INTRODUCTION: It is imperative that scholars investigate social media sites because their policies, configurations, designs, and affordances (the perception of functional attributes of objects by an agent in its environment (Gibson, 1977)) are constantly evolving in order to meet the needs of both their users and investors. The focus of this work is to explore the way in which scholars make use of affordances differently when they are creating personal and professional (as categorized by Amazon Turkers) messages in the microblogging site Twitter. This is important because social media is having an impact on the once invisible backstage activity of scholars, as Priem (Priem, 2014, p. 264) argues, by bringing “the background of scholarship… out onto the [front] stage.”
Twitter, the 9th most visited website in the world, claims over 200 million active users who create over 400 million tweets each day (Wickre, 2013); it is a “global platform for public self-expression and conversation in real time” (Twitter Inc., 2013). Research has shown that approximately 10% to 30% of scholars (Priem, Costello, & Dzuba, 2011; Pscheida, Albrecht, Herbst, Minet, & Köhler, 2013; Rowlands, Nicholas, Russell, Canty, & Watkinson, 2011) make use of Twitter. This social media site presents a variety of unique affordances with which users can create tweets; these include the ‘@’ at symbol used to address messages to particular users (i.e. @user), the retweet feature (i.e., RT @user), the ‘#’ hashtag symbol (#ASIST) to group tweets and to search for tweets, the ability to add URLs, and the ability to add links to media (such as images and video). It is important to examine the way in which these affordances are used because they can influence the way in which the audience frames (Goffman, 1974) and interprets the tweet. This work is guided by the following questions:
• Which affordances are scholars using?
• Do personal or professional tweets vary regarding affordance use?
• To what extent do scholars use affordances?
• Does Twitter activity influence affordance use?
METHODS: The personal Twitter streams of 445 scholars were downloaded in May 2014 using the Twitter API. The sample was derived from a selection of assistant, associate, full and distinguished professors across eight disciplines (Physics, Biology, Chemistry, Computer Science, Philosophy, English, Sociology, and Anthropology) from 62 Association of American University member universities. In sum, the scholars published a total of 585,879 tweets from 2006 to 2014. The sample of collected tweets from May 2014 totals 289,934 and the amount of tweets retrieved per user ranged between 1 and 3,263. A random subset of 75,000 tweets was placed into Amazon’s Mechanical Turk (AMT) application and three Turkers categorized each tweet into one of four categories: personal, professional, non-English, and unknown.
Affordance Use Differences Between Personal and Professional Scholarly Tweets
1. Timothy D. Bowman, Ph.D. Candidate | 2014 ASIST SIG/MET Workshop, Seattle, WA, USA
2. CRC.EBSI.UMONTREAL.CA
WHAT ARE AFFORDANCES?
• affordance - derived from afford,
meaning to make available or provide
naturally (Merriam-Webster, n.d.)
• Gibson (1977) affordance is the
perception of functional attributes of
objects by an agent in its environment
• affordances can vary depending both
on the context (time & space) they are
observed and by the agent doing the
observing
Figure 5: Tree affordance to bird, person, monkey,
and squirrel
3. AFFORDANCES AND SOCIAL MEDIA
• groups gain experience in digital contexts with
affordances and norms develop that enable interaction
(Bradner, 1999)
• feedback loop of personal and social use of affordances
creates consistent behaviors (Chalmers, 2004)
• interaction is moving from space-time constraints to
affordance-based constraints (Hogan, 2008)
• architecture of a particular environment matters; social
media architecture is shaped by their affordances (boyd,
2010)
CRC.EBSI.UMONTREAL.CA
4. WHY CONSIDER “ALTMETRICS” OR “INFLUMETRICS” OR SIMPLY
“SOCIAL MEDIA METRICS”?
- “Altmetrics” is the measure of scholarly communication and
dissemination within social media contexts (Priem & Hemminger,
2010; Priem, Taraborelli, Groth & Neylon, 2010)
- Perhaps a better term is Influmetrics (Rousseau & Ye, 2013) or
CRC.EBSI.UMONTREAL.CA
simply “social media metrics”?
- Social media indicators may measure immediate assessment of
academic impact and social impact (Thelwall, Haustein, Larivière
& Sugimoto, 2013)
- “Products,” not “publications” (Piwowar, 2013)
5. CRC.EBSI.UMONTREAL.CA
AFFORDANCES IN TWITTER
Twitter claims over 200 million active users who create over
400 million tweets each day (Wickre, 2013);
The four widely known affordances in Twitter are:
• @ mention– used to mention, direct messages at, and/or to
reply to user(s)
• # hashtag – used to contextualize or categorize the message
• URL link – used to connect tweet to another information
source
• ReTweet (RT) – used to resend another's tweet
6. SCHOLARS USING TWITTER
- 43% scholars at 2012 STI Conference used
Twitter; it was used privately and professionally,
to distribute professional information, and to
improve visibility (Haustein et al., 2013)
- 80% DH scholars ranked Twitter as relevant for
consumption and 73% for dissemination of DH
information (Bowman et al., 2013)
- differences by discipline found regarding the
way scholars used Twitter (Holmberg &
Thelwall, 2014)
CRC.EBSI.UMONTREAL.CA
7. CRC.EBSI.UMONTREAL.CA
RESEARCH QUESTIONS
1. Which affordances are scholars using?
2. Do personal or professional tweets vary
regarding affordance use?
3. To what extent do scholars use
affordances?
4. Does Twitter activity influence
affordance use?
8. CRC.EBSI.UMONTREAL.CA
PHASE ONE: SURVEY
- 16,862 scholars - associate, assistant, and full professors
from 62 AAU-member universities
- in physics, biology, chemistry, computer science, philosophy,
English, sociology, or anthropology departments
- 60 of the 62 universities rank in top 125 of 2014 CWTS Leiden
Ranking
- survey sent January and February 2014 with a response rate
of 8.5%
- 32% (613) reported having at least one Twitter account
- 289,934 tweets of 585,879 from 445 Twitter accounts (391
scholars) were found and harvested
9. PHASE ONE: 1,910 RESPONDENTS W/TWITTER ACCOUNT ARE:
33%
29%
40%
25%
29%
50%
28%
60%
50%
40%
30%
20%
10%
0%
American
Indian /
Native
American
(n=6)
Asian
(n=79)
Black /
African
American
(n=52)
Hispanic /
Latino
(n=40)
White /
Caucasian
(n=1580)
Pacific
Islander
(n=2)
Other
(n=50)
by ETHNICITY
38%
45%
38%
34% 36%
30%
27%
20%
16%
5%
2%
50%
40%
30%
20%
10%
0%
By SCHOLAR AGE
28% 28%
37% 37%
21%
50%
29%
24%
60%
50%
40%
30%
20%
10%
0%
by ACADEMIC DEPT
43%
36%
39%
41%
25%
40%
45%
40%
35%
30%
25%
20%
15%
10%
5%
0%
Less than 1
Year (n=68)
1 to 3 Years
(n=151)
4 to 6 Years
(n=144)
6 to 9 Years
(n=196)
10 Years of
More
(n=1262)
Not
Academic
(n=5)
by ACADEMIC AGE
10. PHASE ONE: WHO MAKES UP THE 613 ACCOUNT HOLDERS?
5%
10% 10%
15%
59%
Less than
1 Year
1 to 3
Years
4 to 6
Years
6 to 9
Years
10 Years
of More
7%
15%
5%
24%
17%
6%
10%
15%
Anthropology (n=49)
Biology (n=101)
Chemistry (n=35)
Computer Science (n=160)
42%
45%
40%
35%
30%
25%
20%
15%
10%
5%
22%
Personal Both Professional
35% 28%
19%
55%
44%
25%
33%
49%
39%
37% 60%
21%
25%
41%
24% 29% 26%
34%
22% 24% 31% 34%
by ACADEMIC DEPT
by PROFESSIONAL TITLE
by ACADEMIC AGE
SELF-REPORT
29% 29%
42%
0%
Assistant
Professor
Associate
Professor
Professor
11. PHASE ONE: MEAN TPD OF 391 SCHOLARS
by GENDER by DEPARTMENT
1.06
0.53
1.96
1.41
0.67
0.52
0.73
1.18
0.80
1.02
Other Female Male
N=232
SD=2.3
N=122
SD=2.1
N=3
0.89
1.11
1.39
0.67
0.85
I'm Not 10 Years
or More
7 to 9
Years
4 to 6
Years
1 to 3
Years
Less
than 1
Year
by ACADEMIC AGE
N=2
N=207
SD=2.4
N=53
SD=2.2
N=35
SD=2.6
N=39
SD=0.9
N=21
SD=1.1
by PROFESSIONAL TITLE
0.92
0.98
1.03
Professor Associate
Professor
Assistant
Professor
N=116
SD=2.1
N=116
SD=1.7
N=156
SD=2.9
12. PHASE TWO: CATEGORIZATION IN AMT
- scholars were divided into 10 groups based on their mean TPD
- stratified sample of 75,000 tweets from these 10 groups
GROUP 1: 0 < 0.5 | GROUP 2: 0.5 < 1 | GROUP 3: 1 < 1.5 | GROUP 4: 1.5 < 2 | GROUP 5: 2 < 2.5
GROUP 6: 2.5 < 3 | GROUP 7: 3 < 4 | GROUP 8: 4 < 5 | GROUP 9: 5 < 8 | GROUP 10: > 8
- six random tweets were combined with a control question for a total of
CRC.EBSI.UMONTREAL.CA
12,056 AMT HITs
- three turkers were asked to categorize each tweet as either:
Personal for example using incomplete thoughts/sentences, profanity, everyday
events/language, personal opinions, excessive punctuation, informal
Professional for example using academic/scientific/business language or subjects,
correct punctuation, mention job title, referencing professional
organization, formal
Unknown from the text it is impossible to categorize as personal or professional
Non-English the text is not written in English
15. PHASE TWO: PERSONAL & PROFESSIONAL TWEETS WITH AFFORDANCES
67%
69%
PARTIAL AGREEMENT (2/56%
AGREEMENT (3/3)
37%
28%
15% 17% 17%
66%
70%
60%
50%
40%
30%
20%
10%
0%
Mentions URLs Hashtags Retweets
Personal Tweets Professional Tweets
65%
38%
24%
30%
61% 62%
27%
38%
70%
60%
50%
40%
30%
20%
10%
0%
Mentions URLs Hashtags Retweets
Personal Professional
23%
AGREEMENT + PARTIAL
20% 22%
59%
65%
28%
38%
70%
60%
50%
40%
30%
20%
10%
0%
Mentions URLs Hashtags Retweets
Personal Professional
AGREEMENT
Personal Tweets: 27,264
Professional Tweets: 6,810
PARTIAL AGREEMENT
Personal Tweets: 19,403
Professional Tweets: 15,692
DISAGREEMENT
Personal Tweets: 942
Professional Tweets: 833
16. PHASE TWO: FREQUENCY OF AFFORDANCES USED IN PERSONAL & PROFESSIONAL TWEETS
1.38
AGREEMENT (3/3)
1.02
1.29
1.43
1.03
1.46
1.60
1.40
1.20
1.00
0.80
0.60
0.40
0.20
0.00
Mentions URLs Hashtags
Personal Professional
1.48
1.45 1.40
1.03
1.03
1.47
1.60
1.40
1.20
1.00
0.80
0.60
0.40
0.20
0.00
Mentions URLs Hashtags
Personal Professional
1.41
1.03
1.34
1.44
1.03
1.47
1.60
1.40
1.20
1.00
0.80
0.60
0.40
0.20
0.00
Mentions URLs Hashtags
Personal Professional
PARTIAL AGREEMENT (2/3)
AGREEMENT + PARTIAL AGREEMENT
Personal Tweets: 27,264
Professional Tweets: 6,810
PARTIAL AGREEMENT
Personal Tweets: 19,403
Professional Tweets: 15,692
DISAGREEMENT
Personal Tweets: 942
Professional Tweets: 833
17. PHASE TWO: FREQUENCY OF AFFORDANCE USE BY IV GROUP
0.6
0.5
0.4
0.3
0.2
0.1
0
GROUP 1: 0 < 0.5 | GROUP 2: 0.5 < 1 | GROUP 3: 1 < 1.5 | GROUP 4: 1.5 < 2 | GROUP 5: 2 < 2.5
GROUP 6: 2.5 < 3 | GROUP 7: 3 < 4 | GROUP 8: 4 < 5 | GROUP 9: 5 < 8 | GROUP 10: > 8
Hashtags
1 2 3 4 5 6 7 8 9 10
Personal
Professional
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
URLs
1 2 3 4 5 6 7 8 9 10
Personal
Professional
1.4
1.2
1
0.8
0.6
0.4
0.2
0
User Mentions
1 2 3 4 5 6 7 8 9 10
Personal
Professional
13%
11%
15%
7%
% Retweets
8% 8% 8% 8%
9%
15%
22%
10%
17%
9%
8%
9%
7%
5%
9%
5%
25%
20%
15%
10%
5%
0%
1 2 3 4 5 6 7 8 9 10
Personal
Professional
18. PHASE TWO: FREQUENCY OF AFFORDANCE USE BY IV GENDER
0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
Hashtags
Female Male
Professional
Personal
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
URLs
Female Male
Professional
Personal
0.95
0.9
0.85
0.8
0.75
0.7
User Mentions
Female Male
Professional
Personal
23%
% Retweets
68%
23%
66%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Female Male
Personal
Professional
19. PHASE TWO: FREQUENCY OF AFFORDANCE USE BY IV DEPARTMENT
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
Hashtags
1.00
0.90
0.80
0.70
0.60
0.50
0.40
0.30
Personal
Professional 0.00
0.20
0.10
URLs
Personal
Professional
1.40
1.20
1.00
0.80
0.60
0.40
0.20
0.00
User Mentions
Personal
Professional
7%
11%
2%
18%
33%
7% 7%
14%
5%
20%
2%
23%
20%
5%
7%
17%
40%
35%
30%
25%
20%
15%
10%
5%
0%
% Retweets
Personal
Professional
20. CRC.EBSI.UMONTREAL.CA
SUMMARY
• scholars are making use of all the primary affordances of Twitter and there does seem to be
consistency in their practices
• gender, department affiliation, communication type, and time spent on Twitter seems to have a
small impact on affordance use
• URL use is different in personal and professional tweets; there are many more professional
tweets with URLs, but the frequency of URLs used is similar between personal and professional
tweets
• #hashtag use shows variation by department for both personal and professional tweets;
• #hashtag use shows an upward trend as tweet activity increases for professional tweets and a
downward trend for personal tweets as tweet activity increases;
• #hashtag use shows variation by department for both personal and professional tweets;
• #hashtag use shows an upward trend as tweet activity increases for professional tweets and a
downward trend for personal tweets as tweet activity increases;
• @user mentions vary by gender with males using much less mentions in professional tweets
than females
• @user mentions vary by gender with males using much less mentions in professional tweets
than females
21. CRC.EBSI.UMONTREAL.CA
ONGOING WORK
• validity for tweet categorization is being checked currently by
surveying 90 most active scholars using Twitter and asking
them to self-categorize their own tweets
• using linguistic tools, the text of 289,934 tweets will be used to
compare terms used in tweets with scholar’s article titles at the
level of the scholar and discipline
• social network analysis using mentions at the scholarly and
discipline levels
• analysis of particular affordance usage
23. REFERENCES
Bradner, E., Kellogg, W., & Erickson, T. (1999). The Adoption and Use
of “BABBLE”: A Field Study of Chat in the Workplace. In ECSCW’99
(pp. 12–16). Copenhagen, Denmark: Kluwer Academic Publishers.
Retrieved from http://link.springer.com/chapter/10.1007/978-94-011-
4441-4_8
Bowman, T. D., Demarest, B., Weingart, S. B., Simpson, G. L.,
Lariviere, V., Thelwall, M., & Sugimoto, C. R. (2013). Mapping DH
through heterogeneous communicative practices. In Digital
Humanities 2013. Lincoln, NE.
danah boyd. (2010). "Social Network Sites as Networked Publics:
Affordances, Dynamics, and Implications." In Networked Self: Identity,
Community, and Culture on Social Network Sites (ed. Zizi
Papacharissi), pp. 39-58.
Chalmers, M. (2004). A Historical View of Context. Computer
Supported Cooperative Work (CSCW), 13(3-4), 223–247.
doi:10.1007/s10606-004-2802-8
Gibson, J. J. (1977). The Theory of Affordances. In R. Shaw & J.
Bransford (Eds.), Perceiving, Acting, and Knowing: Toward an
Ecological Psychology (pp. 127–143). Lawrence Erlbaum.
Haustein, S., Peters, I., Bar-Ilan, J., Priem, J., Shema, H., &
Terliesner, J. (2013). Coverage and adoption of altmetrics sources in
the bibliometric community. arXiv, 1–12. Digital Libraries. Retrieved
from http://arxiv.org/abs/1304.7300
Holmberg, K., & Thelwall, M. (2014). Disciplinary differences in Twitter
scholarly communication. Scientometrics. doi:10.1007/s11192-014-
1229-3
Letierce, J., Passant, A., Decker, S., & Breslin, J. G. (2010).
Understanding how Twitter is used to spread scientific messages. In
Web Science Conference. Raleigh, NC.
Merriam-Webster. (n.d.). Afford- Definition and More from the Free
Merriam-Webster Dictionary. In Free Merriam-Webster Dictionary.
Merriam-Webster: An Encyclopedia Britannica Company. Retrieved
from http://www.merriam-webster.com/dictionary/afford
Moran, M., Seaman, J., & Tinti-Kane, H. (2011). Teaching, learning,
and sharing: How today’s higher education faculty use social media.
Piwowar, H. (2013). Altmetrics: Value all research products. Nature,
493(159). doi:10.1038/493159a
Priem J., & Hemminger B.M. (2010) Scientometrics 2.0: Toward new
metrics of scholarly impact on the social web. First Monday 15.
Available:
http://firstmonday.org/htbin/cgiwrap/bin /ojs/index.php/fm/article/view/2
874/257. Accessed 2011 December 7.
Priem, J., Taraborelli, D., Groth, P., Neylon, C. Alt-metrics: a
manifesto. 2010. Available from http://altmetrics.org/manifesto/
Priem, J. (2014). Altmetrics. In B. Cronin & C. R. Sugimoto (Eds.),
Beyond bibliometrics: Harnessing multidimensional indicators of
scholarly impact (pp. 263–288). Cambridge, Mass.: MIT Press.
Rousseau, R., & Ye, F. (2013). A multi-metric approach for research
evaluation. Chinese Science Bulletin, 58(3290), 1–7.
doi:10.1007/s11434-013-5939-3
Thelwall M., Haustein S., Larivière V., Sugimoto, C.R. (2013) Do
Altmetrics Work? Twitter and Ten Other Social Web Services. PLoS
ONE 8(5): e64841. doi:10.1371/journal.pone.0064841
Wickre, K. (2013, March 21). Celebrating #Twitter7. Retrieved
September 01, 2013, from https://blog.twitter.com/2013/celebrating-twitter7
24. APPENDIX: 62 AAU-MEMBER UNIVERSITIES
CRC.EBSI.UMONTREAL.CA
Boston University, Brandeis University,
Brown University, California Institute of
Technology, Carnegie Mellon
University, Case Western Reserve
University, Columbia University,
Cornell, Duke University, Emory
University, Georgia Institute of
Technology, Harvard, Indiana
University, Iowa State, Johns Hopkins,
McGill, Michigan State University, MIT,
New York University, Northwestern,
Princeton University, Purdue
University, Rice University, Rutgers,
The State University of New Jersey,
Stanford University, Stony Brook
University-State University of New
York, Texas A&M University, The Ohio
State University, The Pennsylvania
State University, The University of
Chicago, Tulane University, University
at Buffalo, The State University of New
York, University of Arizona, University
of California, Berkeley, University of
California, Davis, University of
California, Irvine, University of
California, Los Angeles, University of
California, San Diego, and University of
California, Santa Barbara ,The
University of Iowa, The University of
Kansas, The University of North
Carolina at Chapel Hill, The University
of Texas at Austin, The University of
Wisconsin-Madison, University of
Colorado Boulder, University of
Florida, University of Illinois at Urbana-
Champaign, University of Maryland,
University of Michigan, University of
Minnesota, University of Missouri-
Columbia, University of Oregon,
University of Pennsylvania, University
of Pittsburgh, University of Rochester,
University of Southern California,
University of Toronto, University of
Virginia, University of Washington,
Vanderbilt University, Washington
University in St. Louis, Yale University
25. APPENDIX: 10 GROUPS OF TWEETERS
Group Name Mean Tweets/Day Total Tweets Percentage Required Member Totals
TEN 8 to 24 29,064 10.02% 9
NINE 5 to 8 25,863 8.92% 8
EIGHT 4 to 5 19,321 6.66% 6
SEVEN 3 to 4 24,532 8.46% 10
SIX 2.5 to 3 25,508 8.80% 10
FIVE 2 to 2.5 22,195 7.66% 10
FOUR 1.5 to 2 23,018 7.94% 13
THREE 1 to 1.5 43,831 15.12% 29
TWO 0.5 to 1 30,463 10.51% 33
ONE < 0.5 46,139 15.91% 317
289,934 100.00% = 75,000 445
Since some of you may be unfamiliar with the term affordance in this context, I wanted to start by explaining the concept.
Let’s start first with the term “afford” It is is a word in the English language that means to make available or provide naturally.
In 1977 J.J. Gibson, an ecological psychologist, challenged the more psychological view that humans simply perceive the qualities (color, height, weight, etc.) that make up the composition of objects;
instead, he believed that affordances (can it be thrown, lifted, rolled, pushed, used for X, etc) are also perceived. He believed that affordances are determined by the agent viewing the object, thus they become available when we see the object in a certain context.
He specifically argued that objects within a context can serve various functions and that these functions are dependent on the agent who views the object within this context.
Because of this dependency on both context and the agent observing, affordances can vary for any particular object.
Affordance is a useful concept to use when studying social media because social media is constantly evolving and the behaviors of its users change with experience.
This work will makes use of Gibson’s idea of affordances in order to examine the ways in which scholars utilize affordances in Twitter to frame communication.
Others have also examined how affordances are used in new media contexts.
Building off of Gibson’s ideas, Bradner (1999, p. 154) examined what he termed social affordances, defining them as “relationship[s] between the properties of an object and the social characteristics of a group that enable particular kinds of interaction”
Chalmers (2004, p. 233) described a social component to the act of tool appropriation in which an agent’s interpretation of the tool and its affordances, combined with reaction to others’ use and interpretation within the community, creates a feedback loop that establish norms for tool use.
Hogan (2008, p. 15) argued that “social life is moving from a focus on space-time social constraints to affordance-based social access.”
boyd (2010, p. 1) believed that “affordances… configure the environment in a way that shapes participants’ engagement. In essence, the architecture of a particular environment matters and the architecture of [social media] is shaped by their affordances.”
As other have discussed, the influx of metrics used to evaluate online contexts has led some to label them as altmetrics, a concept defined as “the measure of scholarly communication and dissemination within social media contexts”
It seems that altmetrics is a term that no longer serves to adequately explain what it is that we are measuring because these indicators are not measuring phenomenon alternative to something else such as citations or journal impact, but instead measure the traces of activity in the context of social media and other tools that were once either unavailable or invisible.
Instead I think of these as simply social media metrics.
One of the appeals of the measure of social media indicators is that it might provide immediate insight into academic and social impact; this has been compared to citations that both take a longer period of time to accumulate and only measure those who cite
Another reason social media metrics are important today is that organizations such as the National Science Foundation in the U.S. are stipulating that scholars submit a list of their “products,” not just a list of relevant “publications”, when applying for funding. This indicates that a scholar’s publications are no longer enough to determine productivity, impact and overall value.
These are just some of the reasons why social media metrics are an important and interesting area of research
Twitter, the 9th most visited website in the world, claims over 200 million active users who create over 400 million tweets each day (Wickre, 2013);
At this time there are four primary affordances including the user mention, hashtag, URL link, and retweet
A user mention is identified through the use of the at symbol
A hashtag is identified by a pound symbol
A URL is shortened and displayed in Twitter
A retweet is identified by the characters RT at the beginning of a tweet
Previous work has shown that somewhere between 10 and 30% of scholars have an account on Twitter.
-In examples of surveys of specific communities Haustein et al find that out of 71 scholars at the 2012 STI Conference, 43% reported using Twitter and that they used it privately, professionally, to distribute professional information, and to improve their visibility
Over 200 Digital Humanities scholars were surveyed with 80% reporting Twitter as relevant for consumption of DH and 73% reported it as relevant for dissemination of DH information (Bowman et al., 2013)
Finally, scholars from 10 different disciplines (astrophysics, biochemistry, digital humanities, economics, history of science, cheminformatics, cognitive science, drug discovery, social network analysis, and sociology) were analyzed and it was found that there were differences in the way they used Twitter (Holmberg & Thelwall, 2014)
Because affordance use in social media is important to examine because it influences the way users make use of social media and helps establish social norms
And because it has been shown that scholars use twitter for personal and professional communications
I was interested in examining the following four research questions:
Which affordances are scholars using?
Do personal or professional tweets vary regarding affordance use?
To what extent do scholars use affordances?
Does Twitter activity influence affordance use?
This work was carried out in two phases:
In the first phase I collected information of 16,862 Associate, Assistant, and Full professors from eight departmental webpages from 62 universities belonging to the Association of American Universities between September 2013 and January 2014.
The faculty belonged to either Physics, Biology, Chemistry, Computer Science, Philosophy, English, Sociology, or Anthropology.
According to the 2014 CWTS Leiden Ranking website that lists universities by scholarly impact, 60 of the 62 universities included in this sample rank in the top 125 of this ranking http://www.leidenranking.com/ranking/2014
A survey was sent to all of the faculty between January and February 2014 with a response rate of 8.5% (1,910 responses).
Of these respondents, 32% (613) reported having a Twitter account
Of the 613 scholars who reported having a Twitter account, 289,934 tweets of a possible 585,879 from 445 accounts were collected.
Note that the Twitter API restricts the collection of tweets to approximately 3,200 of the most recent tweets per account.
The missing 168 accounts were either private or could not be found.
There were 41 scholars with 2 accounts, 11 scholars with three account, and 1 scholar with 5 --- leaving 391 scholars
To begin let me briefly describe the demographic information of the scholars who reported having a Twitter account
There are no differences to be examined from ethnicity, but it seems that the likelihood of having a Twitter account as you age grows less
we see that when comparing respondents by academic age there is a big drop off after 9 years.
There were approximately just over half of the accounts from the natural sciences (%51.53) and just under half from the social sciences (%48.47).
Of the 615 scholars reporting having a Twitter account, 391 Scholars have 445 accounts
I calculated the mean of tweets per day by dividing the total number of tweets the users posted divided by the days since the scholar opened the Twitter account
When we look at the differences by gender, we see that males tweet on average slightly more than females
We see that social scientists tend to tweet on average more than scholars from the natural sciences
Interestingly, we see that scholars in their fourth to sixth years tweeting more than other academic age groups
Finally we see that assistant and associate professors tweet on average more than full professors.
Phase II involved the categorization of tweets using Amazon’s Mechanical Turk environment and qualified Turkers.
When examining the number of tweets by the scholars, it was clear there was a positive, long tail distribution with no clear separations in the data.
Because of this a stratified sampling technique was utilized to obtain tweets for phase II of the project.
Group distinctions were made based on the mean tweets per day calculation, with the groups being broken up at 0.5 Mean TPD intervals for the lowest six groups (<0.5, 0.5<1, 1<1.5, 1.5<2, 2<2.5, 2.5<3), followed by two groups representing 1 TPD interval (3<4, 4<5), one group representing 3 TPD intervals (5<8), and the final group containing everything above a specific threshold (> 8).
A total of 75,000 tweets were taken from the 10 groups and added to 12,056 AMT Human Intelligence Tasks (HITs) within the AMT application environment using a template created in HTML.
Three turkers who met specific criteria were allowed to perform the tasks for payment.
Turkers agreed on the categorization of 34,969 across four categories: personal, professional, non-english (766), and unknown (129)
Turkers partially agreed on the categorization of 37,355 tweets across the four categories: personal, professional, non-english (262), and unknown (1993)
Because of the low numbers of Non-English and Unknown categories and the focus of the research, I want to focus on the Personal and Professional tweets
Within the complete agreement set, we see that professional tweets tend to use more hashtags, many more URLs, and are more often retweets than the personal tweets.
As one might expect, the personal tweets have higher user mentions than the professional tweets, but the difference isn’t as high as expected
As we can see here, surprisingly there is a slightly higher frequency of mentions in professional tweets, even though there are more personal tweets with mentions in them
Otherwise, we see very similar frequency of URLs
And we see slightly more hashtags being used in professional tweets than personal tweets
When we look at affordance use by group, we don’t’ see much a difference with URLs
But when we look at user mentions, it seems that the frequency of user mentions increases in both personal and professional tweets as the scholar is more active on Twitter
Hashtags use seems to show a downward trend in personal tweets and is sporadic in professional tweets
With regards to retweets, we see an interesting jump in the amount of retweets in the most active group in personal tweets
We see a spike in professional retweets in the least active scholars
When we examine differences by gender, we see that females more frequently use user mentions in professional tweets than males
Hashtag use decreases slightly both in personal and professional tweets for males
Otherwise the trends are very similar for both URLs and retweets
When we look at differences by department, we see that chemists use hashtags more frequently in professional tweets than all other departments
Computer scientists and philosophers use hashtags less frequently in professional tweets than others
URL and user mentions seem to be used consistently across all departments
English scholars seem to retweet more as personal tweets than other departments