By Authors: Grace Augustine (The University of Bath); Leanne Hedberg (MacEwan University); Tae-Ung Choi (The Hong Kong University of Science and Technology); Michael Lounsbury (University of Alberta)
Interviewers: Tiffany Yau (University of Southern California, Marshall School of Business); Rotem Rittblat (University of Alberta, Department of Strategy, Entrepreneurship and Management)
Article link: https://journals.sagepub.com/doi/full/10.1177/00018392241287523
Tiffany Yau: What inspired you to focus specifically on recycling coordinators? (Why this unit of analysis?)
Grace Augustine: This project has a little bit more of a personal story. My PhD was on sustainability coordinators, so I studied the second wave of this movement. At around the same time, I had met Mike Lounsbury at a conference, and he was excited to hear that I was looking at how this whole field had evolved over time. He said he had this burning question that he never really got to answer with his work because of the time that he collected his data- as things were unfolding initially with recycling coordinators for his 2001 ASQ paper. This is essentially a story of two ASQs, which is awesome! He’s got the original ASQ on recycling coordinators, and then we have this 25 years later.
Rotem Rittblat: How did the contribution of the downstream process emerge?
Grace Augustine: I’m mainly a qualitative researcher, but a lot of my inspiration for my work comes out of reading quantitative papers, and then seeing where there’s less understanding in the relationship between variables that’s already been established in a quantitative paper. In my PhD, I had read the paper by McDonnell, King, and Soule, from 2015, published in the American Sociological Review Journal. It is a quantitative paper. They show that when activists target companies, those companies respond by establishing new positions that are sympathetic to the activist demands. They found that those positions mattered for future activist targeting (making those firms more open), but they didn’t really have an explanation for why.
The second author, Brayden King, in full disclosure, was my PhD supervisor, so I was familiar with his work, and he kindly worked with me as a qualitative person, but I ended up reading a lot of quantitative stuff in part because he’s mainly quant. I think that paper had me thinking, “Okay, people in these positions clearly matter, but we don’t really know how they go about their work, why they matter, or why they would make organizations more open over time to movement demands.”
As qualitative researchers, it’s still important to have the skill to read, digest, and make sense of quantitative work, because it can be a big source of inspiration for carving out qualitative papers, where findings that are in quant work maybe don’t have the mechanisms fully understood.
“We know so much about decoupling, greenwashing, and everything; it just doesn’t make sense that that’s enough for us to know”
Tiffany Yau: Are there any specific questions that you’re asking or diving into when you’re reading these quantitative papers?
Grace Augustine: I personally think very processually, and that’s not necessarily how quantitative work with traditional statistical models can really be. I am often thinking, “okay, I’ve got the two bookends here, but I want to know what’s going on in between. Or A leads to B, but then what happens?” In this case, the downstream is relevant as well, because there are lots and lots of papers about what leads to companies making concessions to social movements.
One thing we talk about in the paper is that we don’t find that a satisfying end to the story, because we know so much about decoupling, greenwashing, and everything; it just doesn’t make sense that that’s enough for us to know. I really want to know what happens next.
Rotem Rittblat: Why did you decide to use process studies, and how do you understand the difference between a historical perspective and process studies?
Grace Augustine: This question really made me think. I am aware they’re not the same thing. I guess there are certainly process studies not claiming to be historical, but this could be because they are not covering a very long period or not using archival data. Then, there are also historical studies that are not necessarily processual. But I think the two lend themselves well to one another, and if you have a long enough and detailed enough set of data, then you’re able to kind of do both.
This study was more of a historical archival study, but the data just happened to be digitized. It has a similar historical feel to going in an archive and picking up boxes of correspondence, but it just happened to be the early days of online communication, when email correspondence was more common than letters.
Also, I think repeated interviews or multi-wave interviews would be another way to get at both of those things. It’s not that archival data is the only option. But if you’re not doing any kind of overtime data collection, it’s hard to get that processional and that historical view.
A lot of historical studies and their findings are presented in phases, and they’re not necessarily trying to explain the connections between the phases. Those may be more common in some of the historical studies, like market category emergence, but they may not be trying to get at the mechanisms. I’m really passionate about hitting both of those marks, but you need the right kind of data to do so. While working with my PhD students, I am often saying, “I know you’re tempted to make this a process study, but with snapshot interviews, the data may not be rich enough to really understand those processes.”
Tiffany Yau: With historical analysis, there are various factors to consider- for example, source criticism. This is such rich data. How do you consider triangulating your data across different actors to compare and contrast their views? You mentioned that with interviews, it’s just a snapshot, so how do you decide when to use those snapshots?
Grace Augustine: I do think triangulation’s really important. What’s interesting in this study is that we didn’t collect the semi-structured interviews until after maybe the second R&R, they were definitely not in the initial submission, so this can be a bigger conversation about how much a paper can change the review process, and thankfully, ASQ is very developmental and open to things shifting.
With the triangulation, we ended up just doing six interviews, which doesn’t sound like a lot, but we made sure they were really targeted at core actors. We wouldn’t have known who those important voices were until we had really dug into the online forum data and those conversations, and then we could see these are central actors.
The key burning question for us, and why we felt we needed to collect that data to triangulate, is that in our case, the recycling coordinators largely experienced demise of the occupation, but they also had quite a lot of success in establishing the goals of what their occupation was set out to do. So, we were left wondering how they felt about this. Looking back, we asked, “Was this a success? Was this not a success? How would we classify this?” Because the occupation didn’t survive, it worked itself out of a job. We were struggling with putting our own judgment onto the case, and we really wanted to hear from people who experienced it.
I think triangulation was really important, but, interestingly enough, it didn’t come up as necessary until a little further down the line. You may not always start with the right data for triangulating or realizing that you need it. Then sometimes reviewers might want to know more, or they may be skeptical about a certain claim that you don’t have very robust evidence for.
Rotem Rittblat: What I hear is that you need to be open to making changes during the review process, even though this can be very difficult. Do you have any recommendations for scholars dealing with these processes?
Grace Augustine: I think with qualitative data, in the very first submission, you have to show the reviewers and the editor that you have enough data to back up the claims that you’re making, and that you have enough data to potentially go deeper in the analyses. If the data feels quite thin in the initial submission, then that is a riskier submission, because if they’re not crazy about the framing or the conclusions that you’re drawing, they might not have that reassurance that you have enough data to make a contribution. There’s no number I have in mind for interviews or anything, but I’m a senior editor now at Organization Science. I now see that in how reviewers respond initially, and I would say are more willing to give a paper that is just not quite at that refined level in that initial submission a second chance if the data seem robust.
It’s not easy. Most people don’t navigate the review process entirely on their own, so, thankfully, co-authors and advisors who’ve been through it before are hugely valuable as well.
Regarding your decision to apply machine learning, which is, of course, super popular nowadays. Can you elaborate on this aspect of doing research, in your specific context, but also more generally today?
Grace Augustine: This also was not in the first submission, but this relates to a wider shift. We have a problem that qualitative researchers are increasingly encountering, which was not a problem 20 years ago: too much qualitative data. Historically, when you’re reading qualitative methods textbooks, everything is about access to and scarcity of data and building up a sufficient data set. We had the opposite problem in this paper. We had 31,000 messages and needed to reduce them to a manageable number for hand coding, since we wanted to conduct qualitative analysis on them.
One of the challenges we had is that the majority of the messages were not theoretically interesting to us. The majority of the messages were operational, day-to-day trading information kind of messages, which didn’t necessarily have insights into the occupational identity, or goals, or, like, the broader occupational project, or those kinds of things that we were more interested in getting at. We had this really big challenge of how we could identify the messages that we found more theoretically rich, and how we could exclude all those messages that were more like, “what brand of bin do you use,” or “how many times a year do you do dormitory collections?”
We tried a keyword search because we knew the field fairly well. Some keywords were bringing up those richer conversations. Still, it didn’t work that well, to be honest, because it’s the same keywords- for example, that were tied to our first set of findings around operationalizing recycling, unsurprisingly, would bring up all those kinds of information exchange messages. Our first submission was a keyword search. And the reviewers pushed back a bit because we couldn’t really justify the keywords that we chose very well, other than knowledge of this case, which some people use, and sometimes that flies, and sometimes it doesn’t fly.
We had someone on our co-author team who was in my PhD program, Tae-Ung Choi, who’s just super skilled, technically and quantitatively skilled, and he was like, “We could try machine learning,” and I was like, “What’s that?”. It’s ironic that I’m the first author, and when he brought it up, I really had no idea what he was proposing. All the credit for the machine learning approach goes to Tae-Ung.
He was constantly doing checks and doing samples and subsamples, and showing us, this is what’s been excluded, and this is what’s been included. It was a very detail-oriented, meticulous process, but it was brilliant! When we got the sample that it had produced from our initial coding, it was so helpful, and then the reviewers went for it. I think in part because there’s a bit of methodological novelty there at the time.
Tiffany Yau: As you alluded, there are a lot more people utilizing these methods of AI. There are different sorts of critiques around it, so what would your recommendations be to scholars today and in the future in trying to approach these types of methods?
Grace Augustine: I would be really careful to say in this paper that the only thing we used machine learning for was to create the sample. I think nowadays, more of the debate is around the coding itself, and whether or not you can do human-machine coding. I’m probably more of the old-school mentality currently, and still quite a purist in terms of my analysis. I would personally still be quite nervous to integrate any kind of AI in the actual making sense of the data part.
But like I said, I was totally a convert with the sampling approach that I was previously completely unfamiliar with. I’m open to having my mind changed on this. But at the moment, I’m mindful and recommend a highly considered adoption on where to appropriately use it.
I’m keeping an open mind, but I think the journals are also really struggling with developing policies that make sense regarding this. I think they’re following a more conservative model for the time being, because there’s a lot of uncertainty about what the implications for our research will be. Still, I wouldn’t be surprised if we see journals become more open to some kind of integration of these tools – with full disclosure and transparency of course.
Rotem Rittblat: What can we learn from this paper with regard to other roles that are associated with social movements?
Grace Augustine: The main message is really kind of a double-edged sword, and I alluded to it earlier with this idea that, in lots of ways, this occupation was a huge success. They were able to institutionalize something that is so taken for granted now that it’s just almost completely in the background. That’s one definition of institutionalization. You don’t actively notice recycling on a college campus in most contexts. It’s just assumed that it’s there. It’s absolutely taken for granted. But the kind of sad part and what was lost is their really lofty environmental ethos and environmental ambitions, and how they got continually pigeonholed into this operations-only role. And there were some things we talk about in the discussion section; there were some things that they did or could have done differently in terms of learning for similar occupations.
One thing that I really think about a lot with this is around the business case framing, or the business case justification for these types of roles in organizations. The issue selling literature talks about the importance of selling issues in terms of the language that your audience cares about, like talking about the innovation from a diverse workforce, for example. I do think that has a lot of short-term gains, but in this paper, we find that it’s not necessarily the best approach for the long-term implementation of these ideas. We saw that with the recycling coordinators, and the continual framing of their work as cost savings really backed them into a corner. Because then, when they wanted to expand to other areas that didn’t have that immediate return, they had built this justification for themselves and organizations purely around cost savings. I think there are learnings there for other occupations in terms of business case framing. Sarah Kaplan has written on this a bit, too, the kind of backfiring of business case framing, or risks of business case framing, for these types of social or environmental goals. My biggest takeaway from this is the idea that there needs to be more of a layered framing. The business case can be one approach to framing why this work is important.
But I think it’s kind of risky to only frame—for example, diversity in terms of innovation—because then you’re losing the whole sort of moral argument from day one. Like, you’re disregarding the real reason why we should be doing it. My hope would be that these findings might provide more ground for those occupations to expand their work as the situation changes, as they get more of a foothold, and not get pigeonholed in the same way we saw with the recycling coordinators.
“We talked about the recycling coordinators as the environmental conscience in these organizations.”
Rotem Rittblat: As I listen to you, I keep thinking about the implications of AI for future-related occupations. Professionals in these fields may need to embrace such ideas and adapt to this emerging reality when reflecting on questions of right and wrong, as well as broader ethical considerations.
Grace Augustine: I think ethics officers in AI are an absolute contemporary example. There’s societal pressure on AI companies to think more about the ethical implications of the technology.
And so, they create this new occupation, this new role, to show social movements, “Yes, we’re listening, we’re responding, we are, institutionalizing a role, and this person’s job will be to be that conscience”, and we talked about the recycling coordinators as the environmental conscience in these organizations. But, what that person’s authority is, how they might get pigeon-holed, how they might frame their work purely in terms of regulatory effectiveness or something, I’m just guessing, they’re saying, basically, “if you have this role, then you’re gonna be more likely to probably circumnavigate certain regulation (like we saw with affirmative action officers in Lauren Edelman’s work), and then how that could lead in the future to a very restricted mandate or very restricted set of tasks that that person’s able to oversee.”
That’s a really good contemporary example of an occupation that could make similar mistakes along the way. But we often hear from sustainability managers, “If I’m successful, I’ll work myself out of a job.” So this is an interesting idea amongst these movement-oriented occupations: they almost want to embed everything in the organization so much that their particular role is no longer necessary. I do think that is an interesting take on occupational success, and that’s very different from our assumption that success is longevity or success is staying in the role indefinitely, I think some of them would look back and say, actually, a big part of the reason we’re not there anymore is because we actually achieved quite a lot.
Tiffany Yau: Since there are various co-authors on this paper, when you’re working with a larger co-author group, what is your advice for junior scholars?
Grace Augustine: It’s not always easy. I’ve had a couple of four-person papers, and what was nice at the beginning of this project was that we were only at two institutions. That was easier because Tae-Ung and I were at Kellogg, and Mike and Leanne were at Alberta.
When the project started in 2018, life was a little easier when we were just at the two institutions, and we were just 4 hours apart in time zones. It got more complicated towards the end, as I was in the UK, Mike and Leanne were on the west coast of Canada, and Tae-Ung had gotten a job in Hong Kong. There was only one hour in the day that was reasonable for everybody to meet. This is very boring advice but t you’re going to be working, with an ambitious project, you’re going to be working with people for 5 or 6 years, and so being very mindful of those logistics from the beginning, and of course things change. In this case, we were really lucky; we had great dynamics, but the logistics were challenging. We also asked for lots of extensions. I took a whole year of maternity leave under the UK system, and ASQ gave me that whole-year extension on the paper. And I’ve heard stories from other friends who have not gotten those kinds of journal extensions, so it’s really good that ASQ respects authors’ time, and if people need more time, they will give them that time, generally.
If you’re in this kind of environment, then, things might take longer working with 4 people, but if everybody’s kind of open to the fact that I have other projects in my pipeline that are, like, 1- or 2-year projects, ideally, but I can afford to have one project in my pipeline that’s, like, a 5- or 6-year project. I just know it’s going to bumble along, and it’ll hopefully get there eventually. It’s not like everything you’re working on needs to be a six-year project. There are other things that you can be doing in the meantime… it’s okay to know that some projects like this just might take longer but of course it’s worth it when you get there. This is a project all of us can be proud of.
Papers mentioned:
Lounsbury, M. (2001). Institutional sources of practice variation: Staffing college and university recycling programs. Administrative Science Quarterly, 46(1), 29-56.
McDonnell, M. H., King, B. G., & Soule, S. A. (2015). A dynamic process model of private politics: Activist targeting and corporate receptivity to social challenges. American Sociological Review, 80(3), 654-678.
Interviewer Bios:
Tiffany Yau is a doctoral student in the Department of Management and Organizations at USC Marshall School of Business. Her research examines the intersection of entrepreneurship and social evaluation. She studies how industry actors navigate uncertainty and collective dynamics to both work within and reshape institutional environments.
Rotem Rittblat is a postdoctoral fellow in the Department of Strategy, Entrepreneurship, and Management at the University of Alberta. She uses qualitative, comparative, and ethnographic methods to analyze the sociocultural sources and consequences of innovation and entrepreneurship. Approaching these dynamics from a co-constitutive perspective, her work spans varied institutional contexts, with a particular interest in healthcare organizations.
Get new ASQ Blog posts by email