Coding Like a Rockstar: The Role of Social Influence on Action Patterns in GitHub

Abstract

Action patterns occur wherever people perform work. Research has provided important insights about how action patterns form, change and dissolve over time. However, it has neglected the question of why actors choose to enact certain action patterns. In this research-in-progress, we shed light on the socio-cognitive factors underlying the formation of action patterns. We focus on GitHub and argue that so-called ‘rockstars’ exert social influence on other developers’ action patterns over time. As a preliminary step, we showcase the plausibility of this argument by examining an actual project where we see indications that the influence of a rockstar is reflected in the types of actions that are predominantly used. We will continue this research by including (1) more projects and rockstars, and (2) the analysis of action networks to test if action networks converge with those of rockstars over time.

Full text

The Role of Social Influence on Action Patterns
Forty-Second International Conference on Information Systems, Austin 2021
1
Coding Like a Rockstar: The Role of Social
Influence on Action Patterns in GitHub
Short Paper
Thomas Grisold
University of Liechtenstein
Fuerst Franz Josef Strasse,
9490 Vaduz, Liechtenstein
thomas.grisold@uni.li
Michael Gau
University of Liechtenstein
Fuerst Franz Josef Strasse,
9490 Vaduz, Liechtenstein
michael.gau@uni.li
Youngjin Yoo
Case Western Reserve University
10900 Euclid Avenue
Cleveland, Ohio 44106-7235
youngjin.yoo@case.edu
Abstract
Action patterns occur wherever people perform work. Research has provided important
insights about how action patterns form, change and dissolve over time. However, it has
neglected the question of why actors choose to enact certain action patterns. In this
research-in-progress, we shed light on the socio-cognitive factors underlying the
formation of action patterns. We focus on GitHub and argue that so-called ‘rockstars’
exert social influence on other developers’ action patterns over time. As a preliminary
step, we showcase the plausibility of this argument by examining an actual project where
we see indications that the influence of a rockstar is reflected in the types of actions that
are predominantly used. We will continue this research by including (1) more projects
and rockstars, and (2) the analysis of action networks to test if action networks converge
with those of rockstars over time.
Keywords: Action patterns, GitHub, social influence, rockstars, digital traces, routine
dynamics
Introduction
People develop action patterns when they perform work (Feldman and Pentland 2003). Action patterns are
repetitive sets of actions that recur over time. Action patterns form because actors learn to respond to
known stimuli from the past (Feldman and Pentland 2003). Action patterns are not static, however; they
evolve dynamically as actors face changing contexts (Farjoun 2010). To this end, researchers have been
studying how action patterns form, change and dissolve in various organizational environments (Feldman
et al. 2016; Goh and Pentland 2019).
An overriding theme of past research is that it focuses on actions when it studies organizing (Feldman et al.
2016). While extremely powerful in understanding how action patterns form and change, a focus on actions
comes at the expense of actors. Put it bluntly, past research on action patterns pushed actors to the
background. Consequently, little attention has been brought to as to why certain actors enact action

The Role of Social Influence on Action Patterns
Forty-Second International Conference on Information Systems, Austin 2021
2
patterns in one way or another. We know little about the socio-cognitive factors which influence the
formation of action patterns (Rerup and Spencer 2021). Such factors, however, can have a large influence
on actors in organized work, even without their conscious awareness (Bertels et al. 2016). Understanding
how, why and when socio-cognitive influence manifests in work contexts has implications for research
because it helps to refine existing theories about the mechanisms driving the formation and dissolution of
action patterns (Feldman et al. 2016). And it also has implications for practice (Govindarajan et al. 2021)
because managers can design work environments which promote or inhibit social influence in work
settings.
In this research, we bring actors to the foreground. We seek to understand why actors enact certain action
patterns over time. In particular, we focus on the role of social influence that is at play when actors
collectively organize work. We draw on GitHub, a large-scale open-source code development platform, and
investigate the role of social influence in the formation of action patterns. While GitHub has proven to be a
promising context to study action patterns (Lindberg et al. 2016), our research aims to connect two hitherto
unrelated perspectives. On the one hand, we explore how action patterns within such projects change over
time by retrieving action traces from event log data (Lindberg et al. 2016). On the other hand, we analyze
how the presence of ‘rockstars’ –GitHub developers with exceptionally large follower bases– influence the
actions of other developers when they work on projects (Lee et al. 2013). We draw on social learning theory
and propose that rockstars (1) exert influence on the formation and retainment of action patterns in GitHub
projects, and (2) this influence increases with a higher frequency in which rockstars contribute to a project.
We showcase our argument by means of an actual GitHub project. We focus on one project where we assume
that the action preferences of one rockstar have influenced the action patterns of other developers over
time. We propose how we will proceed as we search for large-scale evidence of our claim.
Theoretical Background
Action Patterns in Organizational Contexts
When actors perform work, they develop sequences of recurrent and recognizable action patterns (Feldman
and Pentland 2003). This provides them with stable and reliable structures. At the same time, action
patterns are not static but they are dynamically evolving over time. Actors need to interact and coordinate
when they work on tasks, and their understandings and performances need to align and re-align when
certain aspects of work contexts change. The question of how and why such dynamics occur has been at the
core of several studies (Feldman et al. 2021). For the purpose of this article, we want to highlight two central
themes in this body of research.
First, there is a recent interest to study contexts where actors self-organize. The main interest here is how
actors establish routinized action patterns even in the absence of formal governance. Digital work
environments are particularly interesting because actors cannot rely on conventional means to coordinate,
such as face-to-face communication (Dittrich et al. 2016). Along these lines, studies report that GitHub
developers enact routines and emergent coordination mechanisms to resolve complex and unaccounted
problems (Lindberg et al. 2016). In the context of Wikipedia, for example, communities develop and enact
sequences of routinized action patterns that reflect different requirements related to the development,
review, and repair of articles (Arazy et al. 2020). Overall, we can expect that the study of routinized work in
open source environments will gain more relevance as this represents a next-generation organizational
form that is adopted in a variety of contexts (Kremser and Xiao 2021).
Second, research on action patterns studies observable actions. It decouples the analysis from “invisible
forces” but embraces the specific actions that are visible in organized work (Feldman 2016). This focus, in
turn, has downplayed the role of actors and the socio-cognitive factors which shape the way actors enact
action patterns. While a few studies indicate that the formation of action patterns involves cognitive aspects
(e.g. the interplay of different forms of memory, see Miller et al. 2012), studies fall short to account for
actors’ motivations, intentions, and emotions when they perform tasks (Rerup and Spencer 2021).
Shedding light on the socio-cognitive foundations of action patterns is important, however (Felin et al.
2015); as routines involve multiple cognizing actors, a shift to their underlying perceptions, expectations,
and goals further clarifies the mechanisms driving the formation, change and dissolvement of action
patterns (Rerup and Spencer 2021). In the context of open source development, for example, a focus on

The Role of Social Influence on Action Patterns
Forty-Second International Conference on Information Systems, Austin 2021
3
socio-cognitive factors can shed light on underlying mechanisms to explain why actors adopt certain
patterns (Arazy et al. 2020), or why actors enact drifts and deviations (Pentland et al. 2020).
Social Influence in Open-Source Development
With over 40 million registered users, GitHub has been providing rich insights into various aspects of self-
organized work (Zöller et al. 2020). GitHub is an online code development platform where developers can
jointly work on projects. A large body of research has focused on socio-cognitive factors involved in GitHub
projects because it has been found that success of projects is strongly related to its social organization
(Crowston and Howison 2005; Von Krogh et al. 2012). Studies investigate, for example, how developers
communicate, learn, and socialize when they code (Cosentino et al. 2017).
One key socio-cognitive factor is constituted through the influence of ‘rockstars’. Rockstars are developers
who enjoy an excellent reputation among other developers. Their reputation is reflected by an exceptionally
large follower base (Huang and Chung 2019). Developers become rockstars when they contribute to a large
number of projects in significant ways (Badashian et al. 2014), or when they own projects which gain
popularity (Yu et al. 2014). Developers follow rockstars to stay informed about project developments and
trends and to acquire new skills (Blincoe et al. 2016).
Studies report that rockstars influence other developers in different ways. For example, the presence of one
or more rockstars in a project is a reliable indicator that the project attracts the attention of other developers
(Jiang et al. 2013). Developers are more likely to follow a project when a superstar has started to follow the
same project or –even more so– when he/she performed an action within the project (Blincoe et al. 2016).
Developers are also more likely to contribute to a project after a rockstar has performed an action, as
compared to when a regular developer has performed an action within the same project (Lee et al. 2013).
The frequency in which a rockstar contributes to a project also plays a role: The more often a rockstar
contributes to a project, the more popular becomes the project (Blincoe et al. 2016). Also, the popularity of
rockstars is an indicator of how strong their social influence is; rockstars with a higher number of followers
are more likely to make others contribute to a project (Huang and Chung 2019).
To summarize, multiple studies provide evidence that rockstars exert social influence on other developers
with respect to their following and contributing behaviors. Drawing on these earlier studies, we argue that
rockstars might exert influence in terms of how other developers work on a project.
Theorizing the Social Influence on Action Patterns in Open-Source Development
Our research explores if and to what extent rockstars influence the action patterns of other developers when
they jointly work on projects. We fold research on action patterns (Feldman 2016; Goh and Pentland 2019)
with research on online collaboration in GitHub (Cosentino et al. 2017; Zöller et al. 2020) to propose that
the social influence of rockstars plays out in two key ways.
First, we draw on studies which indicate that social influence is present because rockstars are expected to
point to new developments and trends, and also provide opportunities for others to learn from and adopt
their skills (Blincoe et al. 2016; Dabbish et al. 2012). Rockstars are the sources of learning for other
developers in open source projects. Hence, they have a positive effect on developers’ motivation which has
been found to be a key factor for the success of a project (Von Krogh et al. 2012). Developers are aware of
whom they are collaborating with (Dabbish et al. 2012) and pay closer attention to what rockstars are doing
(Burke et al. 2009; Tsay et al. 2014). Specifically, when developers observe rockstars, they are engaging in
social learning processes (Bandura 1977) – this is because developers use rockstars as role-models and
imitate their behavior over time (Huang and Chung 2019; Lee et al. 2013). Following this line of thinking,
we can expect that rockstars influence other developers not only with respect to what projects, skills, and
topics seem important (Huang and Chung 2019), but also how they approach and work on projects
(Bandura 1977). This aligns with studies reporting that the formation of action patterns depends on
collective learning processes where actors perceive, evaluate and integrate new opportunities into their
work practices (Dittrich et al. 2016; Dittrich and Seidl 2018). Such effects, however, are not immediately
present; they take shape over time (Feldman 2016). Taken together, we argue that developers learn from
and imitate rockstars over time as they work on projects. This leads us to our first proposition:

The Role of Social Influence on Action Patterns
Forty-Second International Conference on Information Systems, Austin 2021
4
Proposition 1: Rockstars influence the action patterns of other developers over time as they work on
projects.
Second, past research shows that rockstars gain more influence as they contribute to projects more strongly
(Blincoe et al. 2016; Lee et al. 2013). This, again, aligns with social learning theory (Bandura 1977) which
suggests that learning effects are stronger when there is more exposure between peers and role-models. In
this regard, research on the formation of action patterns points to two interesting observations. First, the
organization of work practices orients itself towards influential actors who have a higher status within a
collective (Kremser and Blagoev 2020). Second, action patterns form when certain sets of actions are
repeated over time. Actions that repeatedly co-occur together form “ruts in the road” (Goh and Pentland
2019): dominant sets of actions that are likely to be enacted again in the future. Taken together, we assume
that ongoing contributions of rockstars reinforce the enactment of certain action patterns over time. This
leads us to our second proposition.
Proposition 2: A rockstar exerts more influence on action patterns of other developers as he/she
contributes to the project more frequently.
In the following, we showcase how we will investigate these propositions empirically.
An Empirical Showcase of Our Idea
Following previous research, we see three directions to study our propositions (Feldman et al. 2021;
Feldman et al. 2016); (1) by studying sequences of actions and their temporal and logical contingencies over
time; (2) by analyzing the decisions leading to the selection of action patterns, e.g. through communication
in project comments (Dittrich et al. 2016); or (3) by studying the types of actions used and their
distributions, and how they change over time (Goh and Pentland 2019).
In this research-in-progress, we opt for (3). We showcase that the types of actions used change under the
influence of a rockstar. On GitHub, developers take actions when they perform ‘activities’; GitHub offers 16
different types of activities which can be used for different stages and tasks in a project (for example, a
‘PullRequest’ is used to inform others about changes a collaborator suggested to a project). Some activities
can be adopted for multiple purposes. For example, the action ‘IssueCommentEvent’ can be used to report
on issues (e.g. so that other developers can work on them), but it also affords general communication (e.g.
to exchange views on new coding features). When rockstars prefer certain activities over others to
accomplish work, their influence may play out in the sense that other developers espouse similar
preferences over time.
The GitHub platform encompasses a large set of open-source projects. Because the digital trace data of
these projects are publicly available, GitHub provides the opportunity to perform detailed analyses of the
dynamics of action patterns (Lindberg et al. 2016). Against this background, we retrieved trace data
produced in 2019 and 2020 which are available on GH Archive (https://www.gharchive.org/). The trace
data included GitHub events of over 64 million public projects on GitHub.
As a next step, we sought to identify rockstars. To do so, we first identified actors participating in each
project and requested the number of followers on their GitHub profile using the GitHub API
(https://docs.github.com/en/rest/). Following Lee et al. (2013) and others, rockstars have several
thousand followers (according to Lee et al., regular users have an average of 40 followers). Through this
process, we identified 816 actors with more than 5.000 followers. As a preliminary step, for this research-
in-progress, we then selected one of the top ten rockstars.
We showcase our argument by drawing on a rockstar who has over 29.000 followers. Due to the
exceptionally large number of followers, he/she can be expected to be highly influential. The rockstar is
hosting 299 projects on GitHub and made 1997 contributions in the years 2019 and 2020 to different
projects. Then, we searched for an idiosyncratic action pattern of the rockstar by looking at the distribution
of his/her past activities. We discovered that he/she typically ignores ‘IssueCommentEvents’. Following our
theoretical conjectures, we thus expected to see a decline of the occurrences of ‘IssueCommentEvents’
activities over time as the rockstar is contributing to projects.
To showcase the plausibility of our conjecture, we describe one large project where the rockstar has been
contributing since the beginning (‘chrome-launcher’). We collected a complete event log beginning from

The Role of Social Influence on Action Patterns
Forty-Second International Conference on Information Systems, Austin 2021
5
the start of the project until the end of 2020. Importantly, the rockstar did not start the project and was not
the owner of the project; he/she is only contributing to it. Table 1 provides a detailed overview of the project.
Project name: ‘chrome-launcher’
Project statistics
Total number of activities
2366
Number of activities of rock star
444
Forks
128
Contributors
40
Stars on GitHub
781
Project start
09/2017
Event log data
09/2017-12/2020
Table 1. Descriptive statistics
Figure 1 shows the different types of activities that were used in the project, as well as their frequencies from
the beginning until the end of the year 2020.
The two most frequent events are ‘WatchEvent’ and ‘IssueCommentEvent’. A ‘WatchEvent’ action is created
when someone follows a repository but is not directly contributing to the project. An ‘IssueCommentEvent’
action reflects an issue comment and is created when someone creates, edits, or deletes a comment
regarding an issue. Upon visually inspecting Figure 1, we find indications that the occurrences of
‘IssueCommentEvent’ actions (red line) decrease over time. While there are spikes at some stages in the
project, it seems that there is a trend towards a lower number of ‘IssueCommentEvent’ actions.
Figure 1. Distribution of activities
We applied a more formal measure to see if our propositions hold. We compared the ‘IssueCommentEvent’
distribution of the project including all contributors to the distribution of the selected rockstar in the
project. To this end, we applied the Kruskall-Wallis test (Kruskal and Wallis 1952) to determine if there are
statistically significant differences between two or more groups of an independent sample. As shown in
Table 2, the results indicate that the distributions are different from the beginning of the project until the
end of 2019. In the last third of the project, the distribution of the ‘IssueCommentEvent’ activities of the

The Role of Social Influence on Action Patterns
Forty-Second International Conference on Information Systems, Austin 2021
6
overall project converges with the distribution of the ‘IssueCommentEvent’ activities of the rockstar. This
indicates that the use of action patterns of other developers converges with those of the rockstar over time.
Event Type
Event Logs
Chi Square (c2)
P
Description
IssueCommentEvent
09/2017 –
12/2019
10.310
0.00132**
Reject H0: Sample
distributions are not equal.
12/2019 –
12/2020
3.838
0.05010
Fail to Reject H0: Sample
distributions are equal.
Table 2. Kruskall-Wallis ANOVA table
It is important to note that we chose this project to showcase our idea. While there may possibly be other
interpretations of these findings at this stage, the effects observed in this project seem plausible in light of
our theoretical argument. Specifically, the activities of other developers align with those of the rockstar, and
not the other way around.
We have been analyzing other projects where this rockstar has been contributing to and see indications that
our conjectures hold. Effects, however, are subtler in other projects. For example, we see in some projects
that the action distributions of rockstars and other developers do not converge in a linear way; at some
points, the activities converge; then, they drift apart before they converge again more strongly. This is an
important observation highlighting that further analysis needs to be more fine-granular in order to capture
and validate these effects over multiple time frames across a high number of projects.
Further Research
While the approach we presented here serves to illustrate our idea, we will further develop this research by
conducting more thorough and comprehensive analyses. Our key concern is to establish causality that
reflects the uni-directional influence of rockstars on other developers. To this end, we pursue two research
avenues. First, we will expand our data set by including additional open-source projects and rockstars. In
addition to the projects of the rockstar described here, we will add other highly influential rockstars to our
analysis. We will analyze rockstars’ action distributions and compare them to the overall action
distributions by all developers, both within and across projects.
The second avenue will analyze how action networks change under the influence of rockstars. By action
networks, we refer to sequences of actions as they form action patterns. Sequences of actions are an
important indicator for our propositions because previous research has shown that the order of activities is
highly adaptable and changing over time. We are planning to draw on process mining (van der Aalst 2016)
to study the dynamics of action patterns (Grisold et al. 2020). We will create directed action networks of
single projects and single actors collaborating in the project to compare if the action networks converge
between rockstars and other developers over time. Figure 2 illustrates the action network of the overall
project with that of the rockstar we showcased before; the figure shows the types of used activities and the
numbers show how often certain sequences of activities are found in the data. The left image shows the
action network for the overall project; the right image shows the action network of those actions that were
taken by the rockstar. One way to compare these action patterns is to transform the graphical representation
of the action network into matrices. We will apply different mathematical tools to summarize and find
global patterns or similarities between action networks across multiple time frames. Such a procedure will
provide us with a more nuanced view on the influence of rockstars on action patterns.

The Role of Social Influence on Action Patterns
Forty-Second International Conference on Information Systems, Austin 2021
7
Figure 2. Action network of whole project (left) vs. action network of rockstar (right)
Expected contributions
Action patterns are repetitive sets of actions that emerge wherever people perform work. Research so far
has studied the dynamics of action patterns by means of how they form, change and dissolve over time. We
are begging to ask why they form the way they form. By foregrounding socio-cognitive factors involved in
organized work, we focus on the role of rockstars’ social influence on the formation of action patterns. Our
research-in-progress is expected to contribute to the existing body of knowledge in three central ways.
First, we follow recent calls to study why actors choose to enact certain action patterns (Rerup and Spencer
2021). We suggest that social influence as exerted by rockstars has an impact on other developers and how
they perform work. This is a novel perspective because existing studies have been drawing on the implicit
assumption that actors constitute equal elements in social systems. Power dynamics, for example, have only
received limited attention (Kremser and Blagoev 2020). We are shedding light on these aspects and argue
that some actors are more influential than others when action patterns take shape.
Second, our approach implies a deductive study design to translate several key assumptions of routine
dynamics theory into concrete propositions. This is innovative because the majority of research in this field
has followed inductive study designs to develop and refine assumptions regarding the dynamics of action
patterns. Our key move here is to not only focus on observable actions (Feldman 2016) but to integrate
socio-cognitive factors. Our approach seems relevant to other work contexts, too. For example, social
influence may play an important role in self-organized work (Kremser and Xiao 2021), such as holocracy
and agile work.
Third, our findings may be applicable to various digital environments, including social media. One of the
key principles of Twitter, Instagram, and other platforms is that users are judged with respect to their
popularity (as expressed through the number of followers/friends/fans). Investigating how popular and
influential users exert influence on other users can be interesting to see how popular users change the
behavior of other users.
We will present more findings at ICIS 2021 where we include (1) more projects, (2) more rockstars, and (3)
additional techniques to study the change of action patterns over time.
Acknowledgements
This work was funded by the European Union [Erasmus+ Program: 2019-1-LI01- KA203-000169]: “BPM
and Organizational Theory: An Integrated Reference Curriculum Design” and the National Science
Foundation [NSF #1447670, Big data: Multi-level predictive analytics & motif discovery across massive
dynamic patio-temporal networks in complex socio-technical systems: An organizational genetics
approach”].

The Role of Social Influence on Action Patterns
Forty-Second International Conference on Information Systems, Austin 2021
8
References
Arazy, O., Lindberg, A., Lev, S., Wu, K., and Yarovoy, A. 2020. "Emergent Routines in Peer-Production:
Examining the Temporal Evolution of Wikipedia's Work Sequences," ACM Transactions on Social
Computing (3:1), pp. 1-24.
Badashian, A. S., Esteki, A., Gholipour, A., Hindle, A., and Stroulia, E. 2014. "Involvement, Contribution
and Influence in Github and Stack Overflow," CASCON, pp. 19-33.
Bandura. 1977. Social Learning Theory. New York: General Learning Press.
Bertels, S., Howard-Grenville, J., and Pek, S. 2016. "Cultural Molding, Shielding, and Shoring at Oilco: The
Role of Culture in the Integration of Routines," Organization Science (27:3), pp. 573-593.
Blincoe, K., Sheoran, J., Goggins, S., Petakovic, E., and Damian, D. 2016. "Understanding the Popular
Users: Following, Affiliation Influence and Leadership on Github," Information and Software
Technology (70), pp. 30-39.
Burke, M., Marlow, C., and Lento, T. 2009. "Feed Me: Motivating Newcomer Contribution in Social
Network Sites," Proceedings of the SIGCHI conference on human factors in computing systems, pp.
945-954.
Cosentino, V., Izquierdo, J. L. C., and Cabot, J. 2017. "A Systematic Mapping Study of Software
Development with Github," IEEE Access (5), pp. 7173-7192.
Crowston, K., and Howison, J. 2005. "The Social Structure of Free and Open Source Software
Development," First Monday (10:2).
Dabbish, L., Stuart, C., Tsay, J., and Herbsleb, J. 2012. "Social Coding in Github: Transparency and
Collaboration in an Open Software Repository," Proceedings of the ACM 2012 conference on computer
supported cooperative work, pp. 1277-1286.
Dittrich, K., Guérard, S., and Seidl, D. 2016. "Talking About Routines: The Role of Reflective Talk in Routine
Change," Organization Science (27:3), pp. 678-697.
Dittrich, K., and Seidl, D. 2018. "Emerging Intentionality in Routine Dynamics: A Pragmatist View,"
Academy of Management Journal (61:1), pp. 111-138.
Farjoun, M. 2010. "Beyond Dualism: Stability and Change as a Duality," Academy of Management Review
(35:2), pp. 202-225.
Feldman, M. S. 2016. "Routines as Process: Past, Present and Future," in Organizational Routines: How
They Are Created, Maintained, and Changed, J. Howard-Grenville, C. Rerup, A. Langley and H.
Tsoukas (eds.). Oxford, UK: Oxford University Press.
Feldman, M. S., and Pentland, B. 2003. "Reconceptualizing Organizational Routines as a Source of
Flexibility and Change," Administrative Science Quarterly (48:1), pp. 94-118.
Feldman, M. S., Pentland, B., D’Adderio, L., Dittrich, K., Rerup, C., and Seidl, D. 2021. "What Is Routine
Dynamics?," in Cambridge Handbook of Routine Dynamics, M.S. Feldman, B. Pentland, L. D’Adderio,
K. Dittrich, C. Rerup and D. Seidl (eds.). Cambridge, UK: Cambridge University Press.
Feldman, M. S., Pentland, B., D’Adderio, L., and Lazaric, N. 2016. "Beyond Routines as Things:
Introduction to the Special Issue on Routine Dynamics," Organization Science (27:3), pp. 505-513.
Felin, T., Foss, N. J., and Ployhart, R. E. 2015. "The Microfoundations Movement in Strategy and
Organization Theory," The Academy of Management Annals (9:1), pp. 575-632.
Goh, K. T., and Pentland, B. 2019. "From Actions to Paths to Patterning: Toward a Dynamic Theory of
Patterning in Routines," Academy of Management Journal (62:6), pp. 1901-1929.
Govindarajan, V., Srivastava, A., Grisold, T., and Klammer, A. 2021. "Resist Old Routines When Returning
to the Office," Harvard Business Review).
Grisold, T., Wurm, B., Mendling, J., and vom Brocke, J. 2020. "Using Process Mining to Support Theorizing
About Change in Organizations," Proceedings of the 53rd Hawaii International Conference on System
Sciences, Maui, US, pp. 5492-5501.
Huang, Y., and Chung, W. 2019. "Rockstar Effect in Distributed Project Management on Github Social
Networks,").
Jiang, J., Zhang, L., and Li, L. 2013. "Understanding Project Dissemination on a Social Coding Site," 2013
20th Working Conference on Reverse Engineering (WCRE): IEEE, pp. 132-141.
Kremser, W., and Blagoev, B. 2020. "The Dynamics of Prioritizing: How Actors Temporally Pattern
Complex Role–Routine Ecologies," Administrative Science Quarterly), p. 0001839220948483.
Kremser, W., and Xiao, J. 2021. "Self-Managed Forms of Organizing and Routine Dynamics," in Cambridge
Handbook of Routine Dynamics. Cambridge, UK: Cambridge University Press.

The Role of Social Influence on Action Patterns
Forty-Second International Conference on Information Systems, Austin 2021
9
Kruskal, W. H., and Wallis, W. A. 1952. "Use of Ranks in One-Criterion Variance Analysis," Journal of the
American statistical Association (47:260), pp. 583-621.
Lee, M. J., Ferwerda, B., Choi, J., Hahn, J., Moon, J. Y., and Kim, J. 2013. "Github Developers Use Rockstars
to Overcome Overflow of News," in Chi'13 Extended Abstracts on Human Factors in Computing
Systems. pp. 133-138.
Lindberg, A., Berente, N., Gaskin, J., and Lyytinen, K. 2016. "Coordinating Interdependencies in Online
Communities: A Study of an Open Source Software Project," Information Systems Research (27:4), pp.
751-772.
Miller, K. D., Pentland, B., and Choi, S. 2012. "Dynamics of Performing and Remembering Organizational
Routines," Journal of Management Studies (49:8), pp. 1536-1558.
Pentland, B., Liu, P., Kremser, W., and Haerem, T. 2020. "The Dynamics of Drift in Digitized Processes,"
MISQ (44:1), pp. 19-48.
Rerup, C., and Spencer, B. 2021. "Carnegie School Experiential Learning and Routine Dynamics," in
Cambridge Handbook of Routine Dynamics, M.S. Feldman, B. Pentland, L. D’Adderio, K. Dittrich, C.
Rerup and D. Seidl (eds.). Cambridge, UK: Cambridge University Press.
Tsay, J., Dabbish, L., and Herbsleb, J. 2014. "Influence of Social and Technical Factors for Evaluating
Contribution in Github," Proceedings of the 36th international conference on Software engineering,
pp. 356-366.
van der Aalst, W. M. 2016. Process Mining - Data Science in Action. Heidelberg: Springer.
Von Krogh, G., Haefliger, S., Spaeth, S., and Wallin, M. W. 2012. "Carrots and Rainbows: Motivation and
Social Practice in Open Source Software Development," MIS quarterly), pp. 649-676.
Yu, Y., Yin, G., Wang, H., and Wang, T. 2014. "Exploring the Patterns of Social Behavior in Github,"
Proceedings of the 1st international workshop on crowd-based software development methods and
technologies, pp. 31-36.
Zöller, N., Morgan, J. H., and Schröder, T. 2020. "A Topology of Groups: What Github Can Tell Us About
Online Collaboration," Technological Forecasting and Social Change (161), p. 120291.