Content uploaded by Steve M J Janssen
Author content
All content in this area was uploaded by Steve M J Janssen on Dec 05, 2014
Content may be subject to copyright.
Temporal distribution of favourite books, movies, and
records: Differential encoding and re-sampling
Steve M. J. Janssen, Antonio G. Chessa, and Jaap M. J. Murre
University of Amsterdam, The Netherlands
The reminiscence bump is the effect that people recall more personal events from early adulthood than
from childhood or adulthood. The bump has been examined extensively. However, the question of
whether the bump is caused by differential encoding or re-sampling is still unanswered. To examine this
issue, participants were asked to name their three favourite books, movies, and records. Furthermore,
they were asked when they first encountered them. We compared the temporal distributions and found
that they all showed recency effects and reminiscence bumps. The distribution of favourite books had the
largest recency effect and the distribution of favourite records had the largest reminiscence bump. We
can explain these results by the difference in rehearsal. Books are read two or three times, movies
are watched more frequently, whereas records are listened to numerous times. The results suggest that
differential encoding initially causes the reminiscence bump and that re-sampling increases the bump
further.
AUTOBIOGRAPHICAL MEMORY
When people speak of autobiographical memory,
they are referring to the memories a person has
of his or her own life experiences (Robinson,
1986). Autobiographical memory does not only
consist of personal memories that are remem-
bered vividly, but also of autobiographical facts
(Brewer, 1986). Conway and Pleydell-Pearce
(2000; see also Conway, 2005; Conway, Singer, &
Tagini, 2004) proposed an autobiographical me-
mory model called the Self-Memory System
(SMS). The SMS consists of the working self and
the long-term self. The working self refers to the
goals a person currently has. The long-term self
consists of the conceptual self and the autobio-
graphical knowledge base, which comprises three
levels of specificity: lifetime periods, general
events, and event-specific knowledge. Lifetime
periods contain general events, which can be
single or repeated events. Those general events
in turn contain event-specific knowledge. The
SMS integrates episodic memories into the auto-
biographical knowledge base, which is used to
form goals in the working self and self-schemas
in the conceptual self. Arguably, one part of the
autobiographical knowledge base concerns per-
sonal information, such as what one’s favourite
books, movies and records are. We will therefore
look at what the temporal distributions of those
books, movies, and records can tell us about
autobiographical memory.
Several authors have focused on the temporal
distribution of autobiographical memory across
the life span. Such lifetime distributions tend
to have three main components (Conway &
Pleydell-Pearce, 2000; Rubin, Wetzler, & Nebes,
1986; Rybash, 1999). The first main component is
called childhood amnesia (or infantile amnesia).
People recall few events from the first years of
#
2007 Psychology Press, an imprint of the Taylor & Francis Group, an Informa business
Address correspondence to: Steve Janssen, University of Amsterdam, Faculty of Social and Behavioral Sciences, Department of
Psychology, Roetersstraat 15 Room A.605, 1018 WB Amsterdam, The Netherlands. E-mail: s.m.j.janssen@uva.nl
This research was supported by a grant by NWO, the Netherlands Organisation for Scientific Research, to JM. We are grateful to
Professor David Rubin, Professor Dorthe Berntsen, Professor Tom ter Bogt, Dr Chris Moulin, and one anonymous reviewer for
helpful discussions and materials.
MEMORY, 2007, 15 (7), 755 767
http://www.psypress.com/memory DOI:10.1080/09658210701539646
their life. The second main component is the
reminiscence bump (RB). People recall relatively
many events from adolescence and early adult-
hood. The RB is usually followed by a period with
fewer memories. Finally, most recalled events are
from the last (most recent) few years. This is
called the recency effect (or retention effect).
Cermak (1984) considered these recent perso-
nal events to be episodic, while he thought that
most remote memories were semantic. Episodic
memories are generally seen as personal events
bound to a spatial and temporal context, while
semantic memories or autobiographical facts con-
tain knowledge about personal events (Brewer,
1986; Tulving, 1972). Schooler, Shiffrin, and Raaij-
makers (2001) proposed a theory about how
episodic memories could lose their contextual
information and so become semantic memories.
Many theorists assume that episodic memories are
progressively modified in neocortical regions until
they are independent of the hippocampal complex
(e.g., Meeter & Murre, 2004; Murre, 1996, 1997;
Rosenbaum, Winocur, & Moscovitch, 2001). This
mechanism could explain the Ribot gradient in
retrograde amnesia following hippocampal da-
mage. However, Rybash and Monaghan (1999)
presented participants with 18 cue words. The
participants had to describe for each cue word the
memory that came to mind first. Then they had to
indicate whether they remembered the event or
knew that the event had occurred. Finally, they
had to date the events. Rybash and Monaghan
found that both distributions had a RB and a
retention effect. The RB does not only consist of
semantic knowledge, while the recency effect does
not only consist of episodic memories.
Research on the temporal distribution of
autobiographical memory has mainly focused on
the RB and possible effects of age (e.g., Hyland &
Ackerman, 1988; Janssen, Chessa, & Murre, 2005;
also see Murre & Janssen, 2006). The main
problem with comparing the distributions of
different age groups is that the recency effect
coincides with the RB for adolescents and young
adults. To solve this problem, Janssen et al.
proposed a method to remove the recency effect
and uncover the underlying lifetime distribution.
When using this method, Janssen et al. found that
the recency effect obscured the RB in the
temporal distributions of adolescents and young
adults.
There are three general explanations for the
occurrence of the RB. The first general explana-
tion is that in certain calendar years more
memorable events happened than in other calen-
dar years. There is some evidence that there are
external influences on the temporal distribution
of personal events (Conway & Haque, 1999;
Jansari & Parkin, 1996; Rubin & Berntsen, 2003;
Schrauf & Rubin, 2001). However, the RB is a
very robust phenomenon. It is present in the
distributions of many different age groups. The
locations of the RB of these groups fall on
different calendar years, but similar ages (Rubin
et al., 1986). This explanation is therefore insuffi-
cient, although certain public events may affect
the temporal distribution of personal events.
The second general explanation is called re-
sampling. It states that at a certain age people
start reminiscing about events from the period
when they were adolescents or young adults
(Rubin et al., 1986). However, this explanation
has some theoretical shortcomings. It is unclear
why people mainly rehearse events from adoles-
cence and early adulthood and not events from
other lifetime periods. Furthermore, Merriam and
Cross (1982), Hyland and Ackerman (1988), and
Webster and McCall (1999) found no difference
in the amount of reminiscing of younger and
older people. Finally, the re-sampling explanation
is unlikely to be the sole mechanism, because the
distributions of adolescents and young adults
display RBs too (Janssen et al., 2005).
The third general explanation is called differ-
ential encoding. It states that events in adoles-
cence and early adulthood are stored better than
in other lifetime periods (Rubin, Rahhal, & Poon,
1998; Rubin et al., 1986). Four mechanisms are
given for differential encoding, which are not
mutually exclusive. First, there are more novel
events in that lifetime period (cognitive account).
These first-time experiences are encoded more
strongly, because they will be used later on in life
as exemplars when people experience similar
events (Pillemer, 2001; Robinson, 1992). Second,
people form their identity during that lifetime
period (identity formation or self narrative ac-
count). Conway (2005; also see Fitzgerald, 1988)
hypothesises that many self-defining memories,
which are vivid and emotional memories of
personal events that have a large impact on the
identity of a person (Conway et al., 2004), come
from adolescence and early adulthood. Third,
more transitional events occur during that life-
time period (life scripts account). When people
are asked to date their most important personal
events, they recall events, such as graduation,
wedding, and retirement (Berntsen & Rubin,
756 JANSSEN, CHESSA, MURRE
2002; Rubin & Berntsen, 2003). These important
events are often positive, but sometimes negative.
The positive events usually occur in early adult-
hood, whereas the negative events can occur at
any moment in people’s lives (Berntsen & Rubin,
2004; Rubin & Berntsen, 2003). Fourth, the
events are stored better, because the brain works
at an optimum, causing the memory system to
work more efficiently in that lifetime period
(biological account). This mechanism is sup-
ported by results that the RB has been also
found in the temporal distribution of public
events (Howes & Katz, 1988; Rubin et al., 1998;
Schuman, Belli, & Bischoping, 1997; also see
Janssen, Murre, & Meeter, 2007).
Our view, however, is that both differential
encoding and re-sampling influence the RB
(Janssen et al., 2005; Murre & Janssen, 2006).
Janssen et al. found RBs in autobiographical
memory distributions of all age groups, but the
size of the RB became larger as participants
became older. We concluded that personal events
are encoded more strongly during adolescence
and early adulthood, causing the RB in the
distribution of young adults, and that personal
events from adolescence and early adulthood are
re-sampled more frequently than events from
other lifetime periods, causing the increase in
the size of the RB in older participants.
FAVOURITES QUESTIONNAIRE
To examine the role of re-sampling more closely,
one could give people an autobiographical mem-
ory test and ask them how frequently they had
thought of each recalled event. However, people
are very poor at judging their past memory
performance (Arnold & Lindsay, 2002, 2005;
Joslyn, Loftus, McNoughton, & Powers, 2001;
Parks, 1999). Therefore, we set up an experiment
in which the stimuli have a priori different levels
of rehearsal, but still are related to autobiogra-
phical memory. In the experiment, participants
were asked to name their three favourite books,
movies, and records and to indicate when they
had encountered these items for the first time. We
will look at the effect of the different levels of
rehearsal on the recency effect and the RB.
The favourite items can be seen as a part of
autobiographical memory in several ways. They
can be regarded as vivid personal memories or as
autobiographical facts (Brewer, 1986). First, the
items can be seen as event-specific knowledge
about single events in the autobiographical
knowledge base. For example, a person might
prefer the movie Titanic, because she went to see
that movie with the man she later married.
Second, the items can be seen as event-specific
knowledge about repeated events. For example,
people might name Abbey Road as their favour-
ite record, because they often listened to The
Beatles when they were at high school. A third
way is that the items can be seen as representa-
tions of the goals in the working self or self-
schemas in the conceptual self. For example, a
person might prefer the book Martin Eden by
Jack London, because he identifies himself with
the main character of the book, since he has met
similar adversities as the main character. People
might also identify themselves with a certain
music genre, because it represents the beliefs
and attitudes that they have.
The temporal distributions of favourite re-
cords, movies, and books have been examined
separately by Holbrook and Schindler (1989,
1996), Larsen (1996), Sehulster (1996), and Smith
(1994), but to our knowledge there has not yet
been an experiment in which the distributions
of books, movies, and records are compared
directly. Larsen examined the temporal distribu-
tion of favourite books. He asked older adults to
name their favourite book and to indicate when
they had read the book. Larsen found that most
books came from the previous decade, but also
that many other books came from the period in
which the participants were between 10 and 40
years old. The temporal distribution of favourite
records was examined by Holbrook and Schindler
and by Smith. Holbrook and Schindler selected
popular music hits from the period 1932 1986 in
their first study. They asked participants to rate
their preference for each hit on a 10-point scale.
Holbrook and Schindler compared preference
rating with age at the time the song was a hit.
They found a reversed U-shape, which peaked at
24. People preferred music that stems from the
period when they were about 24 years old. Smith
asked participants to rate their favourability
towards different music styles on a 5-point scale.
Smith found a cohort effect on the popularity of
music styles. He found, for example, that blue-
grass, country, and Broadway musicals were the
most popular among participants who were born
between 1920 and 1930, whereas gospel, folk
music, easy listening, and opera were the most
popular among participants who were born be-
tween 1930 and 1940. Holbrook and Schindler
DIFFERENTIAL ENCODING AND RE-SAMPLING 757
and Sehulster examined the temporal distribution
of favourite movies. Holbrook and Schindler used
the same design as in their first study. They
selected films that had won the Academy Award
for Best Picture from the period 1927 1988. They
asked participants to rate their preference for
each winner on a 10-point scale. Holbrook and
Schindler compared preference rating with age at
the time that the picture was released. They found
that participants preferred movies that were
released in the period when they were about
26 years old. Sehulster asked participants to name
their favourite movies, and asked when they
saw these movies for the first time. Sehulster
found that the average age at which people saw
their favourite movie was 27.6 years. Please note
that recent movies can skew the average age at
which favourite movies are watched for the first
time. For example, if a 60-year-old participant
recalls two movies that he saw for the first time
when he was 15 years old and one movie that he
saw for the first time when he was 60 years old,
then the average age at which he saw his favourite
movies is 30 years. However, the mode of the age
at which he saw his favourite movies is 15 years.
In this experiment we ask participants when
they encountered their favourite books, movies,
and records for the first time. We use these dates
to make lifetime distributions. If re-sampling is
the sole cause of the RB, we should not find a RB
in the temporal distribution of books, since books
are hardly rehearsed. If differential encoding is
the sole cause of the RB, we should not find any
differences between the distributions of the three
types of media. Our view is that both differential
encoding and re-sampling influence the RB. We
hypothesise that differential encoding initially
causes the RB in the temporal distribution of
autobiographical memory, and that personal
events from adolescence and early adulthood
are re-sampled more frequently than events
from other lifetime periods (Janssen et al., 2005;
Murre & Janssen, 2006). Because we assume
that favourite items are event-specific knowledge
in the autobiographical knowledge base, repre-
sentations of goals in the working self, or repre-
sentations of self-schemas in the conceptual self,
we expect to find similar recency effects and RBs
in the temporal distributions of favourite items.
We expect to find that participants will prefer
more recent items, which they have encountered
for the first time in the previous 5 years,
compared to the number of items that they
encountered for the first time in any other period.
We also expect to find that participants will name
more recent books than recent movies and
records, because books are re-sampled less
frequently. Furthermore, we expect to find that
participants will name relatively more items
that they encountered for the first time when
they were between 10 and 25 years old than
items that they encountered for the first time in
any other lifetime period (besides the last
5 years), because personal events are encoded
more strongly during that period. We also expect
to find that participants will name more records
from the period in which they were between 10
and 25 years old compared with books and
movies from that period, because records are re-
sampled more frequently. Finally, we expect to
find no differences in the location of the peak of
the RB between the three types of media,
because the location is determined by differential
encoding.
There is possible cross-contamination between
the three types of media in all six directions.
There are soundtracks (e.g., The Bodyguard by
Whitney Houston or Ascenseur pour L’Echafaud
by Miles Davis), and there are records about
books (e.g., Wuthering Heights by Kate Bush and
White Rabbit by Jefferson Airplane). Further-
more, there are books about movies (e.g., Star
Wars: Tatooine Ghost by Troy Denning), and
there are books about records (e.g., Norwegian
Wood by Haruki Murakami and High Fidelity
by Nick Hornby). Finally, there are musicals
(e.g., Grease and The Sound of Music ), but
arguably the most common type of cross-contam-
ination is the movie adaptation (e.g., The Lord of
the Rings and Harry Potter ). Therefore, we ask
participants first what their favourite movies are.
METHOD
Participants
The test, which we called the Favourites Ques-
tionnaire, was administered in Dutch and English
via the Internet, where it is available at http://
memory.uva.nl/. Participants could come into
contact with our website in at least four ways:
(1) through links on other websites, (2) through
search engines, (3) through promotion in tradi-
tional media, such as articles in newspapers and
magazines, which included our web address, or (4)
through word of mouth. At the end of the test,
participants could invite relatives, friends, and
758 JANSSEN, CHESSA, MURRE
colleagues by sending them standardised e-mails.
Furthermore, we invited participants who had
taken other tests on our website to take this test
as well.
The results analysed in this article were gath-
ered between January 2005 and October 2006.
During this period 2161 participants answered the
questionnaire. The majority of the participants
were female (67.89%) and many participants had
a college or university degree or an equivalent of
a college or university degree (59.89%). Most
participants came from the Netherlands (N
1279), followed by the United States (N 406),
United Kingdom (N 104), Belgium (N 66),
Australia (N 64), and Canada (N 54). The
remaining 188 participants came from other
countries. The average age of the participants
was 35.57 years (SD 14.46). We divided the
participants into 12 age groups (see Table 1).
There were more participants in the young
age groups than in the older age groups, x
2
(11,
N 2161) 11.00, p .50. However, this unequal
distribution will not affect the results, because we
will look at the relative proportions within each
age group. We did not ask for socioeconomic
status or ethnicity.
Procedure
Before participants started the test we empha-
sised that the experiment was genuine and
serious. Participants were asked to provide their
informed consent. Then they had to give some
personal information, such as their date of birth,
level of education, how many times a week they
read the newspaper and watched the news, their
e-mail address, and a username and password.
They could also use this username and password
to log in at other tests on our website, such as
the Daily News Memory Test (Meeter, Murre, &
Janssen, 2005) and the Galton-Crovitz test
(Janssen et al., 2005, 2006; Murre & Janssen,
2006). If a participant had registered at those
tests, then he or she could log in directly at this
test.
After the participants had registered or logged
in, they were given the instructions. When they
had read the instructions they were asked to give
the titles and directors or leading actors of their
three favourite movies. To help the participants as
much as possible, we added some links to
websites about movies, so they could recover
the title and the director more easily. Providing
the participants with websites to look up the
names of the directors, authors, or artists made it
easier for us to identify the items, but it might
have affected the participants’ answers to the
questions about the years in which they encoun-
tered the items for the first time in the direction
of the release dates of the items. When the
participants did not know the title, the director,
or the leading actor, they were asked to enter a
question mark. They were also asked to select
from a drop-down menu the year in which they
saw the movies for the first time. The drop-down
menu had options from 1930 to 2005. On 1
January 2006, we also added the option ‘‘2006’’.
When participants did not recall the year in which
they first watched the movie, they could select a
question mark from the drop-down menu, which
was the last option. For each movie, the partici-
pants also had to select from a drop-down menu
how often they had seen the movie. The drop-
down menu had options ranging from ‘‘once’’ to
‘‘more than a 1000 times’’. This last question was
added during the course of the experiment and
was therefore not answered by all participants.
Subsequently, they were asked to name their
three favourite books, the year in which they
read the books for the first time, and how often
they read the books. Finally, they were asked to
name their three favourite records, the year in
which they heard the records for the first time,
and how often they had listened to the records.
After the participants had entered their fa-
vourite movies, books, and records, they were
asked how many new movies they had watched
recently. We explained that we were asking for
movies that they had not seen previously* the
movies did not necessarily have to be released
recently. They could enter the number of movies
TABLE 1
The number of participants for each age group
Age group N
16 20 yrs 365
21 25 yrs 378
26 30 yrs 237
31 35 yrs 214
36 40 yrs 177
41 45 yrs 176
46 50 yrs 203
51 55 yrs 176
56 60 yrs 126
61 65 yrs 56
66 70 yrs 31
71 75 yrs 22
DIFFERENTIAL ENCODING AND RE-SAMPLING 759
and select the corresponding time unit from a
drop-down menu, containing the options ‘‘a
week’’, ‘‘a month’’, and ‘‘a year’’. Furthermore,
they were asked how many new books they had
read recently. Finally, they were asked how many
novel records they had recently bought, received,
or downloaded. We mentioned explicitly that
we were not asking how many new records they
had listened to on the radio. These three ques-
tions were added during the course of the
experiment and were therefore not answered by
all participants.
After the participants had indicated how many
new movies, books, and records they had encoun-
tered recently, they were debriefed. We told them
briefly about the purpose of the questionnaire.
They were given the opportunity to send us
remarks or questions and they could invite
relatives, friends, and colleagues to take the test
as well by sending them a standardised e-mail.
RESULTS
We will examine the role of rehearsal by testing
which medium has the strongest recency effect,
from which lifetime period most items are pre-
ferred, which medium has the strongest RB, and
which medium participants encounter the most at
different ages. The participants reported a total of
19,449 items. However, 816 items were not dated,
leaving 18,633 items. The items consisted of 1646
different movies, 2548 different books, and 3428
different records. The 20 books, movies, and
records that were named most frequently are
given in Tables A1, B1, and C1 of the Appendices.
Recency effect
Participants had to select how often they had
encountered the preferred items. We found
that the participants listened to their favourite
records more frequently (M 82.34) than they
watched their favourite movies (M 5.15) or read
their favourite books (M 2.74). This effect was
highly significant, F(2, 3343) 182.338, MSE
12 540.52, p B.001. Our assumption that records
are listened to more frequently than movies are
watched and books are read was correct.
When we looked at the recency effect, we first
compared the number of items from the last
5 years with the average number of items per
5 years for each age group. We found a recency
effect in the distribution of favourite books,
x
2
(11, N 2649) 3191.64, p .001, favourite
movies, x
2
(11, N 2372) 852.98, p .001, and
favourite records, x
2
(11, N 2238) 788.36, p
.001. About 31.7% of the items the participants
preferred were from the previous 5 years. Second,
we compared the sizes of the recency effect of the
favourite books, movies, and records with paired
samples t-tests. We found that participants pre-
ferred relatively more recent books than recent
movies, t(11) 3.371, p .006, and relatively
more recent books than recent records, t(11)
3.645, p .004, but not relatively more recent
movies than recent records (p .241). Third, we
repeated the analyses without the two youngest
age groups (i.e., 16 20 and 21 25), because in
these age groups the recency effect coincides with
the RB. We found that participants preferred
relatively more recent books than recent movies,
t(9) 2.900, p .018, relatively recent books than
recent records, t(9) 4.774, p .001, and rela-
tively more recent movies than recent records,
t(9) 2.846, p .019.
Location reminiscence bump
The average ages at which the participants
encountered their favourite books, movies, and
records for the first time were 26.06, 25.09, and
24.48 years respectively. The differences between
the average ages are significant, F(2, 18630)
23.901, MSE 163.04, p B.001. However, these
averages can be skewed by the presence of recent
items. Therefore, we looked at the lifetime period
from which most items were preferred. We
calculated the proportion of items from each
period for each age group. We averaged the
proportions of each period. The first four periods
are therefore averaged over the 12 age groups,
the fifth period (i.e., 21 25 years) is averaged
over 11 age groups, and so on. Participants named
most items from the period when they were
between 16 and 20 years old (Mode
books
17,
Mode
movies
18, and Mode
records
18). We did
not find any differences between the distributions
of favourite books, movies, and records. Figure 1
shows the temporal distributions of favourite
books, movies, and records as a function of
the age at which the item was encountered for
the first time. Furthermore, we have added to the
figure the encoding function of autobiographical
memory as found in Janssen et al. (2005), which
was based on the results of a total of 1958
760 JANSSEN, CHESSA, MURRE
American and Dutch participants. All four dis-
tributions are normalised, such that the bins in a
curve add up to 1.
Reminiscence bump
When we looked at the RB, we first compared the
number of items from the period in which the
participants were between 11 and 25 years old
with the average number of items per 15 years
for each age group. Later, we will compare the
number of items from the period when the
participants were between 16 and 20 years old,
which is the period from which they preferred the
most items, with the average number of items per
5 years for each age group.
We found a RB in the distribution of favourite
records, x
2
(11, N 3757) 440.00, p .001, fa-
vourite movies, x
2
(11, N 3587) 357.66, p
.001, and favourite books, x
2
(11, N 3449)
573.95, p .001. About 37.2% of the items that
the participants preferred were from the period
when they were between 11 and 25 years old.
Furthermore, we compared the sizes of the RB
(i.e., between 11 and 25 years) of the favourite
records, movies, and books. We found that
participants preferred more records than books
from adolescence and early adulthood, t(11)
2.977, p .013, and we found that the difference
between the preferred movies and books ap-
proached significance, t(11) 1.992, p .072, but
we did not find that participants preferred more
records than movies from adolescence and early
adulthood (p .193). Finally, we omitted the two
youngest age groups again, because in these age
groups the RB coincides with the recency effect.
When we re-analysed the results, we found that
participants preferred more records than books,
t(9) 3.725, p .005, and more movies than
books from adolescence and early adulthood,
t(9) 2.450, p .037, but we did not find that
they preferred more records than movies from
adolescence and early adulthood (p .187). The
results per age group showed that the proportion
of movies from the period in which they were
between 11 and 25 years old was largest in the
four oldest age groups (i.e., 56 60, 61 65, 66 70,
and 71 75). When we analysed the results of
younger and older participants separately, we
found that the older participants preferred more
movies than books, t (3) 3.669, p .035, and
more movies than records from adolescence
and early adulthood, t(3) 4.341, p .023, but
we did not find that they preferred more records
than books from adolescence and early adulthood
(p .234). The younger participants preferred
more records than books, t (7) 2.798, p .027,
from adolescence and early adulthood, but we did
not find that they preferred more movies than
books (p .386) or more records than movies
(p .105) from adolescence and early adulthood.
Second, we compared the number of items
from the period in which the participants were
between 16 and 20 years old with the average
number of items per 5 years for each age
group. We found a RB in the distribution of
favourite records, x
2
(11, N 1767) 289.47, p
.001, favourite movies, x
2
(11, N 1628) 200.33,
p .001, and favourite books, x
2
(11, N 1528)
244.93, p .001. About 16.1% of the items that
the participants preferred were from the period in
which they were between 16 and 20 years old.
Furthermore, we found that participants pre-
ferred more records than books from the period
in which they were between 16 and 20 years old,
t(11) 4.413, p .001, but we did not find that
participants preferred more records than movies
(p .279) or more movies than books (p .336)
from that period. When we dropped the two
youngest age groups from the analyses, we found
that participants preferred more records than
books from the period in which they were
between 16 and 20 years old, t (9) 3.952, p
.003, but we did not find that participants
preferred more records than movies (p .107)
or more movies than books (p .250) from that
period. The results of the period in which the
participants were between 16 and 20 years old
also showed that the proportion of movies was
largest in the four oldest age groups. Older
participants who were between 56 and 75 years
0.00
0.05
0.10
0.15
0.20
0.25
5
Age at first encounter
Proportion
Books
Movies
Records
Encoding
10 15 20 25 30 35 40 45 50 55 60
Figure 1. The normalised distributions of favourite books,
movies, and records as a function of the age at which the item
was encountered for the first time. Age at first encounter
denotes the end of a 5-year range (e.g., ‘‘20’’ denotes age range
16 20).
DIFFERENTIAL ENCODING AND RE-SAMPLING 761
old preferred more movies than records from that
period, t (3) 4.756, p .018, the difference be-
tween the number of favourite movies and
favourite books approached significance, t(3)
2.619, p .079, but we did not find that they
preferred more records than books (p .344)
from that period. The younger participants who
were between 16 and 55 years old preferred more
records than books, t(7) 5.552, p .001, but we
did not find that they preferred more records than
movies (p .189) or more movies than books
(p .689) from that period.
Novel items
At the end of the test participants were asked to
indicate how many novel books, movies, and
records the participants had encountered re-
cently. The average number of novel books,
movies, and records per year for each age group
is given in Figure 2. Participants had watched
more movies than they had read books, t(1354)
12.828, p B.001, they had listened to more
records than they had read books, t(1354)
9.850, p B.001, and they had listened to more
records than they had watched movies, although
this effect only approaches significance, t(1354)
1.737, p .083. The effects of age group on the
number of novel books, novel movies, and records
were significant, F (11, 1343) 2.961, MSE
1538.13, p .001, F(11, 1343) 5.631, MSE
3674.92, p B.001, and F(11, 1343) 12.090,
MSE 10 463.49, p B.001. However, when we
compared the number of novel items that
adolescents and young adults (i.e., 16 30 years
old) had encountered to the number of novel
items that middle-aged and older adults (i.e.,
31 75 years old) had encountered recently, we
found significant differences for the number of
novel books, t(1353)3.031, p .002, the num-
ber of novel movies, t(1353) 6.779, p B .001, and
the number of novel records, t (1353) 9.705,
p B.001. Adolescents and young adults had
recently read fewer books, but they had watched
more movies and had listened to more records
than middle-aged and older adults recently had.
Participants who were younger than 31 years had
watched more movies than they had read books,
t(499) 13.571, p B .001, they had listened to
more records than they had read books,
t(499) 11.445, p B.001, and they had listened
to more records than they had watched movies,
t(499) 4.221, p B.001. Participants who were
older than 30 years had also watched more
movies than they had read books, t(854) 5.948,
p B.001, and had listened to more records than
they had read books, t (854) 2.239, p .020, but
they had watched more movies than they had
listened to records, t(854)2.576, p .010.
DISCUSSION
In this experiment the temporal distributions of
favourite books, movies, and records were com-
pared directly for the first time. We found a clear
RB in each of the different media types. The
average ages at which participants encountered
their favourite books, movies, and records for the
first time were between 24 and 26 years. These
results are consistent with Sehulster (1996),
whose participants were older than those in this
experiment. However, recent items skew the ave-
rage ages at which people encountered the items
for the first time. We found recency effects in the
distributions of favourite books, movies, and
records. The proportion of books from the last
5 years was greater than the proportion of movies
or records from the last 5 years. It is possible that
participants prefer books that they have read
recently, since we have demonstrated that books
are rehearsed less frequently than other types of
media, such as movies and records.
Because there were many recent items that
skew the average ages at which participants
encountered their favourite items for the first
time, we looked at the temporal distribution
of favourite books, movies, and records, rather
than the average age. We found RBs in the
distributions of favourite books, movies, and
records. All three distributions peaked when the
participants were between 16 and 20 years old.
0
25
50
75
100
125
150
20
Age groups
Number of items per year
Books
Movies
Records
25 30 35
40
45 50 55 60 65 70 75
Figure 2. The average number of novel books, movies, and
records per year for each age group. Age group denotes the
end of a 5-year group (e.g., ‘‘20’’ denotes age group 16 20).
762 JANSSEN, CHESSA, MURRE
These distributions are similar to the temporal
distribution of autobiographical memory (e.g.,
Janssen et al., 2005; Rubin et al., 1986; Rubin &
Schulkind, 1997; Rybash, 1999). We did not find
any difference between the distributions in terms
of location of the peak of the RB, but we did find
that the proportion of favourite records that were
listened to for the first time in the period in which
the participants were between 11 and 25 was
larger than the proportions of favourite books
and movies that were read or watched for the first
time in the period in which the participants were
between 11 and 25 years old. Participants pre-
ferred mainly records that they had listened to
when they were adolescents and young adults,
because we have demonstrated that records are
rehearsed more frequently than other types of
media, such as movies and books.
The results appear to support our view that the
RB is caused by both differential encoding and
re-sampling. Items, such as books, movies, and
records, are stored best between the ages 11
and 25. Items that are not rehearsed frequently,
such as books, are more likely to be forgotten,
causing a large recency effect and a small RB in
their distribution. Items that are rehearsed fre-
quently, such as records, are less likely to be
forgotten, causing a small recency effect and a
large RB in their distribution. If re-sampling was
the sole cause of the RB, we should not have
found a RB in the distribution of books, since
books are hardly rehearsed. If differential encod-
ing was the sole cause, we should not have found
differences between the three types of media in
the size of the RB.
An alternative explanation for the results
could be that difference in the size of the RB is
caused by the difference in the number of novel
items. We found a strong decrease in the number
of novel records as participants became older.
This decrease was stronger than the decrease in
the number of movies. In other words, it is
possible that participants preferred more records
(than movies) from adolescence and early adult-
hood, because they had encountered a larger
proportion of records (than the proportion of
movies) in those periods.
The levels of rehearsal are not the only
difference between books, movies, and records.
Music is associated more with time periods than
literature and cinema. For example, punk music is
associated with the late 1970s, whereas grunge
music is associated with the early 1990s. Smith
(1994) found a cohort effect of popularity of
genres of music * certain styles of music were
popular among people who were born in certain
periods. Music may therefore play a greater
role in the formation of identity. The result that
the proportion of favourite records from adoles-
cence and early adulthood was larger than the
proportion of favourite books and movies was
only found in the distributions of participants
born after 1950. For the participants who were
born before 1950, the proportion of movies from
the period in which the participants were adoles-
cents and young adults was larger than the
proportion of books and records. This is possibly
due to a cohort effect based on the availability of
the different types of media in the 1940s and
1950s. Although we did not ask older adults how
many books they read, how many movies they
watched, or how many records they listened to
when they were between 11 and 25 years old, we
did find that younger participants had recently
listened to more records than they had watched
movies, whereas older participants had recently
watched more movies than they had listened to
records. However, there is no evidence in earlier
research (Larsen, 1996; Holbrook & Schindler,
1989, 1996; Sehulster, 1996; Smith, 1994) or in the
results of this experiment that the three types of
media have their influences on the identity
formation at different ages, since the modes
(i.e., the location of the peak of the RBs) of the
three types of media fall in the same age period.
Imagine yourself being stranded as a castaway
on a desert island. Which three records would you
want to have with you? If you could only take
three books with you, which books would those
be? In this experiment, we looked at what
distributions of favourite books, movies, and
records can tell us about the role of differential
encoding and re-sampling in autobiographical
memory, since the preferred items can be seen
as a part of autobiographical knowledge base.
They can be regarded as event-specific knowledge
about single or repeated events or as representa-
tions of the goals or the self-schemas in the
conceptual self. Even favourite books, although
they are generally read only two or three times,
are stored better in adolescence and early adult-
hood than in prior or subsequent lifetime periods.
This result gives support for the differential
encoding hypothesis. However, people preferred
more records, which are listened to numerous
times, than books from adolescence and early
adulthood. This result gives support to the re-
sampling hypothesis. Therefore we suggest an
DIFFERENTIAL ENCODING AND RE-SAMPLING 763
interaction between differential encoding and re-
sampling. Autobiographical memory is encoded
more strongly in adolescence and early adult-
hood. These personal events thus have a larger
likelihood of being recalled than personal events
from other lifetime periods, making them even
more resistant to forgetting, each time they are
recalled.
Manuscript received 1 August 2006
Manuscript accepted 21 June 2007
First published online 10 August 2007
REFERENCES
Arnold, M. M., & Lindsay, D. S. (2002). Remembering
remembering. Journal of Experimental Psychology:
Learning, Memory, and Cognition , 28(3), 521 529.
Arnold, M. M., & Lindsay, D. S. (2005). Remembrance
of remembrance past. Memory, 13 (5), 533 549.
Berntsen, D., & Rubin, D. C. (2002). Emotionally
charged autobiographical memories across the life
span: The recall of happy, sad, traumatic, and
involuntary memories. Psychology and Aging,
17(4), 636 652.
Berntsen, D., & Rubin, D. C. (2004). Cultural life
scripts structure recall from autobiographical mem-
ory. Memory & Cognition , 32 (3), 427 442.
Brewer, W. F. (1986). What is autobiographical mem-
ory? In D. C. Rubin (Ed.), Autobiographical mem-
ory (pp. 25 49). Cambridge, UK: Cambridge
University Press.
Cermak, L. S. (1984). The episodic semantic distinc-
tion in amnesia. In L. R. Squire & N. Butters (Eds.),
The neuropsychology of memory (pp. 55 62). New
York: Guilford Press.
Conway, M. A. (2005). Memory and the self. Journal of
Memory and Language, 53 (4), 594 628.
Conway, M. A., & Haque, S. (1999). Overshadowing
the reminiscence bump: Memories of a struggle for
independence. Journal of Adult Development , 6(1),
35 44.
Conway, M. A., & Pleydell-Pearce, C. W. (2000). The
construction of autobiographical memories in the
self-memory system. Psychological Review, 107 (2),
261 288.
Conway, M. A., Singer, J. A., & Tagini, A. (2004). The
self and autobiographical memory: Correspondence
and coherence. Social Cognition , 22(5), 491 529.
Fitzgerald, J. M. (1988). Vivid memories and the
reminiscence phenomenon: The role of a self
narrative. Human Development , 31, 261 273.
Holbrook, M. B., & Schindler, R. M. (1989). Some
exploratory findings on the development of musical
tastes. Journal of Consumer Research, 16 , 119 124.
Holbrook, M. B., & Schindler, R. B. (1996). Market
segmentation based on age and attitude toward the
past: Concepts, methods, and findings concerning
nostalgic influences on customer tastes. Journal of
Business Research, 37 ,27 39.
Howes, J. L., & Katz, A. N. (1988). Assessing remote
memory with an improved public events question-
naire. Psychology and Aging, 3 (2), 142 150.
Hyland, D. T., & Ackerman, A. M. (1988). Reminis-
cence and autobiographical memory in the study of
the personal past. Journal of Gerontology, 43 (2),
35 39.
Jansari, A., & Parkin, A. J. (1996). Things that go bump
in your life: Explaining the reminiscence bump in
autobiographical memory. Psychology and Aging ,
11(1), 85 91.
Janssen, S. M. J., Chessa, A. G., & Murre, J. M. J.
(2005). The reminiscence bump in autobiographical
memory: Effects of age, gender, education, and
culture. Memory, 13 (6), 658 668.
Janssen, S. M. J., Chessa, A. G., & Murre, J. M. J.
(2006). Memory for time: How people date events.
Memory & Cognition , 34 (1), 137 148.
Janssen, S. M. J., Murre, J. M. J., & Meeter, M. (in
press). Reminiscence bump in memory for remote
events. European Journal of Cognitive Psychology.
Joslyn, S., Loftus, E. F., McNoughton, A., & Powers, J.
(2001). Memory for memory. Memory & Cognition,
29(1), 789 797.
Larsen, S. F. (1996). Memorable books: Recall of
reading and its personal context. In R. Kreuz &
M. S. MacNealy (Eds.), Empirical approaches to
literature and aesthetics (pp. 585 599). Norwood,
NJ: Ablex.
Meeter, M., & Murre, J. M. J. (2004). Consolidation of
long-term memory: Evidence and alternatives. Psy-
chological Bulletin, 130 (6), 843 857.
Meeter, M., Murre, J. M. J., & Janssen, S. M. J. (2005).
Remembering the news: Modeling retention data
from a study with 14,000 participants. Memory &
Cognition, 33 (5), 793 810.
Merriam, S. B., & Cross, L. H. (1982). Adulthood and
reminiscence: A descriptive study. Educational Ger-
ontology, 8(3), 275 290.
Murre, J. M. J. (1996). TraceLink: A model of amnesia
and consolidation of memory. Hippocampus, 6 (6),
675 686.
Murre, J. M. J. (1997). Implicit and explicit memory in
amnesia: Some explanations and predictions by the
TraceLink model. Memory, 5 (1 2), 213 232.
Murre, J. M. J., & Janssen, S. M. J. (2006, November).
Analyzing the reminiscence bump in autobiographi-
cal memory: First-time experiences, valence, and
emotionality. Paper presented at the 47th Annual
Meeting of the Psychonomic Society, Houston, TX.
Parks, T. E. (1999). On one aspect of the evidence for
recovered memories. American Journal of Psychol-
ogy, 112(3), 365 370.
Pillemer, D. B. (2001). Momentous events and the life
story. Review of General Psychology, 5 (2), 123 134.
Robinson, J. A. (1986). Autobiographical memory: A
historical prologue. In D. C. Rubin (Ed.), Autobio-
graphical memory (pp. 19 24). Cambridge, UK:
Cambridge University Press.
Robinson, J. A. (1992). First experience memories:
Context and functions in personal histories. In M. A.
Conway, D. C. Rubin, H. Spinnler, & W. A.
Wagenaar (Eds.), Theoretical perspectives on auto-
764 JANSSEN, CHESSA, MURRE
biographical memory (pp. 223 236). Dordrecht, the
Netherlands: Kluwer Academic.
Rosenbaum, R. S., Winocur, G., & Moscovitch, M.
(2001). New views on old memories: Re-evaluating
the role of the hippocampal complex. Behavioral
Brain Research, 127 (1 2), 183 197.
Rubin, D. C., & Berntsen, D. (2003). Life scripts help to
maintain autobiographical memories of highly posi-
tive, but not highly negative, events. Memory &
Cognition, 31 (1), 1 14.
Rubin, D. C., Rahhal, T. A., & Poon, L. W. (1998).
Things learned in early adulthood are remembered
best. Memory & Cognition , 26(1), 3 19.
Rubin, D. C., & Schulkind, M. D. (1997). Distribution
of important and word-cued autobiographical mem-
ories in 20-, 35-, and 70-year-old adults. Psychology
and Aging, 12 (3), 524 535.
Rubin, D. C., Wetzler, S. E., & Nebes, R. D. (1986).
Autobiographical memory across the adult life span.
In D. C. Rubin (Ed.), Autobiographical memory
(pp. 202 221). Cambridge, UK: Cambridge Univer-
sity Press.
Rybash, J. M. (1999). Aging and autobiographical
memory: The long and bumpy road. Journal of
Adult Development , 6 (1), 1 10.
Rybash, J. M., & Monaghan, B. E. (1999). Episodic and
semantic contributions to older adults’ autobiogra-
phical recall. Journal of General Psychology, 126 (1),
85 96.
Schooler, L. J., Shiffrin, R. M., & Raaijmakers, J. G. W.
(2001). A Bayesian model for implicit effects in
perceptual identification. Psychological Review,
108(1), 257 272.
Schrauf, R. W., & Rubin, D. C. (2001). Effects of
voluntary immigration on the distribution of auto-
biographical memory over the lifespan. Applied
Cognitive Psychology, 15 (1), 75 88.
Schuman, H., Belli, R. F., & Bischoping, K. (1997). The
generational basis of historical knowledge. In J. W.
Pennebaker, D. Paez, & B. Rime
´
(Eds.), Collective
memory of political events (pp. 47 77). Mahwah,
NJ: Lawrence Erlbaum Associates Inc.
Sehulster, J. R. (1996). In my era: Evidence for the
perception of a special period of the past. Memory,
4(2), 145 158.
Smith, T. W. (1994). Generational differences in
musical preferences. Popular Music and Society,
18,43 59.
Tulving, E. (1972). Episodic and semantic memory. In
E. Tulving & W. Donaldson (Eds.), Organization of
memory (pp. 381 403). New York: Academic Press.
Webster, J. D., & McCall, M. E. (1999). Reminiscence
functions across adulthood: A replication and ex-
tension. Journal of Adult Development , 6(1), 73 85.
DIFFERENTIAL ENCODING AND RE-SAMPLING 765
TABLE A1
The 20 most popular books, the number of the instances, and the mean and mode of the ages at which participants had read the
book for the first time
Title Author N M Mode
1 Harry Potter* J. K. Rowling 176 23.18 19
2 The Da Vinci Code Dan Brown 143 30.50 25
3 The Lord of the Rings* J. R. R. Tolkien 127 19.81 17
4 The Bible 54 16.50 6
5 De Ontdekking van de Hemel Harry Mulisch 49 32.45 19
6 Angels and Demons Dan Brown 45 31.91 23
7 The Secret History Donna Tartt 39 24.51 19
8 Pride and Prejudice Jane Austen 34 18.46 15
9 One Hundred Years of Solitude Gabriel Garcı
´
aMa
´
rquez 29 23.66 22
10 Flowers In The Attic V. C. Andrews 24 21.92 12
11 To Kill A Mockingbird Harper Lee 22 17.18 16
12 Catcher in the Rye J. D. Salinger 21 16.71 17
13 De Donkere Kamer van Damokles W. F. Hermans 21 18.48 16
14 The Hobbit J. R. R. Tolkien 19 18.11 10
15 1984 George Orwell 18 17.11 18
16 Kruistocht in Spijkerbroek Thea Beckman 18 11.94 12
17 Wuthering Heights Emily Bronte¨ 17 20.06 16
18 Sophie’s World Jostein Gaarder 15 25.20 16
19 Le Petit Prince Antoine de Saint-Exupe
´
ry 15 18.27 12
20 Memoirs of a Geisha Arthur Golden 14 21.50 23
*We have collapsed series or sequels in this table, but they are not collapsed in the analyses.
TABLE B1
The 20 most popular movies, the number of the instances, and the mean and mode of the ages at which participants had watched
the movie for the first time
Title Director N M Mode
1 The Lord of the Rings* Peter Jackson 168 30.59 21
2 Titanic James Cameron 89 21.97 13
3 Pulp Fiction Quentin Tarantino 79 21.48 19
4 The Sound of Music Robert Wise 65 13.46 12
5 The Godfather* Francis Ford Coppola 58 22.93 21
6 The Green Mile Frank Darabont 58 26.98 16
7 One Flew Over the Cuckoo’s Nest Milos Forman 57 23.82 21
8 Dirty Dancing Emile Ardolino 55 17.55 13
9 Harry Potter* Chris Columbus 55 29.60 20
10 The Shawshank Redemption Frank Darabont 53 25.66 22
11 Gone with the Wind Victor Fleming 53 15.59 18
12 The Matrix* Andy and Larry Wachowski 51 24.69 14
13 Star Wars* George Lucas 48 16.46 13
14 Grease Randal Kleiser 46 11.54 11
15 Le Fabuleux destin d’Ame
´
lie Poulain Jean-Pierre Jeunet 44 29.05 22
16 Pretty Woman Garry Marshall 39 24.31 20
17 Schindler’s List Steven Spielberg 39 28.31 16
18 West Side Story Jerome Robbins and Robert Wise 37 17.22 15
19 The Sixth Sense M. Night Shyamalan 29 31.83 30
20 Pirates of the Caribbean Gore Verbinski 29 24.07 20
*We have collapsed series or sequels in this table, but they are not collapsed in the analyses.
APPENDIX A
APPENDIX B
766 JANSSEN, CHESSA, MURRE
TABLE C1
The 20 most popular records, the number of the instances, and the mean and mode of the ages at which participants had listened
the record for the first time
Title Artist N M Mode
1 Dark Side of the Moon Pink Floyd 28 21.11 19
2 Jagged Little Pill Alanis Morissette 26 20.38 22
3 The Joshua Tree U2 27 17.52 18
4 Sgt. Pepper’s Lonely Hearts Club Band The Beatles 25 16.28 18
5 The White Album The Beatles 21 15.77 16
6 Hotel California The Eagles 18 20.06 22
7 Come Away with Me Norah Jones 18 22.06 18
8 The Wall Pink Floyd 16 21.13 16
9 Bridge over Troubled Water Simon & Garfunkel 16 23.06 17
10 Abbey Road The Beatles 15 16.07 20
11 Californication Red Hot Chilli Peppers 15 18.53 15
12 Thriller Michael Jackson 14 16.79 13
13 OK Computer Radiohead 14 20.36 20
14 Nevermind Nirvana 14 16.07 18
15 A Rush of Blood to the Head Coldplay 14 21.79 15
16 Ten Pearl Jam 13 15.92 17
17 Songs about Jane Maroon 5 12 24.50 16
18 Little Earthquakes Tori Amos 11 17.91 21
19 Revolver The Beatles 11 16.91 12
20 Blonde on Blonde Bob Dylan 11 17.91 17
APPENDIX C
DIFFERENTIAL ENCODING AND RE-SAMPLING 767