The Effects of Chronic Multitasking on Analytical Writing
Danielle M. Lottridge, Christine Rosakranse, Catherine S. Oh, Sean J. Westwood,
Katherine Baldoni, Abrey S. Mann, Cliff I. Nass
Department of Communication, Stanford University
450 Serra Mall, Stanford, CA 94305-2050
lottridg | being | syoh | seanjw | kbaldoni | asmann | nass @stanford.edu
Chronic multitaskers perform worse on core multitasking
skills: memory management, cognitive filtering and task
switching, likely due to their inability to filter irrelevant
stimuli . Our experiment examines effects of chronic
multitasking with task-relevant and irrelevant distractors on
analytical writing quality. We found a general switch cost
and, when controlling for that cost, effects of chronic
multitasking habits: heavy multitaskers write worse essays
in the irrelevant condition and better essays in the relevant
condition. Our study changes multitasking research
paradigms in two fundamental ways: it studied a realistic
writing scenario including access to both irrelevant and
relevant distractors. We found that the effect of chronic
multitasking is complex; heavy multitaskers are seduced by
unrelated distractors but able to integrate multiple sources
of relevant information.
Multitasking; analytical writing; distractors; chronic
multitasking; media multitasking index.
ACM Classification Keywords
H.1.2. User/Machine Systems: Human factors.
Is multitasking good or bad for you? Proponents (including
many employers) herald the ability to effectively juggle
tasks and pressure workers to respond to messages
immediately. Dissenters bemoan constant interruptions,
distractions and the inability to get engrossed. Can we
improve our ability to multitask? How does multitasking
affect our attention span and the quality of our work?
‘Practice makes perfect’—except when it comes to
multitasking. Chronic multitaskers perform worse on the
core executive functions of multitasking: taking in and
releasing items from short-term memory stores (memory
management), attending to relevant material and inhibiting
irrelevant material (cognitive filtering), and inhibiting the
cognitive structures required for one task and activating
those required for another task (task switching) . These
results suggest a unique and important change in
fundamental information processing. Counterintuitively,
multitasking does not appear to be a problem of attending
to the right things; rather chronic multitaskers’ trouble
appears to be the inability to ignore the wrong things.
Multitasking is studied in two main ways: field studies
prioritizing external validity (e.g., ; what does
multitasking look like in the world?) and controlled
experiments prioritizing internal validity (e.g., ); such
experiments tend to follow cognitive psychology standards,
often with primary tasks related to basic cognitive function
and unrelated distractors. This experiment extends that
research in two fundamental ways: first, it studies
multitasking experimentally in the realistic domain of essay
writing; second, it exposes subjects to information streams
that are either relevant or irrelevant to the task at hand.
The more people are exposed to media, the more they
multitask . High sensation seekers multitask more often
. The impact begins at a young age; in a survey of 3,461
North American girls aged 8-12, multitasking was inversely
face-to-face communication, feelings of social success, and
sleep . When ‘heavy’ (chronic) media multitaskers
(HMM) were compared with ‘light’ media multitaskers
(LMM), HMM were bad at precisely the tasks at which one
would expect them to be good . Habitual and in-the-
moment multitasking both hurt cognitive performance.
Participants completing three tasks simultaneously
performed worse on a subsequent memory test when
compared to participants performing the same three tasks
serially . Multitasking contributes to cognitive overload
through too much information supply and demand,
interruptions, and inadequate infrastructure, thus increasing
needs for planning, monitoring, reminding, and
reclassifying information . A study on word tasks found
that people maximized productivity by switching to
prioritize a continuous rate of return (an information
foraging orientation) and to complete subgoals .
Salvucci et al. model multitasking behavior with cognitive
architecture, threaded cognition and memory-for-goals
theory . Others suggest that people self-regulate their
interruptions to maintain a flow state .
Permission to make digital or hard copies of all or part of this work for
or classroom use is granted without fee provided that copies are not made o
distributed for profit or commercial advantage and that copies bear this notice
and the full citation on the first page. Copyrights for components of this wor
owned by others than ACM must be honored. Abstracting with credit is
permitted. To copy otherwise, or republish, to post on servers or to redistribute
to lists, requires prior specific permission and/or a fee. Request permissions
CHI 2015, April 18 - 23, 2015, Seoul, Republic of Korea.
Copyright 2015 ACM 978-1-4503-3145-6/15/04…$15.00.
Task Interruption & Resumption
CHI 2015, Crossings, Seoul, Korea
The type of task (information versus non-information) and
individual differences can moderate how tasks are
combined for multitasking; for example, higher versus
lower order task, difficulty and information carry-over
affect impact . If both tasks require the same cognitive
framework ('problem state'), it interferes with performance
. It was proposed to be more effective to multitask with
tasks using different parts of the brain . But fMRI
studies reveal that tasks using different parts of the brain
(e.g., perceptual encoding and decision-making) suffered
from temporary attention limits . These costs appear to
increase with age, as older adults direct more attention
toward irrelevant stimuli .
Field studies reveal that multitasking is associated with
poor academic performance . Multitasking students
take more time to achieve the same performance as
monotaskers , or to complete the same task ; texting
students performed worse on exams than non-texters ;
and students with open laptops had worse memory for
lectures than those without . The present research is
however the first to study the causal relationship between
multitasking and writing quality.
The effects of chronic and in-the-moment multitasking on
analytical writing quality were investigated with an essay
task accompanied by relevant or irrelevant distractors.
Undergraduates were granted course credit and categorized
using top and bottom quartiles Media Multitasking Index
(MMI) scores . 40 HMM and 40 LMM participated; 37
were male. Participants were randomly assigned to the
relevant or irrelevant condition. Gender was balanced
across MMI score and condition.
Sessions consisted of groups of 5 students using laptops
separated by divider boards. The task was a 30-minute
GRE-style essay arguing for or against a statement on
whether luxuries and conveniences of modernity strengthen
character. GRE performance is a valid predictor of graduate
grade point average, comprehensive examination scores,
publication citation counts, and faculty ratings . Text
was entered through a simple editor, which included an
input box for the essay and a side column with an ostensible
forum including distractor links to online content.
Participants were invited to post comments and told that the
side column displayed comments by other participants.
Predetermined comments and associated links were
displayed approximately every 2 minutes. In the relevant
condition, comments were related to the writing question,
and included links to relevant articles. In the irrelevant
condition, comments included unrelated links, such as
sports news and funny videos. We recorded the number of
switches between the writing and other applications (e.g.
browsers) and time spent writing.
Essays were scored manually based on a standard
Organization and Coherence university grading scheme.
Organization was judged by presence of 1) an introduction,
body, and conclusion; 2) paragraph structure; and 3)
appropriate arguments. Coherence was judged by 1)
connection between adjacent sentences, 2) building on
ideas, and 3) contributions to over-arching argument. On a
subset of 20 essays (25%), two raters (ignorant of the
conditions) achieved good inter-rater reliability r = .86.
Essays were also scored automatically with the Flesch-
Kincaid, a widely used (e.g., US military, MS Word)
indicator of writing complexity . It is calculated from
numbers of syllables per word and words per sentence. The
lower the score, the more difficult the text to read; for
example, "The cat sat on the mat" scores 116: low
complexity and highly readable.
First, we investigate the correlation between number of
switches between writing software and browser, time spent
writing (includes editing, revising, etc), essay length and
writing quality. Then, we control for number of switches to
isolate effects of chronic multitasking. One MANCOVA
investigates effects of between-subjects variables media
multitasking (HMM/LMM) and condition (relevant and
irrelevant) on time spent writing. A second MANCOVA
controls for length, switches and time spent writing to
examine effects on dependent variables: writing quality and
Flesch-Kincaid (transformed for normality ). In
summary, our hypotheses are:
H1. The more switches, a) the less time spent writing, b) the
worse quality, and c) less complex essays.
H2. HMM will spend more time on distractors than LMM.
H3. HMM will write essays that are a) lower quality and b)
less complex than those written by LMM.
RQ: Will the relevance of the distractors influence H2-3?
We observed a switch cost: the more participants switch,
the less time participants spend writing (H1a: r =-.528,
p<.001) and the worse the quality of their writing (H1b: r =
-.307, p=.006). H1c was not supported. We found that time
spent writing correlated with length (r=.298, p=.01) and
higher complexity (r=-.306, p=.006). Length correlated
with quality (r=.529, p<.001). This suggests that if given
more time to write, students would likely produce longer,
more complex essays and that those who write more
independent of allotted time achieve higher quality essays.
Controlling for number of switches, we found an interaction
effect where HMM spent a quarter of their time following
irrelevant links (almost 7 min.) whereas LMM spent only
16% doing so (approx. 4.5 min.; F[1,79]=6.314, p=.01,
partial η2=0.78). In the relevant condition, both LMM and
HMM spent about one fifth of their time following links.
When controlling for the number of switches and time spent
writing, we observed a significant interaction between
media multitasking and condition on writing quality
Task Interruption & Resumption
CHI 2015, Crossings, Seoul, Korea
Figure 1a. Transformed writing quality
Figure 1b. Transformed Flesch-Kincaid writing
complexity for HMM and LMM in the relevant and
irrelevant conditions; *error bars=95% C.I.
(F[1,79]=5.859, p=.02, partial η2=.074; Figure 1a) and
Flesch-Kincaid (F[1,79]=4.348, p=.04, partial η2=.056;
Figure 1b). Complexity and quality were not correlated (r=-
.121, p=.284). There were no main effects (H2, H3a, H3b).
HMM wrote the highest quality essays in the relevant
condition and the poorest quality essays in the irrelevant
condition, whereas LMM were unaffected (mHMM-R = 4.9,
mHMM-I = 3.9, mLMM-R = 4.1, mHMM-I = 4.4). These effects are
due to chronic multitasking, as in-the-moment multitasking
(switches, time spent following links) is controlled for.
Thus, we find differences based on chronic multitasking
habits: heavy multitaskers in the irrelevant condition
suffered the most in terms of essay quality and simplistic
writing. However, HMM appear to use their multitasking
habits in a positive way in the relevant condition, where
they wrote higher quality essays with more sophisticated
prose. The quality of the LMM essays was unaffected by
condition, though the complexity of the writing was higher
in the irrelevant condition where more time was spent
writing compared to the relevant condition. The apparent
danger is that HMM’s behavior toward irrelevant links
reflects how they tend to write in the real world: a t-test
shows that HMM are much more likely than LMM to self-
report multitasking during homework (p < .001; mHMM = 2.4
(SD = 1.3); mLMM = .64 (SD = .42)).
DISCUSSION AND CONCLUSIONS
In this study, we examined how heavy and light media
multitaskers differ in analytical writing behavior when
faced with relevant or irrelevant distractors. Our study was
designed for external validity with a realistic scenario:
writing essays online with access to both relevant and
irrelevant information. Our study was unique in
reconceptualizing switching as relevancy-dependent rather
than looking only at the effects of irrelevant stimuli.
Our findings support a multi-faceted view of costs and
benefits of multitasking. For both groups, writing quality
suffered as a function of the number of switches between
writing application and browser. Thus, in-the-moment
multitasking carries a cost for writing quality. However,
when controlling for switches and time spent writing, this
negative effect was dependent on one’s chronic
multitasking habits and the type of distractors present.
HMM benefited from the relevant condition and had higher
writing scores. In the irrelevant condition, LMM were able
to easily ignore links and spent more time writing. HMM
were seduced by irrelevant content; they spent significantly
less time writing and wrote poorer quality essays.
The relationship between complexity and quality is
nuanced. Complexity might be sophisticated when paired
with high quality but convoluted when paired with low
quality. Low complexity might mean readable when paired
with high quality, simplistic when paired with low quality.
Our results suggest complexity is expressed differently by
HMM and LMM depending on condition. HMM had high
complexity and quality in the relevant condition and low
complexity and quality in the irrelevant condition,
suggesting that relevancy increased sophistication and
irrelevancy decreased sophistication. LMM writing did not
differ in quality with condition. Lower complexity in the
relevant condition suggests simplicity did not affect quality.
High complexity in the irrelevant condition suggests
intricacy did not improve quality. The relationship between
writing complexity and quality continues to be studied (e.g.
); our results add to this ongoing discussion.
Our study finds chronic multitaskers have advantages when
using multiple streams of task-related content; they appear
to have an ability to quickly integrate relevant information.
This behavior fits within the theoretical framework of
information foraging ; HMM may have developed
better sampling techniques that are utilized when
constrained to relevant material. Yet, as sensation seekers,
chronic multitaskers fail to ignore interesting, irrelevant
material; they lose time by consuming those media, and
their work suffers as a result. As light multitaskers consume
fewer media, the condition neither helped nor hurt.
Multitasking is good and bad for you. Our study found that
HMM benefit when media streams are relevant and suffer
when they are irrelevant. Unfortunately, HMM typically
multitask with irrelevant media streams; our finding that
students multitask during homework is corroborated by
recent research that found students often chat online while
writing essays . Therefore, HMM often do not take full
advantage of their integration capacity to optimize their
performance. Careful construction of a context or digital
Task Interruption & Resumption
CHI 2015, Crossings, Seoul, Korea
environment to constrain HMM’s media intake to relevant
streams may thus greatly benefit their performance.
We thank Google for their support, Erina Dubois for her
help, Ethan Plaut for editing, and our anonymous reviewers
for their generous and constructive critiques.
1. Adler, R.F., and Benbunan-Fich, R. Self-Interruptions in
Discretionary Multitasking. Computers in Human
Behavior 29, 4 (2013), 1441-1449.
2. Beers, S.F., and Nagy, W.E. Syntactic complexity as a
predictor of adolescent writing quality: Which
measures? Which genre? Reading & Writing 22,
3. Borst, J.P., Taatgen, N.A., & Van Rijn, H. The problem
state: A cognitive bottleneck in multitasking. Journal of
Experimental Psychology: Learning, Memory, &
Cognition 36, (2010), 363-382.
4. Bowman, L.L., Levine, L.E., Waite, B.M., and Gendron,
M. Can students really multi task? An experimental
study of instant messaging while reading. Computers &
Education 5, 4 (2010), 927-931.
5. Clapp, W.C., Rubens, M.T., Sabharwal, J., and
Gazzaley, A. Deficit in switching between functional
brain networks underlies the impact of multitasking on
working memory in older adults. Proc. of the National
Academy of Sciences 108, 17 (2011), 7212-7217.
6. Dabbish, L., Mark, G., and Gonzalez, V. Why do I keep
interrupting myself?: Self interruption, habit, and
environment. Proc. CHI (2011), 3127- 3130.
7. Ellis, Y., Daniels, B., and Jauregui, A. The effect of
multitasking on the grade performance of business
students. Research in Higher Education Journal 8,
8. Flesch, R. A new readability yardstick. Journal of
Applied Psychology 32, (1948), 221–233.
9. Foehr, U.G. Media multitasking among American
youth: Prevalence, pairings and predictors. Unpublished
doctoral dissertation (2006), Stanford, CA, USA.
10. Fox, A.B., Rosen, J., and Crawford, M. Distractions,
Distractions: Does Instant Messaging Affect College
Students’ Performance on a Concurrent Reading
Comprehension Task? Cyberpsychology & Behavior,
12, 1 (2009), 51-53.
11. Hembrooke, H., and Gay, G. The Laptop and the
Lecture: The Effects of Multitasking in Learning
Environments. Journal of Computing in Higher
Education 15, 1 (2003), 46-64.
12. Iqbal, S.T., and Bailey, B.P. Leveraging Characteristics
of task structure to predict costs of interruption. Proc.
CHI (2006), 741-750.
13. Jeong, S.H., and Fishbein, M. Predictors of Multitasking
with Media: Media Factors and Audience Factors.
Media Psychology 10, 3 (2007), 364-384.
14. Kirsch, D. A Few Thoughts on Cognitive Overload.
Intellectica 1, 30 (2000), 19-51.
15. Kuncel, N.R., Hezlett, S.A., and Ones, D.S. A
comprehensive meta-analysis of the predictive validity
of the Graduate Record Examinations: Implications for
graduate student selection and performance.
Psychological Bulletin 127, 1 (2001), 162-181.
16. LaLonde, S.M. Transforming variables for normality
and linearity—when, how, why and why not’s. SAS
NESUG (2005), 11-14.
17. Ophir, E., Nass, C., and Wagner, A. D. Cognitive
control in media multitaskers. Proc. of the National
Academy of Sciences 106, 37 (2009). 15583-15587.
18. Payne, S. J., Duggan, G. B., and Neth, H. Discretionary
task interleaving: Heuristics for time allocation in
cognitive foraging. Journal of Experimental
Psychology: General 136, 3 (2007), 370–388.
19. Pea, R., Nass, C., Meheula, L., Rance, M., Kumar, A.,
Bamford, H., Nass, M., Simha, A., Stillerman, B., Yang,
S., and Zhou, M. Media use, face-to-face
communication, media multitasking, and social well-
being among 8- to 12-year-old girls. Developmental
Psychology 48, 2 (2012), 327-336.
20. Reader, W., and Payne, S. Allocating time across
multiple texts: sampling and satisficing. Human-
Computer Interaction 22, 3 (2007), 263-298.
21. Roberts, D.F., and Foehr, U.G. Trends in Media Use.
The Future of Children. Children and Electronic Media
18, 1 (2008), 11-37.
22. Rosen, L.D., Carrier, L.M., and Cheever, N. A.
Facebook and texting made me do it: Media-induced
task-switching while studying. Computers in Human
Behavior 29, 3 (2013), 948-958.
23. Salvucci, D. D., Taatgen, N. A., and Borst, J. P. Toward
a unified theory of the multitasking continuum: from
concurrent performance to task switching, interruption,
and resumption. Proc. CHI (2009), 1819-1828.
24. Spink, A., and Park, M. Information and non-
information multitasking interplay. Journal of
Documentation 61, 4 (2005), 548–555.
25. Tombu, M.N., Asplund, C.L., Dux, P.E., Godwin, D.,
Martin, J.W., and Marois, R. A Unified attentional
bottleneck in the human brain. Proc. of the National
Academy of Sciences 108, 33 (2011), 13426–13431.
26. Vega, V., McCracken, K., Nass, C.I., and Labs, L.
Multitasking Effects on Visual Working Memory,
Working Memory and Executive Control. ICA (2008).
Task Interruption & Resumption
CHI 2015, Crossings, Seoul, Korea