Supporting Improvisation Work in Inter-Organizational Crisis Management

by

Benedikt Ley, Volkmar Pipek, Christian Reuter, Torben Wiedenhoefer

CHI 2012

This paper investigates improvisation in emergency response. They claim that their study revealed a lot of improvisational actions (pg 1535) and that many emergency response actions are collaborative (pg. 1535). In the introduction they ask some rhetorical questions:

But what is the exact nature of improvisation work necessary in crisis response? Where do Communication strategies fail? And how can we provide IT support for improvisation at an inter-organizational level?  (pg. 1530)

They attempt to answer some of these, but they only really answer the last one in a small way.

Improvisation

In the background, they cite many papers about the occurrence and the necessity of improvisation. The authors claim “improvisation becomes necessary when beforehand planned decision-making does not work for any reason. This is especially the case in crises.”(pg. 1530) They quote Mendonca and Wallace (2007), “Decision makers in emergencies must be prepared to improvise.”(pg 1530) In a series of citations, they claim that improvisation adds efficiency and nimbleness and that it should be supported:

Improvisation does not consist of more sophisticated methods of more structured systems (Cibborra, 1996). Instead of trying to eradicate it through automation, the appreciation of flexibility and effectiveness seems more adequate. There “improvisation and preparedness go hand in hand” (Mendonca, Jefferson, and Harrald, 2007): Without improvisation, emergency management loses flexibility, without preparedness, emergency management loses efficiency. (pg. 1530)

Relationships

The authors talk about the personal usage of technology such as smartphones to get information and to resolve information, and that this was sometimes necessary because they needed information, but their organizations didn’t provide such technology (this begs the question as to what has changed).

WIth regards to the formal and informal communication, they claim that both are important because of issues like technology failure, which they later use as part of their argument against centralization. When people, especially like EOC representatives are not available, then the need for this decentralization is critical. This could be really important for shiftwork.

Trust

Trust (Trustworthiness as they refer to it) is also an issue in the paper. Here they emphasize goo relationships and the ability to verify information through these. (pg. 1534) They say: “Because of the need for good relationships a lot of informal contacts  and relationships have been established alongside predefined communication lines.

Emergent Teams and Decision Making

In one of the more interesting sections of the background, they talk about the process of decision making with more than one person turning into a ‘muddling-through’ process, which is suggested in (Lindblom, 1959). Sidenote: I wonder what this means for collaboration and design? Is this similar to why design-by-committee things end up so terribly? Back to the paper, they cite a paper describing the emergent nature of teams in emergency response (Turoff, Van de Walle, Hiltz; 2010).

Synthesis

So they touch on the personal relationships as well as the importance of adaptation in effective work, but then they go back to supporting this in IT infrastructure (1536). One of the key points they touch on is the issue of redundancy and centralization of information. While I don’t necessarily think this is a terrible idea to have things in a central location, it also needs to be somewhere else, and be easy to find. This is sort of the crux of the situation. With cloud systems, especially in the case of the UW EOC one, they are still dependent on UW NetID authentication, which if a major earthquake hits and their servers are out, then their system is unreachable for all of them. The same is true of many of these systems. If there’s a few major fiber cuts, how will things work?

Their focus on the informal mostly relates to volunteer organization in Siegen-Wittgenstein than the more populated Rhein-Erft, but they also mention information failure and outdated lists. They state that actors should be able to readily choose which type of communication channel to use, and the structural knowledge of what a person does is important. They also state how important the personal relationships are.

They stop short of making cohesive suggestions about the IT infrastructure, and about the reasoning for why these are important. They’re also fairly schizophrenic about whether information should be centralized or not.

The Human Infrastructure of Cyberinfrastructure

by

The authors attempt to apply the theories of cyberinfrastructure from “Steps Towards an Ecology of Infrastructure Design and Access for Large Information Spaces” ( Star and Ruhleder, 1996) where they do a figure/ground shift and bring the human infrastructure into focus. They use Star and Ruhleder’s eight properties of infrastructures as a lense. The eight they identify (which I haven’t yet verified) are:

  • Embeddedness
  • Transparency (they use an odd definition)
  • Reach or Scope
  • Learned as part of membership
  • Linked with conventions for practice
  • Embodiment of standards
  • Installed based
  • Visible upon breakdown

The definition that they take of Star and Ruhleder’s is:

An infrastructure occurs when the tension between local and global is resolved. That is, an infrastructure occurs when local practices are afforded by larger-scale technology which can then be used in a natural, ready-at-hand fashion. (Star and Ruhleder, 1996, p. 114)

While there’s certainly many of these I could argue apply to the Emergency Responders interviews, especially in relationship to the main EOCs that we’ve looked at, I’m not sure they all apply. For example, the definition of transparency that they list is “infrastructure invisibly supports tasks without needing to be assembled or reinvented for each task.” However, I’m not totally certain that’s really true. Based on the interviews, I think this is difficult to say especially when so many people talk about how each incident is different. One of the EOCs specifically talked about how they would try to prepare alert messages in advance, but how they always had to be shifted. Does the recognition and realization of the necessity of this constitute transparency? Is the expectation of this work in every situation part of an established process, even if it isn’t explicitly in the protocol?

Another example of this disconnect might be the Learned as part of membership, which suggests that “artifacts and organizational arrangements come to be taken for granted by members.” However, I get the impression from the interviews that they struggle with this regularly. I would love to look into some of their incident logs to see how much of a struggle this really is. They suggest that communication is always the problem, and many of their problems involve them meeting with different people to get these resolved. Some things have pre-determined solutions, but if the emergency manager has to keep telling people to get out of their seat and start working with people at whiteboards, how much of these things are being taken for granted? Is it just that at this highest level, where the organizational representatives are put in a room that everything begins to break down? Is this where we see the last point, where the entire thing is visible upon breakdown?

Teams vs. Working Groups and Transient Membership

One thing from the paper that resonated that I believe I see is the assertion that “teams should not be viewed as the fundamental unit of analysis for human infrastructure” (pg. 485) However, they claim that working groups are different from teams, but I don’t think they really elaborate on this anymore and it’s difficult to know which term should be used with a limited understanding of the CSCW literature. They do say that they’ve found “personal networks which exist alongside traditional collaborative structures…” (pg 485) as well as “people are members of multiple networks, groups, and ‘traditional’ organizational structures.” (pg. 485) I think that I see both of these things in the work, and I would assert that both of these are always true. We may not always investigate them, but it seems like it would be hard to fully understand a collaborative system without taking these into account. Later, the authors note that in their study, the people didn’t even know if they were part of certain teams. This, of course, raises interesting questions about identity in the workplace. How much of it is like traditional identity where people can join and cooperate with things but fail to identify with those groups? Do you withhold your identity as a means of withholding support?

Visibility and Perspective

Later, they talk about visibility within the larger organizations and how those can shift from various positions. I don’t know if they’re suggesting this is always true or just of large collaborations like their own. I would argue that this is always true. Each person has their own view, and these views are formed based on relationships within and outside of the collaboration, and those views shift and color the collaboration in many ways. I don’t see this as being specific to large-scale cyber-infrastructure projects. I think this is evidence for people developing selective views of the entire network as well, everyone has selective insight into the workings of the organization, even on the smallest scales. Maybe it requires at least 3 people, but I can’t imagine this not happening if there are ever discussions between some members and not others. I do agree, however, that this is probably beneficial. They attribute part of this to the turnover rate among other reasons (pg 486 and 487). They do bring up a very interesting point:

… what is remarkable is that the organization continues to function in the absence of this sort of mutual visibility. Participants can successully accomplish work with a partial view of the organizational membership and structure. (pg. 487)

This I believe is totally true. It is amazing that this happens, but I would argue it’s common and can be seen almost everywhere, however this hinges on my assertion that people always have selective views. I think this is common sense. All of our views are shaped by our experiences, and this is especially true in an organization where we have interactions that others don’t have. This should be true even if we ignored people’s past experiences. Their experiences are like flashlights waving in a dark room where everyone, with their own flashlight, will see things from a slightly different angle. One could describe the color of the knob on a radio, while the other could describe the handles on a desk. As they get closer to one another, especially with similar facing directions, they’ll start to see more of the same stuff from similar angles, but not always. One is likely shorter than the other, and they still can’t occupy the same space. I know that’s long-winded and blatantly obvious, but I don’t understand why this is a surprise. I have to be missing something in this. Have others asserted that this is not true? Was it previously accepted?

I do think that we see the last part of that quote occuring in the EOCs, or anywhere in these emergency responses. That’s why we have chains of command and these various structures and organizational representatives. The goal is to pass the information along to these other organizations. That part is clear, and this I suspect is sometimes where the communication breaks down.

Personal Relationships

On the topic of personal networks, the authors say, “Personal networks were used to bring workers to the project, but personal networks are also important for accomplishing work.” (pg. 487) They go on to say that seniority often correlated with the number of named collaborators. In one specific group, they say “researchers also had their own personal networks on which to rely for getting specialized information such as clarifications and explanations of clinical terms or research practices.” (pg. 487) They then assert that their organization benefits from their personal networks, but also that it cultivates new and reinforces networks. They conclude by saying, “Personal networks thus augment collaboration as part of the human infrastructure in addition to more traditional organizational structures.” Later, they’re talking about relationships and collaborations in a similar way, “FBIRN strengthens existing relationships, but it also creates many new research collaborations.” I’m not sure how well established these types of relationships are and how well documented they are, but it’s certainly something that we see.

Working Groups as Spaces for Problems and Blueprints

They say that working groups are not teams (contradicts my earlier concern somewhat), but that “act as a conceptual space for types of problems” (pg. 488). I think this is an interesting way to think of this. I wonder if this only specifically applied to working groups, or did they only observe it there. They then go on to describe a human infrastructure as “more like a blueprint that enables people to figure out the basic who, what and where of conducting big science. A blueprint gives one an idea of how parts of a building fit together and what functions they have.” (pg. 488)

Multimorphous and Recursiveness

This was one section that I really enjoyed. I felt the idea about holding many shapes and changing over time to be particularly salient. I think it also somewhat follows from the perspective idea earlier in the paper. In some sense, it never has a shape, in a Schroedinger’s Cat or electron kind of way. It could be changing rapidly, and would change on every measurement from every angle.

“Rather than being amorphous, human infrastructure, might best be described as multimorphous, holding more than one shape at once and also changing shape over time.” (pg. 490)

“human infrastructure is far more dynamic. As members discover their needs, the infrastructure is modified in a recursive type of relationship. The infrastructure is used by people to negotiate work, and in response to these interactions, the shape of the infrastructure is continually negotiated and changed.” (pg. 490)

“Infrastructure can thus merge recursively.” (pg 490)

I’m not fully certain I understand this last quote. I think they’re claiming that it’s recursive in that it constantly changes the system, thus causing other things to react, and thus further changing the system. Is it recursive or simply connected, reactive, and adaptive? (had another word here, but it’s gone and I think that capture the idea kinda) In the sentence before the last quote, they talk about how working groups can lead to other working groups. In the sense that the work of one, leads to or necessitates another specific working group. It’s somewhat different if a single person creates a working group, it has to come from a need recognized in an existing one.

Ok, I’m done analyzing this. I think I’m a little more confused after reading this. I see a few really great points in here, but some of the others leave me really puzzled. I can see this being useful in cases, but I’m not sure I see all of these things in the EM project. Or at the very least, I don’t have enough data to see them properly.

Random other quotes:

“We see that rather than being taken for granted, artifacts such as cognitive tasks are studied and negotiated.” (pg 489)

“human infrastructure renders artifacts highly visible so that they may be discussed, agreed upon, and used, thus moving the creation of new conventions.” (pg. 489)

“Policies have a direct effect on how BIRN-CC designs the computational infrastructure including specifying requirements and subsequently building tools that provide and restrict access to data, computer time, software, and hardware.”(pg. 489)

“Creating a new ontology for any new domain is a huge undertaking that requires specialized expertise and time, both very experience.” (pg. 489)

“The decision to use and build on existing ontologies was understood to be a compromise” (pg. 489)

“Out of necessity, each test bed is now developing their own ontology by domain scientists and ontology experts.” (pg 490)

“Human infrastructure is integral to the painstaking process of creating a data sharing infrastructure that is both dependent and constrained by conventions of practice, existing standards and the organizations that created them. Each level… standards or tools has ramifications for the other levels and must engage with other similarly multilayered infrastructures. The human infrastructure facilitates the connections and communication among people and practices to lay the groundwork for developing data sharing.” (pg. 490)