,".f "
ELIZABETH F. LOFTUS, MAHZARIN R. BANAJI,
JONATHAN W. SCHOOLER, AND RACHAEL A. FOSTER
WHO REMEMBERS WHAT?: GENDER
DIFFERENCES IN MEMORY
He: We met at 9.
She: We met at 8.
He: I was on time.
She: No. You were late.
He: Ah yes, I remember it well.
He: We dined with friends.
She: We dined alone.
He: A tenor sang. r
She: A baritone.
He: Ah yes, I remember it well.
("1 Remember It Well," sung by Maurice Chevalier
and Hermione Gingold in Gigi)
He remembers meeting at 9, that he was on time, that they dined
with friends, while a tenor sang. She remembers meeting at 8, that
he was late, that they dined alone, while a baritone sang. There is
humor in these differences, which derives in part from highlighting
the fallibility in his memory for this "important" social occasion. But
the example also brings us quickly to the main question we pose
.here: When men and women try to recall the past, who remembers
what?
Psychologists interested in people's memory for complex events
have often been led to ask questions about individual differences.
Our focus here is on one such variable, gender. Who is better, males
or females, at remembering the day they met? Who is better at
remembering the food, or the clothes, or the feelings, the time the
64
"
Loftus, E. F., Banaji, M. R., Schooler, J. W., 
& Foster, R. (1987). Who remembers what?: 
Gender differences in memory. Michigan 
Quarterly Review, 26, 64-85.
~. I,
ELIZABETH F. LOFTUS et at. 65
dinner began, or how long it lasted? Is He or is She? If he and she
leave the restaurant and witness an armed robbery on the way
home, who is better at remembering the event, the way the robber
talked, the way he looked, the direction he ran?
To place these questions in perspective, it will be helpful to know
something about how memory researchers think about memory.
Consequently, in the process of outlining the influence of gender
differences on various aspects of human memory, we will attempt to
familiarize the reader with important concepts and procedures that
have been used to guide memory research. One such concept, the
semantic-episodic distinction, will serve as a springboard for identi-
fying three areas of memory research -traditional episodic, autobi-
ographical, and eyewitness memory studies -that we will subse-
quently consider in our discussion of gender differences.
EPISODIC VERSUS SEMANTIC MEMORY
Information in long-term memory takes a variety of forms. Tulving
posited two distinct classes of memories: semantic memory, or gen-
eral knowledge about the world, such as the fact that a restaurant is
a place to eat, and episodic memory, consisting of our own personal
experiences, such as the fact that I met you at 8:00 p.m. when we ate
at The Broadway Restaurant last Saturday. In his own words: "Epi-
sodic memory receives and stores information about temporally
dated episodes or events, and temporal-spatial relations among these
events. ...Semantic memory is the memQry necessary for the use of
language. It is a mental thesaurus, organized knowledge a person
possesses about words and other verbal symbols, their meaning and
referents, about relations among them, and about rules, formulas,
and algorithms for the manipulation of these symbols, concepts, and
relations" (385-86).
One result of attempting to classify all memory processes within
two broad categories is that it forces researchers to consider the
components of memory. Tulving's distinction between semantic and
episodic memory highlighted the fact that many aspects of episodic
memory had been widely ignored. Maccoby and J acklin's comments
on gender differences in learning and memory clearly reflected this
gap in our knowledge. While Maccoby and Jacklin were able to
discuss how men and women performed on list-learning experi-
ments, they had nothing to say about gender differences for a whole
.
; t,
66 MICHIGAN QUARTERLY REVIEW
range of other important types of episodic memory abilities. Their
review contained no discussion of memory for spatial information or
memory for faces, or memory for real-life, complex events, such as
the day that He and She met. Since Maccoby and Jacklin's seminal
review of gender differences, much research has been devoted to
these frequently neglected aspects of episodic memory. It therefore
seems appropriate to reconsider how the two sexes fare in light of the
substantially broader range of episodic memory research that has
been conducted in recent years.
In undertaking a review of a major individual difference variable
such as gender, it is helpful to identify some underlying principles
that can generate predictions and unite an otherwise disparate col-
lection of individual findings. Maccoby and Jacklin suggest one
principle for understanding gender differences in memory: males
and females do not differ in overall memory ability, although inter-
est, motivation and training could affect the content of what is
remembered. This principle seems reasonable and is one that we
kept in mind as we reviewed the literature. However, it fell short in
one important respect; it failed to make any testable predictions
other than that one sex should not dominate the other in all aspects
of memory. In order to generate some more satisfying predictions we
had to take a different approach. It occurred to us that perhaps the
best predictions regarding differences between men's and women's
memories would come straight from the source, i.e. from men and
women themselves. Accordingly, we developed a brief questionnaire
that listed a number of different memory situations and asked sub-
jects to predict whether or not men and women would differ, and if
so, to estimate the magnitude of that difference. Our hypothesis was
straightforward: men's and women's beliefs about sex differences in
memory may reflect empirically observed differences. Results of our
survey are summarized in Table 1. Having gathered a set of predic-
tions about memory generated by men and women, we then com-
pared them to the findings obtained by researchers in the areas of
verbal, spatial, and face memory, as well as autobiographical and
eyewitness memory.
Throughout our review we kept two hypotheses in mind: 1) that
men's and women's relative memory ability would depend on the
specific type of memory in question; and 2) that these differences
would be predicted, at least to some degree, by men's and women's
beliefs about gender differences.
ci\ ~-,;~
ELIZABETH F, LOFTUS et aI, 67
Table 1. Percentage of Responses for Items x Gender Preference for
Memory
Males Males Females Females
much somewhat somewhat much
better than better than No better than better than
females females difference males males N
1. After listening to a list of 15 .86 8.9 59.0 29.1 2.2 461
common nouns, subjects are
asked to remember as many
words as possible.
,~; 2. After seeing 50 male & female .87 10.4 40.3 43.3 5.2 462
~; faces, subjects are asked to
identify the ones they saw
from a set of 100 faces.
3. Subjects are asked to view a 10.9 56.2 27.8 4.2 .88 457
complex block pattern and
asked to reconstruct it from
memory.
4. After having visited a new 7.54 34.1 40.7 15.1 2.4 464
place only once, subjects are
asked to go back to it 6
months later.
5. After having a conversation .65 5.4 35.5 45.4 13.1 465
with a friend, subjects are
asked to recall what they and
the friend said.
6. After having gone to a .21 1.5 10.1 37.8 49.7 465
wedding, subjects are asked to
describe what the bride's
cousin wore.
7. Subjects are telephoned in the 1.5 16.3 58.8 19.3 4.1 461
morning and asked to recall
where they put their keys the
evening before.
8. After having parked the car 8.0 31.5 45.1 12.8 2.6 461
and shopped for 2 hours,
subjects are asked to find
their way to their car.
9. Subjects witness a murder 9.4 31.9 45.6 12.0 1.1 351
being committed, and are
asked to describe the killer.
10. Subjects view a robbery and 32.1 49.9 16.5 1.4 0 351
are asked to describe the gun
the robber was holding.
11. After seeing a 30-second film 2.8 16.2 72.4 8.0 .57 351
of a bank robbery, subjects
are asked "How long was the
film?"
12. Subjects are asked to 1.7 0 31.5 39.2 26.7 352
remember the first person
they kissed.
13. Do males or females have .74 6.9 68.8 20.9 2.7 407
overall superior memory?
'
1..
r
,
68 MICHIGAN QUARTERLY REVIEW
VERBAL, SPATIAL, AND FACE MEMORY
In our search of the relevant literature, we started with an original
pool of 123 references from the Psych Info database consisting of
journal articles, book chapters, convention presentations, and 1
unpublished dissertations. We supplemented this with hand-
searched literature that we considered relevant, discarding studies
that were inconclusive with respect to sex differences or did not use a
specific memory measure. Our entire literature search started with
research reported post 1978 since Maccoby and Jacklin's work had
summarized the earlier research.
The most influential work published in the area of gender differ-
ences (Maccoby and Jacklin) summarized differences in verbal
memory in the following way:
...verbal content in a memory test may give some advantage to girls,
but it clearly cannot be said that either sex has a superior memory
capacity or a superior set of skills in the storage and retrieval of
information, when a variety of content is considered. (59)
Whereas Maccoby and Jacklin's conclusions sound reasonable, we
discovered upon examination of the literature they reviewed that
out of a total of 22 studies measuring verbal memory, 10 showed
female superiority, 12 showed no difference, and none showed male
superiority. The results of our review closely resembled those of
Maccoby and Jacklin (although, as we shall later discuss, our con-
clusions may differ somewhat). Out of a total of 35 studies that used
a measure of verbal memory, 20 reported female superiority, 13
reported no difference, and 2 reported superior male performance.
The majority of studies reported the use of word lists, and fewer
studies used paired associates, digits, prose, or the Weschler memory
subscales. The dependent measure most commonly used was free
recall, although some used cued recall or recognition, and a small
number of studies used a combination.
Of the 24 studies using an adult population, 15 reported better
performance by females, 7 reported no difference, and 2 favored
males. In the 11 studies with children, 6 showed superior female
performance, 5 showed no difference, and none reported superior
male performance. Our review of the literature on gender differ-
ences in verbal memory suggests a clear pattern: Females, whether
adults or children, appear to do better on tasks involving verbal~
c-
"
ELIZABETH F. LOFTUS et al. 69
material. However, it must be kept in mind that over a third of the
studies showed no sex differences.
Our questionnaire data showed an interesting pattern of results
for the item that asked subjects about their beliefs about whether
males or females would remember verbal information better. The
majority of subjects (59 %) believed that there were no differences in
verbal memory (Table 1, item 1). However, when they favored
males or females, a female preference dominated. Specifically, 31 %
believed that females are better, whereas only 10% favored males.
We might speculate about the reason for the finding that the direc-
tion of the preferences (males better versus females better) matches
the research evidence. It is possible that subjects in memory experi-
ments are acting in accordance with the stereotypes regarding sex
differences in memory. Another possibility is that the research find-
ings are a true reflection of gender differences in memory in the real
world and our subjects' beliefs are based on this experiential data.
Maccoby and Jacklin did not review data on memory for spatial
information (presumably since they focused mainly on gender dif-
ferences in spatial skills in general), and no previous review exists
with which we can compare our pattern of data. The literature
search made available 16 studies in which memory for spatial infor-
mation was measured. Overall, males appear to perform better than
females: In 8 studies, males outperformed females, 6 studies showed
no difference, and 2 studies favored females.
Materials used to test spatial memory included Corsi's Block Tap-
ping test, the Weschler (WISC or WAIS), maps, and blocks. The
memory measure was a free recall or recognition task in most cases.
One study measured position recall and one used a mental rotation
task. In the 10 studies of adults, 5 showed superior male perform-
ance, 3 showed no significant difference and 2 showed better female
performance. Of the 6 studies using children, 3 showed better male
performance and three showed no difference. None showed superior
female performance. When memory for spatial information is
tested, it appears that the gender advantage favors males. However,
a relatively small number of studies were available, and it is diffi-
cult to draw any conclusion regarding the significance of this differ-
ence.
Again, our survey data indicate that people's beliefs about a gen-
der difference in memory for spatial information matches the
,~,:
70 MICHIGAN QUARTERLY REVIEW
research data: they believe that for spatial information, males rather
than females will show superior memory. When memory for a block
pattern was the target information, subjects believed that males
would be somewhat better at the task than females (67 %); only a
corresponding 5 % thought that females would be better than males
(Table 1, item 3). When memory for information about directions
was involved, although 41 % of subjects believed there was no dif-
ference, more than twice the number of subjects believed that males
would be better (42 %) than those who felt that females would be
better (17%) (Table 1, item 4).
For our purpose, a gender difference in memory for faces was
particularly interesting because of the direct relevance of these find-
ings to the domain of eyewitness identification. In a summary of
gender differences in face recognition, Clifford and Bull concluded
that ,c. ..females are better than males at recognizing faces they
have previously seen." They qualified this statement, however, by
cautioning the reader that these findings came from laboratory stud-
ies and may not hold up in real life or simulated life-like situations.
., In most studies on face recognition, subjects were shown photo-
":~ graphs of faces. Later they were asked to identify those faces among
':~'~ a set of randomly presented faces consisting of ones they had seen
::~j before as well as new ones. In our literature review, of 11 studies
~?ii~, dealing with face recognition, 7 showed superior female perform-
i'.-::~; ance, 4 showed no difference between males and females, and none
f';';;-' showed superior male performance. Two studies in which children
irf" are subjects reported no sex difference (Hota; Etaugh and WhittIer).
In a review, Shapiro and Penrod reported a meta-analysis of facial
identification from which they concluded that females are better at
recognizing faces than are males. Although several questions need to
be answered (e.g., Is this effect restricted to recognition of same sex
faces? Does female superiority also surface when the situation is a
real-life, affect-laden one?), it appears that at least in laboratory
situations and with same sex faces, females do perform better than
males.
Why does the research evidence on face memory favor females? A
number of potential interpretations are available, and Clifford and
Bull suggest at least four: (a) superior female performance may be a
function of the greater exposure of females to female faces through
the media, providing them with greater opportunity to learn encod-
~,
."
'I
ELIZABETH F. LOFTUS et at. 71
ing strategies; (b) females' face-recognition superiority may be a
function of their greater social attentiveness; (c) personality and
cognitive factors as well as motivation may be the cause; and (d)
female superiority may have an ontogenetic component.
Interestingly, our survey data indicated that although 40 % of
subjects believed that males and females would be equally accurate
on a face recognition task, 49 % favored females, whereas only 11 %
favored males (Table 1, item 2). Thus, for all three domains (verbal,
spatial, and face memory), although most subjects believe there is no
memory difference between the two sexes, those who give the
advantage to either females or males show intuitions that match the
research data.
AUTOBIOGRAPHICAL MEMORY STUDIES
We asked our group of men and women a number of questions
regarding their relative memory abilities for autobiographical infor-
mation. Women were believed to have better memories than men
for conversations, 59 % of subjects favoring women versus 6 % favor-
ing men. In addition, 87 % of subjects believed that women would
be better at remembering what someone was wearing as compared
to only 2 % of subjects who favored men. The story was somewhat
different for spatial autobiographical memories. For some spatial
memories, such as remembering where one's car is parked, men
were given the upper hand, with 39 % favoring men compared to
15 % favoring women. For other spatial memories such as remem-
bering where they had left their keys the evening before, men and
women were not predicted to differ (Table 1, items 5-8). Armed
with these predictions, we searched the autobiographical memory
literature. Unfortunately, in many cases the nature of autobiograph-
ical studies did not allow us directly to evaluate our subjects' predi9-
tions. Nevertheless, some interesting sex differences did emerge.
In autobiographical memory studies people are typically asked
about their own past personal experiences. Occasionally a very
ambitious researcher will record his or her experiences over an
extended period of time, and will later be tested to see what is
remembered. Psychologist Marigold Linton studied her own mem-
oryevery day for a six year period from 1972-1977. She wrote on
cards individual memories such as "I have dinner at the Canton
Kitchen; delicious lobster dish" or "I land at Orly Airport in Paris."~
,
ELIZABETH F. LOFTUS et al. 73
words (e. g. "book"), activity words (e. g. "open") and affect words
(e.g. "interested"). Robinson's variables included recency of the
event, type of event described, latency to generate an event, and age
of the subject at the time of the incident. Some interesting observa-
tions emerged about how people recall autobiographical memories.
For example, affective words elicited more recent memories than
did object or activity words. Of greatest interest for us, however,
were reasonably reliable sex differences. Specifically, women tended
to generate more recent memories and tended to think of memory
associations more quickly than men. In other respects, however, the
two sexes were quite similar in their performance. Men and women
were equally likely to generate the three most common types of
autobiographical experiences: accidents and injuries, romantic epi-
sodes, and first experiences. In addition, differences in the types of
episodes that were generated in response to the three types of word
prompts were similar for men and women. Thus, while sex differ-
ences in autobiographical memory recollections were observed, they
were not particularly pronounced. Robinson concluded that "For
the present, the sex differences obtained in this study must remain
an intriguing problem which future research m~y clarify" (594).
In order to further investigate differences in men's and women's
recollections of autobiographical memories, we attempted a follow-
up of Robinson's study in which we simply asked people to describe
a memory from their past and then to answer various questions
about that memory. There are various ways in which males' autobi-
ographical memories might be expected to differ from females'. For
example, a common belief among lay people is that women are more
emotionally oriented than men (Broverman, Vogel, Broverman,
Clarkson, and Rosenkrantz). If there is some validity to this intui-
tion, then one might expect women to rate their memories as being
more emotional as compared to men. In addition, in Robinson's
study it was observed that memories produced in response to affec-
tive words were more recent than those generated in response to the
other types of words. It was also observed that women recalled more
recent memories than men. It thus seemed plausible that women
might recall more emotional memories than men, and this might
account for the difference in the recency of their recollections. The
factors that we thought might distinguish the autobiographical rec-
ollections of men and women were: emotionality, abstractness, per-
sonal significance, frequency with which they thought about the
"c,
'"
'1
74 MICHIGAN QUARTERLY REVIEW
Table 2. Mean Evaluations of Self-Generated Autobiographical Memories
Males Females
(n = 94) (n = 113) T value
Emotionality
(1 not emotional, 7 very emotional) 4.45 4.80 1.32*
Abstractness
(1 concrete, 7 abstract) 2.24 2.54 1.55*
Life change associated with event
(1 no life change, 7 major life change) 3.27 3.68 1.38*
Frequency of thinking about the event
(1 never, 7 frequently) 4.39 4.41 .55*
Pleasantness
(1 unpleasant, 7 pleasant) 4.56 4.58 .07*
* = p> .05
memory, and pleasantness. After describing memories from their
past, we asked subjects to rate them on seven-point scales, one for
each of the above dimensions. Our results are presented in Table 2.
As can be seen in Table 2, in general men and women were similar
in their evaluations of their self-generated autobiographical memo-
ries. Although none of these differences reached statistical signifi-
cance, there was a slight trend for women to describe their memo-
ries as being more emotional, more abstract, and representing
greater life change. It thus remains plausible that when queried,
women may spontaneously generate slightly 'more emotional, per-
sonally significant memories than men. Of course such an interpre-
tation must be viewed as suggestive conjecture, since in addition to
not being statistically significant, these differences may be con-
founded by differences in men's and women's willingness to self-
disclose, as well as by variations in their use of interpretation of
ordinal scales.
In order to explore further possible differences in the elicited auto-
biographical memories of men and women, we did a content analy-
sis of a subportion of the memory descriptions. We entertained a
number of specific hypotheses about potential differences in the
focus and content of men's and women's memories. The first
hypothesis that we considered was that women's memories would be
more likely to focus on social situations than men's memories. We
considered three aspects of social situations: 1) whether the memory
~"~
ELIZABETH F. LOFTUS et al. 75
involved other people, 2) whether the memory described a conversa-
tion, and 3) how other people in the memory appeared. An example
of a memory including other people is: "The other day I went out to
the University, messed around with some friends until late at night
and then went home and watched a movie on a video cassette
recorder ." An example of a memory that contains a conversation is
"High school graduation day: the senior luncheon was held that
afternoon. When I arrived home relatives were there from out of
town. My aunt and uncle called to say they wouldn't make it and I
cried. 1 told my parents 1 didn't want to go to the ceremony. 1 went
anyway; it was boring." An example of a memory that describes
how other people appeared is: "I remember making the state quali-
fying time in the 50 free style at districts last year. 1 remember lifting
my hand up and seeing that 1 touched first. But most of all, I
remember my timer's face as she smiled and gave me my time."
A second hypothesis that we considered was that women would be
more likely than men to mention their feelings. An example that
described a feeling is "I reme~ber how 1 felt at a party where I was
meeting new people. 1 felt anxious inside and sometimes nervous. 1
felt impatient at times like I want to have a lot of fun at the immedi-
ate moment."
A third hypothesis that we considered was that men would be
more likely than women to provide spatial information while
describing their memories. An example of a description that con-
tained spatial information is "This morning I walked to class a dif-
ferent way. I left my house, walked dow!) the sidewalk and crossed
the street as usual. Then I walked a short ways and crossed another
street (this I do not usually do). Then I cut across some grass to class
instead of walking on the pathway like I usually do."
As can be seen in Table 3, this content analysis, like the compari-
son of the subjective memory evaluations, revealed few differences
between the memories of men and women. The one significant dif-
ference observed was that men were more likely than women to
include spatial information in their memory descriptions: z = 1.88,
p< .05 (one tailed). This result dovetails with our earlier observa-
tions that men tend to have better memories for spatial details.
Although the comparison of men's and women's self-generated
" memories can demonstrate interesting gender differences in the type
.*)
of autobiographical memory that comes most easily to mind, we
must be cautious in using this type of data to draw conclusions about
,
,
76 MICHIGAN QUARTERLY REVIEW
Table 3. Percentage of Memory Descriptions Containing Various Features
Men Women
Feature (n = 34) (n = 34)
Included another person 64 % 61 %
Described a conversation 12 % 12 %
Mentioned the person's feelings 29 % 21 %
Described someone's appearance 3 % 0 %
Included spatial information 45 % 21 %
differences in memory ability. For example, Robinson's observation
that women tend to recall more recent memories than men does not
demonstrate that women have more accurate memories for the
present, nor that men have a better memory for the past. While it is
possible that differences in memory accessibility account for this
finding, it is equally likely that women are simply more interested in
more recent memories. In short, differences in the unverified memo-
ries that men and women elicit when asked to generate a memory
are more indicative of their memory preferences than their memory
ability.
Since the autobiographical studies mentioned above cannot really
measure accuracy, it is not possible fully to evaluate our subjects'
predictions regarding autobiographical gender differences. Never-
theless, it is worth noting that many of our subjects predicted that
women would be more likely to recall conversations and what peo-
ple were wearing than men, yet no difference emerged in our analy-
sis of the freqQency with which the two sexes mentioned these items.
On the other hand, our subjects predicted that men would be more
likely in many cases to recall spatial details than women, and in fact
spatial details were more often mentioned by men. Since our sub-
jects' intuitions were often correct when it has been possible to vali-
date them, future research might attempt to study autobiographical
memory in ways that permit verification, such as recalling the
details of a lecture that was also videotaped. If our subjects' intui-
tions are accurate, women might be more likely to recall the appear-
ance and conversations involving the lecturer and audience, while
men might be better able to generate a map of the auditorium in
which the lecture was given.
,
j
ELIZABETH F. LOFTUS et al. 77
SURVEY RESEARCH
Although usually very specific in scope, there are some autobio-
graphical memory studies that can be verified for accuracy of mem-
ory. Data from large-scale surveys designed to measure people's rec-
ollections of automobile accidents, hospitalizations, and crime
victimizations can reveal useful information about how people gen-
erally process important events, and how the sexes might differ in
terms of their recall. Use of such data nicely complements the great
majority of psychology experiments that typically involve college
students in university laboratory settings who work on tasks that
bear little resemblance to their everyday activities (Fienberg, Lof-
tus, and Tanur; Loftus, Fienberg, and Tanur).
Gender differences were found, for example, in a survey designed
to evaluate injuries by motor vehicles. The study consisted of inter-
views with approximately 300 Michigan residents known to have
been in motor vehicle accidents during the 12-month period preced-
ing the interview. Information obtained from each respondent dur-
ing the interview was compared with data on official report forms
(Henson, Cannell, and Lawson). While the difference was small
(.77 versus. 73), it was statistically significant in favor of more accu-
rate reporting by females. Unfortunately, the report did not give a
breakdown on which items females reported more accurately and
which, if any, were more accurately reported by males. Thus, it is
entirely conceivable that females are better able to remember some
details (like the number of occupants in. the car) while males are
better able to remember other details (like the weather or the make
and model of the car) but when a combined accuracy score is
derived covering 39 items, the advantage goes to females.
A second survey example comes from the U.S. National Health
Interview Survey, a major data collection program designed to
secure health statistics. One part of the project concentrated on
people's ability to report their own hospitalizations and the hospital-
izations of significant others ("Reporting of Hospitalization"). The
respondents in this survey were approximately 1500 persons who
were interviewed up to a year after a family member had been
discharged from the hospital. The investigators hypothesized that
since it is usually the women's role to care for sick or recuperating
family members, women might be less likely to underreport hospi-
talizations than men. However, this result was not observed. There
'~
...
.\
78 MICHIGAN QUARTERLY REVIEW
was little difference between men and women in reporting either for
themselves or for other members of the family. The percent underre-
porting their own hospitalizations was 11 for men and 10 for
women. The percent underreporting the hospitalizations of their
children was 12 for men and 14 for women. The percent underre-
porting hospitalizations of other adult family members was 16 for
men and 14 for women. When deliveries were excluded from the
analysis, women were slightly poorer reporters for themselves than
were men for themselves. However, women reported somewhat bet-
ter for other adults than did men (8).
EYEWITNESS MEMORY STUDIES
Eyewitness memory studies have an advantage over many autobio-
graphical memory studies in that it is an easy matter to know
whether a given witness is accurate or not. In these studies, subjects
typically view a complex event, such as a film of a crime or accident,
and are later tested.
One common observation is that the way a question is asked can
affect the answer a person gives. Asking people who have seen an
auto accident, "How fast were the cars going when they smashed
into each other?" leads to a higher estimate of speed than the same
question asked with the verb "hit." Another common observation in
these studies is that after an event has been experienced, new infor-
mation about the event sometimes comes to the person's attention
and becomes incorporated into memory. The result is that the mem-
ory is supplemented or altered. When exposed to misleading post-
event information, subjects have misrecalled the color of a car that
was green as being blue, a yield sign as a stop sign, broken glass or
tape recorders that never ex,isted, and even recalled something as
large and conspicuous as a barn when no barn was ever seen (see
Loftus; Wells and Loftus).
The eyewitness studies on the malleability of memory suggest that
some information residing in a person's memory may be altered by
post-event information. This work constitutes evidence for the
greater relative vulnerability of episodic memories about which
Tulving hypoth~ized. This means that even the best laid plans of
men and women for producing accurate recollections may be sty-
mied by a natural and unfortunate byproduct of the reconstructive
and malleable nature of memory. When He remembered that they
..
-,
ELIZABETH F. LOFTUS et al. 79
met at 9, and She said it was 8, perhaps she influenced him to be
wrong.
In terms of suggestibility, there is an ancient belief, mentioned by
Stern among others, that women are more suggestible than men.
Stern wrote that suggestive questions "operate with special force in
the case of young and uneducated persons; more with women than
with men" (273). If Stern had lived another 70 years, he would have
been faced with empirical research that contradicted his intuitions.
An important consideration in terms of whether females will appear
to be more suggestible than males or not is the particular items that
are considered critical in the experiment. Research published some
70 years after Stern's remarks suggests that both women and men
pay more attention to items that catch their interest and conse-
quently store more or better information in memory about those
items. If a subsequent test asks about female-oriented items, women
outperform men. The converse is true if the testing concerns male-
oriented details (Powers, Andriks, and Loftus). In this research sub-
jects saw a simulated incident and some were later exposed to mis-
leading information about four critical items. Two critical items
were male-oriented in that males -in a previous study -had been
more accurate in recalling them. One of these was a question about
the male main character and another was about a nearby automo-
bile. Two critical items were female-oriented; they concerned the
female main character.
The results were clearcut. Males were more accurate and less
suggestible about the male-oriented items while females were more
accurate and less suggestible about the female-oriented items. This
finding provided clear support for the hypothesis that females and
males tend to be accurate on different types of items, perhaps indi-
cating their differential interest in particular items and correspond-
ing differential amounts of attention paid to those items. One conse-
quence of this differential interest is that there is a difference in the
effect on men and women of misleading information about these
specific items. This reasoning is in line with Eagly, who pointed out
in her review of sex differences in the related area of influenceability
of attitudes that "individuals are more readily influenced to the
extent that they lack information about a topic or regard it as trivial
and unimportant" (96). The specific items, then, are critical. The
fact that past research sometimes shows males to be more accurate
--
"""'
1" ,
"
!
.,
.\
80 MICHIGAN QUARTERLY REVIEW
than females may be largely due to the specific items on which those
males and females were tested.
One study suggests that different processes might be involved
when males and females succumb to suggestive information (Tousig-
nant). In this work, males and females watched several simulated
events. Some were exposed to misleading information while others
were not. All were tested. In addition, they were given a variety of
other tests designed to measure 1) their need to present themselves in
a socially desirable manner, 2) the extent to which they habitually
reflect upon themselves and their environment, 3) their beliefs about
their own memory ability, and 4) their visual imagery ability, along
with other tests.
In this particular study females were more susceptible to mislead-
ing information than were males. But of greater interest was the
finding that the two sexes seemed to be influenced by misinforma-
tion for different reasons. For males, those who were good at visual
imagery were less likely to be influenced by misleading information.
For females, those who acquiesced to external sources of informa-
tion and claimed to have a good memory were more influenced by
misleading information. This analysis suggested that males may be
misled because of a distortion in their memory while females may be
misled because of response bias considerations. Clearly more
research is needed to support or refute the hypothesis that different
mechanisms are at work when males and females are exposed to
misleading information.
Probably the best evidence for the hypothesis that neither sex has
superior memory ability, but that they do differ in terms of what is
remembered, comes from studies of general eyewitness accuracy.
The eyewitness literature is filled with studies in which both men
and women participated as subjects. In terms of who performs bet-
ter, the results have been equivocal. Some studies have shown that
females perform better than males (e.g., Ellis, Shepherd, and Bruce;
Lipton). Other studies have shown that males perform better (e. g.,
Trankell). Still others indicate no differences in the accuracy of
women and men (Bird; Cady; Geiselman, Fisher, MacKinnon, and
Holland). A few studies have produced what might be called rather
complicated results. For example, Clifford and Scott showed that
men and women were equally accurate in recalling details of a
nonviolent event, but males were better in recalling the details of a
violent one. Yarmey and Jones showed that males and females were
...
, '
ELIZABETH F. LOFTUS et at. 81
generally equivalent in their ability to recall details of a simulated
assault and to recognize people previously seen. The only significant
accuracy difference between men and women was a greater ten-
dency for women to report that they had not seen an individual
earlier when in fact they had. However, a close examination of the
data reveals an interesting trend. The event that was shown was a
simulated assault and rape. Subjects later tried to identify two
people-the suspect and the victim. When given a test that did not
contain the correct individuals, females made more errors (more
false alarms, fewer correct rejections) than males on the suspect
while males made more errors on the victim. This suggests that men
and women may be paying attention to different things.
It is surprising that women are not routinely better at recognizing
people in the eyewitness studies, given that they outperformed men
in the more traditional face recognition studies. Many people in our
survey anticipated, paradoxically, that women would be better than
men at recalling faces in the traditional recognition studies, while
not holding such an advantage in eyewitness situations. In fact,
people may have overcompensated for the difference between men's
and women's face recognition performance in a traditional and eye-
witness situation by predicting an advantage for men. Approxi-
mately 41 % believed that men would be more likely to recall the
appearance of a murderer, as compared to 13 % favoring women
(Table 1, item 9). It seems likely that this reversal in people's intui-
tions about gender differences in face recognition reflects a belief
that women may be differentially affected by the stress of an eyewit-
ness situation.
There is in fact some evidence indicating that women do respond
to eyewitness situations with a greater degree of stress, and that this
greater stress may, in at least one respect, produce more inaccurate
recollections. Loftus, Schooler, Boone, and Kline showed subjects a
videotape depicting a bank robbery. Women reported higher levels
of anxiety than men and in addition overestimated the duration of
the robbery to a greater extent than did men. In a follow-up study,
subjects witnessed two versions of the bank robbery, one more vio-
lent and stressful than the other. Although both versions lasted the
same amount of time, the more violent version was estimated as
taking more time and induced greater anxiety; thus level of anxiety
was implicated as a source of time overestimation. Taken together,
these studies suggest that women may overestimate the duration of
."
, \
82 MICHIGAN QUARTERLY REVIEW
violent events to a greater degree than men as a result of the higher
anxiety that they experience in response to such events.
Although most subjects in our survey did not believe there would
be a difference in time estimations, the above result is consistent
with their intuitions when they indicated a preference. Approxi-
mately 10 % of those queried believed that men would be more
accurate in recalling the duration of a film depicting a bank rob-
bery, compared to 9 % who believed that women would be more
accurate (Table 1, item 11).
CONCLUSIONS
The current review suggests that neither sex can be said to have a
better memory per se; rather the two sexes differ in terms of what
type of information they remember best. Unfortunately most of the
studies that we reviewed were not designed to explore gender differ-
ences in memory for different types of items. Our review suggests
that interactions between gender and type of memory item would be
a fruitful topic to pursue.
We also demonstrated empirically some interesting aspects of peo-
ple's beliefs about gender differences in memory. First, we showed
that generally people believe there are no gender differences in
memory for most items. When they favor males or females, how-
ever, their beliefs mirror the research data (e.g., females better at
verbal memory and face recognition, males better at spatial mem-
ory). When asked about overall memory, Qver 20 % favor women as
compared to less than 8 % favoring men (Table 1, item 13). This
observation is striking in light of other data that suggest that for both
men and women, males command more attention (Robinson and
McArthur), are more influential (Fry), and are remembered better
when they speak (Gruber and Gaebelein, cited in Bridgewater).
These observations have caused some researchers to claim that peo-
ple expect men to be more competent, intelligent, and knowledge-
able than women (Bridgewater). So why, we ask, do people think
that women have better or equal memories compared to men, but at
the same time fail to respond to women in the same positive way?
Perhaps, memory is not considered to be an important dimension of
competence. (The stereotypical secretary remembers her boss's
appointments; the wife remembers the family birthdays.) Anecdotal
evidence in support of this contention comes from the observation
.
" .
,
, , ELIZABETH F. LOFTUS et al. 83
that people are more than willing to confess "1 do not have a particu-
larly good memory," but they are loathe to admit "I am not particu-
larly intelligent." Thus a slight favoring of women's overall memory
ability may not compensate for people's seemingly lower evaluations
of women in other important respects. Even if people realize that
memory is an important aspect of intellectual functioning, they may
nevertheless be less attentive to or influenced by women as a result
of habit stemming from a long tradition of men holding positions of
greater authority.
Considerable caution should be used in attempting to explain the
gender differences in memory that have been identified in this
review. Other researchers (e.g., Kimura) have noted the inescapable
fact that men and women differ genetically, physiologically, and
psychologically. In addition, differences in the historical social roles
of the two genders have undoubtedly contributed to the develop-
ment of different interests as well as different expectations regarding
the types of activities at which each gender should excel. Thus,
variations between men's and women's memory performance may
be due to their physiological capabilities, their interest, their expec-
tations, or some complex interaction of these factors. Though pres-
ently unexplainable, the gender differences identified in this study
still provide an important insight: If She remembers better than He
that they met at 8, that he was late, that they dined alone while a
baritone sang, there will be something else that He will remember
better than She.
NOTE
This research was supported in part by grants from the National Institute of
Mental Health, the National Science Foundation, and the National Center for
Health Services Research.
WORKS CITED
Bird, C. "The Influence of the Press upon the Accuracy of Report." Journal of
Abnormal and Social Psychology 22 (1927): 123-29.
.~
84 MICHIGAN QUARTERLY REVIEW
Bridgewater, C. A. "When a Man Talks, You Listen." Psychology Today Dec. 1980:
25-27.
Broverman, I. K., S. R. Vogel, D. M. Broverman, F. E. Clarkson, and P. S.
Rosenkrantz. "Sex Role Stereotypes: A Current Appraisal." Journal oj Social
Issues 28 (1972): 59-79.
Cady, H. M. "On the Psychology of Testimony." American Journal oj Psychology 35
(1924): 110-12.
Clifford, B., and R. Bull. The Psychology oj Person IdentiJication. London:
Routledge & Kegan Paul, 1978.
Clifford, B., and J. Scott. "Individual and Situational Factors in Eyewitness Testi-
mony." Journal oJ Applied Psychology 63 (1978): 352-59.
Eagly, A. H. "Sex Differences in Influenceability." Psychological Bulletin 85 (1978):
86-116.
Ellis, H., J. Shepherd, and A. Bruce. "The Effect of Age and Sex Upon Adolescents'
Recognition of Faces." Journal oj Genetic Psychology 123 (1973): 173-74.
Etaugh, C., and T. E. Whittler. "Social Memory of Preschool Boys and Girls."
Psychology oj Women Quarterly 7.2 (1982): 170-74.
Fienberg, S. E., E. F. Loftus, and J. M. Tanur. "Cognitive Aspects of Health Survey
Methodology: An Overview." Millbank Memorial Quarterly 63.3 (1985): 547-64.
Fry, P. S. "Effects of Male and Female Endorsement of Beliefs on the Problem
Solving Choices of High and Low Dogmatic Women." Journal oj Social Psychol-
ogy 96 (1975): 65-77.
Galton, F. Inquiries into Human Faculty and Its Development. 2nd ed. New York:
E. P. Dutton, 1911.
Geiselman, R. E., R. P. Fisher, D. P. MacKinnon, and H. L. Holland. "Eyewitness
Memory Enhancement in the Police Interview." Journal oj Applied Psychology 70
(1985): 401-12.
Gruber, K. J., and J. Gaebelein. "Sex Differences in Listening Comprehension." Sex
Roles 5 (1979): 299-310.
Henson, R., C. F. Cannell, and S. Lawson. EJJects oj Interviewer Style and Ques-
tion Form on Reporting oj Automobile Accidents. Ann Arbor: Survey Research
Center, University of Michigan, 1973.
Hota, N. "Sex, Grade and SES Differences in Categorization in a Recall Task."
Psychological Studies 28.1 (1983): 48-50.
Kimura, D. "Male Brain, Female Brain: The Hidden Difference." Psychology Today
19 (Nov. 1985): 50-58.
Linton, M. "I Remember It Well." Psychology Today July 1979: 81-86.
."Transformations of Memory in Everyday Life." Memory Observed. Ed.
U. Neisser. San Francisco: Freeman, 1982.
Lipton, J. "On the Psychology of Eyewitness Testimony," Journal oj Applied Psychol-
ogy 62 (1977): 90-95.
Loftus, E. F. Eyewitness Testimony. Cambridge, MA: Harvard University Press,
1979.
Loftus, E. F., S, E. Fienberg, and J. M, Tanur. "Cognitive Psychology Meets the
National Survey." American Psychologist 40 (1985): 175-80.
Loftus, E. F., J. W. Schooler, S. M., Boone, and D. Kline. "Time Went by So
Slowly: Overestimation of Event Duration by Males and Females." Applied Cog-
nitive Psychology, in press.
Maccoby, E. E., and C, N. Jacklin. The Psychology oj Sex DiJJerences. Palo Alto:
Stanford University Press, 1974.
Powers, P. A., J. L. Andriks, and E. F, Loftus. "Eyewitness Accounts of Females and
Males." Journal oJ Applied Psychology 64 (1979): 339-47.
-.
ELIZABETH F. LOFTUS et al. 85
Reporting of Hospitalization in the Health Interview Survey (Public Health Service,
DHEW-HSM Publication No. 73-1261). Washington DC: Government Printing
Office, 1965.
Robinson, J. A. "Sampling Autobiographical Memory." Cognitive Psychology 8.4
(1976): 578-95.
Robinson, J., and L. Z. McArthur. "Impact of Salient Vocal Qualities on Causal
Attribution for a Speaker's Behavior." Journal of Personality and Social Psychol-
ogy 43 (1982): 236-47.
Shapiro, P., and S. Penrod. "Meta-analysis of Facial Identification Studies." Psycho-
logical Bulletin 100 (1986): 139-56.
Stern, W. "Abstracts of Lectures on the Psychology of Testimony and on the Study of
Individuality." American Journal of Psychology 21 (1910): 270-82.
Tousignant, J. P. "Individual Differences in Response Bias and Recall: A Characteri-
zation of the Effects of Misleading -Post-event Information." Diss. University of
Washington, 1984.
Trankell, A. Reliability of Evidence. Stockholm: Bechmans, 1972.
Tulving, E. "Episodic and Semantic Memory." Organization of Memory. Ed. E.
Tulving and N. Donaldson. New York: Academic Press, 1972.
Wagenaar, W. "My Memory. A Study of Autobiographical Memory over Six Years."
Cognitive Psychology, in press.
Wells, G., and E. F. Loftus. Eyewitness Testimony: Psychological Perspectives,
Cambridge: Cambridge University Press, 1984.
Yarmey, A. D., and H. P. Jones. "Accuracy of Memory of Male and Female Eyewit-
nesses to a Criminal Assault and Rape," Bulletin of the Psychonomic Society 21.2
(1983): 89-92.