Attention

The 7 Laws of Multitasking

The two main hazards (and two main benefits) of multitasking.

Posted December 16, 2014

The benefits and hazards of multitasking.

You’ve been told that multitasking is a bad idea. Articles and essays appear every day telling you why multitasking is dangerous, makes things take longer, tires out your brain, reduces the quality of your work, and lowers your IQ. And, for the most part, these warnings are all based on good research. ^{1, 2}

But have you ever noticed that you still multitask in many ways, and it usually works out just fine?

The truth is that sometimes multitasking is a very bad idea, sometimes it doesn’t help, but also doesn’t hurt much, and sometimes it brings great benefits. The point of this essay is to help us tell the difference.

And, with that, I present the seven laws of multitasking.

Law One: Some tasks mix well.

In the simplest case, when you multitask, you have a primary task (task A), and a secondary task (task B). The primary task is your priority. The secondary task is tacked on as something else you can enjoy or accomplish at the same time.

Sometimes the two tasks will mix well. Here are some examples:

listening to music while practicing a sports skill
talking with strangers while waiting in line
listening to a talk radio program while driving (sometimes)
listening to music while working.
talking on the phone while cooking
cleaning a room while doing a weight workout
doing dishes while waiting for a video game to load
watching television while folding laundry
listening to an audiobook while exercising
drinking coffee while talking with a friend
eating popcorn while watching a movie

And, contrary to idiom, even chewing gum while walking works out just fine most of the time.

On the other hand, . . .

Law Two: Some tasks mix poorly.

Sometimes tasks don’t mix well. Here are some examples:

texting while driving
monitoring social media while doing high-focus creative work
watching television while doing homework
listening to a podcast while writing
listening to music with lyrics while reading
talking with friends while watching a movie
emailing one person while talking with another
playing angry birds while landing a commercial passenger jet.

These examples, good and bad, come from my own experience, from common experience, and from the research literature. One or two of the examples might fail to resonate for you, but the point remains: multitasking is a mixed bag.

Now let’s see if we can figure out what makes multitasking good in some cases and bad in others.

Law Three: It’s partly about switching costs.

Imagine you have a small kitchen, and you plan to make two omelets and two batches of cookies. In what order should you do your cooking?

When you make an omelet, you must get out the eggs, the cheese, a cutting board, a knife, some spices, some vegetables, a skillet, and a mixing bowl. Then you make the omelet. And, if you’re like me, when you’re finished, you still have a dirty mixing bowl, cutting board and skillet sitting there, with the eggs, cheese, and maybe some extra veggies still sitting on the counter. At some point you will need to clean up and put things away.

So we can break the task of making a single omelet into three parts: 1) setup, 2) make the omelet, 3) cleanup.

And the same goes for baking cookies. You’ll have the same three components to that task: setup, make the cookies, and cleanup.

Now consider two plans for making the two omelets and two batches of cookies:

Plan 1: omelet, cookies, omelet, cookies

Plan 2: omelet, omelet, cookies, cookies.

And consider how much time and effort is required for each plan.

With plan 1 you will need to:

setup for omelet
make omelet
cleanup
setup for cookies
make cookies
cleanup
setup for omelet
make omelet
cleanup
setup for cookies
make cookies
cleanup

Attention Essential Reads

The Secret Power of Mind-Wandering

What Animal Tracking Can Teach You About Self-Awareness

With plan 2 you will need to:

setup for omelet
make omelet
make omelet
cleanup
setup for cookies
make cookies
make cookies
cleanup

When you alternate tasks, you have to clean up task A before you can setup for task B (remember, this is a small kitchen). And it turns out that you can save a lot of work by focusing on one kind of food at a time instead of switching between them. You save two setups and two cleanups by following plan 2 instead of plan 1. This is the benefit of batch processing. It saves on overhead.

The same thing happens in your brain when you multitask. For example, when you sit down to do some homework your mind has to do some setup tasks. It must load certain information into short term memory, build appropriate mental models, erect filters to keep irrelevant information out, and so forth.

Now, if you switch your attention to watching television for a few seconds, you must free up some of the resources being used for homework, and prepare some new mental resources that are needed to track the show you’re watching.

So your mind has to perform some cleanup and setup work every time you switch from homework to television, and every time you switch back as well. These are known as “switching costs”, and they cost you in terms of both time and energy.

The costs might be small for an individual switch, but after an hour of homework/television they add up, and you might find you got only half of the homework done you could have, and you might be more mentally weary as well. ²

And the switching costs will be even greater if your primary task is a complicated creative endeavor, such as constructing a mathematical proof, developing a theory in physics, writing about a complicated topic, or coding a complex algorithm.

Richard Feynman had the following to say about allowing creative work to be interrupted (whether by accident or through multitasking):

“To do high, real good physics work you do need absolutely solid lengths of time, so that when you’re putting ideas together which are vague and hard to remember, it’s very much like building a house of cards and each of the cards is shaky, and if you forget one of them the whole thing collapses again. You don’t know how you got there and you have to build them up again, . . .” ³

The problem is that, when you’re doing complicated work, you often have to build up intricate mental models, and you’re pushing yourself to the edge of your capacity to concentrate. And when you take your attention away from your creative task and attend to an interruption, the mental models dissolve. And you probably won’t be able to build them back up the way they were.

It’s like having your computer crash while writing a paper, and realizing you hadn’t saved your document for half an hour. It will cost you time and energy trying to get everything built back up, and you might not actually get it back the way it was.

The real tragedy here is not that sometimes geniuses lose track of where they were. Geniuses typically value deep concentration and take measures to protect against interruption. The real tragedy is that many chronic multitaskers never bother with deep concentration, and might never discover the genius within them.

Law Four: It’s partly about resource conflict.

Another problem with mutitasking is that task A and task B might need the same mental resource, and they can’t both use it at the same time.

If a person listens to light instrumental music while sending an email message, there is typically little problem. Task B (listening to music) makes use of mental resources not needed for task A (writing the email message). If our email writer sub-vocalizes as she writes, there might be some auditory involvement, but the music won’t require verbal processing, so the degree of conflict is minimal.

If, on the other hand, she talks with a colleague while writing her email, then there is much more conflict over mental resources. Task B requires the person to construct and communicate meaning in sentences, just like task A does. Both require empathy and social strategizing as well. Trying to do both tasks at the same time will cause high switching costs and a greater chance of error.

Texting while driving is an iconic case of resource conflict. Both tasks compete over visual attention. When you switch your gaze from driving to texting, you will no longer be able to see new driving hazards as long as you are looking at your phone. Plus it takes some time to get situation awareness when you look back to the road. That’s why texting and driving is now the number one cause of death for teen drivers. ⁴

In general, when it’s important to do task A well, we should not also take on a secondary task that competes with the primary task for key resources.

Law Five: It’s partly about sweetening the pot.

But there are benefits to multitasking. Sometimes we are staring down a relatively simple task that we just don’t want to do (such as folding clothes). But we know that we would be much happier doing that task if we could do something else pleasant or useful at the same time (such as watching television or listening to an audio book). So we multitask in order to “sweeten the pot”, so we will have the motivation needed to perform the primary task.

We might not fold the clothes in record time. There will be some switching costs. But the alternative, if we’re being frank, is that we won’t fold the clothes at all. And, since the task is relatively simple, the switching costs will be manageable.

Or sometimes the primary task will contain periods of activity interspersed with periods of inactivity, while another task can be broken into small chunks that can fit those gaps. When I workout with weights, I perform sets of exercise with periods of rest in between. When I clean my office I do a series of discrete tasks with natural break points between subtasks (clear the clutter from my desk, empty a trash can, etc.).

That makes these two activities a natural fit. If I arrange to do cleaning tasks during the rest intervals in my workout, that “sweetens the pot” for both tasks. I normally don’t like cleaning my office, and will put it off repeatedly. But, if I can make use of the dead times in my workout, it seems worthwhile, because there’s little else of value I would be able to do during those two-minute rest periods.

Multitasking can help us start a task we don’t want to do, and it can also keep us doing a task when we’ve grown impatient. When the car trip gets boring, we can play twenty-questions. When we get impatient waiting in line, we can strike up a conversation with a stranger.

And pot-sweetening is just one of the two main possible benefits of multitasking.

Law Six: It’s partly about setting picks.

In basketball, it’s easier to score when you’re not being harassed by an opponent. That’s why teammates will sometimes position themselves at a spot on the floor and just stand still. The player with the ball can then dribble close enough to the teammate that the shadowing defender must either run into the teammate, go around the teammate, or switch assignments with the teammate’s defender. Sometimes this allows the player with the ball to get off a clean shot. The teammate in this case is “setting a pick”.

Likewise, a well-chosen secondary task can “set a pick” for the primary task by blocking out potential distractors.

When we work on a task, our minds do many things. Parts of our mind are concerned with executing our primary task. They help us keep the goal in mind, make plans, execute those plans, work around obstacles, keep the right things in memory for easy access, and so on. These are “foreground” processes.

At the same time other parts of our mind are looking out for signs of danger, looping through other problems we are dealing with, monitoring our internal states, or looking for opportunities to switch to more rewarding tasks. These are “background processes”. And background processes have a way of getting us off track at times.

So here’s the thing. This is where we can use multitasking to our advantage. If we choose our secondary task wisely, it can compete for resources with background processes that might otherwise interrupt us. And that means the right task B can actually help us stay focused on task A. Here’s how we might formalize that strategy:

Background Process Interference Strategy: when background processes are likely to interrupt a primary task, try to find a secondary task that will compete for resources with the background processes, but not with the foreground processes.

In other words, use task B to “set a pick” for task A.

If you don’t like doing yard work, and you know that parts of your brain will be looking for more rewarding things to do, and will be sending a constant stream of rationalizations to your mind to try to get you to quit, then you can run interference by listening to a podcast. Listening to the podcast will compete with the background processes for a key resource (strategic thinking), but will not compete substantially for the resources being used by the primary task.

On the other hand, if you’re writing an essay, and you fear your background processes will be trying to get you to quit, setting a pick with a podcast won’t work as well. In that case, task B will interfere not only with the background processes but also with the foreground processes -- like a clumsy teammate who tries to set a pick and knocks over the ball-handler in the process.

Law Seven: It’s all about making trade-offs.

It should be clear by now that we can’t say full stop whether multitasking is good or bad. It all depends on features of task A, features of task B, how A and B interact, and what a person’s goals are.

When it’s important to do the primary task well (driving), we must be extra careful about switching costs and resource conflict (that’s why texting while driving is a terrible idea). When it’s not that important, we can be more relaxed about those costs, and be more open to some of the benefits of multitasking (watching television while folding clothes is probably fine).

Sometimes it will be important to do task A quickly (studying for a test the night before an exam) and sometimes it won’t matter too much how long it takes (folding laundry on an otherwise empty evening). When it’s important to do a task quickly, we must be extra concerned about switching costs (and we might opt for some mid-tempo instrumental music to help us focus and block out distractions while we study -- instead of watching a television program).

Sometimes we are motivated to do task A (playing a new video game), and sometimes we lack motivation (working out). When we lack motivation, a well-chosen task B might just sweeten the pot.

Sometimes we are so familiar with task A, we do much of it on “auto-pilot”. And sometimes task A takes our full attention. That’s why listening to a talk program on radio can be a good idea for an experienced driver, but a bad idea for a student just learning to drive.

Sometimes task A is complicated (writing an essay), and other times it’s simple (folding clothes). Switching costs are usually higher for complicated tasks.

And so, in order to tell whether a given case of multitasking is good or bad, we will have to weigh the costs against the benefits on a case by case basis.

Conclusion: some specific and useful strategies.

We’ve covered a lot of ground. Here are the key lessons in a nutshell:

We’ve seen that multitasking can be a bad idea when:

there are high switching costs
there is resource conflict between task A and task B

And we’ve seen that it can be a good idea when:

task B “sweetens the pot” for task A
task B can “set a pick” for task A

And we’ve seen that the wisdom of multitasking can also depend on other features of task A and task B:

how important it is to do them well
how familiar they are
how important it is to do them quickly
how motivated we are to do them
how complicated they are

Let’s finish with a few specific and useful ways to apply these lessons:

Multitasking can be dangerous. When it’s important to do task A well, we should be very careful about choosing a task B, and err on the side of caution. We should never text and drive, or talk on the phone while driving. And we should take steps to make sure our passengers will not distract us with rowdiness or emotionally challenging conversations (this is of special relevance for those of us with children).
Multitasking can hinder creative productivity. When working on a creative project that pushes us to the limits of our concentration, we should not multitask in ways that will expose us to interruptions of the primary activity (for example, while writing an essay, we should close our facebook, twitter, and email clients, and check them only after we’ve done a good chunk of creative work).
Multitasking can assist creative productivity. Well-chosen secondary tasks can set picks for our creative projects, and can help us maintain focus against background processes that might otherwise interrupt us (for instance while writing an essay in a coffee shop, we might listen to invigorating instrumental music to block out ambient noises, conversations, and internal signals of discomfort).
Multitasking can help us be more patient. We tend to grow impatient when we have a goal and we have just learned that it’s going to cost us more to reach our goal than we originally thought. And our tendency when we are impatient is to either try to find shortcuts or to abandon our goal for another goal. But sometimes the right course of action is to simply stay the course and absorb the extra costs. The right task B can both sweeten the pot, and set picks on those voices in our heads trying to get us to change course. For instance, we might talk with a stranger while waiting in line, so we don’t bolt, or listen to an audiobook while stuck in heavy traffic, so we don’t plot out risky and minimally productive lane-changing maneuvers (For more on impatience, see “The 7 Laws of Impatience”).
Well-chosen music mixes with almost everything. Well-chosen music has a magical ability to both sweeten the pot and set picks for almost any activity -- while avoiding resource conflict and switching costs almost entirely. The music must be chosen carefully, so it does not provide resource conflict (for instance songs with lyrics might not be optimal for reading, and death metal might be a poor choice for meditation). But there is usually a good choice for almost any activity. Some tasks might not mix well with any kind of music, but these will probably be rare for most people.