J.

Experimental Child Psychology 90 (2005) 235–254

www.elsevier.com/locate/jecp

Reminiscence and hypermnesia

in childrenÕs eyewitness memory
David La Rooya,*, Margaret-Ellen Pipeb, Janice E. Murraya
a
Department of Psychology, University of Otago, Dunedin, New Zealand
b
Section on Social and Emotional Development, National Institute of Child Health
and Human Development, Bethesda, MD 20892, USA

Received 23 July 2004; revised 20 October 2004

Available online 15 December 2004

Abstract

Three experiments examined reminiscence and hypermnesia in 5- and 6-year-oldsÕ memory

for an event across repeated interviews that occurred either immediately afterward (Experi-
ment 1) or after a 6-month delay (Experiments 2 and 3). Reminiscence (recall of new informa-
tion) was reliably obtained in all of the experiments, although the numbers of new items
recalled were fewer after a delay than when the interviews occurred immediately afterward.
Hypermnesia (increasing total recall over repeated recall attempts) was obtained only in
Experiment 1 when interviews occurred immediately and 24 h after the event.
Ó 2004 Elsevier Inc. All rights reserved.

Keywords: Memory; Eyewitness testimony; Reminiscence; Hypermnesia; Repeated interviewing

Introduction

Both laboratory-based studies and applied eyewitness memory studies indicate

that repeated retrieval attempts can lead to reminiscence (the elicitation of new
information) and the related phenomenon of hypermnesia (increases in recall across

*
Corresponding author. Fax: +1 301 480 5775.
E-mail address: [email protected] (D. La Rooy).

0022-0965/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jecp.2004.11.002
236 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254

several recall attempts) under certain conditions. Hypermnesia depends on reminis-

cence, but it also depends on the recall of previously recalled information. Insofar as
repeated recall attempts usually contain omissions of previously recalled information
(forgetting) as well as reminiscence, for hypermnesia to be observed, the reminis-
cence of new information must also exceed forgetting. Despite their potential to en-
hance eyewitness accounts, there has been no research that directly examined the
phenomena of reminiscence and hypermnesia in childrenÕs eyewitness memory. This
may be due in part to the controversy surrounding repeated interviews in real-world
contexts. In particular, some researchers have argued that repeated interviews could
be used as a means of introducing suggestive information and that inconsistencies
across repeated tellings of the same event may reduce the credibility of a childÕs tes-
timony. Insofar as interviews composed of open-ended questions typically yield
accurate information from young children (e.g., Fivush, 1994; Jones & Pipe, 2002;
Peterson & Bell, 1996; Pipe, Sutherland, Webster, Jones, & La Rooy, 2004), repeated
open-ended interviews may lead to increased recall in childrenÕs eyewitness memory
as measured by reminiscence and hypermnesia. In the current study, we examined
repeated open-ended interviews following both short and long delays with children
recalling a witnessed event.
Reminiscence and hypermnesia both have typically been found with adults
following a procedure originally developed by Erdelyi and Becker (1974). In a
typical reminiscence and hypermnesia experiment, the participants view a set of
to-be-remembered items presented as either pictures or words. After all of the
items have been presented, the participants are instructed to recall as many of
the items as they can remember. The participants then receive two further tests,
again recalling as many of the items as they can remember. In each test, the par-
ticipants are usually required to make a fixed number of responses even if doing
so means guessing. The results of studies using this procedure have shown that
the correct recall of pictures (Erdelyi & Becker, 1974; Roediger, Payne, Gillespie,
& Lean, 1982; Shapiro & Erdelyi, 1974), and of words if visualized (Erdelyi, Fin-
kelstein, Herrell, Miller, & Thomas, 1976; Henkel, 2004), increases across trials,
whereas errors do not systematically increase. More recently, Kern, Libkuman,
and Otani (2002) found that a greater amount of hypermnesia was obtained when
negatively arousing pictures were used as stimuli than when nonarousing pictures
were used.
However, there may be minor trade-offs against the increases in recall. Henkel
(2004) found that across repeated recall trials, the participants made more source-
monitoring errors in deciding whether the to-be-remembered stimuli were originally
presented as pictures or words, especially when they were unaware that they would
be later tested on memory for source. Shaw, Bjork, and Handal (1995) found retriev-
al-induced forgetting; that is, across repeated tests, the probability of recalling addi-
tional related information was less than that of recalling unrelated information.
Kelley and Nairne (2003) showed that memory for the order in which words from
a list are recalled decreases across repeated testing. Of particular interest to eyewit-
ness memory research is how the costs and benefits observed for repeated testing in
laboratory studies translate into applied contexts.
 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254 237

A few applied studies have used realistic events and provided direct evidence con-
firming the predictions of the laboratory research. One of these applied studies of
hypermnesia focused on the events of the widely televised O. J. Simpson verdict
(Bluck, Levine, & Laulhere, 1999). Eight months after the verdict announcement,
the adult participants were asked to recall the events surrounding the verdict three
times in response to open-ended free recall cues during a 1-h recall session. Bluck
and colleagues (1999) found that the recall of correct information increased between
the first and third trials. There were no simultaneous increases observed for errors
across the recall attempts. Dunning and Stern (1992) investigated hypermnesia in
undergraduate studentsÕ memory for a 4-min video of a violent crime. As in Bluck
and colleaguesÕ (1999) study, the participants were asked for three free recall ac-
counts. The results were clear in that recall increased monotonically as the number
of recall attempts increased, indicative of hypermnesia; errors did not increase signif-
icantly across the successive recall attempts. Scrivner and Safer (1988) also examined
recall across repeated interviews of a violent 2-min video. Their results showed that
the number of details that the participants wrote down increased across each recall
attempt. Although there was a significant increase in the number of errors, it was
small; the mean number of errors increased by only a half error from the first trial
to the fourth trial. Scrivner and Safer concluded that just because initial memory re-
ports are incomplete does not mean that the omitted information has decayed per-
manently from memory. Bornstein, Liebel, and Scarberry (1998) examined
hypermnesia for an emotionally arousing event compared with a nonemotionally
arousing event. Although they did find hypermnesia, they did not find evidence of
a greater amount of hypermnesia for the emotionally arousing event. Errors in-
creased over trials, but although this increase was significant, it was only a 1% in-
crease between the first and third recall tests.
Turtle and Yuille (1994), in contrast, found no evidence of hypermnesia for mem-
ory of a 412- min video of a crime. The absence of hypermnesia in their Experiment 1
might be due to the more stringent criterion that Turtle and Yuille used; whereas
hypermnesia is typically measured simply as an increase in correct recall, they sub-
tracted the amount of information forgotten from the previous recall attempt. The
absolute amount recalled is, in effect, reduced by the amount that is forgotten from
the preceding trial. Using this unusual measure, hypermnesia as defined in other
studies might not have been detected.
With respect to childrenÕs memory, there are only a handful of studies that have
examined hypermnesia directly, whether in the laboratory or in real-world analogs.
Early research conducted by Ballard (1913) and Ammons and Irion (1954) investi-
gated hypermnesia in 12-year-olds. The children in their studies were asked to mem-
orize poetry during a short period of time and then to recall it. Their results
demonstrated that the average number of lines of poetry recalled increased between
an immediate recall test and a recall test repeated 2 days later. As in the adult studies,
these authors demonstrated that there was more information in memory than was
elicited in any single recall attempt. Paris (1978), using a laboratory procedure, dem-
onstrated that 8- and 12-year-oldsÕ memory for a list of words increased between
three recall attempts that were separated by delays of minutes. Howe, Kelland, Bry-
238 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254

ant-Brown, and Clark (1992) also examined memory for word lists in 712- and 10-
year-olds. They observed hypermnesia across four consecutive recall trials separated
by delays of minutes, and they noted that there was no difference in the magnitude of
the hypermnesia effect for the two age groups studied. However, the effect was stron-
ger when children were tested at a delay of 2 days than when they were tested at long-
er delays of 16 and 30 days.
Dent and Stephenson (1979) provided some evidence that hypermnesia may oc-
cur in childrenÕs eyewitness memory across repeated interviews. In their study, 10-
and 11-year-olds recalled more details about a film during an interview conducted
after 24 h than during an interview conducted immediately after seeing the film,
without an increase in errors. However, recall did not increase further in repeated
interviews conducted at 2-week and 2-month delays. In a second experiment, Dent
and Stephenson obtained a similar finding when children were tested in immediate,
24-h, and 48-h interviews. This increase in recall was observed only when children
responded to free recall requests for information or to general questions about
what had happened. Children who were interviewed with specific questions did
not show any increase in recall. Henry and Gudjonsson (2003) examined hyperm-
nesia in the eyewitness memory of 11- and 12-year-olds with and without intellec-
tual disabilities. The eyewitness event was performed in the childrenÕs classrooms
and consisted of a performance by an actor about school life 100 years ago. After
the event, the children were interviewed about what they could remember with an
open-ended interview protocol immediately and 2 weeks later. The results showed
an increase in free recall across the two interviews but not in response to specific
questions.
Even when children do not demonstrate hypermnesia, they may nonetheless recall
new information across repeated recall attempts. When the level of recall remains the
same or decreases over time, it does not necessarily follow that the same information
is simply repeated from interview to interview. Reminiscence (without hypermnesia)
has been observed in childrenÕs reports about past events in many studies specifically
examining the individual pieces of information reported across interviews to see
whether they are new or repeated from previous interviews (e.g., Baker-Ward, Gor-
don, Ornstein, Larus, & Clubb, 1993; Fivush & Hamond, 1989; Hudson & Fivush,
1991; Peterson, Moores, & White, 2001; Pipe, Gee, Wilson, & Egerton, 1999; Salmon
& Pipe, 1997, 2000). These studies have generally found that new information can be
recovered in later interviews and added to childrenÕs accounts. However, a primary
issue surrounding the introduction of newly reminisced information is the accuracy
of the new information (Salmon & Pipe, 1997). A consistent finding has been that
new information is less accurate than information consistently reported across inter-
views. Salmon and Pipe (1997, 2000) found that new information added after a 6-
month delay was approximately 50% as accurate as information reported within a
week. Peterson and colleagues (2001) similarly found that new information recalled
about an injury and subsequent hospital treatment became progressively less
accurate at delays of 6 months, 1 year, and 2 years. By the 1- and 2-year
delays, the accuracy of new information ranged between 44 and 63% on average
for children between 2 and 12 years of age (see also Steward et al., 1996). However,
 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254 239

these studies examined new information added to accounts after a 6-month or 1-year
delay between interviews. Relatively little is known about the accuracy of new
information added to childrenÕs accounts when there are short delays between
interviews.
The three experiments that we present here are unique in that they are the first to
examine systematically how the concepts of both reminiscence and hypermnesia can
help in our understanding of childrenÕs eyewitness memory. We used an event orig-
inally devised by Murachver, Pipe, Gordon, Owens, and Fivush (1996) called ‘‘vis-
iting the pirate’’ and delays of 1 day and 6 months for the interviews. This design
permitted us to examine reminiscence and hypermnesia both when little forgetting
had occurred immediately after the event and when a significant amount of forget-
ting had occurred 6 months after the event (Jones & Pipe, 2002). The childrenÕs recall
was elicited through open-ended verbal recall interviews.
In the analyses, we evaluated the costs and benefits of repeated interviewing by
comparing the number of correct details with the number of errors that were made
in the interviews. Evidence of hypermnesia was measured as an increase in the num-
ber of accurate details recalled across successive interviews. Reminiscence was mea-
sured as the cumulative recall of new details across repeated interviews, that is, the
number of correct details from the first interview plus new details from the subse-
quent interview(s) (Bluck et al., 1999). These two measures allowed us to separately
assess whether the absolute amount of information reported across interviews in-
creased (hypermnesia) as well as whether multiple interviews, taken together, pro-
vided an increasing amount of new information (reminiscence).

Experiment 1

Experiment 1 examined reminiscence and hypermnesia in childrenÕs eyewitness

memory immediately and 24 h after an event when little forgetting had taken place
and recall was expected to be at its greatest. An additional variable considered was
whether being forewarned of an upcoming interview would affect hypermnesia.
Knowing that an interviewer will return to ask for more information may result in
witnesses thinking about further items of information in between interviews, thereby
resulting in a greater amount of hypermnesia. Thus, half of the children were fore-
warned that there would be a repeated interview and half were not.

Method

Participants

The participants were 40 children of European extraction (20 boys and 20 girls)
recruited from local primary schools in Dunedin, New Zealand. The mean age of
the children at the time of the event was 6 years 1 month ðSD ¼ 434 monthsÞ. The
caregivers of the children agreed in writing to their childrenÕs participation, and all
children were willing participants.
240 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254

Materials

Four panels (120 cm wide by 180 cm high) were arranged with a large painting of
a pirate setting comprising the backdrop. Set out in front of these were a sail, a drum
and sticks, a name book, a skeleton pen, a brown box, a waistcoat, a table and
cloth, a water jug, a jar for dye, an eyedropper, a bowl, a paintbrush, map paper,
a red box, a poem, a parrot in a cage, birdseed and a scoop, a telescope, a steering
wheel, a key, a treasure map, a spade, a barrel of polystyrene chips, a treasure chest,
a padlock, gold bars and coins, and a wooden cutout of a boat.

Procedure

A researcher escorted the children individually from their class and introduced
them to the ‘‘friendly pirate’’ who was dressed in blue and white striped pants, a blue
top, a purple waistcoat, and a red sash. After the introductions, the pirate and each
child performed the 20 event-activities together. The event ended when the child
found a treasure chest, inside of which was an inexpensive gift that the child kept
as a token of the pirateÕs appreciation of his or her assistance. During the event,
the pirate did not specifically name the objects and actions used in the activities
but used empty language such as ‘‘Okay, now that we are done with that, letÕs have
a go with this.’’ The entire event lasted between 10 and 15 min.
Children were individually interviewed immediately after the visit to the pirate
(Interview 1) and again 24 h later (Interview 2). Before the first interview, half of
the children received instructions that forewarned them that they would be inter-
viewed again the following day about what they could remember. The remaining
children received interview instructions that did not indicate that they would be
interviewed again. Aside from these instructions, the children received the same
interview protocol in each interview. Each interview began with the child being
asked, ‘‘Tell me everything you can remember about when you visited the pirate.’’
After the child had recalled all that he or she could, the interviewer introduced four
open-ended cues in an attempt to elicit further information:

1. ‘‘I heard that the first thing you have to do is to become a real pirate. I bet you
had to do lots of fun things for that. Tell me what they are.
2. ‘‘It sounds like you had to do some special things to get the map ready. Tell me
what they were.
3. ‘‘What about winning the pirate key? What sorts of things do you have to do for
that?
4. ‘‘I heard that the last thing you have to do is find the treasure. How did you do
that?’’

The interviewer encouraged the child to keep saying what he or she could remem-
ber by using statements such as ‘‘What else happened?,’’ ‘‘Tell me some more things
that happened,’’ and ‘‘That sounds like fun.’’ There were two interviewers (1 man
and 1 woman), and the same interviewer conducted both interviews for each child.
 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254 241

The participants were randomly assigned to each condition within the constraint that
there were equal numbers of boys and girls across conditions. Each interviewer inter-
viewed the same number of boys and girls.
Transcripts of the audiotapes and videotapes of the interviews were coded so that
a child received credit for a correct item by mentioning any of the 55 actions and
objects that were part of the 20 prescribed pirate and child activities. For example,
the statement ‘‘I looked through the telescope’’ received credit for the mention of the
action ‘‘look’’ and of the object ‘‘telescope.’’ Additional credit was not given if the
child mentioned the same detail(s) again later in the interview. Mentions of the ac-
tions and objects that had been provided in the interview cues (e.g., map, chest,
key, unlock) were not credited. Errors were coded as intrusions (mentions of actions
or objects that the child reported as occurring or being present during the event when
in fact they had not occurred or been present) or as distortions (incorrect descrip-
tions of items that had been present). Intrusions and distortions were combined to
form a single category of errors for the purpose of analysis. Two independent raters
coded one third of the transcripts, and interrater reliability was calculated as the
number of coding agreements divided by the total number of agreements and dis-
agreements for each transcript following Tinsley and Weiss (2000). One of the raters
was aware of the conditions in which the children were grouped, whereas the other
rater was not. Interrater reliability was 88.4%.

Results and discussion

Preliminary analyses showed that there was no effect of interviewer on the amount
of correct information or on the number of errors. For clarity, only significant results
that exceed an alpha of .05 are reported in what follows.
To examine hypermnesia (an increase in the number of correct details recalled in
Interview 2), a mixed-model analysis of variance (ANOVA) was performed with the
number of details reported (correct details or errors) and interview (Interview 1 or
Interview 2) as within-subjects factors and with interview instructions (children fore-
warned of second interview or children not forewarned of second interview) as a be-
tween-subjects factor. The results of this analysis showed that children reported a
greater number of correct details than of errors, F (1, 38) = 339.66, p < .001, and that
there was a difference in the number of details reported across interviews,
F (1, 38) = 8.24, p < .01. There was also a significant interaction between these fac-
tors, F (1, 38) = 5.79, p < .05 (Table 1). Further analysis of the interaction showed
that although correct recall increased between the immediate interview (M = 14.65,
SD = 5.21) and the interview 24 h later (M = 16.60, SD = 6.15), F (1, 38) = 7.07,
p < .05, total errors remained constant between the immediate interview
(M = 0.68, SD = 1.14) and the interview 24 h later (M = 0.78, SD = 1.05). These
findings indicate that the effect of repeated interviewing across a delay of 24 h can
be characterized as a growth in the amount of correct information reported in the
interviews but not in the amount of errors. The size of the hypermnesia effect was
an additional 1.95 correct details recalled in the second interview. There was no main
effect or interactions involving the factor of interview instructions.
242 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254

Table 1
Mean numbers of details reported in Experiment 1
Condition Immediate interview 24-h interview Cumulative recalla
Correct
Forewarned 15.75 (6.07) 17.45 (6.09) 22.00 (6.71)
Not forewarned 13.55 (4.76) 15.75 (6.26) 19.95 (5.72)
Overall 14.65 (5.21) 16.60 (6.15) 20.97 (6.24)
Errors
Forewarned 0.65 (1.13) 0.65 (0.93) 1.05 (1.46)
Not forewarned 0.70 (1.17) 0.90 (1.16) 1.45 (2.11)
Overall 0.68 (1.14) 0.78 (1.05) 1.25 (1.80)
Note. Standard deviations are in parentheses.
a
Cumulative recall is the number of details from the immediate interview plus new details from the
24-h interview.

To examine reminiscence (the cumulative recall of new details across Interviews 1

and 2), a mixed-model ANOVA was performed with the cumulative recall of details
reported (correct details or errors) and interview (Interview 1 or Interview 2) as with-
in-subjects factors and with interview instructions (children forewarned of second
interview or children not forewarned of second interview) as a between-subjects fac-
tor. The cumulative recall of details across the two interviews was significant,
F (1, 38) = 171.85, p < .001, and overall there were more correct details than errors,
F (1, 38) = 393.09, p < .001. There was also an interaction between the cumulative re-
call of correct information and errors across the two interviews, F (1, 38) = 103.59,
p < .001 (Table 1). Two further analyses of the interaction showed that cumulative
recall of correct details increased between the immediate interview (M = 14.65,
SD = 5.21) and the interview 24 h later (M = 20.97, SD = 6.24), F (1, 38) = 143.91,
p < .001, and that the cumulative recall of errors increased between the immediate
interview (M = 0.68, SD = 1.14) and the interview 24 h later (M = 1.25,
SD = 1.80), F (1, 38) = 16.45, p < .01. The magnitude of reminiscence was 6.32 new
correct details in the second interview. The cumulative recall in the amount of cor-
rect details was greater in magnitude than the cumulative recall of errors, which
amounted to less than 1 error on average; fully 92% of the new information reported
in the second interview was correct. It seems that reminiscence and hypermnesia can
occur after short delays due to repeated interviewing, as suggested by both labora-
tory and applied studies (e.g., Bluck et al., 1999; Erdelyi, 1996; Howe et al., 1992).
The current study extends the findings of previous research to childrenÕs recall of
an experienced event.
There was no effect of interview instructions in any analyses. Regardless of
whether the children knew that they would be asked again about what they could
remember, reminiscence and hypermnesia occurred. Thus, an explanation that rem-
iniscence and hypermnesia depend on the participants deliberately trying to remem-
ber relevant information between interviews can be tentatively set aside. However, it
remains possible that because the children were recruited from the same school,
those who were not forewarned of the second interview may have nonetheless
 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254 243

suspected that they would be reinterviewed through conversations with their class-
mates (e.g., see Principe & Ceci, 2002).

Experiment 2

Experiment 2 examined hypermnesia in childrenÕs reports of an event after a delay

of 6 months. We followed a procedure similar to that of Bluck and colleagues (1999),
who found hypermnesia after an 8-month delay in adultsÕ free recall of a realistic
event when there were multiple interviews over a short time period. In the current
experiment, children who had previously participated in Experiment 1 were followed
up 6 months later and participated in three interviews separated by 5-min intervals.
The interview protocol also differed from that used in the previous experiment; to be
consistent with Bluck and colleagues, we used only a single free recall instruction,
omitting the four open-ended cues referring to different aspects of the event. During
the 5-min intervals, the children either drew a picture of what they could remember
about their visit to the pirate (event-related drawing interval) or drew a picture about
an unrelated activity (unrelated drawing interval).
The drawing manipulation was predicted to enhance both reminiscence and
hypermnesia. Instructing participants to focus their thinking on the to-be-remem-
bered material between recall attempts has been found to increase the amount of
hypermnesia in laboratory studies with adults (Erdelyi & Becker, 1974) and has been
used in applied studies to facilitate hypermnesia (Bluck et al., 1999; Bornstein et al.,
1998). Laboratory research has also shown that there is greater hypermnesia for the
recall of pictures than for the recall of words (Erdelyi & Becker, 1974) and that recall
of words that are visualized produce hypermnesia (Erdelyi et al., 1976). Further-
more, with respect to childrenÕs recall, drawing has been shown to benefit recall by
serving as a unique and individual retrieval cue (Butler, Gross, & Hayne, 1995; Gross
& Hayne, 1999). We anticipated that asking the children to draw a picture of what
they could remember about their pirate visit would help them to both think about and
visualize information that they could remember from the event—information that
would then be available for reporting in a subsequent interview. In the unrelated
drawing condition, drawing served as a distracter that prevented the children from
thinking about and visualizing what had happened between recall attempts.

Method

Participants

In Experiment 2, 35 children (19 boys and 16 girls) of the 40 children who orig-
inally participated in Experiment 1 were available for testing and were randomly as-
signed into the two drawing conditions (event-related drawing and unrelated
drawing) with the constraints that there were approximately equal numbers of boys
and girls in each drawing condition and that there were equal numbers of children
from each condition in Experiment 1 in each of the drawing conditions in Experi-
244 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254

ment 2. The mean age of the participants available at follow-up was 6 years 1 month
ðSD ¼ 412 monthsÞ at the time of the event. Caregivers gave written consent for their
children to participate, and all children assented to participate in the interviews when
they were called on to do so.

Materials and procedure

The children were interviewed about the event they had experienced as part of
Experiment 1. After a 6-month delay, the children were interviewed three times
about what they could remember about their visit to the pirate. Each interview con-
sisted of a single open-ended request for the children to recall all that they could
(e.g., Bluck et al., 1999). Although the children did not receive the four open-ended
cues as they had in Experiment 1, general encouragers such as ‘‘Tell me some more
things that happened’’ were used. Each interview ended when the child stopped
responding. There was a 5-min interval between each interview. There were two
interviewers, and each interviewer interviewed approximately equal numbers of boys
and girls in each condition.
After the first interview, the children in the event-related drawing condition were
given 5 min to draw a picture about what they could remember from when they had
visited the pirate. After the second interview, the children were again given 5 min to
draw another picture of what they could remember about the friendly pirate. The
children who participated in the unrelated drawing condition drew pictures about
what they did during their holidays. During the first 5-min interval, they drew a pic-
ture about what they did during their Christmas holiday. During the second 5-min
interval, they drew a picture about what they did during their midyear school holi-
day. The children drew their pictures on white sheets of paper (210 by 297 mm) with
crayons. While the children were drawing, the interviewer sat at a nearby table
attending to ‘‘paperwork’’ that he or she needed to do. If a child spoke, the inter-
viewer told him or her to continue drawing and that he or she (the interviewer)
would be ready to continue shortly. The interviews were taped, transcribed, and
coded as in Experiment 1. Two independent raters coded one third of the transcripts,
and interrater reliability was 86.4%. One of the raters was aware of the conditions
the children were in and one was not.

Results and discussion

Preliminary one-way ANOVAs performed on the number of correct details and

errors in the first, second, and third interviews revealed that there was no effect of
interviewer. A second set of analyses showed that there were no unpredicted effects
of the manipulation in Experiment 1 (forewarned or not forewarned) on the first,
second, and third interviews for correct information or for errors. Only significant
results that exceed an alpha of .05 are reported in what follows.
To examine hypermnesia (an increase in the number of details recalled across
interviews), a mixed-model ANOVA was performed with the number of details re-
ported (correct details or errors) and interview (first, second, or third) as within-sub-
 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254 245

jects factors and with drawing condition (event-related drawing or unrelated draw-
ing) as a between-subjects factor. The only significant effect was that children re-
ported more correct details than errors, F (1, 66) = 90.93, p < .001 (Table 2).
To examine reminiscence (the cumulative recall of new details across the three
interviews), a mixed-model ANOVA was performed with the cumulative recall of de-
tails reported (correct details or errors) and interview (first, second, or third) as with-
in-subjects factors and with drawing condition (event-related drawing or unrelated
drawing) as a between-subjects factor. Children reported more correct details than
errors, F (1, 66) = 86.16, p < .001, and cumulative recall of details differed as a func-
tion of interview, F (1, 66) = 80.74, p < .001. The interaction between these factors
was also significant, F (1, 66) = 10.96, p < .001 (Table 3). Further analysis revealed
that the cumulative recall increased monotonically across the first interview
(M = 8.11, SD = 5.31), second interview (M = 11.40, SD = 5.80), and third interview
(M = 12.89, SD = 5.97) for correct details, F (1, 66) = 61.29, p < .001, and across the

Table 2
Mean numbers of details reported at the 6-month follow-up in Experiment 2
Condition Interview 1 Interview 2 Interview 3
Correct
Pirate drawing 7.81 (6.46) 8.94 (4.09) 7.44 (4.60)
Unrelated drawing 8.37 (4.28) 8.47 (5.54) 9.05 (5.04)
Overall 8.11 (5.31) 8.69 (4.28) 8.31 (4.84)
Errors
Pirate drawing 0.88 (1.20) 1.63 (2.47) 2.25 (2.59)
Unrelated drawing 0.37 (0.60) 1.05 (1.13) 1.53 (1.39)
Overall 0.60 (0.95) 1.31 (1.86) 1.86 (2.03)
Note. Standard deviations are in parentheses.

Table 3
Mean cumulative recall of details reported at the 6-month follow-up in Experiment 2
Condition Interview 1 Cumulative recall at Cumulative recall at
Interview 2a Interview 3b
Cumulative correct
Pirate drawing 7.81 (6.46) 11.75 (5.85) 12.75 (5.83)
Unrelated drawing 8.37 (4.28) 11.11 (5.90) 13.00 (6.25)
Overall 8.11 (5.31) 11.40 (5.80) 12.89 (5.97)
Cumulative errors
Pirate drawing 0.88 (1.20) 2.19 (2.99) 3.87 (4.78)
Unrelated drawing 0.37 (0.60) 1.21 (1.03) 2.16 (1.80)
Overall 0.60 (0.95) 1.66 (2.18) 2.94 (3.54)
Note. Standard deviations are in parentheses.
a
Cumulative recall at Interview 2 is the number of details from Interview 1 plus new details from
Interview 2.
b
Cumulative recall at Interview 3 is the number of details from Interview 1 plus new details from
Interviews 2 and 3.
246 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254

first interview (M = 0.60, SD = 0.95), second interview (M = 1.66, SD = 2.18), and

third interview (M = 2.94, SD = 3.54) for errors, F (1, 66) = 20.33, p < .001. The in-
crease in the cumulative recall of correct details was greater than that for errors.
For correct details, the increase amounted to an extra 4.78 pieces of information
and was twice that of the cumulative recall of errors (2.34). Of the new information
reported by the third interview, 67% was correct.
There was no evidence of hypermnesia within the three testing periods at the 6-
month follow-up, and there was no effect of the interpolated event-related drawing
activity. These results are inconsistent with the findings of Bluck and colleagues
(1999), who found hypermnesia after a delay of 8 months in adults. Clearly, from
the analysis of the cumulative recall, new details were added to childrenÕs accounts
in the subsequent interviews, although the amount of reminiscence was less than that
observed in Experiment 1 due to forgetting. A practical consideration is that the ab-
sence of a hypermnesia effect might be because children did not repeat enough infor-
mation from the previous interviews. It is possible that the children thought that
because they had already provided details a few minutes earlier, there was little need
to repeat them and so instead they focused on recalling new details. In addition, chil-
dren were not prompted in any recall session. Experiment 2 used only free recall
instructions, whereas Experiment 1 (in which hypermnesia was found) used four
open-ended prompts after free recall to further elicit more complete narratives. It
is possible that additional prompting facilitated the reporting of previously recalled
details in additional interviews. This possibility was investigated in Experiment 3.

Experiment 3

Experiment 3 examined forgetting, reminiscence, and hypermnesia in childrenÕs

eyewitness memory after a 6-month delay, following the interview protocol used
in Experiment 1, in which hypermnesia was observed. Unfortunately, it was not pos-
sible to compare recall at 6 months in Experiment 2 with that soon after the event in
Experiment 1 due to the different interviews used and the different intervals between
the interviews. Therefore, in Experiment 3, all of the interviews were identical. The
participants received a baseline interview immediately after their participation and
were interviewed twice at the 6-month delay, with 24 h separating these two inter-
views. This design allowed us to examine forgetting and reminiscence between the
immediate baseline interview and the 6-month initial interview and to examine rem-
iniscence and hypermnesia across the 6-month initial and repeated interviews.

Method

Participants

The participants were 21 children (10 boys and 11 girls) recruited from local pri-
mary schools. The mean age of the children at the time of the event was 6 years
212 months (SD = 5 months). Caregivers gave written consent for their childrenÕs
 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254 247

participation, and all children assented to participate in the interviews when they
were called on to do so.

Materials and procedure

The materials and the ‘‘visiting the pirate’’ event were identical to those used in
Experiments 1 and 2. The children were interviewed immediately after their partici-
pation in the event, and two further interviews were conducted after a 6-month de-
lay. At the 6-month delay, the interviews were 24 h apart. The interviews were
identical in format to those in Experiment 1. At the 6-month delay, the children were
forewarned that the interviewer would be returning and that they would be reinter-
viewed. All interviews were audiotaped and then transcribed.
Two interviewers conducted the immediate interviews. At the 6-month delay, one
of the interviewers who had conducted the immediate interviews and one new inter-
viewer were used. In the 6-month delay interviews, the interviewer who had previ-
ously interviewed children in the immediate interviews now interviewed children
who he or she had not interviewed previously. Two independent raters coded one
third of the transcripts, and interrater agreement was 88.3%.

Results and discussion

Preliminary analyses showed that there was no effect of interviewer on the amount
of correct information recalled or on the number of errors.

Forgetting and reminiscence across the 6-month delay

To establish whether significant forgetting had occurred and whether the effect of
delay was consistent with previous research (e.g., Jones & Pipe, 2002), a within-sub-
jects ANOVA was performed with the number of details reported (correct details or
errors) and interview (immediate interview or 6-month initial interview) as factors.
The analysis showed that children reported more correct details than errors,
F (1, 20) = 67.63, p < .001, and that the total number of details reported overall de-
creased across the 6-month delay, F (1, 20) = 13.70, p < .001. There was also an inter-
action between the number of details reported and the interview, F (1, 20) = 43.92,
p < .001. Two further analyses confirmed that the interaction could be characterized
as a decrease in total amount of correct information between the immediate inter-
view (M = 15.66, SD = 6.68) and the 6-month initial interview (M = 9.72,
SD = 6.94), F (1, 20) = 33.75, p < .01, and as an increase in errors between the imme-
diate interview (M = 0.57, SD = 0.68) and the 6-month initial interview (M = 2.00,
SD = 2.70), F (1, 20) = 6.44, p < .05 (Table 4).
To examine reminiscence across the 6-month delay (the cumulative recall of new
details across the immediate interview and the 6-month initial interview), a within-
subjects ANOVA was performed with the number of details (correct details or er-
rors) and interview (immediate interview or 6-month initial interview) as factors.
This analysis showed that children reported more correct information than errors,
248 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254

Table 4
Mean numbers of details reported in the immediate interview and the 6-month initial interview in
Experiment 3
Condition Immediate interview 6-month initial interview Cumulative recall at the
6-month initial interviewa
Correct 15.66 (6.68) 9.72 (6.94) 18.14 (7.22)
Errors 0.57 (0.68) 2.00 (2.70) 2.52 (2.99)
Note. Standard deviations are in parentheses.
a
Cumulative recall at the 6-month initial interview is the number of details from the immediate
interview plus new details from the 6-month initial interview.

F (1, 20) = 35.03, p < .001, and that the cumulative recall of details increased across
the 6-month delay, F (1, 20) = 105.36, p < .001. However, there was no interaction
between these factors indicating a similar increase in the cumulative recall of new er-
rors and correct information (Table 4) across the 6-month delay. The magnitude of
the increase was 2.48 correct details and 1.95 errors; just over half (56%) of the new
information reported after the 6-month delay was correct.

Reminiscence and hypermnesia across the 6-month interviews

To examine hypermnesia after a 6-month delay (an increase in the number of de-
tails recalled in the 6-month repeated interviews), a within-subjects ANOVA was
performed with the number of details (correct details or errors) and 6-month inter-
view (6-month initial interview or 6-month repeated interview) as factors. The anal-
ysis showed only that children reported more correct details than errors,
F (1, 20) = 31.22, p < .001 (Table 5).
To examine reminiscence (the cumulative recall of new details across the 6-month
initial and 6-month repeated interviews), a within-subjects ANOVA with the number
of details (correct details or errors) and 6-month interview (6-month initial interview
or 6-month repeated interview) as factors showed that children reported more cor-
rect details than errors, F (1, 20) = 28.57, p < .001, and that there was an increase
in the cumulative recall of details across the interviews, F (1, 20) = 30.86, p < .001.
The interaction between these variables was also significant, F (1, 20) = 5.98,
p < .05 (Table 5). Two further analyses showed that the cumulative recall of correct

Table 5
Mean numbers of details reported in the 6-month initial and 6-month repeated interview in Experiment 3
Condition 6-month initial interview 6-month repeated interview Cumulative recall at the
6-month repeated interviewa
Correct 9.72 (6.94) 10.57 (6.73) 12.62 (7.05)
Errors 2.00 (2.70) 1.81 (1.72) 3.14 (3.52)
Note. Standard deviations are in parentheses.
a
Cumulative recall at the 6-month repeated interview is the number of details from the 6-month initial
interview plus new details from the 6-month repeated interview.
 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254 249

information increased between the 6-month initial interview (M = 9.72, SD = 6.94)

and the 6-month repeated interview (M = 12.62, SD = 7.05), F (1, 20) = 19.07,
p < .001, and that the cumulative recall of errors increased between the 6-month ini-
tial interview (M = 2.00, SD = 2.70) and the 6-month repeated interview (M = 3.14,
SD = 3.52), F (1, 20) = 15.87, p < .001. The magnitude of the cumulative recall of cor-
rect information was 2.90 pieces of information, which was twice that of the increase
in errors (1.14) across the same interviews; nearly three quarters (72%) of the new
information provided in the 6-month repeated interview was correct.

General discussion

In these experiments, we examined reminiscence and hypermnesia in 5- and 6-

year-oldsÕ verbal recall of an event when repeated interviews occurred soon after
the event or following a long delay. Reminiscence (the recall of new information)
proved to be reliable across all three experiments. In each experiment, open-ended
recall instructions were sufficient to elicit the new information. However, repeated
interviewing did not consistently produce hypermnesia. Hypermnesia occurred only
in Experiment 1, with the total amount of correct information recalled increasing by
13% across the two interviews conducted immediately and 24 h after the event,
respectively. In Experiments 2 and 3, when the repeated interviews took place follow-
ing a 6-month delay, hypermnesia was not observed. The total number of errors that
was reported within an interview did not increase across repeated interviews sepa-
rated by 5 min or 24 h in any of the experiments, whereas the cumulative recall of
new errors did increase.
Although there was forgetting, it is interesting that from the perspective of eyewit-
ness memory, a second interview at the 6-month delay elicited new information. New
details were added regardless of whether the interviews were separated by a 5-min
interval (following the procedure used by Bluck et al., 1999), a 24-h delay, or a 6-
month delay. Previous research has shown that children also report new information
when there are long delays between repeated interviews, for example, the first inter-
view conducted soon after the event and the second interview conducted 6 months
after the event (e.g., Pipe et al., 1999; Salmon & Pipe, 1997, 2000). Together with
these past studies, the current findings suggest that reminiscence is a reliable and ro-
bust phenomenon in childrenÕs eyewitness recall of past events over both short and
long delays. However, delay duration does exert an influence. The number of new
details reported when the interviews occurred after a 6-month delay was fewer than
that reported when the interviews occurred soon after the event.
An inconsistency of our findings compared with those in applied adult hypermne-
sia research is that childrenÕs reports included increasing numbers of new errors
across interviews in all three experiments. In Experiment 1, when the interviews oc-
curred immediately and 24 h after the event, the magnitude of the increase was less
than a single error and accounted for approximately 8% of all the new information.
Compared with the much larger increase in the recall of correct details, this relatively
minor increase in errors would not appear to compromise the accuracy of new
250 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254

information elicited. However, at the 6-month delay, when interviews were con-
ducted 5 min or 24 h apart, errors represented approximately 30% of all the new
information recalled. In Experiment 3, errors represented 44% of the new informa-
tion reported between the immediate baseline interview and the 6-month initial inter-
view. However, the higher percentage of errors should be considered in the context
of an overall decrease in the amount of new information reported after the 6-month
delay.
The finding of hypermnesia in Experiment 1, coupled with the greatest amount of
reminiscence, is consistent with the recommendation that eyewitness interviews with
young children should be conducted as soon as possible before too much forgetting
has occurred (Jones & Pipe, 2002; Pipe et al., 1999; Pipe & Wilson, 1994; Salmon &
Pipe, 2000) and suggests additional advantages. After short delays between experi-
encing an event and recalling it, childrenÕs accounts are generally more accurate
and contain few errors when open-ended cues are used. The current research suggests
that if children receive multiple interviews while their memories are ‘‘fresh’’ immedi-
ately after the event of interest, hypermnesia occurs without an increase in the total
amount of errors across interviews. The hypermnesia finding is consistent with that
of Dent and Stephenson (1979) and extends their findings to encompass increases in
recall for younger children used in our study. The improvement in recall across re-
peated interviewing in Experiment 1 also did not depend on whether or not the chil-
dren had been forewarned that they would be reinterviewed. In a real-life situation,
interviewers might not want to set up the expectation that there will be a second
interview before the first one has begun. In particular, interviewers might not know
whether a second interview will be required until the first one has been completed.
We had expected that, based on previous research, repeated interviewing would
result in both reminiscence and hypermnesia in all three experiments. The trace
integrity theory (Brainerd, Reyna, Howe, & Kingma, 1990; Howe et al., 1992), how-
ever, does provide a partial explanation for why we observed a reduction in the
amount of reminiscence over time and, hence, the absence of hypermnesia. Brainerd
and his collaborators (1990) argued that reminiscence and hypermnesia can occur
due to retrieval relearning (a process in which recall cues become increasingly effec-
tive) or as a result of reintegration of the memory trace (a process in which the mem-
ory is restored). In the current study, when children were interviewed immediately
after the event, little forgetting had occurred and retrieval relearning might have
facilitated reminiscence and hypermnesia. Because no forgetting had taken place, re-
peated retrieval attempts during the interviews were highly effective at accessing
additional information to the point where increasing amounts of information could
be recalled. In contrast, by 6 months, the children had forgotten a significant amount
of what they had originally encoded about the event. Before correct details can be
retrieved, the memory trace must be reintegrated from related information, and this
may be a more effortful and time-consuming mental process. However, even if it
were more difficult to reminisce new information after forgetting has occurred, the
trace integrity theory does not help us to understand why we did not obtain hyperm-
nesia. What remains to be explained is why, after a 6-month delay, there appears to
be trial-to-trial forgetting of correct information that is similar in magnitude to the
 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254 251

reminiscence of new information, with the end result being neither an increase nor a
decrease of correct details reported across repeated interviews.
The event used in this study, although realistic and enjoyable for the children, did
not carry with it the personal significance of events that are known to be well remem-
bered, and it is possible that hypermnesia might be observed and reminiscence might
be enhanced when the event is of personal significance. For example, Peterson and
her colleagues examined memory for a personal injury and subsequent hospital treat-
ment of children between 2 and 13 years of age in interviews consisting largely of free
and cued recall. After an initial interview, interviews were repeated at delays of 1
week, 6 months, 1 year, 2 years (Peterson, 1999; Peterson & Bell, 1996), and 5 years
(Peterson & Whalen, 2001). Peterson and Bell (1996) found that the percentage of
relevant information reported by the children decreased up to the 6-month interview.
At a 2-year delay, Peterson (1999) observed an effect of delay only on details about
the visit to the hospital. When these children were followed up and given a final
interview 5 years later, effects of delay were once again found. At this long 5-year
delay, the memory decrements were now confined to recall of peripheral information
about the injury and memory for the visit to the hospital (Peterson & Whalen, 2001).
Peterson also reported that new information added to the childrenÕs accounts about
the salient aspects of the events was generally accurate. Thus, at the longest delays (2
years and 5 years), memory for the injury was maintained; therefore, there may have
been new details that could be reported across multiple interviews.
Fivush, McDermott Sales, Goldberg, Bahrick, and Parker (2004) interviewed 3-
and 4-year-olds about what they could remember about Hurricane Andrew between
2 and 6 months after the hurricane and again 6 years later. The interviews consisted
of open-ended requests for children to tell all they could remember about the stress-
ful event. The results showed that across the 6-year delay, the amount of information
that the children reported doubled and that what was reported in the 6-year inter-
view contained very few details that had been reported earlier. Fivush and colleagues
suggested that the children reconstructed what they had remembered about the
events over the course of time and that the stories they told changed in accordance
with what was relevant to their own lives. It would have been interesting to see what
would have happened to the level of reminiscence if the children had been inter-
viewed 2 days in a row at the 6-year delay. Given that they were able to report so
much new information about the event, a second interview may have resulted in
hypermnesia.
Repeated interviewing may also be a sufficient way of maintaining memory
across long delays. Pipe and colleagues (2004) followed up children who were orig-
inally interviewed immediately or after a delay of 1 day, 1 week, 1 month, and 6
months as part of an earlier study. Pipe and colleagues reinterviewed these children
1 and 2 years after the event and included a control group whose members were
interviewed only at the 1- and 2-year delays. The results suggested benefits to
long-term recall when an intervening interview occurred at the 6-month delay
rather than shortly after the event. It appeared that the single interview at the
6-month delay attenuated further forgetting. In Pipe and colleaguesÕ study, children
received only a single interview at each delay, and it is unknown whether repeated
252 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254

interviewing across short delays would have been of even greater benefit for long-
term memory.
Clearly, there is a need for future research to further explore the effects of repeat-
ing open-ended interviews on childrenÕs reports of past events. Relatively few studies
have explored the effects of repeating open-ended interviews as a means of enhancing
accurate recall in children. Our study has shown that there are advantages to repeat-
ing interviews with young children, with the greatest benefits occurring when the
interviews take place soon after the event (Experiment 1). However, although our
interview protocol was open-ended, it remains to be discovered whether free recall
instructions alone would have been enough to obtain hypermnesia. When hypermne-
sia was obtained, the children also received four open-ended cues targeting specific
aspects of the event, and it is possible that it was these cues that facilitated hyperm-
nesia. Therefore, a short delay might be necessary, but whether it is sufficient remains
to be determined. Moreover, children are not always interviewed soon after alleged
events have occurred, and our results also show that new errors are likely to be re-
ported after a delay of 6 months. It is also important to be mindful that the exper-
iments reported here examined childrenÕs memory for an event under conditions that
facilitate accurate recall. When children are interviewed in real-world contexts, they
are not necessarily asked exclusively open-ended questions, and the effects of repeat-
ing specific and leading questions may differ from those observed here. What is clear
is that children have more to tell than simply what they report in a single interview
and that there may be advantages to repeating open-ended interviews designed to eli-
cit accurate recall.

Acknowledgments

The authors are grateful for the assistance of Linda Barclay, Deirdre Brown, Sa-
sha Farry, Keith Garraway, and Natasha Pomeroy for interviewing, coding, data en-
try, pirate acting, and logistics. Our sincere thanks go to the parents, teachers, and
children of the Dunedin primary schools who donated their time and made this study
possible.

References
Ammons, H., & Irion, A. L. (1954). A note on the Ballard reminiscence phenomenon. Journal of
Experimental Psychology, 48, 184–186.
Baker-Ward, L., Gordon, B. N., Ornstein, P. A., Larus, D., & Clubb, P. (1993). Young childrenÕs long-
term retention of a pediatric examination. Child Development, 56, 1103–1119.
Ballard, P. B. (1913). Oblivescence and reminiscence. British Journal of Psychology, 1, 1–82.
Bluck, S., Levine, L. J., & Laulhere, T. M. (1999). Autobiographical remembering and hypermnesia: A
comparison of older and younger adults. Psychology and Aging, 14, 671–682.
Bornstein, B. H., Liebel, L. M., & Scarberry, N. C. (1998). Repeated testing in eyewitness memory: A
means to improve recall of a negative emotional event. Applied Cognitive Psychology, 12, 119–131.
Brainerd, C. J., Reyna, V. F., Howe, M. L., & Kingma, J. (1990). The development of forgetting and
reminiscence. Monographs of the Society for Research in Child Development, 55, Serial No. 222.
 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254 253

Butler, S., Gross, J., & Hayne, H. (1995). The effect of drawing on memory performance in young children.
Developmental Psychology, 31, 597–608.
Dent, H. R., & Stephenson, G. M. (1979). An experimental study of the effectiveness of different
techniques of questioning child witnesses. British Journal of Social and Clinical Psychology, 18,
41–51.
Dunning, D., & Stern, L. B. (1992). Examining the generality of eyewitness hypermnesia: A close look at
time delay and question type. Applied Cognitive Psychology, 6, 643–657.
Erdelyi, M. H. (1996). The recovery of unconscious memories: Hypermnesia and reminiscence. Chicago:
University of Chicago Press.
Erdelyi, M. H., & Becker, J. (1974). Hypermnesia for pictures: Incremental memory for pictures but not
words in multiple recall trials. Cognitive Psychology, 6, 159–171.
Erdelyi, M. H., Finkelstein, S., Herrell, N., Miller, B., & Thomas, J. (1976). Coding modality vs. input
modality in hypermnesia: Is a rose a rose a rose? Cognition, 4, 311–319.
Fivush, R. (1994). Young childrenÕs event recall: Are memories constructed through discourse?.
Consciousness and Cognition, 3, 356–373.
Fivush, R., & Hamond, N. R. (1989). Time and again: The effects of repetition and retention interval on 2
year oldsÕ recall. Journal of Experimental Child Psychology, 47, 259–273.
Fivush, R., McDermott Sales, J., Goldberg, A., Bahrick, L., & Parker, J. (2004). Weathering the storm:
ChildrenÕs long-term recall of Hurricane Andrew. Memory, 12, 104–118.
Gross, J., & Hayne, H. (1999). Drawing facilitates childrenÕs verbal reports after long delays. Journal of
Experimental Psychology: Applied, 14, 163–179.
Henkel, L. A. (2004). Erroneous memories arising from repeated attempts to remember. Journal of
Memory and Language, 50, 26–46.
Henry, L. A., & Gudjonsson, G. H. (2003). Eyewitness memory, suggestibility, and repeated recall sessions
in children with mild and moderate intellectual disabilities. Law and Human Behavior, 27, 481–505.
Howe, M. L., Kelland, A., Bryant-Brown, L., & Clark, S. L. (1992). Measuring the development of
childrenÕs amnesia and hypermnesia. In M. L. Howe, C. J. Brainerd, & V. F. Reyna (Eds.),
Development of long-term retention (pp. 56–102). New York: Springer-Verlag.
Hudson, J. A., & Fivush, R. (1991). As times go by: Sixth grade children recall a kindergarten experience.
Applied Cognitive Psychology, 5, 347–360.
Jones, C. H., & Pipe, M-E. (2002). How quickly do children forget events? A systematic study of childrenÕs
event reports as a function of delay. Applied Cognitive Psychology, 16, 755–768.
Kelley, M. R., & Nairne, J. S. (2003). Remembering the forgotten? Reminiscence, hypermnesia, and
memory for order. Quarterly Journal of Experimental Psychology, 56, 577–599.
Kern, R. P., Libkuman, T. M., & Otani, H. (2002). Memory for negatively arousing and neutral pictorial
stimuli using a repeated testing paradigm. Cognition and Emotion, 16, 749–767.
Murachver, T., Pipe, M.-E., Gordon, R., Owens, J. L., & Fivush, R. (1996). Do, show, and tell: ChildrenÕs
event memories acquired through direct experience, observation, and stories. Child Development, 67,
3029–3044.
Paris, S. G. (1978). Memory organization during childrenÕs repeated recall. Developmental Psychology, 14,
99–106.
Peterson, C. (1999). ChildrenÕs memory for medical emergencies: 2 years later. Developmental Psychology,
35, 1493–1506.
Peterson, C., & Bell, M. (1996). ChildrenÕs memory for traumatic injury. Child Development, 67,
3045–3070.
Peterson, C., Moores, L., & White, G. (2001). Recounting the same events again and again: ChildrenÕs
consistency across multiple interviews. Applied Cognitive Psychology, 15, 353–371.
Peterson, C., & Whalen, N. (2001). Five years later: ChildrenÕs memory for medical emergencies. Applied
Cognitive Psychology, 15, S7–S24.
Pipe, M. E., Gee, S., Wilson, J. C., & Egerton, J. M. (1999). ChildrenÕs recall 1 or 2 years after an event.
Developmental Psychology, 35, 781–789.
Pipe, M. E., Sutherland, R., Webster, N., Jones, C. H., & La Rooy, D. (2004). Do early interviews affect
childrenÕs long-term recall?. Applied Cognitive Psychology, 18, 823–839.
254 D. La Rooy et al. / Journal of Experimental Child Psychology 90 (2005) 235–254

Pipe, M-E., & Wilson, J. C. (1994). Cues and secrets: Influences on childrenÕs event reports. Developmental
Psychology, 30, 515–525.
Principe, G. F., & Ceci, S. J. (2002). ‘‘I saw it with my own ears’’: The effects of peer conversations on
preschoolersÕ reports of nonexperienced events. Journal of Experimental Child Psychology, 83, 1–25.
Roediger, H. L., Payne, D. G., Gillespie, G. L., & Lean, D. S. (1982). Hypermnesia as determined by level
of recall. Journal of Verbal Learning and Verbal Behavior, 21, 635–655.
Salmon, K., & Pipe, M. E. (1997). Props and childrenÕs event reports: The impact of a 1-year delay.
Journal of Experimental Child Psychology, 65, 261–292.
Salmon, K., & Pipe, M-E. (2000). Recalling an event one year later: The impact of props, drawing, and a
prior interview. Applied Cognitive Psychology, 14, 99–120.
Scrivner, E., & Safer, M. A. (1988). Eyewitnesses show hypermnesia for details about a violent event.
Journal of Applied Psychology, 73, 371–377.
Shapiro, S. R., & Erdelyi, M. H. (1974). Hypermnesia for pictures but not words. Journal of Experimental
Psychology, 103, 1218–1219.
Shaw, J. S., Bjork, R. A., & Handal, A. (1995). Retrieval-induced forgetting in an eyewitness-memory
paradigm. Psychonomic Bulletin & Review, 2, 249–253.
Steward, M. S., Steward, D. S., Farquhar, L., Myers, J. E. B., Reinhart, M., Welker, J., et al. (1996).
Interviewing young children about body touch and handling. Monographs of the Society for Research
in Child Development, 57, Serial No. 248.
Tinsley, H. E. A., & Weiss, D. J. (2000). Interrater reliability and agreement. In H. E. A. Tinsley & S. D.
Brown (Eds.), Handbook of applied multivariate statistics and mathematical modeling (pp. 95–124). San
Diego: Academic Press.
Turtle, J. W., & Yuille, J. C. (1994). Lost but not forgotten details: Repeated eyewitness recall leads to
reminiscence but not hypermnesia. Journal of Applied Psychology, 79, 260–271.