With the growing volume of user-generated classification systems arising from media
tagging-based platforms (such as Flickr and Tumblr) and the advent of new
crowdsourcing platforms for cultural heritage collections, determining the value and
usability of crowdsourced, “folksonomic,” or
user-generated, “freely chosen keywords”
 [21st Century Lexicon] for libraries, museums and other cultural heritage organizations becomes
increasingly essential. The present study builds on prior work investigating the
value and accuracy of folksonomies by: (1) demonstrating the benefit of
user-generated “tags” – or unregulated keywords typically meant
for personal organizational purposes – for facilitating item retrieval and (2)
assessing the accuracy of descriptive metadata generated via a game-based
crowdsourcing application. In this study, participants (N = 16) were first tasked
with finding a set of five images using a search index containing either a
combination of folksonomic and controlled vocabulary metadata or only controlled
vocabulary metadata. Data analysis revealed that participants in the folksonomic and
controlled vocabulary search condition were, on average, six times faster to search
for each image (M = 25.08 secs) compared to participants searching with
access only to controlled vocabulary metadata (M = 154.1 secs), and
successfully retrieved significantly more items overall. Following this search task,
all participants were asked to provide descriptive metadata for nine digital objects
by playing three separate single-player tagging games. Analysis showed that 88% of
participant-provided tags were judged to be accurate, and that both tagging patterns
and accuracy levels did not significantly differ between groups of professional
librarians and participants outside of the Library Science field. These findings
illustrate the value of folksonomies for enhancing item
“findability,” or the ease with which a patron can access
materials, and the ability of librarians and general users alike to contribute valid,
meaningful metadata. This could significantly impact the way libraries and other
cultural heritage organizations conceptualize the tasks of searching and

1. Introduction

Classification is a basic, integral and historically significant human function.
Defined by Golder and Huberman as an act “through which meaning emerges”
 [Golder and Huberman 2006, 200], the practice of classification represents one of the primary intellectual
foundations of library and information sciences. Useful classification systems both
accurately reflect the contents of a particular collection and allow for the
effective and efficient retrieval of items or information. Given the subjective
nature of human classification, examining meaning has no universal procedure. From
the birth of the library science field until the late 1800s, during which the first
edition of Dewey Decimal Classification was published in 1876, followed quickly by
the Cutter Expansive Classification System and the introduction of Library of
Congress Subject Headings [Stone 2000], all libraries and centers of
information each used an independent, unstandardized systems of organization,
otherwise know as “local classification.”
With the rise of such classifications as Dewey, Cutter and LC Classification, local
classification was mostly retired in the late 19th century in favor of a united
system that would allow for understanding across all libraries. This was, of course,
until the rise of Internet tagging-based platforms such as Flickr, Twitter, Tumblr,
and Delicious presented a challenge to these standardized, centralized systems.
Currently information professionals are facing an unprecedented amount of
unstructured classification. Classification systems generated by such tagging-based
platforms are referred to as “folksonomy,” or “a type of classification system for online content, created
by an individual user who tags information with freely chosen keywords; also,
the cooperation of a group of people to create such a classification
 [21st Century Lexicon].

While folksonomies represented an increased diversity of classification, they were
perceived mostly as sources of entertainment and documentations of casual
colloquialisms rather than a formal system of documentation. However, in 2006 Golder
and Huberman conducted a study of folksonomic data generated by users of the website
Delicious demonstrating that user-provided tags not only formed in coherent groups,
but also accurately described the basic elements of items that were tagged, such as
the “Who,”
“Why,” and “How” of items, proving their
ability to supplement formal records.

Golder and Huberman’s analysis also revealed two highly subjective categories that
may diminish the potential value of crowdsourced metadata: “Self Reference”
(i.e., any tag including the word “my,” such as “mydocuments”) and “Task
Organizing” (i.e., tags denoting actions such as “readlater” or
“sendtomom”). In examining the overall accuracy and reliability of tags,
Golder and Huberman concluded: “Because stable patterns emerge in tag proportions, minority
opinions can coexist alongside extremely popular ones without disrupting the
nearly stable consensus choices made by many users”
 [Golder and Huberman 2006, 199].
This research was expanded upon in 2007, when Noll and Meinel compared tags from
websites Delicious and Google against author-created metadata and found that the
former provided a more accurate representation of items’ overall content. Bischoff et
al. [2009] also examined folksonomic metadata within the context of the music
industry and found that tags submitted by users at the website Last.fm were
comparable with professional metadata being produced at AllMusic.com. Syn and Spring
[2009] examined folksonomic classifications within the domain of academic publishing,
and also found authoritative metadata to be lacking when compared with its
user-generated counterparts. As stated by Noll and Mienel, “tags provide additional information which is not directly
contained within a document. We therefore argue that integrating tagging
information can help to improve or augment document classification and
retrieval techniques and is worth further research.”
 [Noll and Meinel 2007, 186]. Together, these studies indicate that folksonomies can easily and usefully be
stored alongside classical, controlled vocabularies.
While some may have reservations regarding the “mixing” of
folksonomic and controlled vocabularies, these two classification systems need not be
viewed as mutually exclusive. Both systems have inherent advantages as well as
potential flaws. Controlled vocabularies are reliable and logically structured, but
can be somewhat inaccessible to the casual user [Maggio et al. 2009].
Furthermore, they can be time-consuming and expensive to produce and maintain. For
example, if one’s local library employs two catalogers at an average salary of
$58,960/year [American Library Association – Allied Professional Association 2008], it sets aside $117,920 for implementing
controlled vocabularies on a limited number of new collection items alone. In
contrast, folksonomies represent a relatively quicker and more cost-effective
alternative [Syn and Spring 2009], with greater public appeal and accessibility,
as evidenced by their overwhelming usage on social media. Mai noted: “Folksonomies have come about in part in reaction to the
perception that many classificatory structures represent an outdated worldview,
in part from the belief that since there is no single purpose, goal or activity
that unifies the universe of knowledge”
 [Mai 2011, 115]. As many users have become accustomed to the level of service and interaction
styles offered by current popular search engines, traditional searches are “unlikely to be very successful” and are becoming “less frequent as patrons’ behavior is shaped by keyword
search engines”
 [Antell and Huang 2008, 76]. As researcher Heather Hedden points out, if an artifact “[in today’s culture] if it cannot be found, it may as well
not exist”
 [Hedden 2008, 1]. However, their lack of centralization renders folksonomies prone to issues of
potential data contamination, such as an unorganized, unstructured plurality of
subjects and the likelihood of data duplication between users. Regardless,
folksonomies and traditional systems of organizations may be used in tandem to
address the shortcomings of their respective features, allowing for a more diverse
and organized form of classification.

Our research aimed to provide new empirical evidence supporting the value of
folksonomies by: (1) directly testing the benefits of adding user-generated
folksonomic metadata to a search index and (2) comparing the range and accuracy of
tags produced by library and information science professionals and non-professional
users. The three main questions guiding this work were:

  • RQ1: Will users exhibit reduced search times and greater hit rates when
    retrieving images with a search index that includes folksonomic metadata
    contributed by previous users?
  • RQ2: Will general users and information science professionals differ in the
    type and quality of metadata they provide in a free-form tagging game?
  • RQ3: Will users only provide information useful to them (e.g., Self Reference
    and Task Organization tags), or will they attempt to provide metadata that is
    useful on a wider scale?

2. Background and Overview of Present Research

The present research employed a hybrid form of usability testing utilizing the
Metadata Games platform [http://www.metadatagames.org], an online, free and open-source suite of
games supported by the National Endowment for the Humanities and developed by
Dartmouth College’s Tiltfactor Laboratory. The Metadata Games Project, launched in
January 2014, aims to use games to help augment digital records by collecting
metadata on image, audio, and film/video artifacts through gameplay [Flanagan et al. 2013]. Current Metadata Games media content partners include
Dartmouth College’s Rauner Special Collections Library, the British Library, the
Boston Public Library, the Sterling and Francine Clark Art Institute Library, UCLA,
and Clemson University’s Open Parks Network. Inspired by other successful
crowdsourcing efforts, the designers endeavored to create a diverse suite of games
that could enable the public to engage with cultural heritage institutions and their
digital collections, invite them to contribute knowledge to those collections, and
set the stage for users to create and discover new connections among material within
and across collections. The Metadata Games platform currently includes a palette of
games that cater to a variety of player interests, including both single-player and
multi-player games, competitive and cooperative games, and real-time and turn-based
games, available for browsers and/or mobile devices. Despite their variety, all the
games in the suite are united by a common purpose: to allow players to access media
items from a number of cultural heritage institutions’ collections and provide them
with the opportunity to contribute new metadata, in the form of single-word or
short-phrase tags, within the context of an immersive, enjoyable game experience. The
end result is that institutions benefit from increased engagement from a variety of
users and acquire a wider array of data about their media collections.
This research represents a collaborative project between the first author, who chose
to use the Metadata Games platform as the focus for an independent study project at
the Simmons College of Library and Information Sciences, and the co-authors from
Dartmouth College’s Tiltfactor Laboratory. To be clear, the goal of the reported
study was not to provide a validation of the Metadata Games platform, but rather to
study the value of folksonomic metadata more generally; that is, the focus of this
research was on the data itself, and the tool employed was intended to be largely
incidental and peripheral to the study’s aims. At the time, the Metadata Games
project was one of the few open-source metadata gathering tools available for
cultural heritage institutions.[1] Thus, while the reported study
results are specific to datasets gathered by the Metadata Games
platform, the conclusions drawn from this study are intended to be generalizable to
any organization currently making use of services such as CrowdAsk, LibraryThing or
Scripto or considering a crowdsourced metadata project.

3. Methods

According to Nielsen, the number of participants needed for a usability test to be
considered statistically relevant is five [Nielsen 2012b]; however,
because of the additional collection of quantitative data in our hybrid study,
sixteen individuals (eight men and eight women; six of whom were aged 18-24, eight
aged 25-44, one aged 45-60, and one over 60 years of age) were recruited to
participate individually in 30-40 minute sessions. In order to discern any
differential patterns of results between librarians and non-librarians, and to
separate out any potential advantage that users in the field of Library and
Information Sciences might have with content search and metadata generation, a mixed
sample (with nine participants recruited from LIS-related fields and seven from
non-LIS-related fields) was used for the study.
The study was divided into two main tasks. In the first task, participants were
presented with physical facsimiles of five images from the Leslie Jones Collection of
the Boston Public Library and instructed to retrieve these items using the Metadata
Games search platform. The images presented to participants were divided into the
following categories: Places, People (Recognizable), People (Unidentified or
Unrecognizable), Events/Actions, Miscellaneous Formats (Posters, Drawings,
Manuscripts etc.), as seen in Figures 1-5 below. Upon
being given each physical facsimile, participants were timed from the moment they
entered their first search term until the moment they clicked on the correct digital
item retrieved from the Metadata Games search platform. This practice was adapted to
reflect the digital items that would most commonly exist in a typical
humanities-based collection, (i.e., photographs, manuscripts, postcards, glass plate
negatives and other miscellanea). This design mirrored the common everyday occurrence
of patrons attempting to retrieve a specific media item that they have in mind when
using a library search index. According to a 2013 PEW Research Study, 82% of people
that used the library in the last 12 months did so looking for a specific book, DVD
or other resource [PEW Internet 2013].
For this image search component of the study, participants were randomly assigned to
one of two search index conditions: one with access to controlled vocabularies and
folksonomic metadata (i.e., the “folksonomy condition”) and the other with
restricted access only to controlled vocabularies (i.e., the “controlled vocabulary
condition”) [See Figure 6 for a schematic
representation of the study design]. Searches were conducted using two different,
customized versions of the search index on the Metadata Games website [play.metadatagames.org/search]. The
folksonomic metadata was generated by previous users of the Metadata Games platform,
whereas the controlled vocabularies attached to the items were generated by Boston
Public Library staff. The process of inputting the controlled vocabularies into both
versions of the search index required some reformatting. For example, because the
version of the search platform used in the study did not allow for special characters
such as the dash “-” or the comma “,”, terms such as “Boston Red Sox
(baseball team)” were imported as two individual tags: “Boston Red Sox”
and “baseball team.”

Figure 1. 

Search Image 1; Category: Places.

Figure 2. 

Search Image 2; Category: People (Recognizable).

Figure 3. 

Search Image 3; Category: People (Unrecognizable).

Figure 4. 

Search Image 4; Category: Events/Actions.

Figure 5. 

Search Image 5; Category: Misc. Formats.

Figure 6. 

Schematic representation of study design for Task 1 (Image Search).

Additionally, the system returned exact matches only, which meant that if a
participant searched for “sailboat” and the only term present in the system was
“sailboats,” the search would be unsuccessful. This aspect of the study
design was necessitated by the technical specifications and functionality of the
version of the Metadata Games search index used in the study, rather than a strategic
methodological choice. The frequency of preventable “exact match”
retrieval failures is discussed below in Section 4.1.

To further illustrate the differences between the two search index conditions,
consider the case of a participant in each condition attempting to retrieve image 3
(see Figure 3 above). In the controlled vocabulary
condition, the only search terms that would yield a successful retrieval were:
“marblehead harbor” and “glass negatives.” In contrast, in the folksonomy
condition, a participant would successfully retrieve this item by entering any of the
following search terms: “harbor,”
“sail boats,”
“marblehead harbor,”
“girl looking at boats,”
“glass negatives,”
“sailboats on water” and “yacht.”
Immediately following the image search task, participants were instructed to play
three different single-player tagging games from the Metadata Games suite: Zen Tag, NexTag, and Stupid Robot. In the “free-tagging” game
Zen Tag (Figure 7),
users are able to input as many tags as they wish for four separate images. NexTag (Figure 8), uses the
same game mechanic as Zen Tag, but utilizes a more
minimalist user interface and presents a more robust image to players. Finally, in
Stupid Robot (Figure
), a novice robot asks users to help it learn new words by tagging images. The
game presents one image to users and gives them two minutes to input tags, with the
constraint that they may only enter one tag for each given word length (i.e., one
four-letter word, one five-letter word, and so on). In playing a single session of
each game, participants in the present study tagged the same nine images (four images
each in Zen Tag and NexTag
and one in Stupid Robot). Tags from these three
games were compiled and compared against the traditional metadata provided by staff
from the Boston Public Library.

Figure 9. 

Stupid Robot.

3.1 Scoring

Tags were scored by the lead author using a revised version of the Voorbij and
Kipp scale used by Thomas et al. [Thomas et al. 2009]. This scale, which
was chosen due to its overall similarity to the Library of Congress Subject
Headings hierarchy, includes the following categories for scoring the level of
correspondence between a folksonomic tag and a tag included in the controlled
vocabulary for the same item:

  1. Exact match to controlled vocabulary
  2. Synonyms
  3. Broader terms
  4. Narrower terms
  5. Related terms
  6. Terms with an undefined relationship
  7. Terms that were not related at all
A score of one was thus reserved for an exact match between a user-provided tag
and the professional metadata, including punctuation. To illustrate, consider the
sample image provided in Figure 10 and the
corresponding professional and folksonomic metadata provided in Tables 1 and 2 below. With
this example, the user-provided term “Hindenburg Airship” would not be deemed
an exact match because, as indicated in Table 1, the controlled vocabulary term
encloses “Airship” in parentheses. Scores three through five were based on
judgments made by the Library of Congress in their subject heading hierarchy. For
example, “dog” would represent a broader term of the controlled vocabulary
term “Golden retriever,” whereas the tag “biology” would represent a
narrower term than the controlled vocabulary term “Science.” We reserved
“related terms” (a score of 5) for tags referring to objects or concepts
that were represented in the image but not expressed in the professional metadata.
A score of six was only to be awarded if, after research, the conclusion was made
that the term was unrelated to the image or any of the terms included in the
controlled vocabulary. A score of seven, though rare, was reserved for useless
“junk” tags, such as any term that was profane, explicit,
nonsensical or anything that would not qualify as useful to libraries (e.g.,
anything under the “Self Reference” or “Task Organization” classes
mentioned previously).

Figure 10. 

Hindenburg explodes. (Leslie Jones Collection; Boston Public Library)

Aircraft Accidents
Hindenburg (Airship)
1934-1956 (approx.)

Table 1. 

Professional Metadata for Figure 9.

Folksonomic Metadata Score (Voorbij and Kipp Scale) Notes
Hindenburg (Airship) 1 exact match to “Hindenburg (Airship)”
Hindenburg 2 synonym for “Hindenburg (Airship)”
Accidents 3 broader term of “Aircraft accidents”
Zeppelin 4 narrower term of “Airships”
Flames 5 Present in photograph; related to “Explosions”
Painting 6 this is a photograph
omgreadlater 7 junk tag

Table 2. 

Folksonomic Metadata and Scores from Voorbij and Kipp Scale for Figure

4. Findings and Analysis

4.1 Findability and Searching

Search Time. On average, participants in the
controlled vocabulary index condition took six times longer to search for each
image (M = 154.1 secs, SD = 98.84) compared to
participants in the combined index condition (M = 25.08 secs,
SD = 19.39) (see Figure 11 below). A
one-way Analysis of Variance (ANOVA) confirmed that this difference was a
statistically significant one, F (3, 15) = 3.94, p successfully
retrieved items in both conditions, an ANOVA confirmed that participants in the
controlled vocabulary condition still exhibited a significantly longer search time
per item (M = 111.36, SD = 89.24) compared to participants in the folksonomy
condition (M = 19.94, SD = 9.04), F (1, 12) = 8.51, p
Items Found.
Because participants were allowed to “give up” on finding
any particular item, each participant was assigned a numeric score from 0 to 5 to
represent the number of items they successfully found. A one-way
ANOVA showed that the average number of “found items” was
significantly higher in the folksonomy index condition (M = 4.88,
SD = .35) than in the controlled vocabulary index condition
(M = 1.38, SD = 1.06), F (1, 15) =
78.40, p Figure 12

Figure 11. 

Average Group Search Time (in secs).

Figure 12. 

Average Items Found per group.

When taking the failed searches of the controlled vocabulary condition into
account, there were a total of 131 completely “preventable”
failures overall. Importantly, a majority of these failures were due to the entry
of folksonomic synonyms for tags that were included in the controlled vocabulary
index, which clearly demonstrates the value of folksonomies for improving search
effectiveness and efficiency. Additionally, of the total 379 searches that
participants in the controlled vocabulary condition conducted, only a small
fraction of search failures (13 searches or 3.4%) were caused by the exact match
parameters set forth by the Metadata Games search system. Thus, the differential
levels of search time and item retrieval exhibited by participants in the two
search index conditions are primarily attributable to the specific tags that were
accessible in the two indexes – and not to the particular of the search system
used in the present study.

4.2 Tag Analysis

Recall that all tags generated by participants in the gameplay portion of the
study were coded using the Voorbij and Kipp scales; Figure
(below) depicts the breakdown of scores assigned to the 811 tags
generated by the participants.

Figure 13. 

Breakdown of overall Voorbij and Kipp scores.

As shown above, a score of five (“related terms”) accounts for the largest
segment of recorded tags, meaning that 50% of all of the tags entered were valid
classifications not included in traditional metadata. This implies a fundamental
semantic gap between traditional classification and folksonomies.

Figure 14. 

Distribution of Voorbij and Kipp scores for tags generated by LIS

Figure 15. 

Distribution of Voorbij and Kip scores for tags generated by non-LIS

As illustrated in Figures 14 and 15 (above), the distribution of Voorbij and Kipp
scores was constant and nearly identical between the LIS and non-LIS subsamples.
This suggests that, when given the same instructions, both librarians and
non-librarians can and do produce the same types of useful, accurate data.
Additionally, a score of seven, for a so-called “junk” tag, was
equally rare in both subsamples’ data. By comparing every participant’s percentage
of exact matches and synonyms versus undefined and unrelated terms (Figure 16 below), it is clear that most participants
(88%) show an inclination towards folksonomic correctness. It is worth noting that
the two highest scoring participants were a horticulturalist between the ages of
18-25 (participant 16) and a librarian between the ages of 45-60 (participant 14).

Figure 16. 

Score percentages for perfect matches and synonyms vs. undefined or
unrelated terms.

4.3 Best Subjects for Crowdsourcing

Another concern for cultural heritage institutions is determining what media
subjects might work best to collect new metadata through crowdsourcing. As
previously mentioned, the images that participants tagged in the present study
were divided into five key subject groups (see Figure

Figure 17. 

Average Search Time per Image Type (in seconds)

Results revealed that the images garnering the highest number of unique tags were
those that fell into the categories People (Unrecognizable) and Miscellaneous
Formats (in this case, a digitized newspaper). The image that generated the fewest
tags was Image 4, a picture of Thomas Edison, Harvey Firestone, and Henry Ford.
Few people recognized the inventors and many simply input tags such as “old
men,” although it is important to note that several participants did express
some level of familiarity with the figures in the image (e.g., one participant
uttered, “I feel like I should know this.”). These results suggest that the
best subjects for crowdsourced metadata might be media items that requires no
prior knowledge. For example, the unrecognizable person and the digitized
newspaper were some of the only instances in which the intent of the photograph
was either completely subjective (unrecognizable person) or objectively stated
(digitized newspaper). Many other images of famous historical figures and events
simply caused the participants to become frustrated with their own lack of
knowledge. In light of this fact, crowdsourcing platforms may be well-advised to
provide users with the tools to perform their own research about the content of
the media to fill in any gaps in knowledge or recollection that they experience.
This is a challenge that Metadata Games has begun to address, with the addition of
a Wikipedia search bar to encourage users to research what they do not know about
a particular media item.

5. Conclusions

As of now, there remains debate about the comparative value of traditional and
folksonomic metadata as organizational systems for today’s information needs.
Nonetheless, there is growing recognition of the fact that folksonomies offer
libraries with an ideal return-on-investment scenario [Syn and Spring 2009] with
minimum cost (much of which can be off-set by digital humanities grants), maximum
output of data [Bischoff et al. 2009] [Noll and Meinel 2007], as well as the chance to increase community engagement
with their patrons. As the findings of the present study demonstrate, folksonomic
metadata, when used in tandem with traditional metadata, increases findability,
corrects preventable search failures, and is by and large accurate. Furthermore, the
data suggest that given the same tagging conditions, librarians and non-librarians
produce a surprisingly similar distribution of useful metadata. Collectively, these
findings point to the potential to change the way we search for and organize our most
treasured media.


The research team would like to thank all study participants, Tom Blake, Linda
Gallant, the Boston Public Library, The Digital Public Library of America, Mary
Wilkins Jordan, Jeremy Guillette, and the National Endowment for the Humanities.

Works Cited

Antell and Huang 2008 Antell, K., and Huang, J.
(2008). Subject Searching Success Transaction Logs, Patron Perceptions, and
Implications for Library Instruction. Reference and User
Services Quarterly 48, no. 1 (Fall 2008):
68-76. Academic Search Complete,

Bischoff et al. 2009 Bischoff, K., Firan, C.S.,
Paiu, R., Nejdl, W., Laurier, C. and Sordo, M. (2009). Music Mood and Theme
Classification-a Hybrid Approach. In Proceedings of the 10th
International Society for Music Information Retrieval Conference (ISMIR
pages 657-662.

Golder and Huberman 2006 Golder, S., and Huberman, B.
(2006). Usage patterns of collaborative tagging systems. Journal
of Information Science, 32(2),

Grey and Hurko 2012 Grey, A., and Hurko, C. (2012). So
You Think You’re an Expert: Keyword Searching vs. Controlled Subject Headings. Codex:
The Journal of the Louisiana Chapter of the ACRL,
Retrieved March 1, 2014 from http://journal.acrlla.org/index.php/codex/article/view/47
Gross and Taylor 2005 Gross, T. and Taylor, A. (2005).
What have we got to lose? The effect of controlled vocabulary on keyword searching
results. College & Research Libraries, May 2005.
Retrieved May 1, 2014 from http://crl.acrl.org/content/66/3/212.full.pdf+html.

Heintz and Taraborelli 2010 Heintz, C., and
Taraborelli, D. (2010). Editorial: Folk Epistemology. The Cognitive Bases of
Epistemic Evaluation. Review of Philosophy and Psychology,
. doi:10.1007/s13164-010-0046-8

Heo-Lian Goh et al. 2011 Heo-Lian Goh, D., Ang, R.,
Sian Lee, C., and Y.K. Chua, A. (2011). Fight or unite: Investigating game genres for
image tagging. Journal of the American Society for Library
Science and Technology, 62(7),
1311–1324. doi:10.1002/asi.21478

Kipp 2006 Kipp, M. E. I. (2006). Complementary or
Discrete Contexts in Online Indexing: A Comparison of User, Creator and Intermediary
Keywords. Canadian Journal of Information and Library Science,

Lu et al. 2010 Lu, C., Park, J., Hu, X., and Song, I.-Y.
(2010). Metadata Effectiveness: A Comparison between User-Created Social Tags and
Author-Provided Metadata (pp. 1–10). IEEE.

Maggio et al. 2009 Maggio, L., Breshnahan, M., Flynn,
D., Harzbecker, J., Blanchard, M., and Ginn, D. (2009). A case study: using social
tagging to engage students in learning Medical Subject Headings. Journal of the Medical Library Association, 97(2), 77.

Moltedo et al. 2012 Moltedo, C., Astudillo, H., and
Mendoza, M. (2012). Tagging tagged images: on the impact of existing annotations on
image tagging (p. 3). ACM Press.

Noll and Meinel 2007 M. G. Noll, C. Meinel (2007).
Authors vs. Readers: A Comparative Study of Document Metadata and Content in the
World Wide Web. Proceedings of 7th International ACM Symposium
on Document Engineering, Winnipeg, Canada, August 2007,
pp. 177-186, ISBN

Oomen 2011 Oomen, J., and Aroyo, L. (2011).
Crowdsourcing in the Cultural Heritage Domain: Opportunities and Challenges. In
Proceedings of the 5th International Conference on
Communities and Technologies
(pp. 138–149). New York, NY, USA: ACM.

Syn and Spring 2009 Syn, S.Y. and Spring, M.B. (2009).
Tags as Keywords – Comparison of the Relative Quality of Tags and Keywords, In Proceedings of ASIS&T Annual Meeting, 46(1). November
6-11, 2009, Vancouver, BC, Canada.

Syn and Spring 2013 Syn, S.Y., and Spring, M.B. (2013).
Finding subject terms for classificatory metadata from user-generated social tags.
Journal of the American Society for Library Science and
Technology, 64(5),
964–980. doi:10.1002/asi.22804

Thomas et al. 2009 Thomas, M., Caudle, D. M., and
Schmitz, C. M. (2009). To tag or not to tag? Library Hi Tech,
411–434. doi:10.1108/07378830910988540

U.S. Constitution U.S. Constitution, pmbl.

Voorbij 1998 Voorbij, H. J. (1998). Title keywords
and subject descriptors: a comparison of subject search entries of books in the
humanities and social sciences. Journal of Documentation,
466–476. doi:10.1108/EUM0000000007178

DHQ: Digital Humanities Quarterly: By the People, For the People: Assessing the Value of Crowdsourced, User-Generated Metadata.

Source: DHQ: Digital Humanities Quarterly: By the People, For the People: Assessing the Value of Crowdsourced, User-Generated Metadata