Social media is increasingly used in higher education settings by researchers, students and institutions. Whether it is researchers conversing with other researchers, or universities seeking to communicate to a wider audience, social media platforms serve as a tools for users to communicate and increase visibility. Scholarly communication in social media and investigations about social media metrics is of increasing interest for scientometric researchers, and to the emergence of altmetrics. Less understood is the role of organizational characteristics in garnering social media visibility, through for instance liking and following mechanisms. In this study we aim to contribute to the understanding of the effect of specific social media use by investigating higher education institutions’ presence on Twitter. We investigate the possible connections between followers on Twitter and the use of Twitter and the organizational characteristics of the HEIs. We find that HEIs’ social media visibility on Twitter are only partly explained by social media use and that organizational characteristics also play a role in garnering these followers. Although, there is an advantage in garnering followers for those first adopters of Twitter. These findings emphasize the importance of considering a range of factors to understand impact online for organizations and HEIs in particular.
Selaginella: features, morphology ,anatomy and reproduction.
Drivers of higher education institutions’ visibility: a study of UK HEIs social media use vs. organizational characteristics
1. Julie M. Birkholz1,*, Marco Seeber1 & Kim Holmberg2
*Julie.Birkholz@UGent.be
1Centre for Higher Education Governance Ghent &
Research Unit for the Sociology of Education, Ghent University, Belgium
2Research Unit for the Sociology of Education, University of Turku, Finland
Drivers of higher education
institutions’ visibility:
a study of UK HEIs social media use
vs. organizational characteristics
2.
3. Higher education institutions
are increasingly using social
media platforms as tools to
communicate to prospective
and current students, alumni
and society at large
5. Core organizational attributes
matter in explaining online
communication; where
status, reputation and size
are important predictors of
hyperlink connections and
centrality (Seeber et al. 2012,
Lepori et al. 2013).
6. Is online visibility affected
by social media use or by
other organizational
characteristics?
What is the relative
contribution of
organizational
characteristics and social
media use in explaining
social media visibility (n of
followers on twitter)?
7. We investigate to what
extent the number of twitter
followers is predicted by the
use of Twitter and by the
organizational characteristics
of the Higher Education
Institutions (HEIs) in the UK.
8. Social media visibility can be explained by:
• Hypothesis 1: the social media use of the
organization
• Hypothesis 2: a HEIs organizational characteristics
related to organizational size, status and reputation
• Hypothesis 3: both the HEI’s social media use and
organizational characteristics
Hypotheses
9. Data about
137 UK HEIs
*European Micro Data dataset
(Eumida) - a database containing
the structural characteristics of
2,457 Higher Education institutions
in twenty-eight European countries
(Bonaccorsi et al. 2010; Eumida 2009).
*
1 2
10. Collected from Twitter profiles
Dependent: Social media visibility (Twitter)
• Number of followers
• Total number of tweets sent
• The number of users that the HEIs are
following as a measure of their activity
• Date of first tweet
• and also whether the HEIs use Twitter to
share general news or to reach out to
students specifically
Measures (1/3)
11. Independent: Organizational characteristics (Owen-
Smith & Powell 2008):
• size of the university (number of staff units
and undergraduate students)
• reputation in the core activities of research,
measured through the scientific productivity
and the research intensity, and teaching,
measured through the teaching burden
• status, as measured through the relational
centrality of the university in the system
Measures (2/3)
12. Control variables
• the discipline profile, as some disciplines
may attract more attention than others
because of the societal salience of the topics
addressed
• the geographical context, in terms of the
urban centrality of the city where the
university is located.
Measures (3/3)
13. Mean Median Maximum Minimum
Standard
Deviation
size - units of staff 2.001 1.665 9.498 68 1.675
size - undergraduate students 13.826 13.356 33.640 351 8.462
reputation - scientific productivity 274,66 72,50 1.828,00 0,00 389,03
reputation - research intensity 0,04 0,02 0,27 0,00 0,05
reputation - teaching burden 8,14 7,89 28,03 1,78 3,80
status - coreness 68 66 173 0 45
urban centrality 2,2 0,0 9,0 0,0 3,5
number of followers 17.189 15.900 46.200 1.233 10.085
number of tweets 6.792 5.598 19.000 300 4.220
days on twitter 1.918 2.019 2.644 305 342
number of following 1.312 832 12.700 107 1.506
Table 2. Variables’ descriptive statistics
14. Method: Negative binomial regression
We find that HEIs’ visibility on Twitter are only
partly explained by social media use and that
organizational characteristics also play a role in
explaining the social media visibility of HEIs. There is
also an early-adopter (of social media) advantage.
Results (1/6)
16. Table 4 - Negative Binomial regressions models
Estimate S.E. Pr(>|z|) Estimate S.E. Pr(>|z|) Estimate S.E. Pr(>|z|) Estimate S.E. Pr(>|z|)
Intercept 9,752 0,054 <2e-16 *** 8,862 0,089 <2e-16 *** 8,371 289,300 <2e-16 *** 7,671 0,230 <2e-16 ***
size - undergraduate students 0,000023 0,000007 0,0007*** 0,000018 0,000006 0,0035**
research intensity 2,774 1,088 0,01* 3,416 1,013 0,0007***
coreness 0,005 0,001 0,0002*** 0,005 0,001 0,0004***
Tweets 0,00004 0,00001 0,0003*** 0,00003 0,00001 0,0004***
days twitter 0,001 0,000 0,0003*** 0,00053 0,00011 0,000002***
orientation: news and events 0,296 0,113 0,009**
orientation:students -0,312 0,183 0,09 .
Null deviance 145,96 252,25 183,82 304,75
Residual 142,49 142,25 144,15 141,37
AIC: 2898,6 2822,1 2871,3 2798,4
log-likelihood: -2894,6 -2812,1 -2859,3 -2784,4
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Empty Model Organizational variables Model Social media use Model Full Model
on 136 df
on 136 df
on 131 df
on 136 df
on 132 df
on 136 df
on 133 df
on 136 df
Results (3/6)
17. delta: standard
deviation
proportion in number of
followers
1 size - undergraduate students 8 462 1.16
2 research intensity 0.049 1.18
3 Status - coreness 45 1.24
4 Tweets 4 220 1.15
5 days twitter 342 1.20
Table 5 - Negative binomial regression model:
comparing the impact of the variables.
Results (4/6)
18. Results (5/6)
The overall predicting
capability of the full
model is fairly good.
Values below and above
the median of 15,900
Twitter followers (Full
model) were predicted
correctly for 79%
(positive predictive
value) and 92%
(sensitivity) of the cases
respectively. Figure 1. Results from Full model
19. Results (6/6)
Some outliers (based on number of followers) were
excluded from the initial test, but later tested:
• Open University (129,825 predicted vs. about 100,000 actual
followers)
• University of Oxford (60,180 vs. about 175,000)
• University of Cambridge (85,692 vs. about 151,000)
• London Business School (12,624 vs. about 69,800)
While Open University is well predicted by the model, the
other three are not. All four attract a larger number of
followers than predicted by the model.
20. Discussion (1/3)
The results emphasize the importance of considering a
range of organizational and social media use variables
in order to fully understand online visibility.
This topic is of
interest for
scientometrics, as
it is an additional
avenue from
bibliometrics to
evaluate potential
impact of a HEIs.
21. Discussion (2/3)
In addition…
The results advances our
understanding of metrics
validity and sheds light on
the practical questions of
how organizations can garner
visibility online, but also of
the limitations to do that, as
organizational characteristics
strongly affect a HEIs visibility
potential online.
22. Discussion (3/3)
Future research could be
interested to investigate how
the content of tweets may
influence online visibility
The existence of a few outliers
suggest that few actors attract a
disproportionally high attention
from the public. Future research
may investigate why this occurs.
23. İlginiz için teşekkür ederiz
Kim Holmberg
kim.j.holmberg@utu.fi
http://kimholmberg.fi
@kholmber
24. Natural
Sciences
Engineering &
Technology
Medical
Sciences
Agricultural
Sciences
Social
Sciences
Humanities
Rotated Component Matrixa
Component
1 2 3
sci ,657 -,095 -,131
eng ,554 ,309 ,239
med ,456 -,688 -,271
agr ,021 -,053 ,942
sosc ,228 ,806 -,221
hum -,975 -,020 -,047
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
a. Rotation converged in 6 iterations.
We run a factor analysis on the UK sample discipline compositions
and find 3 factors.
In the EUMIDA dataset we have data on the share of
academic staff in six disciplinary areas:
The discipline profile
Hinweis der Redaktion
Social media is increasingly used in higher education settings by researchers, students and institutions. Whether it is researchers conversing with other researchers, or universities seeking to communicate to a wider audience, social media platforms serve as a tools for users to communicate and increase visibility. Scholarly communication in social media and investigations about social media metrics is of increasing interest for scientometric researchers, and to the emergence of altmetrics.
Higher education institutions (HEIs), in particular, are increasingly
using social media platforms as tools to communicate to prospective and current students,
alumni and society at large (Gibbs 2002; Helgesen 2008; Hemsley-Brown & Oplatka 2006).
Social media is increasingly used in higher education settings by researchers, students and institutions. Whether it is researchers conversing with other researchers, or universities seeking to communicate to a wider audience, social media platforms serve as a tools for users to communicate and increase visibility. Scholarly communication in social media and investigations about social media metrics is of increasing interest for scientometric researchers, and to the emergence of altmetrics.
Higher education institutions (HEIs), in particular, are increasingly
using social media platforms as tools to communicate to prospective and current students,
alumni and society at large (Gibbs 2002; Helgesen 2008; Hemsley-Brown & Oplatka 2006).
The delineation of this mechanism advances our understanding of metrics validity and sheds light on the practical questions of how organizations can garner visibility online.