What prevents scientists from being more productive and if we knew, could we do anything about it?
I’d like to look at an often overlooked, but huge productivity inhibitor — bad multitasking. Many people put “excellent multitasker” on their resume as a badge of honor. We laud the efficiency of a good multitasker — they are rarely idle — someone that busy must be getting a lot of work done, right?
What Multitasking Looks Like, Visually
Let us visually consider the impact of multitasking on task completion, and see why I’m tempted to put “excellent multitasker” resumes into the round file.
In the figure below, we have 3 tasks which each take the same amount of time to complete — say 15 days. If we perform each task sequentially without interruption, then tasks A, B, and C will be completed at day 15, 30, and 45 respectively.
Consider then, a more typical multitasking workflow where we alternate between tasks A, B, and C doing, say 5 days of work on each before switching to the next one, alternating until all tasks are finished. As shown below, task A is completed by day 35, task B by day 40, and task C by day 45.
For most projects it is fair to say that the benefits are not realized until the project is complete. That is, the money, time, and resources invested in project execution is not bringing a return until project completion. In the first scenario, the benefits of project A could be realized after 15 days, and project B after 30 days. In the multitasking scenario, the benefits of project A and B were not realized until day 35 and beyond — more than double the time of serially tackling the tasks one by one.
It gets worse. Context switching between tasks causes loss of flow — that state of uninterrupted concentration where we work at our best. And let’s also acknowledge that if 10 days elapses before returning our attention to a task, that we are going to have to spend additional time recalling context and getting back up to speed from where we left off.
Let’s conservatively add a day of switching costs between each task. With this assumption, task A would be done by day 41, task B by day 47, and task C by day 53. We have almost tripled the time to see the results from task A!
Instead of 3 tasks, consider the dozens of things most of us have going concurrently in our lives — it is no wonder we never finish many of the things we start.
Why Researchers Are the Worst of Multitaskers
I contend that generally, researchers are the worst of multitaskers — it would be unusual to find one with less than 20 different irons in the fire. Like a butterfly or honeybee hopping from flower to flower, a researcher delights in nothing more than starting something new, getting distracted by something else novel, and hopping from one thing to the next. The productivity implications are devastating. Why then do we do this?
A very important need is satisfied by being a honeybee: cross-pollination of ideas. Researchers are driven by curiosity, by learning. Seeking diversity and synergizing ideas across multiple domains is a well-used learning strategy.
But this strategy puts a researcher into the following dilemma:
In order to be a successful researcher, one must both learn and get a lot done, which for those in our field usually means completing research projects and publishing results. To learn and satisfy the natural drive for curiosity (B), a researcher naturally jumps from area to area in their intellectual exploration (multitasking) (D). On the other hand, if a researcher wants to publish and get stuff done (C), they should seek to focus on few projects and see them through to completion (minimize multitasking) (D).
So how does any doing get done? Professors have a great labor saving device, employed at least since the time of Isaac Newton: the graduate student! Yet other than for the über-funded, ideas for new projects can be generated faster than any army of graduate students could hope to keep pace with. Can you say “10-year Ph.D.”? The multitasking gets transferred to them!
Multitasking and the Reiss Profile
Back to my opening question — if we knew better would we do better? I have known about the killer implications of bad multitasking for over 7 years, yet I have only begun to mend my ways. The curiosity urge is too strong.
I think the core reason we multitask boils down to one simple cause: at this moment I would rather do something else. What then determines what we would rather do?
A psychology professor, Steven Reiss, did a factor analysis on hundreds of things people said they value and came up with 16 reasonably orthogonal values that characterize goal-seeking behavior in humans (interestingly their basis can be traced to evolutionary survival needs). He developed a test, the Reiss Profile, to determine the extent the 16 values are higher or lower than average for a given person. The values are: acceptance, romance/beauty, curiosity, eating, honor, family, idealism, independence, order, physical exercise, power, saving, social contact, status, tranquility, and vengeance. Each of us have lower or higher levels of drive towards performing actions that fulfill the above needs. Further, our most dominant behaviors are driven by attraction to doing more of what we value highly and aversion to doing more of what we value minimally. I hypothesize that high curiosity is overrepresented among researchers, and high order is overrepresented among administrators. This might explain why there is often a fight between these two camps over filling out reports!
We have many values that drive our behavior. To the extent we can do work that fulfills most or all of our needs simultaneously, we don’t need to multitask. When this is not the case, we multitask to fill up our value “tanks” as time elapses when we haven’t had enough of something else we value, or to stop filling a “tank” of a value that is satiated. The real challenge is to construct a career that maximizes action along the lines of what we value, and minimizes or delegates actions we don’t value to those who do value those activities. There are also ways of reframing so that unvalued activities are seen in light of how they contribute at a higher level context to something we value — reducing the misery and stress of that stuff we just have to do but don’t really want to.
Learning vs. Doing
So as researchers, would we rather be learning than doing? But wait a minute — is that really possible?
The generation of knowledge occurs in a continuous cycle of forming theories or models that describe the way the world works, and then testing those models in the arena of action, learning better ones through trial and error. That is the essence of the scientific method discussed at length in a recent webinar of mine. It is easy to build “castles in the clouds” — ideas disconnected from reality and action if we only think and rarely do.
Taking the learn/do conflict to an international scale, consider the positioning of America as a “knowledge economy”. While we may differ on what we mean by this term, lurking in there is an idea something like “America will do the thinking and other economies will do the doing.” That is, we will specialize in the “superior” high-value knowledge work, and offshore workers will roll up their sleeves and do the “dirty” work of manual labor.
There is a key fallacy in positioning a nation exclusively as a knowledge economy. If you disconnect the learning from the doing, you slow down both of them. Learning occurs fastest with the instant feedback of doing. Therefore, the economies that roll up their sleeves, make things, and bounce models off of reality will learn an order of magnitude faster than those who do so through others across an ocean and 12 time zones. The doers quickly catch up and become higher value “situated knowledge” economies.
Consider that in science, if researchers stop short of trying to falsify their hypotheses in the laboratory of reality, they don’t really learn. This was discussed at length in my post: “Dammit Jim, I’m a doctor, not a bioinformatician!” When you abstract a real-world problem into a model, you reduce the degrees of freedom and then it is more manageable. According to Ashby’s law of requisite variety, if you want to control a system, the controller has to have more degrees of freedom than the controlee. Our models have low degrees of freedom, and messy reality has high degrees of freedom. So it actually takes more degrees of freedom, more mental horsepower to translate theory into the world of action. This is the learning loop of modeling reality and testing against reality.
So to sum up: there is no such thing as learning without doing — it is a false dichotomy!
Obeying the maxim “Physician heal thyself,” I’m working on the multitasking thing — along with 20 other projects. When I make some progress, I’ll let you know. Fundamentally I think the challenge is making the paradigm shift from: “productivity=activity” to “productivity=task completion,” and figuring out how to do what we love. And I believe that there is a way to break the assumption that the only way to learn is through unfettered diversity exploration. There is a 60+ year old body of knowledge, TRIZ, that breaks this assumption, but requires us to bypass our normal (cherished) modes of thinking. That too represents a paradigm shift. So at the end of the day, the real challenge is how to make paradigm shifts and not have to wait a generation for them to happen.
Nevertheless, there are simple things we can do short of making a paradigm shift that can give us a significant productivity boost. Besides our butterfly mind, some of the biggest additional contributors to bad multitasking are external factors that interrupt our concentrated work. It is pretty bad when we have institutionalized a culture of automated interruption though email pop-ups with sound reminders on our computers and smartphones. I shut these off years ago, and it has made a huge difference in creating conditions for uninterrupted flow.
Too much praise has been sung for the “open door policy” and “management by walking around” (interrupting) and not enough attention to setting blocks of time aside that co-workers know are uninterruptible short of a real emergency — and this needs to include both physical and virtual (e.g. text message) interruptions. If the productivity boost is not motivation enough, I’ve discovered after weaning myself from the addiction to continuous interruption, that the real payoff for uninterrupted flow is more happiness.
And who doesn’t want more of that?