190
Handbook of Psychology of Investigative Interviewing
video recording of the moment of the crash on television. Whilst it has never
been established whether such a fi lm exists, it has certainly never been shown
on television. Nevertheless, 44% of Ost et al. ’ s participants claimed to have
seen it. Furthermore, many participants even went as far as to say on which
television channel they had seen it. Ost et al. found that participants who
scored higher on a measure of ‘ eagerness to please ’ (the self - monitoring scale;
Snyder, 1974 ) were more like to claim to have seen the fi lm than participants
who scored lower on this measure.
In a replication of the Crombag et al. study, Granhag, Str ö mwall & Billings
(2003) asked participants whether they had seen a (nonexistent) fi lm of the
sinking of the Estonia ferry. Again they found that 52% of their participants
claimed to have seen the fi lm. Granhag et al. (2003) were interested in whether
these false reports were susceptible to social infl uence. Thus, in a novel twist,
they had a confederate present when the participants completed the question-
naire. This confederate either claimed to have seen the nonexistent fi lm or
claimed not to have seen it. Granhag
et al. found that participants either
increased, or decreased, their levels of false reporting in line with the social
infl uence exerted by the confederate. When the confederate claimed to have
seen the fi lm, the number of participants also claiming to have seen it increased,
and vice versa (see Ost, Hogbin & Granhag, 2006 , for a replication).
The ‘ crashing memory ’ effect appears to be robust, with between 36% and
66% of participants in any given study claiming to have seen the nonexistent
video footage (Jelicic, Smeets, Peters, Candel & Merckelbach, 2006 ; Wilson
& French, 2006 ; but see Jelicic, Smeets, Candel, van Suijdam & Merckelbach,
2006 ). These rates, however, can be reduced by phrasing the misleading ques-
tion in an unambiguous rather than ambiguous manner (i.e., asking about ‘ a ’
fi lm, rather than ‘ the ’ fi lm; Smeets et al. , 2006 ).
Parental m isinformation m ethod
The studies mentioned above show that some individuals will, when misled
by subtle changes in wording, or by a confederate, come to report that they
remember events, or details of events, that they did not witness. However, it
is diffi cult to generalize these fi ndings to cases of allegedly false or recovered
memories of childhood events. The events in the Crombag et al. , Ost et al.
and Granhag et al. studies were all relatively recent and had occurred whilst
the participants in their studies were adults. In contrast, most delayed claims
of childhood abuse concern events that, by defi nition, occurred many years
beforehand (see Pendergrast, 1996 ). Can individuals be misled to report false
events from their childhood?
Loftus & Coan (cited in Loftus & Pickrell, 1995 ) describe a study in which
a 14 - year - old boy, Chris, was asked to recall details over fi ve days regarding
four events involving family members. One of the events was false and other
three were true (as verifi ed by the family). Chris was interviewed in the pres-
ence of a sibling (a confederate of the investigators) about these events. The
Recovered Memories
191
sibling provided verbal corroboration that all the events (including the false
event) had taken place. Over time Chris began to report more about the four
events, even rating the false event (becoming lost in a shopping mall as a child)
as more likely to have occurred than all but one of the three true events.
Loftus & Pickrell (1995) replicated this effect with a larger sample of under-
graduate students. In this study participants were asked to complete a booklet
concerning four events, the third of which was false (becoming lost in a shop-
ping mall). Parents who confi rmed that their child had never become lost in
a shopping mall as a child also provided details of the three real events.
Participants were interviewed three times over three weeks and also asked,
between interviews, to write down in their booklets anything that came to
mind about the events. Loftus & Pickrell (1995) found that after three weeks
six out of 24 participants (25%) erroneously believed part or all of the false
event. However, as Pezdek, Finger & Hodge (1997) argue, becoming lost in
a shopping mall is a fairly common occurrence for which many people would
have a ‘ script ’ (or ‘ schema ’ ). Pezdek et al. contend that it would be relatively
easy for an individual to create a convincing report of an event like this,
although as Loftus ( 1997b : 180) argues, the important point is not that par-
ticipants might be able to construct a general narrative of such an event, but
that they report specifi c details suggested by the experimenter. In these studies
participants ‘ were not asked about ANY experience of being lost. They were
asked to remember being lost around the age of fi ve, in a particular location,
with particular people present, being frightened, and ultimately being rescued
by an elderly person ’ ( ibid. ) In fact, Pezdek et al. (1997) demonstrated that
an event that was lower in plausibility and for which participants were less
likely to have ‘ script - relevant ’ knowledge (in this case receiving an enema as a
child) was less likely to be implanted (although see Scoboria, Mazzoni, Kirsch
& Relyea, 2004 , for a discussion of plausibility).
Hyman, Husband & Billings (1995) attempted to address the criticisms of
Pezdek et al. (1997) by suggesting to their participants that they had experi-
enced more unusual events. Following the methodology devised by Loftus &
Pickrell (1995) they asked their participants to try to remember three events
(two of which had occurred and one of which had not). Hyman et al. con-
ducted two studies in which they suggested that their participants had expe-
rienced one of the following false events when they were children:
• an overnight hospitalization with a suspected ear infection (study 1);
• a birthday party with a visit by a clown and pizza (study 1)
• spilling a punch bowl at a wedding (study 2);
• evacuating a grocery store when the sprinklers went off (study 2);
• releasing the handbrake of a car in a car park and colliding with something
(study 2).
Their participants were interviewed three times over a three - week period
and after each interview participants were asked to think about the events and
192
Handbook of Psychology of Investigative Interviewing
try to remember more details before the next session. By the third and fi nal
interview 89 – 95% of the ‘ real ’ events were recalled along with 25% of the ‘ false ’
events (see also Hyman & Loftus, 1997 ). Thus, participants can be misled to
report more unusual events from their childhood. However, critics argue that
reporting that you remember spilling a punchbowl at a wedding does not
compare to reporting that you remember being abused as a child – the latter
is a much more negative, traumatic and emotionally - char
ged event. Nevertheless,
one study has examined whether it is possible to mislead participants to report
that they falsely remember negative, traumatic and emotionally - charged events
from their childhood.
Porter
et al. (1999) , using a similar methodology to Loftus
& Pickrell
(1995) and Hyman
et al. (1995) , suggested to participants that they had
experienced serious negative events as children, such as:
• a major medical procedure;
• getting lost;
• being seriously harmed by another child;
• a serious animal attack.
Porter et al. also verbally encouraged their participants to remember the
events ( ‘ most people are able to retrieve lost memories if they try hard enough ’ ,
Porter et al. , 1999 : 522), as well as asking them to think about this for fi ve
minutes every night between interviews. By the third and fi nal interview,
Porter et al. found that 54% of participants reported a ‘ full ’ or a ‘ partial ’ false
memory. In the fi nal interview, Porter et al. (1999) also asked participants to
deliberately fabricate an account of an event that did not occur. All three types
of memory report (real, false and fabricated) were rated using a technique
called the Memory Assessment Procedure (MAP) to investigate possible quali-
tative differences between them. Porter et al. found that real and fabricated
memories were rated as more vivid, more coherent and were given higher
confi dence ratings than false memories. Fabricated memories were also rated
as more stressful and contained more details than both real and false memories
(Porter et al. , 1999 ; see also Pezdek & Talyor, 2000 ; Davies, 2001 ; Heaps &
Nash, 2001 ; Loftus & Bernstein, 2005 ). Taken together, these studies indicate
that, when misled by information provided by siblings or parents, participants
will report that they remember unusual, negative, emotionally - charged and
traumatic childhood events.
In the studies described above, some kind of verbal instructions was used
to encourage participants to try to remember the events (the only exception
to this is Hyman et al. , 1995 , experiment 1). However, the extent to which,
in the case of false reports of childhood events, the behaviour of the interviewer
infl uenced the manner of recall is unclear. For example, Loftus & Pickrell
( 1995 : 722) noted that ‘ the interviewers maintained a pleasant and friendly
manner whilst pressing for details ’ . In order to explore the possible role of
social pressure in the development of false memories Ost, Foster, Costall &
Recovered Memories
193
Bull (2005) followed a similar methodology to these previous studies, but
trained their interviewers to interview in an appropriate and non - pressuring
manner. Participants were asked about both positive and negative false events,
similar to those used in previous studies, for example:
• becoming lost;
• a trip to the hospital;
• a serious accident;
• an eventful birthday party;
• winning a contest.
Levels of social pressure in the interviews were monitored by the participants
(who were asked to rate, amongst other details, how pressured they felt to
remember the event(s)) and by independent judges (who rated videotapes of
the interviews). Overall levels of social pressure reported by both participants
and independent judges were low. Despite the low levels of social pressure,
Ost et al. (2005) found that seven out of 31 participants produced a ‘ full ’ or
‘ partial ’ report of a childhood event that did not occur. This study suggests
that even minimal social pressure and repeated interviewing are suffi cient to
lead some individuals to come to report events from their childhood that never
occurred.
Methodological l imitations of p arental m isinformation s tudies. There are
several criticisms of the above studies that limit their generalizability to
cases of delayed reports of childhood trauma. As already noted, Pezdek et al.
(1997) argue that most of the false events that participants are asked to recall
are events for which participants are likely to have a ‘ script ’ ; an event that is
lower in plausibility and for which participants are less likely to have ‘ script -
relevant ’ knowledge (such as childhood sexual abuse) is less likely to be falsely
reported.
Pezdek et al. (1997) tested this hypothesis by suggesting to Jewish and
Catholic participants that they had taken part in both a Catholic ritual (receiv-
ing Communion) and a Jewish ritual (Shabbot), neither of which had actually
occurred. Pezdek et al. (1997) argue that the plausibility of having taken part
in a Catholic ritual would be low for Jewish participants, that they would have
less script - relevant knowledge to draw on, and vice versa. They found, in line
with their predictions, that seven of the Catholic participants but none of the
Jewish participants reported the false Catholic event and three Jewish partici-
pants and one Catholic participant reported the Jewish false event. This shows
that participants were more likely to remember the plausible false event than
the implausible false event (i.e. Jewish participants were more likely to remem-
ber the false Jewish ritual than the false Catholic ritual). In a second experi-
ment, Pezdek et al. (1997) replicated the study by Loftus & Pickrell (1995)
and extended it by suggesting to participants that they had (a) been lost in a
shopping mall (a plausible event) and (b) that they had received an enema
194
Handbook of Psychology of Investigative Interviewing
as a child (an implausible event). Pezdek et al. (1997) found that whilst three
out of 20 participants falsely reported becoming lost in a shopping mall, none
of the participants falsely reported receiving an enema. These experiments,
therefore, suggest that the probability that participants can be misled to
report false events from their childhood is likely to be a function of the plau-
sibility of, or familiarity with, the event concerned (although see Scoboria
et al. , 2004 , for a discussion of the differences between plausibility and script
consistency).
However, certain ‘ scripts ’ are more familiar in our culture than might at
fi rst be imagined. Lynn & Pezzo ( 1994 ; cited in Lynn & Kirsch, 1996 ) found
that participants were able to construct very convincing narratives of having
been abducted by aliens even when given relatively little warning or further
information. Lynn & Kirsch (1996) argue that narratives of alien abduction
are so common in our culture that individuals are likely to have access to
‘ script - relevant ’ knowledge (see also Arndt & Greenberg, 1996 ). Given the
large number of self - help books, media programmes, news articles and the like
dealing with abuse, it is not inconceivable that some individuals could con-
struct a false narrative of having been abused.
A second, often cited limitation with parental misinformation studies is that
the events participants are misled to report are not of an abusive or traumatic
&
nbsp; nature. However, Porter
et al. (1999) showed that participants could be
misled to report negatively charged and traumatic events, such as being victim
to a serious animal attack. Although untestable in the laboratory, the possibil-
ity remains that participants in situations other than a psychology experiment
could be misled to report abusive events that did not occur (see also Hyman
& Loftus,
1997 ). In fact, evidence from retractors (individuals who have
repudiated their earlier claims of abuse) suggests that this does occur (see de
Rivera, 1998 ; Ost, Costall & Bull, 2001; 2002 ; Ost & Nunkoosing, in press ).
Sometimes they became so convinced of the truth of their false beliefs that
they initiated legal proceedings against their alleged abusers (de Rivera, 1998 ;
Ost et al. , 2001; 2002 ). Worryingly, retractors reported that they experienced
levels of social pressure and inappropriate questioning techniques not dissimi-
lar to those that can lead to false confessions in police interrogations
(Wrightsman & Kassin, 1993 ; Kassin, 1997 ; Ost, et al. , 2001 ; Gudjonsson,
2003 ).
A third limitation with parental misinformation studies is that it is not clear
what degree of social pressure is required in order to lead participants to make
false claims about the past (Ost et al. , 2005 ). For example, as mentioned above,
Loftus & Pickrell ( 1995 : 722) note that their interviewer ‘ maintained a pleas-
ant and friendly manner, whilst pressing for details ’ yet do not provide details
of how participants were ‘ pressed for details ’ . Porter et al. (1999) also state
that they employed a degree of verbal encouragement (
‘ most people can
remember details if they try really hard ’ , p. 522). Malinoski & Lynn (1999)
found that positive verbal encouragement led participants to report earlier (and
more implausible) memories. Ost et al. (2005) found that, even when social
Recovered Memories
195
pressure was kept to a minimum (by appropriately training interviewers not to
pressure participants), a number of participants still reported false events from
their childhood (see also Erdmann, Volbert & B ö hm, 2004 , for similar fi nd-
ings with child witnesses). The role of social pressure in the genesis and devel-
Handbook of Psychology of Investigative Interviewing: Current Developments and Future Directions Page 37
