Blog by Mirjam van Praag
Does this sound familiar? Cycling through town, listening to a nice tune on your iPhone. Carrying a bag of documents you read the night before, while glancing at a news programme on the television, and listening to your children. Your phone rings. You replace the music with the caller, who asks your opinion about a topical theme. The caller also reminds you that someone else is still waiting for a response to a question and that you need to confirm an appointment. You cycle on, open your mail to send the confirmation and see you have 24 unread emails. You have a quick look, dodging other cyclists and trams. Thoughts pop into your head: what shall we have for dinner, when do I go shopping? Or: how will I respond to that article? It’s what I normally do, switching between tasks, or multitasking. It makes you restless, but it seems that the constant stream of fresh stimuli is addictive. Perhaps it’s productive?
What does organisational economic research tell us about the effectiveness of multitasking? Over the past few decades, a number of well-received articles have been published by researchers, including Holmstrom and Milgrom (1991). Impressively, most of these studies date from before the smartphone (and even internet and email). Unfortunately, multitasking in those studies is associated with a different issue, namely measuring the payment of variable rewards for employee performance.
Many employees perform tasks that cannot be measured or where individual contributions to value creation go unobserved. Take for instance researchers who must combine teaching and research tasks, but whose performance is measured mostly based on their research. This is easier, but means their teaching tasks go unnoticed. Or, recently in the news, researchers whose careers depend on the so-called publication credits indicator, but who can score with two different types of effort: the ‘cheap’ and unwanted manner (making up data) or the ‘expensive’ and wanted manner (collecting data). These studies about multitasking demonstrate the challenges of performance rewards in a multidimensional world. Having to give account based on measurements that do not reflect all of a task’s dimensions will often, even if unintentionally, lead to behaviour that will effectively increase measurement scores, but not the value to the organisation. ‘What gets measured gets done’ but which tasks are used to achieve this, and which ones are ignored, is of secondary interest. It explains why performance stimuli in performance contracts are often not as pronounced as principal-agent models in a one-dimensional world would predict.
In the (organisation) economy, I believe there´s been only one (!) study, a very recent one, about the productivity of modern-day multitasking. Buser and Peter (2012) invited students to do word games and sudokus in the lab. The students were paid based on their performance. They were divided randomly into three study groups. The first group (‘single’) was asked to do one task first, and then the other, of twelve minutes or so each. Students in the second group (‘multi’) were asked to switch tasks every four minutes or so. Handy if you’re stuck, but a nuisance if you’re on a roll. The third group (‘choice’) was asked to switch between the two tasks at random.
The result? Multitasking is not productive! The ‘choice’ group showed that people switch between tasks more than is necessarily productive. And that’s exactly what I experience: switching looking for stimuli. The Amsterdam researchers also distinguished between men and women. It’s a well-known assumption that the latter group is much better at multitasking. But that didn’t appear to be the case at all: men and women both did equally well. Of course, this is just one experiment and it is therefore difficult to draw general conclusions. Psychologist have done more studies into multitasking. The results are the same and seem to have some general validity: multitasking is productive for neither men nor women (Borst et al., 2010; Mäntylä, 2013).
We could use more research by economists in this little explored field. The social relevance is clear due to the possibly large influence on individual (labour) productivity and the organisation of work processes. And my new year’s resolution: no more multitasking for me!
Borst, J., N. Taatgen en H. van Rijn (2010) The problem state: a cognitive bottleneck in multitasking. Journal of Experimental Psychology, 36(2), 363–382.
Buser, T. en N. Peter (2012) Multitasking. Experimental Economics, 15(4), 641–655.
Holmstrom, P. en B. Milgrom (1991) Multitask principal-agent analyses: linear contracts, asset ownership and job design. Journal of Law, Economics and Organization, 7(1), 24–52.
Mäntylä, T. (2013) Gender differences in multitasking reflect spatial ability. Psychological Science, to be published.