MEASURING BOUNDARY SPANNING SUCCESS IN ENTERPRISE SOCIAL MEDIA
 By Ya-Wen Yu
      A THESIS Submitted to Michigan State University in partial fulfillment
 of the
 requirements for
 the
 degree of Media and Information Ñ Master of Arts 2015 ABSTRACT
 MEASURING BOUNDARY SPANNING SUCCESS IN ENTERPRISE SOCIAL MEDIA
 By Ya-Wen Yu
 The study aims to explore empirical proxies of boundary spanning success by developing 
measurement for 
hierarchical alignment, goal alignment, and time lag for three boundary 
spanning activities occurring in enterprise social media. 
The study employs Generalized Linear 
Models (GLMs) and Linear Models (LMs) to test the effect of distinct 
boundary
- spanning types 
and tie strength on each of the success measures. The findings show that 
strongly tied individuals 
who engage in information 
search activity can achieve boundary spanning success, as it is much 
easier for them to get the response in a short time compared to more weakly tied individuals. 
Trust as measured by symmetric (strong) tie thus can achieve boundary spanning success. 
Impli
cations for theory and practice are also discussed.
           !!"""!ACKNOWLEDGMENTS
  I completed my master thesis under the guidance from professor Wietske van Osch. IÕm 
so honored to have 
her full support and time. In addition, I appreciate the other two 
committee 
members, Dr. Charles Steinfield and Dr. Constantinos K. Coursaris, who provide me with the 
valuable suggestions to refine the thesis. The whole process of writing a thesis made me pursue 
the PhD study in the incoming semester.
 I spent two years s
tudying in the Department of Media and Information at Michigan State 
University. During the period, I have met knowledgeable professors, who enlightened the start in 
my research road. Moreover, I appreciate all the friends I have in East Lansing. Without t
hem, I 
could not have felt the warmth. Furthermore, there is no much for feeling thankful to my family. 
Without my parents, I would not have started the adventure of my American dream.
            !!"#!TABLE OF CONTENTS
 LIST OF TABLES
 .......................................................................................................................... v LIST OF FIGURES
 ....................................................................................................................... vi KEY TO SYMBOLS
 .................................................................................................................... vii
 CHAPTER I INTRODUCTION
 ..................................................................................................... 1 CHAPTER II LITERATURE REVIEW AND HYPOTHESES
 .................................................... 6 Boundary Spanning Type and Hierarchical Alignment
 .............................................................. 6 Boundary Spanning Type and Goal Alignment
 .......................................................................... 7 Tie Strength and Goal Alignment
 ............................................................................................... 8 Tie Strength and Time Lag
 ......................................................................................................... 9 CHAPTER III RESEARCH DESIGN
 .......................................................................................... 11 Case Organization
 ..................................................................................................................... 11 Data Collection
 ......................................................................................................................... 12 Operationalization
 ..................................................................................................................... 12 Dependent Variables
 ............................................................................................................. 12 Individual
-level Variable
 ...................................................................................................... 13 Dyadic
-level Variable
 ........................................................................................................... 14 Control Variable
 ................................................................................................................... 14 CHAPTER IV RESULTS
 ............................................................................................................. 16 CHAPTER V DISCUSSION
 ........................................................................................................ 19 CHAPTER VI IMPLICATIONS FOR RESEARCH AND PRACTICE
 ..................................... 21 CHAPTER VII LIMITATIONS AND FUTURE RESEARCH
 ................................................... 23 REFERENCES
 ............................................................................................................................. 24       !!#!LIST OF 
TABLES
  Table 1 The Variables of Research Model
ÉÉÉÉÉÉÉÉÉÉ........................................... 14 Table 2 The Summary of R
esults
ÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉ.......................... 18                                       !!#"!LIST OF FIGURES
  Figure 1 The Network Predictors of Boundary Spanning S
uccess
............................................... 10                                        !!#""!KEY TO SYMBOLS
  ESMÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉEnterprise Social Media
!!$!CHAPTER I INTRODUCTION
 !The study aims to investigate how trust and distinct boundary spanning activities influence 
boundary spanning success in 
Enterprise Social Media 
(hereafter as ESM). Such success refers to 
the targeted audiences reached, the effectiveness, and the efficiency for boundary spanning 
activity enacted by a poster and reciprocated by a responder. T
he study tests hierarchical 
alignment, goal alignment, and
 time lag as the proxies of boundary spanning success. Moreover, 
the study looks at tie strength
Ña network metric
Ñas a proxy of trust. Accordingly, the study 
would like to explore how tie strength and the type of boundary spanning activity influence 
hierar
chical alignment, goal alignment, and time lag to measure the boundary spanning success.
 Groups of ESM serve as open innovation communities. Such communities highly depend 
on social capital built by trust as the communities
Ñsimilar to communities of practi
ceÑexist 
independent of formal organizational norms and structures despite ESM being a corporate tool. 
Additionally, 
cooperation and coordination between individuals and groups inside the 
organizations at different levels and ways (Jones and George, 1998; 
Noteboom and Six, 2003; 
Stephenson, 2005). The various collaborative and coordination activities between distinct groups 
inside ESM could be described as intra
-organizational or team boundary spanning (
van Osch, 
Steinfield and Zhao, 2015)
. Studies on ESM s
uggested that such information communication 
technologies encourage boundary spanning through their affordances by rendering the 
interactions between individuals as well as the information and content shared more visible 
(Steinfield et al., 2009; 
van Osch 
and Steinfield, 2015a). 
Team boundary spanning refers to an 
activity that encourages effective organizational collaboration through maintaining and 
developing the relationship with external linkages 
beyond the team boundary
 (Ancona, 1990; 
Ancona and Caldwe
ll, 1992, 
Tushman, 1977). Additionally, boundary spanners serve as 
!!%!Òguardians who redirect crucial informationÓ from the external linkages (
Tushman and Scanlon, 
1981; Fleming and Waguespack, 2007, p.166)
.  As above mentioned, trust is an important antecede
nce to effective collaboration and 
coordination. 
Gulati (1994) has proposed tie strength as a proxy for trust. At the dyadic level, tie 
strength
Ñbroadly defined as the distinction between strong versus weak ties
Ñrefers to the 
frequency of interaction for a
 relationship between a knowledge seeker and source (Granovetter, 
1973; Hansen, 1999). 
ESM helps create opportunities for new social capital developed by 
enabling repeated social interactions, even between distinct member inside the organization, and 
thus 
may contribute to the development of strong ties which in turn is likely to increase 
individualsÕ willingness to contribute to ongoing boundary
-spanning activities on the system 
through enabling access to expertise and resources 
(Steinfield et al., 2009)
. Accordingly, trust
Ñas measured through
 tie strength based on the frequency of interaction between a poster and 
responder
Ñmay influence the likelihood and promptness of the boundary spanning interaction.
  Team boundary spanning, i.e., the management of exte
rnal linkages and interactions in 
co-located and virtual teams of ESM, aims to pro
cure resources such as Òmonetary, 
informational, or social
Ñas well as coordination and reputational benefitsÓ
 (van Osch & 
Steinfield, 2015
). Additionally, such team boundary 
spanning distinguishes such group
-level 
activities, namely representation, coordination, and information search (
Ancona and Caldwell, 
1992). Of the boundary spanning activities, representation activities in ESM refer to group 
membersÕ strategic actions for
 impression management, which aims to draw seniors amangersÕ 
attention. In other words, the targeted audiences for representation boundary
-spanning activities 
are the members with more influence and power in the groups or organization (Ancona and 
Caldwell,
 1992; van Osch, Steinfield, & Balogh, 2015). 
Accordingly, the 
posters tend to content 
!!&!aimed to reach the high
-level targeted audiences in an attempt to gain support, legitimization or 
access to important monetary team resources. Hence, part of the success
 of representational 
activities should be determined by the extent to which the posting is read and reciprocated by 
users that display a higher hierarchical level than the original poster. In other words, the 
hierarchical alignment in the interaction betwe
en a poster and responder 
  In addition, to the important of the hierarchical level of the responder, the extent to which 
the strategic goal of the original boundary spanning activity (i.e., posting) is reciprocated by a 
response of the same nature is anot
her measurement of success, which refers to goal alignment. 
Such goal alignment serves as a predictor of effectiveness for boundary spanning success. In the 
case of representational activities, the attempts by the poster to lobby up the corporate hierarchy
 need to yield responses by senior managers that indicate legitimization or the commitment of 
additional resources in order to be considered effective and thus successful. Given the open and 
distributed nature of ESM, questions asked by posters and subsequ
ent respondersÕ answers 
constitute an information exchange between a large amounts of distributed individuals 
(Kearsley 
1976; Ram 1991
; Wasko and Faraj 2005).
 Beyond the importance of theses metrics, this study 
further explores how trust 
measured by indivi
dualsÕ tie strength in dyadic relationships 
influences the hierarchical and goal alignment of postings for each of three boundary spanning 
activities
Ñrepresentation, information search, and coordination
Ñrespectively.
  For the second boundary
-spanning activ
ity, referred as information search, 
van Osch and 
Steinfield (2015) proposed in their study of the interaction between the postersÕ hierarchical 
positions and post type that general group members tend to engage in information search 
activities whereas the 
higher level leaders or managers are more likely to create representational 
postings. Additionally, 
information search and coordination boundary spanning activities in ESM 
!!'!aim to 
gain access to relevant information or knowledge and engage in task coordinat
ion 
(Ancona and Caldwell, 1992). These two horizontal boundaries spanning activities
Ñas opposed 
to representational activities
Ñrequire hierarchical alignment, as general members in the same 
level of hierarchical positions would be more likely to engage in 
the activities mentioned above 
rather than senior managers
. Such hierarchical alignment refers to the alignment for the 
hierarchical position between poster and responder. 
However, hierarchical alignment is not 
important for representation boundary spannin
g activities while the information search and 
coordination boundary spanning activities require higher hierarchical alignment. 
The study thus 
would like to explore how the type of boundary spanning activity influences hierarchical and 
goal alignment respec
tively in 
ESM in terms of their engagement in the representation, 
information search, and coordination boundary spanning activities.
  In addition to the targeted audience reached (i.e. hierarchical alignment) and 
effectiveness (i.e., goal alignment) for boundary spanning success, time lag is crucial as well. 
Time lag refers to 
the time between the creation of the original response and th
e comment by a 
responder to the original posting. Although hierarchical and goal alignment are proxies of the 
effectiveness, time lag could be viewed as a proxy for the efficiency of boundary spanning 
activity. Subsequently, the efficiency of response may 
be more critical depending on the urgency 
of the original posting though overall effectiveness is more likely to be important for boundary 
spanning success. Existing literature shows that 
strong ties result in faster information exchange
 due to the existen
ce of a common ground characterized by a shared language (Chiu et al., 2006) 
and mutual relationship heuristics
 (Uzzi, 1997).
 Accordingly, individuals could get quick 
response from the strong ties they connect with. Therefore, the study wants to investigat
e how tie 
strength influences time lag as one of predictors for boundary spanning success.
 The study 
!!(!leverages individualsÕ tie strength of dyadic relationships with others as a proxy for trust to 
measure the boundary spanning success through hierarchical 
and goal alignment, and time lag in 
the postings in terms of three distinct types of boundary spanning activities.
                     !!)!CHAPTER II LITERATURE REVIEW AND HYPOTHESES
 !The study examines boundary
-spanning success in ESM, measured through the 
hierarchical alignment, goal alignment, and time lag, based on a series of network predictors
. The network predictors are 
the type of boundary spanning activity and tie strength (See the Table 
1). Boundary Spanning Type and Hierarchical Alignment
 !Informat
ion search and coordination boundary spanning activities in ESM aim to 
gain 
access to relevant information or knowledge and engage in task coordination (Ancona and 
Caldwell, 1992). These two horizontal boundaries spanning activities require hierarchical 
alignment, as general members in the same level of hierarchical positions would be more likely 
to engage in the activities mentioned above rather than senior managers
. Such hierarchical 
alignment refers to the alignment of the hierarchical position between p
oster and responder. 
However, hierarchical alignment is not important for representation boundary spanning activities. 
Representation activities in ESM refer to group membersÕ strategic actions for impression 
management, which aims to draw the attention of
 those higher up the corporate ladder. Although, 
an acknowledgement by peers may be a intrinsic reward for the person engaging in a 
representational activity, it could never satisfy the goal of representation boundary spanning 
activities that aim for the l
egitimization or monetary support. So the poster expects the response 
given by someone who is at least one level higher. Therefore, this leads to the following 
hypotheses:
 H1: A posterÕs type of boundary spanning activities in ESM 
influences the alignment 
for 
hierarchical position between a poster and responder.
 !!*!H1a: A posterÕs representational 
boundary spanning activity
 in ESM negatively 
influences the 
alignment for hierarchical position between a poster and responder.
 H1b: A posterÕs 
coordination boundary
 spanning activities
 in ESM positively 
influences the 
alignment for hierarchical position between a poster and responder.
 H1c: A posterÕs 
information search
 boundary spanning activities
 in ESM positively 
influences 
the alignment for hierarchical position b
etween a poster and responder.
 Boundary Spanning Type and Goal Alignment
 !Van Osch and Steinfield (2015) 
investigate the interactions between 
the hierarchical 
positions of posters and boundary spanning activities that they engage in ESM by looking at the 
post type. The finding revealed that original post and the response had distinct objectives; for 
instance, an information search posting is met
 with acknowledgement or legitimization rather 
than an answer to the original question. 
Accordingly, 
this study will look at the goal alignment of 
postings provided between the parent posting and child response across the entire 
ESM in terms 
of their origi
nal objective defined by representation, information search, or coordination. For 
example, a poster publishes a post looking for a coordination opportunity but he or she gets a 
response in which the responder tries to brag about his or her achievements. Su
ch 
representational response would show a lack of alignment with the original posterÕs goal 
understanding the effectiveness of the boundary
-spanning activity in terms of goal alignment.
  On the other hand, if a poster publishes a posting looking for a repr
esentation or 
promotion opportunity, he or she receives a response from the responder who does not respond 
to his or her achievement and rather randomly gives an irrelevant information instead, in which 
two parties does not attain goal alignment. Specifica
lly, when posters ÒtailerÓ or ÒframeÓ their 
words in the postings to represent themselves, they tend to use ambiguous meaning of words to 
!!+!cover their aims. For instance, they would not use ÒachieveÓ but rather ÒshareÓ. However, such 
strategy of equivocalit
y does not lead to the desired outcome, as the equivocality suggests an 
implication for an attempt to deceive in nature. In other words, equivocality would result in the 
conflict between an aim and a outcome. Unlike the words used in the information or coo
rdination 
boundary spanning activities, the meaning of words is much more obvious and easy for a 
responder to interpret. Under high equivocality, the two parties would extensively exchange 
information to clarify the ambiguities till everything reaches the 
agreement (
Daft and Lengel, 
1986). Therefore, this leads to the following hypothesis:
 H2: 
A posterÕs type of boundary spanning activities that an individual engages in ESM
 influences whether or not the type of boundary spanning activity in which a poster o
r responder 
engages is identical.
 Tie Strength and Goal Alignment
 !If the goal of responder in the child posting matches the one of poster in the parent 
posting, the goal alignment between these two individuals would be achieved. In other words, 
such goal 
alignment implies the quality of response is positive. On the contrary, if the goals of a 
poster and responder do not match, the quality of response is negative in terms of achievement of 
his or her boundary spanning goal. 
From the perspective of BlauÕs so
cial exchange theory (1964), 
individuals are motivated to gain more expected benefits than exert efforts by contributing in the 
social interaction in ESM, such as information or knowledge sharing. 
In dyadic level, tie strength, 
strong versus weak tie, refe
rs to the frequency of interaction for a relationship between a 
knowledge seeker and source (Granovetter, 1973; Hansen, 1999). Individuals are more likely to 
gain trust that can safeguard against uncertainty 
with their s
trongly tied others (Krackhardt, 
1992). Individuals thus are more willing to engage in effective as well as efficient 
!!,!communication with strongly tied others 
due to the obligation of reciprocity 
(Hung
, Chen, and 
Lin, 
2009), which is crucial for coordination boundary spanning activities. 
Ther
efore, this leads 
to the following hypotheses:
 H3: The tie strength between a poster and responder in ESM influences whether or not the type 
of boundary spanning activity in which a poster or responder engages is identical.
 H3a: Strong tie strength between
 a poster and responder in ESM positively influences whether or 
not the type of boundary spanning activity in which a poster or responder engages is identical for 
the coordination 
boundary spanning activities
.  On the other hand, individuals are more likel
y to gain access to the sources of novel 
information with their weakly tied others (Granovetter, 1973), which is beneficial for engaging 
in 
information search 
boundary spanning activities
. Therefore, this leads to the following 
hypothesis:
 H3b: Weak tie between a poster and responder in ESM positively influences whether or not the 
type of boundary spanning activity in which a poster or responder engages is identical for 
information search 
boundary spanning activities
. Tie Strength and Time L
ag !Granovetter (1973) proposes that weak ties are the ideal resources for obtaining novel 
information. However, when it comes to the efficiency of information seeking via the questions 
posted on ESM, as measured through the time lag between post and respo
nse, is crucial for 
information exchange as well, in addition to the effectiveness of boundary spanning as measure 
through goal alignment. Specifically, the literature shows that strong relationship are more likely 
to result in faster responses due to the 
existence of a common ground characterized by a shared 
language (Chiu et al., 2006), which would facilitate faster information exchange. Additionally, 
!!$-!strong emotionally tied dyads are more motivated to help each other because of the relationship 
heuristic
s (Uzzi, 1997), which leads to trust. Accordingly, the study would like to explore the 
interaction between tie strength and response speed of answer as time lag. So this leads to the 
following hypothesis:
 H4: Tie strength between a post and responder in ES
M positively influences time lag, the 
duration of time when a responder to post the comment for a poster
  Figure 1 The Network Predictors of Boundary Spanning S
uccess
                ./01!01"234536
!7/8390:;!
<=03
3"32!6;=5
!>"5!
<6:532
6?!@"5:0:A?"A01!01"234536
!>"45
!102
!!!$$!CHAPTER III RESEARCH DESIGN
 !The study leverages individualsÕ tie strength of dyadic relationships with others as a 
proxy for trust to measure boundary spanning success via the network predictors, hierarchical 
alignment, goal alignment,
 and time lag in terms of three distinct types of boundary spanning 
activities. The case, data collection, operationalization, and data analysis are discussed in the 
following sections.
 Case Organization
 !This research is conducted in a 
worldwide provider 
(hereafter referred to as The Company)
 of technological products, furnishings, and research and consulting services to workplace offices. 
The Company has over 11,000 employees around the world. Its headquarters is located in the US 
with offices and divisio
ns in nearly 40 countries including Americas, Europe, Asia, Africa, 
Oceania, 
and the Middle East. 
The Company launched an enterprise social media system 
(hereafter: the
 System) built based on the Jive platform in March, 2012. Jive is one of the 
providers o
f corporate social technologies for supporting employeesÕ communication and 
collaboration activities as well as their business connections by providing a built
-in platform 
with the functionalities for group chat, blogging, and social bookmarking. The adopt
ion of the 
System has grown rapidly due to strong involvement by the executives of company.
 The usage of 
System is voluntary and the groups of System serve as open forums in which there are no 
explicit written policies or hard guidelines. Although the Syst
em is not normally treated as a 
project management tool since the Company has an official project management tool in place, it 
has been used as the platform for informal communication and social interactions within and 
between the groups of employees, whic
h is beneficial for the early stages of collaboration as well 
as the awareness of the organization for distributed collaboration. Moreover, the shared content 
!!$%!enables individuals to engage in inter
-unit and inter
-group connections by blog postings 
(referre
d as a posting).
 Data C
ollection
  The study examines the success of boundary spanning activities, measured through the 
hierarchical alignment, goal alignment, and time lag as explained in the theory section, based on 
the network predictors, the type of bou
ndary spanning activity and tie strength. Data for the 
analysis was collected from the system log data in which individual usersÕ group interactions 
within all group types (i.e. open, members only, private, and secret group) and outside of the 
groups are c
aptured. Since the complete population of groups and individual group members is 
used, no sample approach is used.
 Operationalization
 !The study explores trust in terms of individualsÕ tie strength to measure hierarchical 
alignment, goal alignment, and time lag for three boundary spanning activities, as the proxies of 
boundary spanning success. Data to test all dependent and independent v
ariables are collected 
using the system log data of the System.
 Dependent V
ariables
 !The study tests the hierarchical alignment, goal alignment, and time lag as the dependent 
variables. All three variables are operationalized at the dyadic level. The hierarchical alignment 
refers to the alignment of hierarchical position between a poster an
d responder. If the 
hierarchical level of the poster is not equal to the responder but lower, the alignment would be 
measured as 0. If the hierarchical level of the poster and responder is identical, the alignment 
!!$&!would be measured as 1. If the hierarchica
l level of poster is higher than the responder, the 
alignment would be measured as 2. 
  Goal alignment refers to the score for whether or not the type of boundary spanning 
activity in which a poster or responder engages is identical. If the type of boundar
y spanning 
activity in which the poster or responder engages is identical
Ñe.g., a representational post that is 
met with recognition
Ñthe alignment would be measured as 1. If the type of boundary spanning 
activity in which a poster or responder engages is n
ot identical, the alignment would be measured 
as 0.
  Finally, time lag refers to the duration of time that passed between the original publishing 
of posterÕs boundary
-spanning activity and the comment by a responder. Time lag is measured as 
a continuous va
lue in minute units
. Individual
-level V
ariable
 !The study examines the type of boundary spanning activity as one of the independent 
variables. The type of activity refers to the type of boundary spanning activity that an individual 
engages through parent o
r child posting. If type of boundary spanning activity that an individual 
engage in parent or child posting is representation, the type would be measured as 0. If type of 
boundary spanning activity that an individual engages in parent or child posting is c
oordination, 
the type would be measured as 1. If type of boundary spanning activity that an individual 
engages in parent or child posting is information search, the type would be measured as 2. The 
boundary spanning type was determined using a machine lear
ning approached as is documented 
in Van Osch, Steinfield and Zhao (2015).
  !!$'!Dyadic
-level V
ariable
 !Tie strength was measured as a dyadic
-level variable. The tie strength refers to the 
strength of the dyadic tie between poster and responder. If there is no t
ie between the poster and 
responder, the tie would be measured as 0. If there is an indirect (i.ei., weak) tie, the tie would be 
measured as 1. If there is a no
-reciprocal/asymmetric tie, the tie would be measured as 2. If there 
is a reciprocal/symmetric (
i.e., strong) tie, the tie would be measured as 3. It is important to note 
that the nature of tie is only one dimension of tie strength and that other dimensions
Ñsuch as the 
frequency of personal interactions or the emotional nature of support
Ñare other ch
aracteristics 
of tie strength. However, given the nature of our date set, measuring the directionality of the tie 
was the most objective metric of tie strength and has been supported by previous literature 
(Gilbert and Karahalios, 2009)
 Control V
ariable
 !In addition to testing the effects of boundary spanning type, tie strength, the study 
employed a group
-level variable, group type, as the control variable. The group type refers to the 
type of group that the poster or responder belongs to, such as open or s
ecret group. If the group 
type is open, the group would be measured as 0. If the group type is open, the group would be 
measured as 0. If the group type is secret, the group would be measured as 1. Such scale thus 
represents increasing levels of privacy of
 the group and its content and interactions, where 0 is 
visible and 1 is invisible.
 Table 1 The Variables of Research Model
 Variable
 Type
 Definition
 Level
 Measurement/Operation
 Type of 
boundary 
spanning 
activity
 IV, DV
 The type of 
boundary spanning 
activity engaged in 
parent or child 
posting
 Individual
 0=representation
 1=coordination
 2=information search
  !!$(!Table 1
 (contÕd)
 Tie strength
 IV The strength of dyadic 
tie between poster and 
responder
 Dyadic
 0=no tie
 1=weak tie (indirect)
 2=asymmetric tie
 3=symmetric tie
 Hierarchical 
alignment
 DV The alignment for 
hierarchical position 
between a poster and 
responder
 Dyadic
 0=poster!=(not equal 
to) responder
 1=poster=responder
  Goal alignment
 DV The score for whether or 
not the type of boundary 
spanning ac
tivity in 
which a poster or 
responder engages is 
identical
 Dyadic
 0=non-aligned 
1=aligned
  Time lag
 DV The duration of time that 
passed between the 
original publishing of 
posterÕs boundary
-spanning activity and the 
comment by a responder.
 Dyadic
 Continuous value by 
the unite of minute
 Group type
 Control
 The type of group that 
the poster or responder 
belongs to
 Group
 0=open
 1=secret
          !!!!!!!!!!!!$)!CHAPTER IV RESULTS
 !The study employs Generalized Linear Models (GLMs) and Linear Models (LMs) for 
testing the hypotheses regarding the effects of boundary spanning type and tie strength on 
hierarchical alignment, goal alignment, and time lag to measure the boundary spanning 
success, 
summarized in Table 2 and discussed as follows. GLMs are best used for the analysis in which 
dependent variables are categorical and binomial while LMs are best used for the analysis in 
which dependent variables are continuous, which is only used 
for testing time lag in this study.
  With respect to the effect of boundary spanning type on hierarchical alignment, the study 
found that the type of boundary spanning activity significantly affects hierarchical alignment 
(!
=0.180, p=0.023; where 
! is the 
GLMsÕ coefficient, and p is p
-values for this coefficient), 
which supports H1. Additionally, the study expected a negative significant effect of 
representational boundary spanning activity on hierarchical alignment (
!
=-.18, p=0.023), which 
supports H1a. Mo
reover, the study expected the significant effect of coordination 
boundary 
spanning activity
 on hierarchical alignment but found the effect to be non
-significant (
!=0.245, p=0.219), which disapprove H1b. Furthermore, the results found that 
information sear
ch boundary spanning activity
 significantly enacts 
hierarchical alignment 
(!=0.392, p=0.031), which supports H1c.
  With the respect to the effect of boundary spanning type on goal alignment, the study also 
discovered that coordination boundary spanning 
activity significantly affects goal alignment 
(!2=0.658, p2=0.030). Similarly, information search boundary spanning activities significantly 
predict goal alignment (
!3=1.881, p3=0.012). However, as predicted, representational boundary 
spanning activity doe
s not significantly predict goal alignment (
!1=0, p1=0). Therefore, the 
!!$*!study found that the boundary spanning type affects goal alignment for only information search 
and coordination activities, which supports H2.
  With respect to the effect of tie streng
th on goal alignment, the study found that the 
results show that weak tie significantly enacts the strongest goal alignment (
!
1=1.498, p1=0.051), 
and asymmetric tie also performs strong (
!
2=1.402, p2=0.069) effect. However, symmetric 
(strong) ties signific
antly enacts less strong on goal alignment (
!
3=1.278, p3=0.092) compared 
to weak tie does. Moreover, the results show that no tie has the smallest effect on goal alignment 
(!
0=0, p0=0). Thus, suggesting a curvilinear effect where asymmetric and weak ties h
ave the 
strongest effect on goal alignment, with no tie having the weakest effect and strong ties a small 
effect.
  With respect to the interaction effect of boundary spanning type and tie strength on goal 
alignment,
 symmetric (strong) tie does not signific
antly affect coordination 
boundary spanning 
activity
 (!
5=12.901; p5=0.962), which disapproves H3a. In addition to symmetric tie, the other 
two ties, weak (
!
1 =13.286; 
p1=0.960) and asymmetric (
!
3=14.600; p3=0.957), both do not 
significantly enact coordinat
ion 
boundary spanning activity
. Therefor, tie strength does not 
significantly enact coordination 
boundary spanning activity
 coordination boundary spanning 
activity. The results also reveal that weak (
!
2=2.987; p2=0.088) and asymmetric ties (similar to 
weak tie; 
!
4=3.459; p4=0.067) show a more significantly positive effect on information search 
boundary spanning activity than symmetric (strong) tie 
(!
6=2.648; p6=0.100) 
does, which 
supports H3b. Accordingly, tie strength 
significantly impacts on information s
earch 
boundary 
spanning activity but not on representational and coordination boundary spanning activities. 
Moreover, 
the results show that no tie type reveals a negative significant effect on goal alignment 
(!0=0, p0=0). !!$+! With respect to the effect of tie
 strength on time lag, we found that the results also show 
that symmetric (i.e. strong) and asymmetric tie significantly enacts the least time lag (
!
2=-4622; 
p2=0.05) and (
!
3=-4534; p3=0.05
) effect, with stronger ties slightly outperforming asymmetric 
ties. Weak tie still perform better than no tie, but do incur a longer response time (
!
1=-42912; 
p1=0.000). Moreover, the results show that no tie reveals a negative significant effect on time lag 
(!
0=-.429, p=0.000). Therefore, tie strength significantly affe
cts time lag, supporting H4.
 Table 2 The Summary of R
esults
 Hypothesis
 Poisson Regr. Coefficient/p
-value/Std. Err.
 Supported
 H1 !
=0.180; 
p=0.023; Std. Err.=0.079
 Supported
 H1a
 (rep.)
 !
=0; 
p=0; Std. Err.=0
 Supported
 H1b 
(coor.)
 !
=0.245; 
p=0.219; Std. Err
.=0.199 Not supported
 H1c
 (info.)
 !
=0.392; 
p=0.031; Std. Err.=0.182
 Supported
 H2 !
1(rep.)=0; p1=0; Std. Err.=0 
 !
2 (coor.)=0.658; p2=0.030; Std. Err.=0.304
 !
3 (info.)=1.881; 
p3=0.012; Std. Err.=0.281
 Supported
 H3 !
0 (no tie)=0; 
p0=0; Std. Err.=0
 !
1 (weak)=1.498; 
p1=0.051; Std. Err.=0.769
 !
2 (asym.)=1.402; p2=0.069; Std. Err.=0.770
 !
3 (sym.)=1.278; p3=0.092
; Std. Err.=0.759
 Supported
 H3a 
(sym.*coor
.) !
0 (no tie)=0; 
p0=0; Std. Err.=0
 !
1 (weak*coor.)=13.286; 
p1=0.960; Std. Err.=267.709
 !
2 (weak*info
.)=2.987; p2=0.088; Std. Err.=1.750
 !
3 (asym.*coor.)=14.600; p3=0.957
; Std. Err.=267.710
 !
4 (asym.*info.)=3.459; p4=0.067
; Std. Err.=1.888
 !
5 (sym.*coor.)=12.901; p5=0.962
; Std. Err.=0.759
 !
6 (sym.*info.)=2.648; p6=0.100
; Std. Err.=1.608
 Not supported
 H3b
 (weak*inf
o.) !
0 (no tie)=0; 
p0=0; Std. Err.=0
 !
1 (weak*coor.)=13.286; 
p1=0.960; Std. Err.=267.709
 !
2 (weak*info.)=2.987; p2=0.088; Std. Err.=1.750
 !
3 (asym.*coor.)=14.600; p3=0.957
; Std. Err.=267.710
 !
4 (asym.*info.)=3.4590; p4=0.067
; Std. Err.=1.888
 !
5 (sym.*coor.)=12.901; p5=0.092
; Std. Err.=0.759
 !
6 (sym.*info.)=2.648; p6=0.1
; Std. Err.=1.608
 Supported
 H4 !
0 (no tie)=0; p0=0; Std. Err.=0
 !
1 (weak)=
-42912; p1=0; Std. Err.=12045
 !
2 (asym.)=
-4622; p2=0; Std. Err.=12092
 !
3 (sym.)=
-4534; p3=0; Std. Err.=117
57 Highly supported
 !!$,!CHAPTER V DISCUSSION
 !With respect to hierarchical alignment, individualsÕ information search boundary 
spanning activities are best able to reach the targeted audiences. Little is known about whether or 
not an individualÕs coordination boundary spanning activity is able to reac
h the audience. Such 
situation might result from the limited number of such activities, occurring in ESM. ESM is not 
popular for individuals to complete tasks or projects compared to email, Skype, instant 
messaging, or any task
-oriented communicative tools
. As for representational boundary spanning 
activities in ESM, the study found that individuals tend not to seek hierarchically aligned 
targeted audiences.
  With respect to goal alignment, individualsÕ coordination and information search 
boundary spanning 
activity respectively result in goal alignment, which leads to greater 
effectiveness of boundary spanning, whereas representational boundary spanning activity is 
unable to achieve goal alignment, which refers to lower effectiveness of boundary spanning. Th
e messages in coordination and information boundary spanning activities may be much more clear 
and easy for responders to be interpreted than the ones in representation boundary spanning 
activity. Individuals who engage in representation boundary spanning 
activity tend to ÒtailorÓ the 
words in subtle way but they are thus not easily understood and interpreted. Hence, future 
research explore the role of message ambiguity in obtaining goal alignment, where it appears that 
representational postings
Ñdue to thei
r ÒbraggingÓ nature
Ñmay be more ambiguous in nature 
than information and coordination postings.
  Moreover, weakly and asymmetrically (similar to weakly) tied individuals who engage in 
information search boundary spanning activity are the most likely to ach
ieve goal alignment, 
though symmetrically (strongly) tied ones also positively affect goal alignment but revealing a 
!!%-!smaller relationship. Accordingly, overall, individualsÕ information search boundary spanning 
activity can lead to greater effectiveness an
d the reach of targeted audiences to achieve boundary 
spanning success.
  Last, with respect to time lag, the stronger the tie between the dyadic individuals, less 
time between the responder comment and the original posting created on the posting created by
 poster. Therefore, strongly tied individuals who engage in information search boundary spanning 
activities can achieve boundary spanning success, as it is much easier for them to get the 
response in a short time compared to more weakly tied individual who
 engage in information 
search boundary spanning activity. Trust operated by symmetric (strong) tie thus can achieve 
boundary spanning success.
              !!%$!CHAPTER VI IMPLICATIONS FOR RESEARCH AND PRACTICE
 !Representational boundary spanning is a dominant 
activity due to visibility (Treem and 
Leonardi, 2012) afforded by ESM but was found to be ineffective because of the lack of goal 
alignment. 
Additionally, individuals engaging in representational boundary spanning tend not to 
seek hierarchically aligned au
diences, th
ough the visibility afforded by ESM seems promising 
for individuals to engage in these type of interactions. Moreover, coordination boundary 
spanning activity is not affected by tie strength and was insignificant in terms of triggering 
hierarchi
cal alignment, though it leads to goal alignment, which creates the effectiveness of 
boundary spanning success. Furthermore, information search boundary spanning activity is able 
to reach the targeted audiences because of hierarchical alignment. Informatio
n search boundary 
spanning activity is also effective because of goal alignment when a poster and responder have a 
weak or asymmetrical (similar with weak tie) tie. Additionally, tie strength is important for 
information search boundary spanning activity. 
Although strong tie leads to effective information 
search boundary spanning activity due to its goal alignment, weak and asymmetrical ties 
outperform the symmetric (strong) tie, which responders to GranovettorÕs the strength of weak 
tie. However, tie stren
gth is not important for representational and coordination boundary 
spanning activities. Moreover, strongly tied individuals tend to respond to the question faster 
than weakly tied ones, which further validates the importance of common ground with shared 
language (Chiu et al., 2006) and relationship heuristics
 (Uzzi, 1997) possessed in the pair of 
individuals with strong relationship. These finding imply that different strategic goals require 
different types of relationships. For instance, if one needs a hi
gh-quality answer, weak ties 
provide more ÒeffectiveÓ responses. However, if one requires a fast response
Ñmaybe in an 
urgent situation
Ñstrong ties offer better support.
 !!%%! Yet the main theoretical contribution of this study lies in offering a perspective of 
the 
dyadic interaction between a poster and a responder and using the dyadic perspective to delineate 
boundary spanning success. This approach is novel as the existing literature has only focused on 
the boundary spanning activities as enacted by the indivi
dual without considering the 
reciprocating actions of the responding party which is crucial in determining the success of such 
activities.
  In addition for advancing theories of boundary spanning, the study offers practical 
contribution. The results can pr
ovide the managers or higher executive levels individuals with 
the improvement for information exchange by employing the effect of tie strength on the 
attainment of quality and quick responses for information search boundary spanning activity in 
ESM. In ot
her words, the managers should encourage individuals who would like to get the best 
quality of response to ask the questions in open group, as the potential audiences would be much 
more diverse than the ones in closed group. On the other hand, if an indivi
dual who wants to get 
quick response, he or she could post the questions in the groups where people with whom he or 
she has ever worked with, as there are more strongly tied individuals there.
         !!%&!CHAPTER VII LIMITATIONS AND FUTURE RESEARCH
 !The study aims to explore 
trust as measured through an individualsÕ tie strength to 
explore its effects on hierarchical alignment, goal alignment, and time lag
Ñas the proxies of 
boundary spanning success
Ñfor the three boundary spanning activities. The stud
y does not rule 
out other predictors for trust and only considers tie strength as a proxy of trust from a network 
perspective. Other antecedences involve individualÕs geographical diversity, groupÕs maturity, 
and organizational structure. Similarly, the st
udy only looks at boundary spanning success via 
activities. The study only concludes hierarchical alignment, goal alignment, and time lag for 
boundary spanning success. Future research could explore external outcome metrics unrelated to 
ESM use per se, suc
h as overall team performance and innovativeness as well as job satisfaction. 
Additionally, future research could broaden the scope to individual
-level and group
-level factors, 
such as individualÕs experience and history of using social media as well as gr
oup leadership.
            !!%'!            REFERENCES
           !!%(!REFERENCES
  Ancona, D.G. (1990). Outward bound: Strategies for team survival in an organization. 
Academy
 of Management Journal
 33: 334
-336.  Ancona, D. G., & Caldwell, D. F. (1992). Bridging the 
boundary: External activity and
 performance in organizational teams. 
Administrative science quarterly (37).
 pp. 634-665.  Allen, T. (1977). Managing the Flow of Technology. 
MIT Press
. Cambridge, MA.
  Arrow, K. J. (1974). Limited knowledge and economic 
analysis. 
American Economic Review
, 64(1), 1-10.  Blau, P. M. (1964). Exchange and Power in Social Life, New York, USA: Wiley.
  Brzozowski, M. J. (2009). Watercooler: exploring an organization through enterprise social
 Media. 
GROUP Õ09: Proceedings of the 
2009 international conference on supporting group
 work, 
pp. 219Ð228.  Bulgurcu, B., van Osch, W., & Kane, G. (2015). The Rise of the Promoters: Understanding a
 New Class of Users and their Participation in Enterprise Online Communities. Unpublished
 manuscr
ipt.
  Burt, R. (2005). Brokerage and closure: An introduction to social capital. 
Oxford: Oxford
 University Press.
  Chiu, M., Hsu, H., and Wang, T. (2006). Understanding Knowledge Sharing in Virtual
 Communities, 
Decision Support Systems, 2006, 42:3,
 pp. 1872-1888.  Daft, R. L., & Lengel, R. H. (1986). Organizational information requirements, media richness
 and structural design. 
Management science
, 32(5), 554-571.  Davies, E. (2003). IETF problem statement.
 http://www.ietf.org/Internet
-drafts/draft
-ietf-problem
-issue
-statement
-02.txt
.  Fleming, L., & Waguespack, D. M. (2007). Brokerage, boundary spanning, and leadershi
p in
 open innovation communities. 
Organization science
, 18(2), 165-180.  Fulk, J. and Yuan, Y. C. (2013). Location, motivation, and social capitalization via enterprise
 social networking. 
Journal of Computer
-Mediated Communication, (19:1)
, pp. 20Ð37.  Gibb
s, J. L., Eisenberg, J., Rozaidi, N. A., & Gryaznova, A. (2014). The ÒmegapozitivÓ role of
 enterprise social media in enabling cross
-boundary communication in a distributed Russian
 organization. 
American Behavioral Scientist
, 0002764214540511. !!%)! Gilbert, E.
, & Karahalios, K. (2009). Predicting Tie Strength with Social Media. In Proceedings
 of the 27th International Conference on Human Factors in Computing Systems, pp.
 211-220.  Granovetter, M. (1973). The Strength of Weak Ties. 
American Journal of Sociology
, 78:1360
-1380.  Hansen, M. T. (1999). The search
-transfer problem: The role of weak ties in sharing knowledge
 across organization subunits. 
Administrative science quarterly
, 44(1), 82-111.  Hung, S.
-W., Chen, C.
-J., and Lin, M.
-J. J. (2009). Fostering the 
Determinants of Knowledge
 Sharing in Professional Virtual Communities. 
Computers in Human Behavior,
 25:4, pp. 929-939.  Jones, G.R. & George, J.M. (1998). ÔThe Experience and Evolution of Trust: Implications for
 Cooperation and TeamworkÕ. 
Academy of 
Management Review. 23(3)
. pp. 531Ð546  Kearsley, G. P. (1976). Questions and Question Asking in Verbal Discourse: A Cross
- Disciplinary Review, 
Journal of Psycholinguistic Research, 5:4,
 pp. 355-375.  Kogut, B., & Zander, U. (1992). Knowledge of the firm, 
combinative capabilities, and the
 replication of technology. 
Organization science
, 3(3), 383-397.  Krackhardt, D. (1992). The strength of strong ties: The importance of philos in organizations.
 Networks and organizations: Structure, form, and action
, 216, 239.  Lawler, E. J., and Thye, S. R. (1999). Bringing Emotions into Social Exchange Theory. 
Annual
 Review of Sociology, 25
, pp. 217-244  Majchrzak, A., Faraj, S., Kane, G. C., & Azad, B. (2013). The contradictory influence of social
 media
 affordances on online communal knowledge sharing. 
Journal of Computer
- Mediated Communication, (19: 1)
, pp. 38Ð55.  Molm, L. D. (1994). Dependence and Risk: Transforming and Structure of Social Exchange.
 Social Psychology Quarterly
, 57: 163
-176.  Noteboom
, B. & Six, F. (2003). The Trust Process in Organizations. 
Edward Elgar Publishing
. Northhampton, MA.
  Ram, A., ÒA Theory of Questions and Question Asking,Ó Journal of the Learning Sciences, 1991,
 1:3, pp. 273
-318.  Steinfield, C., DiMicco, J.M., Ellison, 
N.B., and Lampe, C. (2009). Bowling online: social
 networking and social capital within the organization. 
Proceedings of the fourth
 international conference on Communities and technologies (C&T '09), ACM
, New York,
 !!%*!NY, USA, pp. 245
-254.  Stephenson Jr, M. 
(2005). Making humanitarian relief networks more effective: operational
 coordination, trust and sense making. 
Disasters
, 29(4), 337-350.  Treem, J. W., & Leonardi, P. M. (2012). Social media use in organizations: Exploring the
 affordances of visibility, ed
itablity, persistance, and association. 
Communication Yearbook
 (36).  Tushman, M. (1977). Special boundary roles the innovation process. 
Admin. Sci. Quart. 22
 587-605.  Tushman, M. L., & Scanlon, T. J. (1981). Boundary spanning individuals: Their role in
 information transfer and their antecedents. 
Academy of Management Journal. 24(2)
. pp. 289-305.  Uzzi, B. (1997). Social structure and competition in interfirm networks: The paradox of
 em-beddedness. 
Administrative Science Quarterly, 42.
 35-67.  Van Osch
, W, & Steinfield, C. W. (2013). Boundary Spanning through Enterprise Social
 Software: An External Stakeholder Perspective. 
Proceeding of 2013 International
 Conference on Information Systems (ICIS 2013)
.  Van Osch, W., Steinfield, C. W., & Balogh, B. A. 
(2015). Enterprise Social Media: Challenges
 and Opportunities for Organizational Communication and Collaboration.
 In System Sciences (HICSS), 2015 48th Hawaii International Conference on
 (pp. 763-772). IEEE.
  Van Osch, W., & Steinfield, C. W. (2015). 
Intra
-Organizational Boundary Spanning: Strategic
 Implications for the Implementation and Use of Enterprise Social Media. Unpublished 
manuscript.
  Van Osch, W., Steinfield, C. W., and Zhao (2015). 
Team Boundary Spanning through Enterprise
 Social Media: Explorin
g the Effects of Group
-Level Diversity Using a Data Science
 Approach. 
Unpublished manuscript.
   Wasko, M. M., Teigland, R., and Faraj, S. (2009). The provision of Online Public Goods:
 Examining Social Structure in an Electronic Network of Practice, 
Decision Support
         Systems, 47:3,
 pp. 254-265.   !