Eyal Ophir on the Science of Multitasking

Photo: Eyal Ophir with his daughter Sahar, courtesy of the subject.

Eyal Ophir was primary researcher on the pioneering Stanford Multitasking study. He now designs information interfaces for the browser RockMelt.

Avi Solomon

How did you get to studying multitasking at Stanford?

Eyal Ophir

While I was at Stanford, Cliff Nass (my advisor, and a global expert on human-computer interaction) introduced me to some great ethnographic work done by Ulla Foehr and Donald Roberts at the Dept. of Communication looking at media consumption among youth. They saw that young people were reporting more media-use hours than actual hours, and figured out these same young people must be consuming multiple streams of media simultaneously in order to fit it all in. This is where I was introduced to the concept of Media Multitasking. I came from a cognitive psychology background, and I was inspired by Anthony Wagner's work on memory and cognitive control (Anthony was my reference for all things cognitive, and ended up being the third author on the paper). So for me, the interesting question was simply how these kids are managing to process and control so much information all at once.

Because looking at historical data on processing multiple information streams, you wouldn't think it was possible. Research in the 1950s showed that when people listen to two different streams of audio, they focus on one, and can only do the most basic processing of the other. So if you're listening to a voice in one ear, you can tell if the speech to the other ear has changed gender, or if it calls your name. But you can't tell if it's changed language, or suddenly starts playing in reverse. More recent research suggests that humans have a cognitive bottleneck, which forces us to only really be able to engage in one decision-making process at any one instant. So you'd think, given this data, that we're serial processors. Intuitively, that certainly rang true for me. So - how are these kids watching TV, while talking on the phone, and chatting on their computer, and browsing the web, and listening to music, and doing their homework, all at once? How have their brains adapted to allow them to process so much simultaneous, disparate information?


How do you define multitasking? Is true multitasking even really possible?


I specifically look at media multitasking. To me, that means engaging with multiple information streams simultaneously. What's interesting to me about this form of multitasking is that the media are so available, and so engaging, that they make media multitasking so easy. The media are actually so dedicated to, and so effective at, attracting our attention, and they are so omnipresent, that one actually has to work at not media multitasking.

Is it possible? The passive nature of many types of media makes it very easy to surround yourself with multiple information streams, and so in that sense, it's possible and even easy. As to how much you're actually processing these multiple media and effectively engaging with them -- that is precisely what motivated me to do this research. That's the question I tried to begin to answer.


What is the real attentional cost of task-switching and interruptions?


Every task you take on requires a certain set of mental rules, or context. Psychologists call this the "task set" Recent research suggests that the cost of task switching is rooted in cognitive interference from the irrelevant task set: interference from all those thoughts about the task you're NOT doing. Every task you do competes for your mental resources, even once you think it's no longer relevant. The more you do, the more you increase this competition. So that momentary interruption is still fighting for some of your mental resources even when you'd like to focus back on your main task. The more competing tasks you take on, the more interference you must overcome to fully dedicate yourself to what's really important. What may be worse is that over time you may be training yourself NOT to focus. You teach yourself that something more exciting might be just around the corner - behind that notification, or the app on your mobile phone, or the email you haven't checked. If you prioritize the unknown, but potentially exciting, over what's in front of you, you'll have a hard time controlling your own focus. You may transition from a top-down model of attention allocation, where you decide what to focus on, to a bottom-up model, where any new notification or alert will dictate what you focus on. We've done some early exploratory work on this, and it does seem that heavy media multitaskers place high value on new information, and may be more impulsive and responsive to rewards.


Is multitasking addictive?


This is certainly a hot topic, and one of the big questions that still has to be answered. I am no expert on addiction, but I think media in general, and media multitasking in particular, especially the tendency to constantly look for the next unknown, exciting tidbit, may be addictive. The relationship between media, media multitasking, and the brain's most basic pleasure and reward mechanisms is an area of very active research that may provide an answer.

Test Your FocusTest How Fast You Juggle Tasks


What multitasking experiments did you design and test people on?


I wanted to find out how heavy media multitaskers were processing so much information all at once. I wanted to find the secret to why they were regularly doing something that traditional research seemed to suggest was impossible. So we surveyed hundreds of students about their media use habits, and invited the heaviest and lightest media multitaskers to come to the lab. Then I ran them through what was basically a cognitive obstacle course- testing all the basic cognitive control abilities that we thought could in some way be relevant to media multitasking. We looked for those abilities that were significantly different between the heavy and light media multitaskers.

We found several significant differences. The first was that heavy media multitaskers were not as effective at ignoring irrelevant information. I asked them to perform a visual task where some of the objects they saw were clearly marked as irrelevant, and told them to ignore these objects. And yet the more of these objects were present, the worse they performed on the task. Light media multitaskers, on the other hand, performed equally well regardless of the number of irrelevant objects - they just blocked them out. I tested this result again in another, different experiment, and found the same result - heavy media multitaskers generally performed worse whenever there was irrelevant information around. They had trouble ignoring it.

I then asked them to perform a task where they had to keep information around in their head, and use it to perform a task. But at each given moment, they only had to think about a few bits of information. Heavy media multitaskers had trouble letting go of older information - it kept interfering with their task. So once again we found trouble with filtering. But this time, it was not about filtering irrelevant stimuli from the environment. Here, they had trouble filtering irrelevant information in their own mind - irrelevant memories. They had greater trouble keeping all the different pieces of information sorted in their mind, and knowing what was relevant and what was irrelevant.

Lastly, I tested their ability to switch between tasks. Here, the heavy media multitaskers were slower switching from one task to the next. This again points to trouble filtering - if you consider evidence that difficulty switching tasks is driven by interference from the irrelevant task, this result meant that heavy media multitaskers were having difficulty not thinking about the task they weren't doing. They had difficulty filtering out thoughts about the irrelevant task.


What surprised and alarmed you about your findings?


I actually expected the media multitaskers to filter less. How can you multitask if you block everything out? But what I did not expect was the task switching result.

Humans don't really multitask - we task switch. Our brains are serial machines, so we just switch very quickly between tasks, and it feels like we're multitasking. So when we found that heavy media multitaskers performed worse on a classic test of task-switching, it was like finding out that these heavy media multitaskers were worse at multitasking.

I said those words at one of the first presentations of the findings. It became a very attractive soundbite - that "multitaskers are worse at multitasking" (even Conan O'Brien repeated it) - but that statement on its own is dangerously non-specific.


How do you explain this paradox, that heavy multitaskers were actually less effective at multitasking?


I have to preface my answer by saying that I think that's a problematic statement. What does it mean to be "effective at multitasking"? If multitasking is being able to get through writing a paper while handling interruptions, is effectiveness the ability to sustain your focus on the paper? If multitasking is driving while talking on a mobile phone (NOT a recommended pairing), does being effective mean not focusing so heavily on the conversation that you still notice when the driver in front of you slams on the brakes?

I think heavy multitaskers are not less effective - they simply have a different goal. A different set of priorities. Where you might say traditionally we value the ability to focus through distractions, they are willing to sacrifice focus in order to make sure they don't miss an unexpected, but rewarding, surprise. As a result, they might do worse in the office scenario I described, but they might also be the first to slam on the brakes in the car/mobile phone scenario. So who is more effective?

As a researcher, I don't want to be in the business of passing judgment. I want to find counter-intuitive new facts and use them to make valuable predictions. I think heavy media multitaskers have a tendency for bottom-up attentional control: compared to light media multitaskers, they tend to be more reactive than pro-active. They seem to be less in control of their own attention, and they do this so they can be quicker to discover the unexpected.


Are there gender or generational differences?


That certainly is a question that comes up a lot. It's not something we tried to address with our work. But shortly after I finished this research I spoke to a very prominent psychologist about the intersection of gender and multitasking. According to him, after the issue came up between him and his wife, he went and tried to find evidence of gender differences in multitasking ability. And he found none.

There is more evidence of age-related differences, especially in cognitive control ability. But the more interesting question of generational differences will probably take more time to answer. In general, our paper did not address causality. We can't say if media multitasking causes these cognitive tendencies, or if people with these cognitive tendencies gravitate to media multitasking. But there has been some fantastic research, such as the work on gaming from Daphne Bavelier's group, showing how media activity can affect change in the brain. And media multitasking seems to me to be such an extreme environment, and we subject ourselves to it so frequently and regularly, that it would surprise me if it didn't have some very dramatic effects on our minds.

Given that, I think the difference in exposure to media multitasking between generations is probably going to drive some interesting generational differences. I look forward to finding out.


Why is listening to instrumental music while performing a complex cognitive task different?


We use different cognitive resources for processing different stimuli. For example, studies have demonstrated that it's much easier to remember words while performing a spatial task than while performing a linguistic task. I'd expect that listening to music while trying to perform a musical task would be quite difficult. But most of what we do when we measure our multitasking ability probably doesn't overlap heavily with processing music. You'd probably also see differences depending on the type of music. Instrumental music may be less disruptive than music with lyrics. Familiar music, where you're less actively processing the lyrics, or which is less likely to surprise you, might also turn out to be less disruptive.

From Ophir E, Nass CI, Wagner AD (2009) Cognitive control in media multitaskers.


What has been the impact of your landmark multitasking paper?


I think we finally brought some data to bear on an argument that has been raging in bars, offices, and bedrooms for years. It seems both multitaskers and non-multitaskers are quite committed each to their perspective, and to the inefficiency of the other. So when the paper came out, it caught on not only in academic circles, but in popular circles - we were one of the top 10 most shared and most popular stories on Yahoo News and CNN the day we published. So the first thing the paper accomplished was to take the work of Roberts and Foehr, and the whole topic of Media Multitasking, and bring it one step closer to the public sphere. To elevate the public discussion with a bit of empirical data.

The second effect was to feed into the backlash against digital media. Statements like "Multitaskers are bad at multitasking" seemed to pit science against today's connected culture. I think that was both good and bad. I think media multitasking is an extreme request to make of our brains, and I strongly expect it has consequences. Discussing this openly, and allowing it to inform our media consumption, is a good thing. That said, I think judgment precludes possible evolution. Media multitaskers may be making what is a very sensible adaptation to an extreme environment. Do we try to reverse the trends that generated this environment? Do we adapt as they did? Or is there just too much we still need to learn before deciding on a course of action? I think we don't need to turn reactionary, but rather use this new knowledge to shape our informational environment, and maintain our individual control within it.


What do we still not understand about multitasking?


Most of what you've asked me. This is a young phenomenon, and an even younger field of study. Causality, addiction, generational and gender differences, underlying neural systems, and of course, the best way to navigate this information environment, are all open questions. Many of them will unfortunately take more time to answer than we'd like.

Around the time our paper came out, we hosted a seminar on media multitasking, with representatives from multiple academic disciplines, as well as from education and the media and tech industries. I was most surprised by the sense of urgency in the room. There were so many pressing questions that we were just beginning to answer. Participants began playing less the role of professionals and researchers and more the role of concerned parents, wondering about how best to help their children navigate this new reality. I was inspired by the level of the discussion, but there is a ton we don't know. Though I think that's the mark of an exciting and relevant field of research.


What life-strategies can we adopt to counter distraction? Is there a fundamental prerequisite understanding that needs to sink in before adopting these strategies?


I think we need to realize that attention is a zero-sum game. If you truly focus on two things, they can each only get part of your attention.

We also need to realize that the attractiveness of the little distractions provided by media, and our motivation to attend to them, is probably out of proportion to their actual importance to us. This of course may vary greatly based on context -sometimes, when we make bad decisions about where to focus our attention, distractions are vital (such as when texting while crossing the street, etc.). But I think there's another reason we might feel so inclined to respond to every notification.

Historically, when someone tapped on our shoulder, they were necessarily physically next to us. So they knew if we were already holding a conversation with someone else, and could adjust their behavior, or withhold their request. We, in turn, felt compelled to respond to the tap on the shoulder when it came. But now, the incoming chat message, the phone call, and the television announcer, all tap on our shoulder in a sense, trying to get our attention. They are entirely oblivious to each other, and solicit our attention as if they were the only ones. We, on the other hand, feel the same obligation to respond. It may be that our social norms and instincts are not scaling at the rate of communication channels. In this way, media may have brought about a new tragedy of the commons - by aggressively trying to grab our undivided attention, they have threatened the very notion of undivided attention.

I think the key might be control. We should start making decisions about how we allocate attention. If focus is important to us, then we should protect it. This may mean creating periods where we silence the notification channels, and make it easier for ourselves to keep our focus where we want it. Work for 30 minutes, then check your email, or your favorite websites, and get back to work (or to the dinner conversation). Get in the habit of focusing without the underlying expectation of something more exciting coming along.

I actually think there's potential here for media to play a larger role. One layer above the content, we have meta-interfaces that are situated at the junction of multiple information streams, that can help us navigate our information environment. Examples are the car, which can start thinking about when the mobile phone should ring, and when it shouldn't; your computer operating system; or your web browser. In my role as a designer for the browser RockMelt, this is a very exciting opportunity. I'm working on a browser that standardizes all my information channels - and thereby makes staying on top of that information more manageable. For example, I recently designed a 'Quiet Mode' button, to help users regain their control. So now RockMelt puts all your Facebook, Twitter, RSS, and web mail notifications in a single place, and when you're ready to focus, you just click a single button and all that noise goes away. When you're ready to see what's new, you can turn it back on. It's my attempt at putting control back in users' hands.


How has your multitasking research changed your personal behavior in relation to managing your attention?


The funny thing is that I'm a horrible multitasker. When I think about something, I automatically block out everything else. My wife knows that if she doesn't have my full attention, I have no idea what she's saying - I probably don't even hear her.

Apart from becoming a real influence and a source of inspiration on the way I design information interfaces, these findings also impacted me personally. My daughter will soon be 3. When I'm with her, I block out the media. It's something that demands conscious effort, because that drive to quickly check the latest update has been internalized. So she hardly ever sees me with a laptop, or a mobile phone. I've also had friends that, since my research was published, have tried to take back control of their attention. So far, they've been very pleased with the results. As a creative professional, my ability to focus and create space for complex thought is important to me. As a father and husband, my ability to give my family the attention they deserve is vital. So I've become more aware of the way in which media might compete for control of my cognition, and I've made the decision that I'd rather be proactive than reactive. I want to decide what I think about. But working in media, there's a lot of exciting stuff happening out there -- more than I could ever process. So it's a work in progress.



  1. Fascinating interview!

    I wonder how these findings compare to what’s known about ADHD. People with ADHD generally have a lot of trouble with task-switching, and can have a lot of trouble filtering out irrelevant input — but are also capable of hyperfocusing on some sufficiently stimulating input/task and ignoring ALL other input, even if it is relevant. Can the heavy media multitaskers have that kind of single-task hyperfocus?

    I also think that media multitasking, looking for the next new piece of interesting information, can definitely be addictive. There is clearly an emotional reward in finding something novel, interesting, and exciting. The internet provides that, but at unpredictable intervals — so you want to keep trying.

  2. The results he describes give me the ugly feeling that “multitasking”, at least in the media context, is less about “how these kids are managing to process and control so much information all at once” and more about which kids are incapable of processing and end up being controlled by so much information available all at once…

  3. always a fascinating topic. several months ago I tried to scrape the surface of how it’s conceivable that heavy “connectedness” to media, could actually yield *more* meaningful relationships (counter intuitive). http://one.valeski.org/2010/08/ambient-intimacy-humans-machines.html I need to update my thoughts in a fresh blog post, but I do believe there’s something there.

    I like Eyal’s highlight about the passion multitaskers vs. non-multitaskers have about their points of view.

    obviously multitasking (or ability to switch tasks efficiently; that’s a nit in my mind; tomato tomato) can be done incredibly well. consider a fighter pilot managing untold numbers of inputs and data and perceptions, who has to process all of that information very quickly.

    obviously multitasking can be done incredibly horribly also. from my POV, each human brain handles it differently. some thrive on switching tasks frequently, and some do not.

    for me, it all depends on the context and setting. there are times when I enjoy it (usually in the office), and there are times when it drives me up the wall (spending time with the kids who are bouncing all over the place from topic to topic; rapidly).

    there are some tasks that shouldn’t be interrupted. complex brain surgery comes to mind. software engineering. trapeze. etc.

    there are some tasks that can be interrupted regularly and frequently. light conversation with friends. mindless watching of video content.

    there’s no doubt in my mind that heavy task switching impedes deeper efforts to accomplish something though. task switching has a high cost, and nothing is free. if you task switch heavily, it comes out in the wash somewhere (stress, hindered productivity, fatigue).

    I love the comment about us seeking excitement behind the next notification. so true. this is one thing I’m constantly aware of.

  4. Riveting piece.  There has been a common claim that our attention span is diminishing these days, but perhaps it’s just that “media multitaskers” are currently in the ascendency so the perception is that they are dominant?

  5. Also — try the tests linked in the middle of the interview. Apparently I’m awesome at focusing on red rectangles and ignoring blue ones. But I’m terrible at switching between classifying letters as vowels/consonants and numbers as odd/even. Much worse than the heavy media multitaskers in the study.

    I found it interesting that on the task-switching test, I got mentally tangled up because “vowel” and “odd” were assigned to the same button (and “consonant” and “even” were both assigned to the other button). I think of vowels as being like even numbers, and consonants as being like odd numbers. So not only did I have to switch my focus between numbers and letters, but I also had to break my vowel/even association to press the correct button. I wonder how much my time would change if I retook the test with “vowel” and “even” on the same button. (Others who take the test — are vowel/odd always paired, or do they switch it up?)

    1. I had the same responses as you did.

      Perfect score on the red rectangles, but kept having trouble remembering that vowels were on the left but so were odd numbers. That did not compute for me, for whatever reason.

  6. “We’ve done some early exploratory work on this, and it does seem that heavy media multitaskers place high value on new information, and may be more impulsive and responsive to rewards.”

    No shit! I’m not sure I’m seeing what’s so amazing about this research. It’s pretty obvious if you deal with anyone in their mid-twenties and below that not only are their attention spans shot, but they also lack any ability to focus critically on something, mostly because their shitty attention spans never gave them the mental space to get to the point where they could deconstruct something long enough to really analyze it fully. With art, too- how are you ever going to really appreciate what the artist / musician / writer is doing if you don’t spend quality time with the work? You get hooked on the superficial elements to it and move on to something else before you can really start to appreciate all the hidden craft.

    I know boing boing gets hard for all things technology, and you’re automatically an “out of touch old man luddite” if you don’t praise all technological advances we create, but whatever- the writing is pretty much on the wall. Despite Ophir “not wanting to judge,” his own research judges things quite clearly. If you want to wait for the research to catch up to the obvious before making a decision, go ahead- but in the meantime, kids are growing up in a brave new world of trending topics. My kids aint gonna get no information overload, that’s for darn sure.

  7. Think back to a deep historical level. Hundreds of thousands of years ago, our ancestors were on the African veldt, tracking a gazelle to kill it. At the same time they had to be aware of their surroundings (behind them, to either side, and away from the gazelle), just in case a lion was tracking them to kill them.

    Multitasking in this sense is not new at all. And I suspect that what goes on is we set some level of awareness to a ‘monitor this, let me know if anything changes in the pattern that might be cause for concern.’ Studying chemistry while watching TV is like this. We put primary concentration on the chem text, and just monitor TV in the background. TV shows occur in recognizable patterns, most of them, with occasional bursts of ‘something interesting and worth seeing.’ When such a burst occurs, our monitor alerts us, we refocus on the TV, and when that burst of interest passes, return to the chem text.

    We all have different levels of senses (hearing, smell, sight, tactile) and different areas of sensibility (near and far, ahead and behind, so forth). Our ancestors had to manage all these in order both to focus on the task at hand, and to alert us in case of dangerous or important occurrences. 

    The difference comes as noted in the interview, when the media involved are artificial and occur in the same channel, such as two auditory streams. But I imagine our old mechanisms can handle the difference between someone talking off to the right of us, and the background music of the trio on stage.

  8. Huh, I got the same exact identification of vowels with even numbers as earlier commenters and my times were really bad for the switching task as well.  My performance on the first task was quite satisfactory though.  I tend to hyperfocus and ignore distractions (or if I can’t ignore them, get very frustrated and anxious) much as Eyal Ophir describes himself in his last response in the interview.

    Repeating the switching task and reminding myself to speed through, I got much better times, though still much worse than both light and heavy multitaskers. OTOH my “additional switching time” beat the crap out of both. I’m not completely sure how to interpret this, except that:
    a) it seems very sensitive to the even/vowel odd/consonant synesthesia we’ve been talking about in this thread — I constantly had to remind myself to invert which buttons to use whenever the switch happened
    b) it also seems very sensitive to how much you prioritize being right over being fast — I’m just generally willing to take more time to make sure I do things right and I think that may be the major factor in the test. (I read fairly slowly but very thoroughly and usually took a lot of time with difficult school assignments.)

    I find that even my very amateurish, short attempts at mindfulness meditation have been very helpful in giving me some control over my attention.  I’d recommend it to anyone who would like to do so as a free, easy, rewarding, effective method.

Comments are closed.