REPORTS

Acknowledgments: The present study was funded by designed the experiments and wrote the manuscript. SOM Text
NIH grants RO1 MH086563 to W.A.S. and Y.N. Y.N. performed the experiment and analyzed Figs. S1 to S7
and RO1 MH058847 to W.A.S. We thank E. Wang, the data. Tables S1 to S5
A. Shang, and N. Nystrom for expert animal care References (24–35)
and E. Hargreaves, M. Yanike, and M. Shapiro Supporting Online Material
for helpful comments. The authors declare no www.sciencemag.org/cgi/content/full/333/6043/773/DC1 11 April 2011; accepted 21 June 2011
competing financial interests. Y.N. and W.A.S. Materials and Methods 10.1126/science.1206773

Google Effects on Memory: participants had encountered a series of ques-

tions to which they did not know the answers,
t(68) = 3.26, P < 0.003, two-tailed. It seems that
Cognitive Consequences of Having when we are faced with a gap in our knowledge,
we are primed to turn to the computer to rectify
Information at Our Fingertips the situation. Computer terms also interfered
somewhat more with color naming (M = 603 ms,
SD = 193 ms) than general terms (M = 559 ms,
Betsy Sparrow,1* Jenny Liu,2 Daniel M. Wegner3 SD = 182 ms) after easy questions, t (68) =
2.98, P < 0.005, suggesting that the computer
The advent of the Internet, with sophisticated algorithmic search engines, has made accessing may be primed when the concept of knowl-
information as easy as lifting a finger. No longer do we have to make costly efforts to find the edge in general is activated.
things we want. We can “Google” the old classmate, find articles online, or look up the actor Comparison using a repeated measures anal-
who was on the tip of our tongue. The results of four studies suggest that when faced with

Downloaded from http://science.sciencemag.org/ on June 25, 2018

ysis of variance (ANOVA) of specific search
difficult questions, people are primed to think about computers and that when people expect to engines (Google/Yahoo) and general consumer-
have future access to information, they have lower rates of recall of the information itself and good brand names (Target/Nike) revealed an
enhanced recall instead for where to access it. The Internet has become a primary form of external interaction with easy versus hard question blocks,
or transactive memory, where information is stored collectively outside ourselves. F(1,66) = 5.02, P < 0.03, such that search engine
brands after both easy questions (M = 638 ms,
n a development that would have seemed people in dyads or groups form transactive mem- SD = 260 ms) and hard questions (M = 818 ms,

I extraordinary just over a decade ago, many

of us have constant access to information. If
we need to find out the score of a ball game,
ory systems (2, 3). The present research explores
whether having online access to search engines,
databases, and the like, has become a primary
SD = 517 ms) created more interference than
general brands after easy questions (M = 584 ms,
SD = 220 ms) and hard questions (M = 614 ms,
learn how to perform a complicated statistical transactive memory source in itself. We investi- SD = 226 ms) (Fig. 1). Simple effects tests showed
test, or simply remember the name of the actress gate whether the Internet has become an ex- that the interaction was driven by a significant
in the classic movie we are viewing, we need ternal memory system that is primed by the need increase in RT for the two search engine terms
only turn to our laptops, tablets, or smartphones to acquire information. If asked the question after the hard question block, F(1,66) = 4.44,
and we can find the answers immediately. It has whether there are any countries with only one P < 0.04 (Fig. 1). Although the concept of knowl-
become so commonplace to look up the answer color in their flag, for example, do we think edge in general seems to prime thoughts of
to any question the moment it occurs that it can about flags or immediately think to go online computers, even when answers are known, not
feel like going through withdrawal when we to find out? Our research then tested whether, knowing the answer to general-knowledge ques-
can’t find out something immediately. We are once information has been accessed, our internal tions primes the need to search for the answer,
seldom offline unless by choice, and it is hard to encoding is increased for where the information and subsequently computer interference is par-
remember how we found information before the is to be found rather than for the information ticularly acute.
Internet became a ubiquitous presence in our itself. In experiment 2, we tested whether people
lives. The Internet, with its search engines such In experiment 1, participants were tested remembered information that they expected to
as Google and databases such as IMDB and the in two within-subject conditions (4). Partic- have later access to—as they might with infor-
information stored there, has become an external ipants answered either easy or hard yes/no mation they could look up online (4). Partic-
memory source that we can access at any time. trivia questions in two blocks. Each block was ipants were tested in a 2 by 2 between-subject
Storing information externally is nothing par- followed by a modified Stroop task (a color- experiment by reading 40 memorable trivia state-
ticularly novel, even before the advent of com- naming task with words presented in either ments of the type that one would look up online
puters. In any long-term relationship, a team blue or red) to test reaction times to matched (both of the new information variety, e.g., “An
work environment, or other ongoing group, peo- computer and noncomputer terms (including ostrich’s eye is bigger than its brain,” and infor-
ple typically develop a group or transactive mem- general and brand names for both word groups). mation that may be remembered generally, but
ory (1), a combination of memory stores held People who have been disposed to think about a not in specific detail, e.g., “The space shuttle
directly by individuals and the memory stores certain topic typically show slowed reaction times Columbia disintegrated during re-entry over Texas
they can access because they know someone (RTs) for naming the color of the word when the in Feb. 2003.”). They then typed them into the
who knows that information. Like linked com- word itself is of interest and is more accessible, computer to ensure attention (and also to pro-
puters that can address each other’s memories, because the word captures attention and inter- vide a more generous test of memory). Half the
feres with the fastest possible color naming. participants believed the computer would save
1
Department of Psychology, Columbia University, 1190 Am- Paired within-subject t tests were conducted what was typed; half believed the item would be
sterdam Avenue, New York, NY 10027, USA. 2Department on color-naming reaction times to computer and erased. In addition, half of the participants in
of Psychology, University of Wisconsin–Madison, 1202 West general words after the easy and difficult ques- each of the saved and erased conditions were
Johnson Street, Madison, WI 53706, USA. 3Department of
Psychology, Harvard University, 33 Kirkland Street, Cambridge, tion blocks. Confirming our hypothesis, com- asked explicitly to try to remember the infor-
MA 02138, USA. puter words were more accessible [color-naming mation. After the reading and typing task, par-
*To whom correspondence should be addressed. E-mail: RT mean (M) = 712 ms, SD = 413 ms] than ticipants wrote down as many of the statements
[email protected] general words (M = 591 ms, SD = 204 ms) after as they could remember.

776 5 AUGUST 2011 VOL 333 SCIENCE www.sciencemag.org

 REPORTS
A between-subjects 2 (saved or erased) by In Experiment 3, we tested memory for where recognized the accuracy of a large proportion of
2 (explicit memory instructions versus none) to find information that one might look up on- statements. But for those statements they be-
ANOVA revealed a significant main effect for line. Participants again read and typed in items lieved had been erased, participants had the best
only the saved/erased manipulation, as those of memorable trivia, this time in three within- memory (erase M = 0.93, SD = 0.09, pairwise
who believed that the computer erased what subject conditions (4). For one-third of the ques- comparisons to both saved conditions P < 0.05)
they typed had the best recall, omnibus F(3, 56) = tions, participants were shown “Your entry has compared with the statements participants be-
2.80, P < 0.05 [Erase M = 0.31, SD = 0.04, and been saved.” For another one-third of the ques- lieved they would continue to have access to
Erase Remember M = 0.29, SD = 0.07, paired tions, participants were shown “Your entry has (saved generically M = 0.88, SD = 0.12, and
comparisons of erased conditions not significant been saved into the folder X” (where X is one saved specifically to a folder M = 0.85, SD =
(ns)] compared with those who believed the com- of six folders named FACTS, DATA, INFO, 0.12, pairwise ns), repeated measures omnibus
puter would be their memory source (Save M = NAMES, ITEMS, or POINTS—generic inter- F (1, 27) = 4.01, P < 0.03.
0.22, SD = 0.07 and Save Remember M = 0.19, changeable names to which the statements had However, the opposite pattern was found
SD = 0.09, paired comparisons of saved condi- previously been randomly assigned). For the fi- for the question, “Was this statement saved or
tions ns). This finding corresponds to previous nal one-third of the questions, participants were erased?” Participants accurately remembered what
work on directed forgetting, showing that when shown “Your entry has been erased.” Participants they had saved (saved generically M = 0.61, SD =
people don’t believe they will need information were given the expectation that they would have 0.21, and saved into a folder M = 0.66, SD = 0.20,
for a later exam, they do not recall it at the same access to what they saved through a supposed pairwise ns) more than they accurately remem-
rate as when they do believe they will need it (5). “practice” trial in which they had access to the bered what they had erased (M = 0.51, SD = 0.19,
Participants apparently did not make the effort to file folders during a “recall” task. Thus, generi- pairwise comparisons with both saved conditions,
remember when they thought they could later cally saved, saved in a specific folder, and erased P < 0.04), repeated measures ANOVA omnibus
look up the trivia statements they had read. Be- trials were created for all participants. F (1, 27) = 5.34, P < 0.03. Thus, it appears that

Downloaded from http://science.sciencemag.org/ on June 25, 2018

cause search engines are continually available to Participants were then given a recognition believing that one won’t have access to the in-
us, we may often be in a state of not feeling we task. They saw all 30 statements, half of which formation in the future enhances memory for the
need to encode the information internally. When had been altered slightly (names or dates altered). information itself, whereas believing the in-
we need it, we will look it up. Participants had to judge yes or no whether the formation was saved externally enhances mem-
The main effect of the instruction to explicitly statement they were now shown was exactly ory for the fact that the information could be
remember or not was not significant, which is what they had read, whether the statement had accessed, at least in general.
similar to findings in the learning literature on been saved or erased, and finally, if the statement In this recognition task, when asked “If the
intentional versus incidental studying of material, had been saved to a folder, which folder it had information was saved, what folder was it saved
which generally finds that there is no difference of been saved into (they were given the folder into?” participants did remember more that the
explicit instruction (6, 7). Participants were more names, and also had “no specific folder” and information was erased (M = 0.54, SD = 0.19,
affected by the cue that information would or “erased” as answer options to this last question). pairwise comparisons with both saved condi-
would not be available to them later, regardless of Overall, in answer to the question “Was this tions, P < 0.001) than specifically whether the
whether they thought they would be tested on it. statement exactly what you read?” participants information was generically saved or which
folder it was saved into (saved generically M =
0.30, SD = 0.20, and saved into a specific folder
Fig. 1. Accessibility of brand names M = 0.23, SD = 0.14, pairwise comparisons ns),
(as measured by color-naming reac- repeated-measures ANOVA omnibus F (1,27) =
tion time) after blocks of easy or hard 21.67, P < 0.001. This result is a reminder of the
test items. Error bars, mean T SEM. experience of remembering something you have
read online that you would like to see again or
share but no longer remembering where you
saw it or what steps you took to find it in the
first place, or even knowing that a file is saved
onto your hard drive but having to use the search
feature to find it. The fact that some of the state-
ments were saved in a general folder was im-
portant to include to rule out increased memory
demands in the two saved conditions but does
not parallel the continuous access to information
we experience with current technology, in that
there is no nameless depository of leftover in-
Fig. 2. An if/then analysis of mem- formation we would check after searching the
ory for what the information is and obvious places. In addition, recognition is not
where to find it. Scale is measured usually the task we are charged with when an-
in proportion recalled. Error bars, swering someone’s question. We need to recall
mean T SEM. the information we have gathered.
Experiment 4 was conducted to see if people
would recall where to find information more
than the information itself. All participants ex-
pected trivia statements that they read and then
typed to be saved to a specific folder with a
generic name (“FACTS,” etc., as in the previous
experiment, although in this case there were no
practice trials and the names and number of

www.sciencemag.org SCIENCE VOL 333 5 AUGUST 2011 777

REPORTS
folders was never explicitly called to the parti- conducted giving participants separate scores for active memory who knows what in our families
cipants’ attention) (4). Participants were then giv- whether they (i) recalled both the statement and and offices, we are learning what the computer
en a recall task, in which they were given 10 min the folder where it was saved, (ii) recalled the “knows” and when we should attend to where
to write down as many of the statements as statement but not the folder, (iii) didn’t recall the we have stored information in our computer-
they could remember. Participants finally were statement but recalled the folder, or (iv) recalled based memories. We are becoming symbiotic
given an identifying feature of the statement that neither the statement nor the folder. with our computer tools (8), growing into inter-
they read (and that had been saved), and they Participants were particularly poor at re- connected systems that remember less by know-
had to answer with the folder name in which it calling both statement and folder (M = 0.17, ing information than by knowing where the
was saved. For example, for the statement “An SD = 0.16) and recalling the statement but not information can be found. This gives us the ad-
ostrich’s eye is bigger than its brain,” the ques- the folder (M = 0.11, SD = 0.08, pairwise com- vantage of access to a vast range of information,
tion would be “What folder was the statement parison, ns). They were significantly more likely although the disadvantages of being constantly
about the ostrich saved in?” Participants had to to recall nothing (M = 0.38, SD = 0.24), but “wired” are still being debated (9). It may be no
type into a dialog box called “Items” to recall surprisingly equally likely to recall the folder, more that nostalgia at this point, however, to wish
this particular folder correctly. Folder names were when they didn’t recall the statement (M = 0.30, we were less dependent on our gadgets. We have
not mentioned again, past the original typing SD = 0.16, pairwise ns), repeated measures become dependent on them to the same degree
period, and participants were never explicitly told ANOVA omnibus F (1, 31) = 11.57, P < 0.003 we are dependent on all the knowledge we gain
there were six folder names where the items were (Fig. 2). It would seem from this pattern that from our friends and co-workers—and lose if
saved. people don’t remember “where” when they they are out of touch. The experience of losing
Overall, participants recalled the places where know “what” but do remember where to find our Internet connection becomes more and more
the statements were kept (M = 0.49, SD = 0.26) the information when they don’t recall it. This like losing a friend. We must remain plugged in
better than they recalled the statements them- is preliminary evidence that when people ex- to know what Google knows.

Downloaded from http://science.sciencemag.org/ on June 25, 2018

selves (M = 0.23, SD = 0.14), between-subject pect information to remain continuously avail-
t(31) = 6.70, P < 0.001 two-tailed. These results able (such as we expect with Internet access),
seem unexpected on the surface, given the mem- they are more likely to remember where to find References and Notes
1. D. M. Wegner, in Theories of Group Behavior, B. Mullen,
orable nature of the statements and the unmem- it than to remember the details of the item. G. R. Goethals, Eds. (Springer-Verlag, New York, 1986),
orable nature of the folder names. Also, these One could argue that this is an adaptive use of pp. 185–208.
recall results are notable in comparison with the memory—to include the computer and online 2. D. M. Wegner, Soc. Cogn. 13, 319 (1995).
dismal level of recognition of which folder the search engines as an external memory system 3. V. Peltokorpi, Rev. Gen. Psychol. 12, 378 (2008).
4. Materials and methods are available as supporting
statement was saved into in Experiment 3. How- that can be accessed at will.
material on Science Online.
ever, several caveats need to be mentioned. Par- Relying on our computers and the infor- 5. R. A. Bjork, in Coding Processes in Human Memory,
ticipants did have a cue to memory (a word from mation stored on the Internet for memory de- A.W. Melton, E. Martin, Eds. (Winston, Washington, DC,
the trivia statement) with the folder recall that pends on several of the same transactive memory 1972), pp. 217–235.
the statements themselves did not have. We processes that underlie social information-sharing 6. F. I. Craik, E. Tulving, J. Exp. Psychol. Gen. 104, 268
(1975).
were not able to counterbalance the trivia and in general. These studies suggest that people 7. T. S. Hyde, J. J. Jenkins, J. Exp. Psychol. 82, 472 (1969).
the folders trials such that the folders were as share information easily because they rapidly 8. A. Clark, Natural-Born Cyborgs: Minds, Technologies,
numerous as the statements, which would be think of computers when they find they need and the Future of Human Intelligence (Oxford Univ.
necessary to counterbalance the uncued and knowledge (experiment 1). The social form of Press, New York, 2003).
9. N. Carr, Atlantic 302, 56 (2008).
cued recall tasks. information storage is also reflected in the find- Acknowledgments: We acknowledge the financial support of
However, if we look at the pattern of what ings that people forget items they think will be Columbia University to B.S. and NSF grant BNS-0841746
was remembered, the results do suggest that available externally and remember items they to D.M.W. Raw data are available from B.S.
“where” was prioritized in memory, with the think will not be available (experiments 2 and
advantage going to “where” when “what” was 3). Transactive memory is also evident when Supporting Online Material
forgotten. You might expect that, with the ad- people seem better able to remember where an www.sciencemag.org/cgi/content/full/science.1207745/DC1
Materials and Methods
vantages of cued recall, participants would most item has been stored than the identity of the SOM Text
remember the folder where statements were item itself (experiment 4). These results suggest References (10–13)
saved if they were cued both by our question that processes of human memory are adapting to 2 May 2011; accepted 27 June 2011
and by their recalling the statement in the first the advent of new computing and communica- Published online 14 July 2011;
place. To examine this, an if/then analysis was tion technology. Just as we learn through trans- 10.1126/science.1207745

778 5 AUGUST 2011 VOL 333 SCIENCE www.sciencemag.org

Google Effects on Memory: Cognitive Consequences of Having Information at Our Fingertips
Betsy Sparrow, Jenny Liu and Daniel M. Wegner

Science 333 (6043), 776-778.

DOI: 10.1126/science.1207745originally published online July 14, 2011

Downloaded from http://science.sciencemag.org/ on June 25, 2018

ARTICLE TOOLS http://science.sciencemag.org/content/333/6043/776

SUPPLEMENTARY http://science.sciencemag.org/content/suppl/2011/07/14/science.1207745.DC2
MATERIALS
http://science.sciencemag.org/content/suppl/2011/07/13/science.1207745.DC1

RELATED http://science.sciencemag.org/content/sci/333/6043/671.6.full
CONTENT
http://science.sciencemag.org/content/sci/333/6040/277.full

REFERENCES This article cites 7 articles, 0 of which you can access for free
http://science.sciencemag.org/content/333/6043/776#BIBL

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Use of this article is subject to the Terms of Service

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of
Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive
licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. The title
Science is a registered trademark of AAAS.