Académique Documents
Professionnel Documents
Culture Documents
Yogo
210
Hearing the News of the 9/11 Attacks
measurement (e.g., exact location, ongoing new situations and that this mechanism has
activity, type of informant, idiosyncratic adaptive value (e.g., Brown & Kulik, 1977;
details often including perceptual aspects; for Conway, 1995; Finkenauer et al., 1998). The
an example, see Finkenauer et al., 1998). The appraisal of novelty is then followed by a non-
second perspective is adopted in studies valenced emotional state of surprise. The ap-
relying on a test-retest design (e.g., Conway et praisal of importance/consequentiality devel-
al., 1994) and emphasizes consistency between ops almost simultaneously with the appraisal
initial test (usually shortly after encoding) and of novelty and has been proposed as a critical
retest. Conway (e.g., 1995) argues that predictor for memory for the reception context
flashbulb memories should be restricted to this (Brown & Kulik, 1977; Christianson, 1989;
latter definition. According to him, in order to Christianson & Engelberg, 1999; Conway et
be considered a flashbulb memory, a al., 1994; Finkenauer et al., 1998; Rubin &
recollection involves not only a “live” quality Kozin, 1984). The appraisal of impor-
accompanied by recall of minutiae, but also tance/consequentiality then leads to emotional
preserves details of the reception events, and feeling states such as sadness, anger or fear
remains unchanged over long periods of time. (Conway et al., 1995; Finkenauer et al., 1998;
Acceptance of these last two conditions means Oatley & Johnson-Laird, 1987). The intensity
that it is not technically possible to assess of emotional feeling states predicted memory
flashbulb memories shortly after encoding, for the reception context in various studies
because a delayed retest is required to assess (Bohannon, 1988; Bohannon & Symons, 1992;
the preservation and consistency of details Christianson, 1989; Finkenauer et al., 1998;
over time. It is possible that shortly after a high Pillemer, 1984; Rubin & Kozin, 1984).
impact event, the majority of people will have Rehearsal is another critical predictor of
detailed memories of the reception event. Only memory for the reception context (Bohannon
some of these memories will persist (i.e, & Symons, 1992; Christianson, 1989;
become flashbulb memories), whereas others Finkenauer et al., 1998; McCloskey, Wible, &
will fade or become distorted. Cohen, 1988; Neisser, 1982; Wright, 1993).
Consistent with Conway’s arguments, our Rehearsal includes ruminations, an urge to
study is aimed at investigating memory for communicate with other people about the
reception context, a necessary but not emotional circumstances and the experienced
sufficient condition for flashbulb memories. feelings and reactions, and the following of the
Shortly after the 9/11/01 terrorist attacks, we media (TV, radio, newspapers, internet).
assessed memory for first hearing the news of The rememberer's prior knowledge
the events. The aim of the study was to related to the news is also of central
examine group differences in memory for the importance for the memory for the reception
reception context and for its determinants. context (Conway et al., 1994; Finkenauer et
al., 1998). For Conway et al. (1994), flashbulb
PREDICTORS OF MEMORY FOR memories do not only represent specific
RECEPTION CONTEXT features of the reception context, but also
conceptual and abstract knowledge related to
Research has identified several pre- the original event. Hence, they proposed that
requisites for the formation of vivid memories the recipient's prior knowledge of issues
of the reception context: novelty, surprise, im- related to the original event is crucial for
portance/consequentiality, emotional feeling forming memories of the reception context.
states, rehearsal, background knowledge, atti- People’s attitudes also affect the way
tudes, and memory for event-related facts. We theyapproach and react to situations. Prior
will briefly discuss each of these. knowledge and attitudes are predicted to
Models generally support that memo- facilitate the organization and assimilationof
ries for the reception context are initialized by the incoming information into existing
211
O. Luminet, A. Curci, E. J. Marsh, I.Wessel, T. Constantin, F. Gencoz, and M. Yogo
212
Hearing the News of the 9/11 Attacks
213
O. Luminet, A. Curci, E. J. Marsh, I.Wessel, T. Constantin, F. Gencoz, and M. Yogo
214
Hearing the News of the 9/11 Attacks
the news by assessing how much they felt (a) were run on each set of indicators, and, for
shaken, (b) affected, and (c) indifferent. each set, the one-dimensionalsolution
Social sharing. Respondents described wasconsidered.
(a) how many times they had discussed the In CatPCA, a respondent’s score on a
events, (b) how long after hearing the news given indicator variable is weighted with re-
they first spoke about it, (c) with how many spect to the distribution of frequencies for that
people they had discussed the news. indicator variable (Greenacre, 1993). Optimal
Following the mass media. Participants scores are computed. Optimal scores are stan-
rated how often in the past three days they had dardized coefficients, and, comparable to ordi-
followed the news via (a) TV, (b) radio, (c) nary correlation coefficients, they may be posi-
newspapers, and (d) internet. tive or negative in conformity with the direc-
Rumination. Participants rated the tion of the correlation of the corresponding
frequency over the last three days of thoughts, observed scores with the dimensional solution.
memories or images related to the event. Optimal scores indicate the position of respon-
Background knowledge. Participants dents along a continuum represented by the
were assessed on 11 questions. Questions were latent construct. A positive score indicates that
for instance about (a) the number of floors in that respondent scored above the mean for the
the Twin Towers or (b) the name of the whole sample on a given construct, a negative
administrations located at the Pentagon. score indicates that the respondent scores be-
Attitudes. Non-US respondents rated to low the mean. For example, a positive optimal
what extent they liked (a) the United States as score on memory for event-related facts means
a country, and (b) United States citizens. that this country had a better memory for the
Memory for event-related facts. Memory event than a country who had a negative score.
for the event was assessed by nine questions Given the high number of statistical tests,
concerning for instance (a) the number of for both main effects and planned contrasts,
planes hijacked or (b) the time at which the and in order to reduce the possibility of type I
Pentagon was hit. errors, comparisons significant at p > .005
level were not considered.
Procedure
Structural Analyses by Citizenship
Questionnaires were distributed among First, the sample of US respondents was
university students, university faculty and split into three groups (West coast, Central
staff, and experimenters’ acquaintances one to regions, East coast), to test whether location
six weeks after the attacks within the US affected the results. One-way
ANOVAs were run on the CatPCA Optimal
RESULTS Scores for US respondents, with location
(West coast vs. Central regions vs. East coast)
Measurement Issues and Structural Analyses as the between subjects factor. Results showed
In order to aggregate each set of that the US sample was very homogeneous for
indicators (memory for the reception context, the vast majority of the variables investigated
novelty, surprise, importance, emotional and was considered as a whole .
feeling states, rehearsal, background Then, in order to compare responses from
knowledge, attitudes, memory for event- US and non-US participants, a one-way
related facts) in single composite scores, the ANOVA was also run on the CatPCA optimal
SPSS CatPCA (Categorical Principal scores for each set of indicators, with
Component Analysis) procedure was used. citizenship (US vs. non-US respondents) as the
CatPCA is an exploratory analysis aimed at between-subjects factor, and elapsed time
identifying the latent dimensions underlying a between 9/11/01 and questionnaire completion
set of categorical variables. Separate analyses as a covariate (see table 1). All comparisons
215
O. Luminet, A. Curci, E. J. Marsh, I.Wessel, T. Constantin, F. Gencoz, and M. Yogo
Table 1. ANOVAs by citizenship (US vs. non-US) on the CatPCA Optimal Scores
were significant at .001 alpha level. Thus, US (elapsed time) was significant for all variables
respondents scored higher than non-US except for novelty, surprise, and importance.
respondents on all considered variables. The This means that the ratings of novelty,
largest effect sizes (> .10) were found for the surprise, and importance were not affected by
ratings of importance and memory for event- when the questionnaire was completed.
related facts. Moderate effect sizes (< .10 but The next step in the analysis was to
above > .05) were found for the comparisons consider differences among the non-US
on background knowledge about the WTC and respondents. One-way ANOVAs were run on
Pentagon and terrorist acts. Finally, low effect the CatPCA optimal scores, with citizenship
sizes (< .05) were observed for all the other (French-speaking countries, Italy, Japan,
variables (memory for the reception context, Romania, The Netherlands, Turkey) as the
novelty, surprise, emotional feeling state, between-subjects factor, and covaring the time
social sharing, following the mass media, and elapsed between 9/11/01 and the completion of
rumination). The effect of the covariate the questionnaire. Table 2 displays the results
216
Hearing the News of the 9/11 Attacks
Citizenship
Composite F Effect M (French- M M M M (The Ne M
variables (df) size speaking (Italy) (Japan) (Romania) dherlands) (Turkey)
ES countries)(SD) (SD) (SD) (SD) (SD) (SD)
Memory for the 3.394* .006 .006 .002 -.041 -.119a .108 a .090
Reception Context (5, 2891) (.980) (1.262) (.944) (.924) (.987) (.512)
Novelty 19.86** .036 -.089a .044b .127a,c .002d -.212c,e -.644a,b,c,d,e
(5, 2663) (.960) (.843) (1.102) (.856) (1.159) (1.008)
Surprise (1) 16.69** .039 .076a .014b -.170a,c .150c,d -.216a,b,d -.516a,b,c,d
(5, 2437) (1.056) (1.003) (.944) (.986) (.959) (.758)
Importance (2) 10.36** .020 -.293a .141a,b -.138a,b -.148b -.309b -.232b
(5, 2484) (.906) (.890) (.973) (1.075) (.822) (.920)
Emotional feeling 114.70** .180 .128a .506a,b -.694a,b,c .172b,c,d -.110b,c,d,e -.653a,b,d,e
state (5, 2611) (.956) (.875) (.677) (1.024) (.897) (.693)
Social sharing (2) 23.40** .047 .230a .488a,b -.254a,b,c -.367a,b,d .371c,d,e -.560a,b,c,e
(5, 2473) (.828) (.880) (.957) (1.085) (.799) (.890)
Following the mass 11.89** .024 -.128a .225a,b .066a,c -.285b,c,d .276a,c,d,e -.250b,c,e
media (2) (5, 2450) (.915) (1.045) (.997) (.870) (1.056) (.645)
Rumination (2) (3) 24.56** .046 2.930a 3.402a,b 2.766b,c 2.606a,b,d 3.587a,c,d,e 2.410a,b,c,e
(5, 2520) (1.192) (1.040) (1.081) (1.133) (1.195) (.935)
Background
knowledge about 111.15** .161 .530a -.410a,b -.286a,c -.475a,c,d .293a,b,c,d,e -.210a,e
WTC and Pentagon (5, 2891) (.884) (.808) (1.092) (.842) (.870) (.917)
Background
knowledge on 50.85** .091 .264a -.471a,b -.335a,c -.351a,d .115b,c,d .427a,b,c,d
terrorist acts (2) (5, 2535) (1.088) (.792) (.831) (.823) (1.020) (.503)
Attitude 43.86** .072 -.175a .104a,b .132a,c .310a,c -.492a,b,c -.326b,c
(5, 2841) (.793) (1.006) (.954) (1.212) (.686) (.824)
Memory for event- 24.84** .047 .384a -.088a,b -.385a,b,c -.630a,b,c,d .365b,c,d,e -.075a,c,d,e
related facts (2) (5, 2535) (.720) (.980) (1. 230) (1.024) (.693) (.688)
* p < .005; ** p < .001. (1) Results were controlled for delay from the event, significant at p < .005, (2)
Results were controlled for delay from the event, significant at p < .001, (3) Raw scores, with range 1-7
Note: Means in a row sharing subscripts are significantly different at least at .005 alpha level. For each
composite variable, higher positive means indicate higher scores on that variable.
of these comparisons. All comparisons were WTC and Pentagon, and attitudes. This means
significant at .001 alpha level, except that that the ratings corresponding to these
memory for the reception context was constructs did not seem to depend on when the
significant at .005 alpha level. Overall, the questionnaire was completed.
effect sizes for the comparisons were not high, A ceiling effect was expected for
except for the ANOVAs on emotional feeling memory for the reception context, novelty, and
states (.108), and background knowledge about surprise. On memory for the reception context,
WTC and Pentagon (.161), thus indicating that the only difference was that Romanians scored
the differences among countries were strong significantly lower than Dutch respondents,
only for those variables. The effect of the but the effect size was extremely low (.006).
covariate (elapsed time) was found to be Large differences were only observed, as
significant for all variables except for memory expected, for Turkish people, who had the
for the reception context, novelty, emotional lowest coefficients for novelty and surprise.
feeling states, background knowledge about Italians rated the event as significantly more
217
O. Luminet, A. Curci, E. J. Marsh, I.Wessel, T. Constantin, F. Gencoz, and M. Yogo
important than participants from the other of the details of how they heard the news. This
countries. The lowest ratings of importance ceiling effect is similar to other studies
were found for Dutch, French-speaking, and assessing immediate memory for the reception
Turkish respondents. As predicted, participants context (e.g., Christianson & Engelberg, 1999;
from individualistic countries (French- Conway et al., 1994; Curci et al., 2001;
speaking countries, Italy, and The Pezdek, 2003; Schmolk et al., 2000; Wright,
Netherlands) scored higher than participants 1993).
from collectivistic countries (Japan and We will now turn to the major aim of the
Turkey) on emotional feeling states, social present study, i.e., examining group
sharing, and rumination – planned differences in predictors of flashbulb
comparisons were significant at .001 alpha memories. Moderate and large differences will
level. As for rehearsal through the mass media, be examined first, followed by small ones.
planned contrasts were consistent with the Three findings fall into the category of
expectation that Romanian and Turkish moderate to large differences. First,
respondents would have had the lowest access Americans rated the events as much more
to the media. As a consequence, Romanian and important to them. More specifically,
Turkish respondents were also expected to Americans rated the events as having greater
exhibit very low levels of background national and personal importance than did non-
knowledge about the WTC and Pentagon. This US respondents. US respondents were more
prediction was confirmed through a planned likely to have known someone killed or
contrast, significant at .001 alpha level. injured, and the attacks involved more changes
However, Italians also scored significantly low in daily activities in the US than in any other
on this variable. On the other hand, Turkish country. Finally, financial and political
participants showed the highest background consequences of the events were more
knowledge about terrorist acts compared to prevalent in the US (e.g., bankruptcies of
participants from the other involved countries. airline companies, the new priorities of the US
As expected, Turkish respondents exhibited government towards military exercises and
very low scores on attitudes towards the US. expenses).
However, the lowest scores were found for US respondents showed greater
Dutch respondents. background knowledge, on both relevant
Finally, the planned contrasts confirmed composite scores. These differences can be
the prediction of low memory for event-related explained by different factors. Even prior to
facts for respondents from Turkey and the attacks, US respondents likely knew more
Romania. However, Japanese people also basic facts about the WTC and the Pentagon.
showed significantly lower memory for event- Such facts were also salient in media coverage,
related facts compared to respondents from the as they were crucial for understanding event
other involved countries. implications (e.g., information about the
number of floors in the WTC was key to
DISCUSSION understanding missing peoples’ fates). Item
effects support the hypothesis that people
Comparison between US and non-US learned some of these facts as events unfolded.
Countries For example, people were much better at
Confirming our predictions, only very listing the number of floors in the WTC (59%
small differences emerged for the memory for correct) than its height (30-32%). These two
the reception context (the effect size was facts refer to the same thing: the size of the
below 1%). There was a ceiling effect across building. However, media coverage framed
the two samples, with response rate above 95% the planes’ points of impact in terms of floors,
in the US group and above 90% in the non-US not feet or meters. Interestingly, US vs. non-
one. Both groups reported remembering most US differences were larger for number of
218
Hearing the News of the 9/11 Attacks
floors (72 vs. 55.5%) than for height of the feeling states, attitudes and background
towers (33 vs. 23.7%). Again, Americans knowledge. Only very small differences were
were more likely to know people who worked observed, however, for memory for the
in the buildings (and knew their location in reception context, novelty, surprise, and
terms of floors) and thus needed to rapidly importance. For each of these variables, ceiling
assess where these people were located in levels were observed -- indicating that the 9/11
relation to the critical floors. Finally, some attacks had a major international impact that
knowledge was likely learned after the events. was unaffected by economic and cultural
In particular, knowledge about terrorist acts differences. For each country or group of
likely increased after the attacks. US citizens countries, we will now examine the relation of
were probably more motivated than non-US the results to predictions.
ones to learn about such issues in the As predicted, the group of Western
aftermath of their national tragedy. European countries (French-speaking
The largest difference between US and countries, Italy and The Netherlands) had high
non-US respondents was in memory for event- scores on almost all the variables investigated.
related facts. US respondents found the events For each of these groups, only one or two
more important, and had more background predictions out of 12 were not confirmed (i.e.,
knowledge - , two variables that the literature French-speaking respondents rated importance
suggests are antecedents of memory for event- as low, Italians showed low levels of
related facts (e.g., Finkenauer et al., 1998). background knowledge and citizens of The
Higher event-memory could thus be the result Netherlands found the events less important
of a cumulative advantage of US respondents than other Western Europeans and also had
for these antecedents. less positive attitudes towards US). Thus, we
Both groups assessed the events as very can conclude that, as expected, Western
novel and surprising. These results are European citizens were similar to Americans
consistent with media labels of the 9/11 attacks in their reactions, although the intensity of
as distinctive and unpredictable. The their reactions was sometimes lower.
terrorists sent no explicit warnings, and no Due to their lack of experience with
events in particular suggested such an attack terrorism, and their cultural similarity,
was in the works. US and non-US groups also Romanians was expected to react like Western
rehearsed the event to a similar extent, again Europeans in their ratings of novelty, surprise
pointing to the unique nature of the events. and emotional feeling states. These results
Both groups also had similar emotional were confirmed. However, Romanians were
reactions to the attacks. Although raw scores expected to differ from Western Europeans on
for emotional feeling states were high, they did variables more directly related to their
not reach a ceiling level, as novelty and economic status, such as following the media,
surprise did. This can be explained by background knowledge and for the two types
variations within the items assessing the of memories (event-related facts and reception
emotional feeling states. While overall means context). Again, these predictions were fully
were slightly higher for US than for non- US supported. The only unexpected finding was
citizens, it is interesting to note that large related to low levels of social sharing and
variations were observed across countries on rumination. This finding may be due to the
specific items. delayed data collection in Romania ;
Romanians responded on average 30 days after
Comparison between non-US Countries the attacks whereas Western Europeans
Major differences were observed responded after 15 days. As these questions
across non-US countries for three variables specifically required assessment of behavior in
that have a direct or an indirect effect on the the past three days , it is not surprising that
formation of flashbulb memories: emotional reports were lower among Romanians.
219
O. Luminet, A. Curci, E. J. Marsh, I.Wessel, T. Constantin, F. Gencoz, and M. Yogo
Most predictions were confirmed for the assimilate the news about the attacks (see
Japanese group. They responded similarly to Conway et al., 1994; Finkenauer et al., 1998),
Western Europeans on variables related to leading to average levels observed for memory
economic development (following the media for the reception-context and for memory for
and attitudes). In contrast, Japanese event-related facts. Cultural and economic
respondents reported lower levels of emotional differences help us to understand the other
feeling states, social sharing, and rumination. differences. For example, the collectivistic
These results are consistent with findings nature of the country might explain low scores
showing that collectivistic countries report less on emotional feeling states, social sharing and
intense immediate and long-term emotional rumination (e.g., Basabe et al., 2000, 2002;
reactions (e.g., Basabe et al., 2000, 2002; Fernandez et al., 2000). Additionally, this
Fernandez et al., 2000). Unexpectedly, country shares a different religion and different
however, Japanese respondents reported less values as compared to Western countries,
background knowledge and memory for event- which might have exacerbated differences.
related facts. One explanation is that the Turkey is also less economically developed,
Japanese may have been less interested in the likely reducing media coverage of the events.
9/11 events than citizens of other countries. Turkey has also suffered economically from
This hypothesis is supported by differences in US imposed prohibitions on trade with
the mean level of “interest” in the news, which countries accused of helping terrorist acts
respondents rated on a 7-point scale (from 1, immediately after the 9/11 attacks ; this likely
“not at all”, to 7, “very much”). The mean plays into the less positive attitudes held by the
level for Japanese respondents was only 4.77, Turkish respondents.
while all other respondents scored close to or
above 6. Limitations and Strengths of the Study
Finally, as expected, Turkish respondents Studies of natural events are difficult to
displayed the most differences from the other do; the data must be collected soon after the
respondents Turkish respondents had surprising event and thus there is little time to
particularly low scores on novelty, surprise, plan what data are of interest and how they
emotional feeling states, social sharing, should be obtained. Methodologically, the
rumination, following the media, attitudes, and biggest weakness in the current study was the
background knowledge about the WTC and the occurrence of a lag between when the event
Pentagon. Smaller differences were observed occurred and when the data were collected,
for memory for the reception context and and that this lag varied across countries.
event-related facts. Overall, the pattern of However, by introducing elapsed time between
results can be explained by considering the 9/11/01 and questionnaire as a covariate, we
Turkish experience. That is, Turkey had a controlled for this difference in the analyses.
number of terrorist attacks in the last few Methodologically, the strength of the study
years, and also a serious earthquake. In both was its use of a questionnaire validated by its
cases, many thousands were killed, likely use in previous studies (Curci et al., 2001;
habituating Turkish citizens to such Finkenauer et al., 1998). One advantage of
catastrophes (and thus reducing their responses this questionnaire is that it assesses each
to the 9/11 attacks). Turkey’s history of predictor with multiple items. It thus allows for
terrorist attacks (and its geographical computing composite scores, which are more
proximity to many terrorist bases) likely reliable than single items. In addition, in
played a role in their high level of knowledge comparison to previous studies investigating
about terrorist groups (M = 83.8% correct, as memory for the reception context, the present
compared to a sample mean of 31.6%). one has the advantage of assessing a larger set
Turkish respondents’ knowledge about of predictors. To our knowledge, no previous
terrorism may have help them to encode and study investigated 12 predictors
220
Hearing the News of the 9/11 Attacks
simultaneously as we did in the present study. importance reaching high levels (e.g., Conway,
Even fewer studies have examined group 1995), while personal importance has been
differences on the different predictor variables. defined as the key feature (Brown and Kulik,
For example, only few have examined group 1977). In contrast to many other studies in
differences in memory for event-related facts which personal importance was either low
(see Curci et al., 2001 for an exception). Such (e.g., Schmolk et al., 2000; Wright, 1993) or
differences can be critical as memory for not assessed (e.g., Bohannon, 1988;
event-related facts is a direct predictor for Christianson, 1989; McCloskey et al., 1988;
flashbulb memories (e.g., Finkenauer et al., Neisser & Harsch, 1992; Neisser et al., 1996),
1998). our subjects clearly rated the events of 9/11 as
The other major methodological strength personally important (Ms above 5 on scales
is of course the number and nature of the ranging from 1 to 7). Fourth, mean levels for
comparison groups. While previous studies emotional reactions and rehearsal were
usually involved only one or two comparison remarkably high, while some previous studies
groups (Brown & Kulik, 1977; Conway et al., failed to find high levels on these variables
1994; Curci et al., 2001), eight non-US (e.g., Conway et al., 1994; Pillemer, 1984).
national groups were involved in the present Fifth, memory for event-related facts reached
study. Also previous comparison studies high levels in the US sample. Out of 10
involved other national groups that were close variables investigated, four had a correct
economically and culturally (e.g., UK vs. US answer rate above 90% and seven were above
and Denmark for Conway et al., 1994; France 70%. In the US sample, background
and Belgium for Curci et al., 2001). In the knowledge was the only predictor for which
present study, we had a first group of countries some ratings were low-- but the overall mean
economically and culturally close to the US level was still in the middle range. The present
(French-speaking countries, Italy and The event is thus unique as it is characterized by
Netherlands), one country culturally close but high ratings on all the required characteristics
economically different (Romania), one country for the formation and maintenance of flashbulb
economically close but culturally different memories.
(Japan), and one country with both large Finally, we wish to note an empirical
cultural and economic differences (Turkey). strength of the current study. A statistical
Another major improvement relates to the procedure called CatPCA was employed that
nature of the event investigated. Previous allowed a scoring system analogous to the
studies did not use events with high scores on standard WAS procedure used by Neisser &
all the major predictors of the formation of Harsch (1992) and Conway and his colleagues
flashbulb memories. In contrast, our data (1994; 1995). This procedure assigns different
confirm that all predictors were high for weights to the different attributes of the
Americans remembering 9/11 . Almost all reception context and, in general, to all
American respondents had a vivid memory for indicator variables in the questionnaire. The
the circumstances in which they learned about procedure is aimed at creating composite
the news . Similarly, they were at ceiling level scores corresponding to the considered
in ratings of novelty and surprise, which is in theoretical constructs. Unlike the WAS
sharp contrast with previous studies procedure, the weights assigned to the
investigating the formation of flashbulb indicator variables by CatPCA were not
memories (e.g., Christianson & Engleberg, decided a priori by the researcher, but came
1999; Conway et al., 1994; Curci et al., 2001; from the empirical distributions of the
Pillemer, 1984). Third, many studies that have response categories for each indicator variable.
examined the formation of flashbulb memories
were criticized for their lack of personal
relevance, with only national or international
221
O. Luminet, A. Curci, E. J. Marsh, I.Wessel, T. Constantin, F. Gencoz, and M. Yogo
222
Hearing the News of the 9/11 Attacks
223