mean, there were very esoteric readings that were still being done, but otherwise there was not a lot of contact.
Sam was then invited to join the current, private group, which he did not realize existed until he received the invitation. He hypothesized that “someone must have said, ‘oh hey, what about [Sam], … we [should] tell him, or ask him?’” The new group had many of the same people, but not all the same people.
Melissa stated that Facebook had
kind of [drawn] a lot of people out of [LibraryThing Group C] … usually they’re sharing a lot of their news on Facebook, and mainly we’ve kept [the group] up for playing games, book discussions, things like that."
She discussed another point earlier in Group C’s lifecycle:
Originally, I think someone else had started it because of a love of the Narnia series, and, that person eventually disappeared, I think when the group was kind of in, at its highest point.
Betty made mention that another online community she was part of had migrated over to Facebook in the last couple of years, but was “not as tight knit” or “as active anymore.” When I made brief mention of the concept of lifecycles to her, she discussed the case where people may leave one community and join another:
[I]f you’ve got people who are acquaintances … and you share something in common—I mean, over time people’s interests change, and they’ll go join a different community.
This chapter reviewed the findings from this research study, including discussion of the nature of phenomena observed in messages examined in the content analysis phase; participants’ responses to the survey; and the roles of phenomena in interviewees’ use of LibraryThing and Goodreads as individuals and as part of groups and communities. Chapter 5 will provide a discussion and synthesis of these findings across the three methods and in relation to the literature, with an eye to drawing conclusions and considering implications from the data for theory, research, design, and practice in digital libraries and social informatics.