The Past,Present,and Future of Organizational Learning Research

Why don’t we first start by asking Linda about her early career? You have become a very prominent management scholar. I wonder how your career evolved.

Well, I started college in ’71 at Tulane University, which was in my home city of New Orleans, thinking that I would become a math teacher. Then I took a psychology course, taught by Gordon Gallup, which seemed fascinating, because it had the rigor of math, but with very interesting questions. So I became more and more interested in psychology. I had an opportunity to study abroad through a scholarship to go to the University of Nottingham, England, in 1973-1974. That was an era, those of you that were around then can remember, in which there were many changes in women’s roles and women’s opportunities. I returned to Tulane thinking, “Well, maybe I could do something different than being a math teacher.” I did an honor’s thesis in psychology under the supervision of Ed O’Neal and Peter McDonald and then went off to the Organizational Psychology program at Michigan for graduate work. I had become aware, in part, from the teaching that I had done in schools, that the organizations were doing the same work but with very different outcomes, and it seemed to me that there were these organizational factors that would be so fascinating to try to understand. I went off to Michigan, not knowing much about what it took to get a PhD. And maybe that was good [Audience laughter]. At Michigan, I benefited from the opportunity to work on a project through the Institute for Social Research that was led by Basil Georgopoulos and also opportunities to interact with Dan Katz, Jim House, Hazel Markus, and Denise Rousseau, who kindly served on my qualifier and/or dissertation committees.

In your earlier works, you studied the structural contingency of organizations. Then your work moves into robotics, and further into organizational learning. We were wondering how this change in research happened?

When I went to Carnegie Mellon in 1979, the university had just launched the Robotics Institute. Building on my original interests, I thought I would apply structural contingency theory to new task and technology characteristics that were engendered by the robots. Originally, I saw the work as an extension of structural contingency theory. Paul Goodman and I did a couple of case studies (e.g., Argote & Goodman, 1986; Argote, Goodman, & Schkade, 1983) to begin the research. As I got more into that area, I realized—and this was the early ’80s—that there really were not that many robots being used yet. To be able to study them, it was going to take many years, which is certainly not the length of time that one has early in a career, at least not for getting tenure. So while the project was disappointing in terms of the initial goal, it was very useful because the work actually took me into plants, mainly manufacturing plants in automotive and aerospace industries, where I saw the phenomenon of an organizational learning curve. As a psychologist, I had seen learning curves at the individual level, but I had not yet been exposed to organizational learning curves, which seemed to be very interesting phenomenon. I encountered one plant that had fired a manager because he was not moving down the learning curve fast enough. There was the normative belief that he should achieve an 80% learning curve, which meant that every time cumulative output at the plant doubled, costs would decline to 80% of their previous value. The manager had not achieved this goal. But when I started looking at the literature, there seemed to be a lot of variation in the learning rates that people didn’t understand very well and that seemed like an interesting puzzle to pursue. Also the whole issue of “forgetting,” or whether the knowledge in the learning curve was cumulative or whether it depreciated, was unexplored. Although a few papers had mentioned that “forgetting” might occur, all of the empirical work at the organizational level assumed that knowledge was cumulative and persisted through time. Of course, there is considerable evidence of forgetting at the individual level. But organizations are different from individuals. So understanding whether knowledge decays or depreciates at the organizational level seemed an interesting issue as well. So the work in the plants to study robotics exposed me to these phenomena that I found fascinating. So I started to work on them.

Your work on learning curves started in the early 1990s (Argote, Beckman, & Epple, 1990; Argote & Epple, 1990; Epple et al., 1991), and we found that you frequently mentioned this assembly plant of trucks. I wonder what kind of plant it was, what kind of experiences you had, what you learned there, and how it influenced your subsequent work?

Great, so this was when we (Dennis Epple [Linda’s husband, an economist she met at Carnegie Mellon] and Linda) came back to Carnegie Mellon after spending 1 year at Stanford University as visiting professors (1984-1985). While at Stanford, I enjoyed the opportunity to participate in a seminar on organizational learning that Jim March organized. Given the central role that Jim has played and continues to play in the area of organizational learning, this was a wonderful opportunity. I also benefited from interacting with colleagues in the Department of Industrial Engineering and Engineering Management where I was a visitor. Upon our return to Carnegie Mellon, Dennis and I had an opportunity to get involved with a project at three truck assembly plants, and the angle we were interested in, not surprisingly, was learning. I think three things were especially valuable about our interactions with individuals at the plants. One was the data they provided. The second was just being there and getting to talk with the managers and engineers. I think that is what crystallized the questions we later worked on. We interviewed the managers to try to be sure we understood all the factors that could affect learning curves. A lot of what the managers said had been discussed in the literature, maybe not studied empirically, but at least had been discussed. But our respondents said something that had not been mentioned before, and that was “knowing who is good at what,” and assigning tasks accordingly. The managers did not use the term transactive memory, but around the same time, Daniel Wegner (1986) was publishing his work on transactive memory, which was a similar concept but at the dyad level. Wegner applied the concept to people in close relationships and how they specialize. For example, one person might be the expert on how to get things fixed, and the other person the expert on birthdays on both sides of the family.

Can you tell us more about the state of the learning curve research at the time you started it, and which pieces were missing in the literature? Can you give us examples from your research about the findings that you have found about learning curves?

One issue that we saw was variation in learning rates. So these three truck plants we studied were all producing the same product and were all part of the same organization but had very different rates of learning. That suggested to us that there is something going on in the plant, because the macro factors were constant and the product factors were constant. Understanding that variation was important, and other people have pursued the issue as well (Edmondson, Bohmer, & Pisano, 2001; Henderson & Clark, 1990).

This naturally leads to your second research topic, which is knowledge transfer. I guess the most famous piece was the pizza . . .

The pizza paper . . . Right, that is how our paper has come to be called [Laughter]!

Can you talk more about that paper with us?

That research is joint work with Eric Darr (Argote & Darr, 2000; Darr et al., 1995) and Dennis Epple. Eric was a student in my course on organization theory at the time—this was early 90s—and I had mentioned how we had been doing the work in the truck assembly plants. But there were only three plants producing the same product. So our “N” was too small to do anything at the organizational level of analysis. We wanted a setting with lots of fairly standardized organizations, where we could meaningfully compare their productivity, and where there would be interesting variation. Eric had the brilliant idea of pizza franchises, which we came to refer to as the fruit fly of organizational research [Laughter]. With his leadership, we obtained data from 36 pizza stores in one geographic area. Several of these stores were mom and pop stores, where one owner had just one store, and the maximum in our sample was one owner having 11 stores.

I wonder what kinds of mechanisms were found for knowledge transfer in your interviews.

The things our respondents mainly mentioned were observing another store, which was more likely to happen among common ownership, and communicating with other stores. I think personnel movements were also incredibly important. We saw that this happened more within commonly owned franchises than between them. When we did the work, we tried to catalog the 14 different innovations that our respondents had mentioned. We catalogued whether they were embedded in individuals, in tools, or in routines.

Thank you. Can you share some of your recent works on knowledge transfer?

One of the things that we have tried to look at is the role of identity, which we think is very important. If you share an identity with other individuals, you are more likely to give them a chance to explain their knowledge and to try to learn from them. If you don’t share an identity, you are more likely to reject their ideas. I am not saying there is blind adoption if you share an identity with others, or that you will do whatever they say. But you will give them a chance to explain or demonstrate what they know and see if you can learn from them. I think the whole issue of identity is fascinating. Aimee Kane, John Levine, and I went into the laboratory to investigate identity, because we felt that we could isolate the effect of identity there. I am glad to see more lab work being done in strategic management. The lab work nicely complements fieldwork.

Why don’t we move on to your next research area, transactive memory?

Can you tell us a bit more about the elements of the transactive memory system? What is it actually?

We—Dick Moreland, Ranjani Krishnan, Diane Liang, and I—went into the lab for the first study. We wanted to see whether transactive memory positively affected team performance and also to identify indicators of transactive memory. One indicator was specialization, where different people were doing different things. This is a form of memory differentiation. If you know who was good at what, you know whom to go to. So there was an element of coordination built in.

Given the conference we are at, we would also like to talk about the linkages between organization learning and strategy. We noticed a shift towards the strategy field in both the research you were relating to, and also the publications that are citing your work. That shift seems to coincide with the influential paper that you wrote with Paul Ingram published in 2000 (Argote & Ingram, 2000). Can you tell us more about how you got in touch with the strategy community?

As you pointed out, when Paul Ingram arrived at Carnegie Mellon in 1994, I became more deeply engaged in strategy-related research. That was the same year that Sid Winter invited me to a conference in Vienna. Dan Levinthal, Richard Nelson, David Teece, Dick Rumelt, Rebecca Henderson, Gabriel Szulanski, and other wonderful researchers were there. It was a great conference to get to know more strategy researchers and to start realizing that there was mutual ground. I could learn a lot from them, and they were kind enough to say that what we (Linda’s research group) were doing had implications for them. So I started reading more research in strategic management, including work by people such as Constance Helfat in addition to those at the Vienna conference. Other strategic management researchers such as Marvin Lieberman had been working on learning, too. There was a lot of really interesting work in the field of strategy that I have certainly enjoyed reading and tried to connect to.

How do you see the organization learning field evolve in the near future?

In my own work, I am getting a bit more micro trying to understand what leads to the adoption of routines or exploration–exploitation at a fairly micro-level. In a project with Dorthe Hakonsson, Jacob Eskildsen, Dan Monster, Rich Burton, and Borge Obel, we examine emotion and performance as predictors and consequences of the decision to explore a new routine versus exploit an existing routine.

There is a recent trend in studies of cognitive neuroscience, meaning that people start using eye tracking devices, brain imaging, and so on. How do you see this line of research influencing organizational learning or learning in groups?

We published an interesting piece related to this by Senior, Lee and Butler in Organization Science in 2011. I think cognitive neuroscience has the potential to say something about the mechanisms behind organizational learning. For example, we differentiate between mindful and less mindful processing in theories of organizational learning. There is a fair amount of evidence on where mindful processing occurs in the brains of individuals. If we could get pencil and paper measures and then validate them with a brain scan measure, that would give us greater faith in the measures and findings.

Why don’t we open up the stage for a couple of more questions from our audience?

I have been reading the literature on learning and there is also a discussion on group learning and group heuristics. Could you please help me understand a little bit how you differentiate the learning of heuristics in terms of group heuristics and transactive memory?

My take on it is that heuristics are more of a shared understanding. You develop a heuristic in response to a given situation. Transactive memory is more differentiated. There is an element of sharing, because you know who is good at what, but it is more differentiated.

I was interested in your comment about identity and the relationship between identity and learning. I am particularly interested in the type of situations in which identities lead to learning and not learning. I wonder, if you could elaborate on the range of identities that you think are more likely to lead to learning or not. Perhaps you can talk a little bit about some of the mechanisms.

I can talk about that in two specific ways. One way is coming more out of the social psychology literature. In terms of whether people share an identity or feel that they are part of the same group, that is the way that Aimee Kane, John Levine, and I used identity when we did that initial study (Kane et al., 2005). If you felt like you are part of the same group, you would be more willing to share information. That doesn’t mean that you would blindly adopt an idea from someone with whom you shared an identity, but that you would want to listen and learn from them.

Before we let Linda go, I would like to thank Linda for the time that she spent with us today. Thank you.