Features

XIX.4 July + August 2012
Page: 50
Digital Citation

Taking action in your research


Authors:
Gillian Hayes

How can you make a real difference in the world? Whether through volunteer work, joining a service profession like teaching, or just in our everyday behaviors, many struggle to answer this question all the time. To some, science and research are divorced from this inherent human need to make positive changes in the world—if not divorced, at least distanced. However, more and more designers and researchers in HCI are challenging that long-held view. We can, in fact, create positive social change and simultaneously do good research. It’s just a matter of slightly changing the way we think about both scholarship and participation in research. It helps to borrow from other disciplines, such as education and the social sciences, that have been doing it for years. To that end, I have found action research (AR) to be helpful in ensuring high-quality change and high-quality scholarship in cooperative and participatory research.

interactions readers are no strangers to the idea of work having broader impacts. The CHI Social Impact Award in many ways applauds just these kinds of efforts. Judy Olson recently presented on this topic in her Athena Award keynote at the CSCW conference. The NSF has used broader impacts as criteria for grants for years. Papers published at CHI and other HCI venues often focus on social issues, such as healthcare, education, sustainability, and green IT; HCI solutions for developing and conflict-ridden nations; and so on. And of course, HCI has a strong tradition of participatory, cooperative, and democratic design orientations, such as Scandinavian and participatory design.

Anyone who does this kind of work is familiar with the critique that it is not sufficiently systematic, generalizable, or, dare I say, scientific. At the same time, we interventionists are not immune to being overly critical at times, and can be found lamenting the impracticality and infeasibility of many research solutions. While we will never satisfy fundamentalists at either end of the scholarly distance versus local cooperative pragmatism debate, action research offers a systematic collaborative approach to conducting research in HCI that satisfies the need for both scientific rigor and promotion of sustainable social change. Action research is learning by doing, a form of research that necessitates taking some action for the dual purpose of addressing a problem and learning something from that action. AR “aims to contribute both to the practical concerns of people” in problematic situations and to the academic goals of science “by joint collaboration with a mutually acceptable ethical framework” [1].

To help clarify what action research entails, here are responses to a few common questions researchers have about it.

What does this mean in terms of the practical execution of the research activities? AR requires iterative “planning, action, and fact-finding about the result of the action” [2,3]. That is to say, AR does not have a prescribed set of steps to follow and is not a method in the traditional sense.

If AR is not a method, then what is it? AR is an approach to research that commits a necessarily interdisciplinary research team to act democratically and collaboratively. Research must be undertaken with people experiencing real problems in their everyday lives, not for, about, or focused on them.

If AR is not a method, what methods does it use? AR can incorporate multiple methods and welcomes the use of both qualitative and quantitative methods. The only methods not applicable to an AR approach are those that distance researchers from problems and questions of inquiry to ensure “objectivity” or avoid “contamination.”

AR builds on the many democratic and inclusive approaches to research to focus on highly contextualized, localized solutions (see [4] for a review of the history). In the end, scholarship in AR is about the ability to transfer knowledge (and sometimes solutions themselves) between contexts and domains, not to generalize knowledge or solutions to a larger case. Given the tradition of cyclical design methods and the impulse to make solutions that really work, HCI research is well primed to make use of AR as a platform to conduct research that is simultaneously socially meaningful and scientifically rigorous.

How Can HCI Researchers Get Started?

HCI researchers not already engaged in AR should have an easy time of taking it up, because HCI researchers already do a lot of the things required of AR, such as the following:

  • working with community partners
  • working in the field, where real problems reside
  • designing and developing solutions iteratively.

There are still some things from traditional HCI work that need to change, however, to take an AR approach. First, we must learn to act not as researchers, nor even scientific advisors, but as “friendly outsiders” [5]. That is, we must explicitly reject the notion of scientific distance and objectivity. We must instead open up lines of communication and design solutions with, not for, community partners. Perhaps most important, we must take the radical view that the contributions of these community partners are as important and as scholarly as our own scientific and professional knowledge. AR is not quite as simple as following seven easy steps, but they will get you started:

Step One: Find a community partner. Sometimes researchers start with an idea about a problem they want to tackle and find a community partner (e.g., teachers, nonprofit organizations, public advocacy groups) with whom they can work whose expertise is in this problem area. Sometimes a community partner might seek out the researchers. Both approaches work well as long as the relationship is established early and nurtured throughout the project. This partnership sits at the heart of any AR project and is easily the most important (and riskiest) step in the process. Therefore, it is key that teams take time to get to know one another and to understand one another’s goals, limitations, and resources.

Step Two: Formulate a problem statement and some research questions. Research questions should be developed collaboratively with the community partner you wish to engage. This approach tends to create inherently interdisciplinary research with contributions to multiple fields, as well as to the community problem. Research questions may also arise that the team does not have the expertise to answer, opening the door to collaborations with other researchers who might also like to become involved with the project at this point.

Step Three: Plan and execute some action. After thoughtful consideration, AR teams create interventions, which often include a mix of social, process, and technology changes. Adopting an attitude that focuses on the outcome of learning something, regardless of the “success” of the design or intervention, frees up the team to attempt interventions that may be risky or underdetermined.

Step Four: Evaluate and plan again. Evaluation in AR, just like problem definition and intervention design, is a value-laden enterprise. It is important to ensure that the evaluation addresses both the research questions and pragmatic concerns of the community partners. Based on the results of the evaluation, teams should begin to plan for additional intervention and possibly develop new research questions—in other words, “rinse and repeat.” This is, after all, a cyclical process.

Step Five: Share what you learn. Throughout the iterative cycles of a long-term AR project, team members should be sharing their results with the outside world, and they should be doing so as a group, with full inclusion of the community partners. Of course, how results get shared can vary wildly from project to project, and may include internal reports for the community organization, scholarly works in the domain area or in HCI, presentations to community organizations or government, and so on.

Step Six: Don’t forget to stop to enjoy things from time to time. Because there is no clearly defined ending point in most cases, it is also important to recognize intermediate moments of success throughout the project. AR requires sustained long-term engagement with research sites and community partners. This kind of relationship and effort can be exhausting to all involved. Thus, using moments of celebration to demarcate the beginning of new phases and the end of old ones can help build more collaborative teams and reinvigorate everyone involved.

Step Seven: Step back and trust in the sustainable change. Once the research facilitators leave, the community partners should be able to maintain the positive changes that have been made. In my experience, this is one of the most difficult parts of a successful AR project. Leaving behind the technology artifacts is not enough. The partners must be empowered to use and maintain them.

But Is It Science?

The best way to understand something is to try to change it. —Kurt Lewin

AR is about finding knowledge through action. In the case of HCI research, the “action” is a sociotechnical solution to some human need. Given the amount of change AR introduces into an environment, those who advocate for generalizability and scientific distance sometimes struggle with AR as a scholarly pursuit. AR demonstrates scientific rigor through trustworthiness [6], not generalizability. Trustworthiness, as defined by Lincoln and Guba, comes from a combination of credibility, transferability, dependability, and confirmability and can usefully be applied within the domain of HCI [7]. The long-term nature of AR and its emphasis on participant language and multiple perspectives help build credibility, but the biggest measure of credibility is that the solutions actually work to address real problems in the lives of the research participants [5]. These solutions are inherently contextualized and localized to the situation in which the AR project took place. Thus, the goal is transferability from that local context to another, not generalizability to all contexts. To accomplish this goal, data must be collected, analyzed, and described as transparently as possible (supporting dependability). Furthermore, enough evidence must be presented to confirm the events transpired as described (supporting confirmability). Others can then examine the similarities and differences of their own situations as they relate to the published AR project and replicate appropriate parts of the solution while changing others. Thus, even with its emphasis on local solutions, AR provides a rigorous framework for generating and sharing knowledge about a solution to allow for knowledge transfer.

Discussion

The HCI community strongly supports iterative user-centered design and collaborative and democratic processes, such as participatory design. Systems development and design communities also regularly engage in cyclical product-development processes that privilege intense engagement with the future users of their products. User-centered design uses a cyclical approach of gathering formative information about user needs and experiences, designing and testing prototype systems, and creating new designs and understanding based on these past experiences.

However, it is important to note that although software design and development bears many similarities to AR—particularly participatory and iterative approaches to these activities—it is not the same as AR. Most notably, the end product of AR is the same as the end product of any research: learning and scholarly knowledge, whereas the end product of design and development is typically a technological artifact. Additionally, designers are often tasked by some power or authority to create a solution, whereas action researchers tend to attempt to align themselves and co-create solutions with those in positions of limited power who primarily are empowered to act locally.

AR is focused on local solutions to local problems. Research projects in HCI also often result in local solutions to local problems. Why then do we still write up these projects with a focus on generalization? The truth is that HCI work that looks like traditional “science” (e.g., statistically significant surveys, carefully controlled lab studies, fieldwork conducted on research subjects, not with research participants, etc.) still dominates HCI journals and conferences. Using some of the language of AR with a focus on transferability, not generalizability, we may be able to change the ways collaborative research must be written in order to be published. In the meantime, there are publication venues that are quite friendly to discussions about democratically and collaboratively developed knowledge (notably, those focused on participatory design and action research).

AR is not for everyone, and AR is not for every project. Many of us want to do good in both our lives and work. One need only look at past issues of interactions or the annual CHI conference to see how much we in HCI like to do good in the world. So, it is important to mention a couple of quick caveats. First, “Too often people engage in meaningful participatory and democratizing change processes” and then claim to be doing AR, but there is no “research” element to the project [5, p.99]. In starting an AR project, researchers must ask themselves not only where the action is, but also where the research is. Second, AR raises additional ethical questions beyond the usual questions we might encounter in our design and research practice. The relationships involved in an AR project are necessarily close. These kinds of close relationships mean that community partners sometimes acquiesce to researcher interests in the name of being polite. (The reverse can also be true.) These close relationships also muddy questions about who chooses the action to be undertaken and how it is chosen. Finally, becoming personally and professionally involved with members of a research site may make it difficult for researchers to leave and allow the sustainable change to set in. It feels good to be needed and to be part of a team that is doing something to effect positive change in the world. Why would anyone want to leave that setting? For that matter, as a community partner, why would you want to let an important team member leave? In the end, however, the ethics of AR require that community partners are willing and able to carry on the work without the research partners.

This article is not about making everyone in HCI into action researchers. For that matter, although I strongly believe in AR, a lot of my projects are not AR projects. As I mentioned, AR is not for everyone or every project. There are many ways to involve community partners, to promote participatory and democratic design processes, and to do good works through HCI without using AR. My hope here is not to advocate for AR per se, but to bring awareness of its commitments, processes, and vocabulary to people who are doing research that involves substantial community engagement and collaboration. This kind of substantial long-term democratic engagement with community partners during research is incredibly difficult to do, much less to do well, but it is well worth doing when the fit is right. For those interested in this kind of work, I hope this article will stimulate enough interest to prompt them to read some of the excellent books on the subject (see sidebar), or even my own longer article in ToCHI [4]. Finally, I hope this article provides some guidance on getting started to those new to AR and that I have provided some vocabulary that may be useful to those already doing this work who are fighting to ensure scientific rigor while truly collaborating with community partners.

References

1. Rapoport, R. Three dilemmas in action research. Stronger Families Learning Exchange Bulletin 23, 6 (1970), 499–513.

2. Lewin, K. Action research and minority problems. Journal of Social Issues 2, 4 (1946), 34–46.

3. Lewin, K. Resolving Social Conflicts. Harper, New York, 1948.

4. Hayes, G.R. The relationship of action research to human-computer interaction. ACM Trans. Comput.-Hum. Interact. 18, 3 (August 2011), article 15.

5. Greenwood, D.J. and Levin, M. Introduction to Action Research (2nd Edition). Sage Publications, Thousand Oaks, CA, 2007.

6. Lincoln, Y.S. and Guba, E.G. Naturalistic Ingulry. Sage Publications, Beverly Hills, CA, 1985.

7. Stringer, E.T. Action Research. Sage Publications Newbury Park, CA, 2007.

Author

Gillian R. Hayes is an assistant professor of informatics in the School of Information and Computer Sciences and in the Department of Education at UC Irvine. Her research interests lie in HCI, ubiquitous computing, assistive and educational technologies, and health informatics. She is particularly interested in the ways in which people keep records, document their everyday lives, and are documented by the record-keeping systems around them.

Figures

UF1Figure. AR is commonly represented as a spiral of planning, action, and reflection.

Sidebar: Further Reading on Action Research

  • Action Research by Ernest P. Stringer
  • The Action Research Dissertation: A Guide for Students and Faculty by Kathryn G. Herr and Gary L. Anderson
  • Participatory Action Research by Alice McIntyre
  • Introduction to Action Research: Social Research for Social Change by Davydd Greenwood and Morten Levin
  • All You Need to Know About Action Research by Jean McNiff and Jack Whitehead
  • Handbook of Action Research: Participative Inquiry and Practice by Peter Reason and Hilary Bradbury-Huang
  • Participatory Action Research: Theory and Methods by Jacques M. Chevalier and Daniel J. Buckles
  • Participatory Design: Principles and Practices by Douglas Schuler and Aki Namioka

©2012 ACM  1072-5220/12/0700  $10.00

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.

The Digital Library is published by the Association for Computing Machinery. Copyright © 2012 ACM, Inc.

Post Comment


No Comments Found