Article 2004 The Cognitive, Emotional And Social Impacts Of

O. Luminet, A. Curci, E. J. Marsh, I.Wessel, T. Constantin, F. Gencoz, and M. Yogo

THE COGNITIVE, EMOTIONAL AND SOCIAL IMPACTS OF THE SEPTEMBER 11TH ATTACKS: GROUP DIFFERENCES IN MEMORY FOR THE RECEPTION CONTEXT AND ITS DETERMINANTS Olivier Luminet University of Louvain at Louvain-la-Neuve, Belgium, and Belgian National Fund for Scientific Research [email protected]

Antonietta Curci University of Bari, Italy

Elizabeth J. Marsh Washington UniversitySt Louis, MO, USA

Ineke Wessel Maastricht University, The Netherlands

Ticu Constantin University of Iasi, Romania

Faruk Gencoz Middle East Technical University, Ankara, Turkey

Masao Yogo Doshisha University, Kyoto, Japan

ABSTRACT This study examines group differences in cognitive, emotional, and social responses to hearing the news of the September 11th attacks. Although all groups were hypothesized to remember the reception context immediately after the events, of interest was whether groups differed initially in predictors of long-term memory. One to 6 weeks after 9/11/01, a questionnaire investigating memory for the reception context and a set of critical predictors wasdistributed to 3665 participants in Belgium, France, Italy, Japan, Romania, Switzerland, The Netherlands, Turkey, and the US. US vs. non-US respondents showed large differences in memory for event-related facts and in self-rated importance of the news. Moderate-level differences between groups were found for background knowledge, and small differences were observed for memory for the reception context, ratings of novelty, surprise, emotional feeling states and rehearsal. Within non-US groups, large differences were only found for emotional feeling states, whereas moderate differences 210

were obtained for background knowledge and attitudes towards the US. Finally, non-US groups did not differ much in memory for the reception context, memory for event-related facts, or ratings of novelty, surprise, importance, and rehearsal. INTRODUCTION For more than two decades, emotion and memory researchers have been interested in people’s memory for the circumstances in which they first heard about major public events, such as political assassinations (e.g., Brown & Kulik, 1977; Christianson, 1989) or natural or man-made disasters (e.g., Christianson & Engelberg, 1999; Neisser & Harsch, 1992; Wright, 1993). Brown and Kulik (1977) introduced the term “flashbulb memories” for these particularly vivid and apparently long-lasting recollections. Two different views of flashbulb memories co-exist in the literature. The first one emphasizes the detailedness of the recollection, often defined by the number of canonical categories endorsed at one time of

Hearing the News of the 9/11 Attacks measurement (e.g., exact location, ongoing activity, type of informant, idiosyncratic details often including perceptual aspects; for an example, see Finkenauer et al., 1998). The second perspective is adopted in studies relying on a test-retest design (e.g., Conway et al., 1994) and emphasizes consistency between initial test (usually shortly after encoding) and retest. Conway (e.g., 1995) argues that flashbulb memories should be restricted to this latter definition. According to him, in order to be considered a flashbulb memory, a recollection involves not only a “live” quality accompanied by recall of minutiae, but also preserves details of the reception events, and remains unchanged over long periods of time. Acceptance of these last two conditions means that it is not technically possible to assess flashbulb memories shortly after encoding, because a delayed retest is required to assess the preservation and consistency of details over time. It is possible that shortly after a high impact event, the majority of people will have detailed memories of the reception event. Only some of these memories will persist (i.e, become flashbulb memories), whereas others will fade or become distorted. Consistent with Conway’s arguments, our study is aimed at investigating memory for reception context, a necessary but not sufficient condition for flashbulb memories. Shortly after the 9/11/01 terrorist attacks, we assessed memory for first hearing the news of the events. The aim of the study was to examine group differences in memory for the reception context and for its determinants. PREDICTORS OF MEMORY FOR RECEPTION CONTEXT Research has identified several prerequisites for the formation of vivid memories of the reception context: novelty, surprise, importance/consequentiality, emotional feeling states, rehearsal, background knowledge, attitudes, and memory for event-related facts. We will briefly discuss each of these. Models generally support that memories for the reception context are initialized by

new situations and that this mechanism has adaptive value (e.g., Brown & Kulik, 1977; Conway, 1995; Finkenauer et al., 1998). The appraisal of novelty is then followed by a nonvalenced emotional state of surprise. The appraisal of importance/consequentiality develops almost simultaneously with the appraisal of novelty and has been proposed as a critical predictor for memory for the reception context (Brown & Kulik, 1977; Christianson, 1989; Christianson & Engelberg, 1999; Conway et al., 1994; Finkenauer et al., 1998; Rubin & Kozin, 1984). The appraisal of importance/consequentiality then leads to emotional feeling states such as sadness, anger or fear (Conway et al., 1995; Finkenauer et al., 1998; Oatley & Johnson-Laird, 1987). The intensity of emotional feeling states predicted memory for the reception context in various studies (Bohannon, 1988; Bohannon & Symons, 1992; Christianson, 1989; Finkenauer et al., 1998; Pillemer, 1984; Rubin & Kozin, 1984). Rehearsal is another critical predictor of memory for the reception context (Bohannon & Symons, 1992; Christianson, 1989; Finkenauer et al., 1998; McCloskey, Wible, & Cohen, 1988; Neisser, 1982; Wright, 1993). Rehearsal includes ruminations, an urge to communicate with other people about the emotional circumstances and the experienced feelings and reactions, and the following of the media (TV, radio, newspapers, internet). The rememberer's prior knowledge related to the news is also of central importance for the memory for the reception context (Conway et al., 1994; Finkenauer et al., 1998). For Conway et al. (1994), flashbulb memories do not only represent specific features of the reception context, but also conceptual and abstract knowledge related to the original event. Hence, they proposed that the recipient's prior knowledge of issues related to the original event is crucial for forming memories of the reception context. People’s attitudes also affect the way theyapproach and react to situations. Prior knowledge and attitudes are predicted to facilitate the organization and assimilationof the incoming information into existing 211

O. Luminet, A. Curci, E. J. Marsh, I.Wessel, T. Constantin, F. Gencoz, and M. Yogo semantic structures in memory, thereby leading to more detailed and consistent memories for the reception context (Conway et al., 1994; Finkenauer et al., 1998). Finally, memory for event-related facts seems to impact directly on memory for the reception context (Finkenauer et al., 1998). These authors suggested that when learning about the original event, all currently activated information--including the reception context, sensory information, and the original event--is encoded in memory (Tulving & Kroll, 1995). Thus, both the original event and the reception context would immediately be encoded. Subsequently, during rehearsal of the original event, the reactivation of this information in memory would spread to any associated information, and thus to flashbulb memories. GROUP DIFFERENCES IN MEMORY FOR THE RECEPTION CONTEXT One impressive finding in the literature is how frequently people report vivid memories of the reception context shortly after unexpected events. For example, Wright (1993) reported a 100% rate of memories for the reception context for the Hillsborough disaster one month after the event, or Schmolck et al. (2000) reported a 98.4% rate of memories for the reception context for the OJ Simpson verdict 3 days after it was announced. It is not clear, however, that those rates would have been as high with different subject populations. Accordingly, there are studies showing more memories for the reception context in one group of participants than another (e.g., Brown & Kulik, 1977; Conway et al., 1994; Curci et al., 2001, Neisser et al., 1996). These comparison studies suggest that the salience of a major news event and its personal importance may vary for different cultural or national groups resulting in differences in memories for the reception context. In the next section, we will turn to the improvements in the current study that allow for a better understanding of group differences in the predictors of memory for the reception context. 212

IMPROVEMENTS OF THE PRESENT STUDY There are two major differences between the current study and previously published studies on memory for the reception context: the nature of the target event, and the nature and number of the comparison groups. We will discuss each of these differences in turn. One important limitation of previous studies is that they did not use events hypothesized to yield high scores on all the predictor variables (e.g., Christianson & Engelberg, 1999; Conway et al., 1994; Curci et al., 2001). The events of 9/11 are tragically unique in this regard, and as such present an important opportunity to better understand memory for reception context. The participants in the current study were not only Americans, but also citizens of eight other countries. This is in sharp contrast to previous studies, which had only a few comparison groups. In addition, the comparison groups in the current study are less similar to one another than is typical of such studies. For example, in the study on Thatcher’s resignation (Conway et al., 1994), non-UK comparison countries were the US and Denmark. In the present study, comparison groups included not only Western European countries, but also Romania, Turkey, and Japan. Of interest was the effect of the countries’ differences in cultural and economic backgrounds on the predictors of memory for the reception context. OVERVIEW AND HYPOTHESES On September 11, 2001, more than three thousand people died in the worst terrorist attacks ever-committed on American soil. In co-ordination, four commercial airliners were hijacked and turned into missiles aimed at buildings that symbolized American prosperity. Two planes hit the north and south towers of the World Trade Center (WTC), leading to the eventual collapse of both buildings. One plane hit the Pentagon, and the

Hearing the News of the 9/11 Attacks fourth plane crashed in rural Pennsylvania. These attacks were allegedly planned and executed by members of Al Qaeda, a terrorist group based in Afghanistan. The aim of the present study is to examine memory for hearing the news of the 9/11 attacks (memory for reception context) and its predictors. Predictor variables are 1) novelty, 2) surprise, 3) personal consequentiality and importance, 4) valenced emotional feeling states, 5) rehearsal, 6) background knowledge, 7) attitudes towards the target country and 8) memory for eventrelated facts. For Americans, the 9/11 attacks were assumed to yield high ratings on all these predictors (except for attitudes, which were not relevant for the US group). It was not assumed, however, that those rates would be as high in other countries. The simultaneous testing of large groups of both US and non-US respondents allowed examination of possible group differences in the predictor variables. Due to the large sample sizes, differences were likely to be observed for the vast majority of variables. Thus, we use three levels of effect sizes to categorize the differences. Small differences were related to predicted effect sizes below .05, moderate ones were above .05 but below .10 and large ones were above .10 . First, we predicted only a small difference between US and non-US groups for memory for the reception context. In addition, we expected a ceiling level across countries for the appraisal of novelty. Novelty has been shown to directly influence the emotional state of surprise, which in turn directly strengthens memory for the reception context (Finkenauer et al., 1998). For the other variables, however, we predicted larger differences between US and non-US respondents. Most importantly, we predicted large group differences for the ratings of importance. Americans were more likely to have known someone killed in the attacks, or to know people who were otherwise directly involved. Also, the 9/11 attacks directly impacted American daily lives (e.g., security at public events and airports). Following the emotional-integrative model (Finkenauer et al., 1998), differences in impor-

tance/consequentiality should be reflected in variables that are supposed to be direct outcomes. Thus, we predicted moderate to large group differences for valenced emotional feeling states, rehearsal and memory for eventrelated facts. As Americans were likely more knowledgeable about the WTC and the Pentagon before the attacks, moderate differences were also predicted for background knowledge. Moderate differences were also predicted for prior knowledge of terrorist groups, as Americans were more motivated to have sought out information related to attack perpetrators. The large number of comparison groups allowed us to make a 2nd set of predictions, involving differences among the non-US countries. The eight comparison countries can be divided into four groups. Citizens of Western European countries (French-speaking countries - Belgium, France, Switzerland 1 - The Netherlands and Italy) were expected to react similarly to the Americans. These European nations are socio-economically and culturally similar to the US. These countries have previously experienced terrorist attacks, but were not as directly involved in the 9/11 events, as were Americans. Romania was considered as similar to the other European countries except for its level of economic development, which might affect access to the media. Less access to the media might have resulted in lower background

1 Due to low response frequencies, respondents from Switzerland and France were aggregated with respondents from Belgium. This resulted in a new category for the variable citizenship, corresponding to respondents from French-speaking countries. These three countries did not seem to differ significantly on the variables considered, except for the prior knowledge about the event. Swiss respondents had a better background knowledge about the event than did Belgians and the French (F (2,660) = 7.26, p = .001). Respondents from these three countries are thus considered together in the paper and are refered as “French-speaking countries”.

213

O. Luminet, A. Curci, E. J. Marsh, I.Wessel, T. Constantin, F. Gencoz, and M. Yogo knowledge and in poorer memory for attack details. Due to its similarities on socio-economic factors, Japanese respondents were predicted to react similarly to Western Europeans on variables like rehearsal through the media, attitudes towards the US, and background knowledge. Some differences were predicted, however, in the intensity of emotional feeling states, and the frequency of social sharing of emotion and rumination. Cross-cultural studies show that in individualistic countries such as the US or Western Europe, the public reporting of one’s emotional state is socially desirable, while it is not in collectivistic countries such as Turkey or Japan (e.g., Fernandez et al., 2000). Also, people’s focus of attention is more selfcentered in individualistic countries, which may result in more rumination and social sharing (e.g., Basabe et al., 2000, 2002). Finally, Turkey was considered as the country most likely to show differences from both the US and other non-US countries. Turkey has been the site of numerous terrorist attacks, and the habituation of Turkish citizens to terrorism should led to lower ratings of novelty, surprise, importance and emotional feeling states. The Turkish people also have a different cultural and religious background, which is partly reflected in a lower level of individualism (see above). This should also result in lower emotional feeling states and additionally in less frequent social sharing and rumination. Moreover, Turkish people were economically affected by US-dictated prohibitions on trade with countries accused of helping terrorist groups. Less positive attitudes towards the US are thus expected. Finally, less access to mass media due to Turkey’s lower socio-economic status could contribute to lower levels of rehearsal through the media, to lower background knowledge and to poorer memory for the attacks.

214

METHOD Participants A total of 3665 participants from nine different countries participated in the study. Participants resided in Belgium (n = 546), France (n = 62), Italy (n = 296), Japan (n = 683), Romania (n = 716), Switzerland (n = 55), The Netherlands (n = 348), Turkey (n = 191), and the USA (n =768). Measures The measures used for the present study were very close to the ones used by Finkenauer et al. (1998) and by Curci et al. (2001). Only substantial differences with the previous version are reported here. See Finkenauer et al. (1998) and Curci et al. (2001) for more details. Memory for the reception context. Eight items assessed recall of the circumstances in which participants first learned about the terrorist attacks. Questions were for instance about the date (day, month, year), the day of the week, (or the informant (TV, radio, newspaper, another person). These questions are similar to the ones used by previous research (Bohannon, 1988; Brown & Kulik, 1977; Conway et al., 1994; Finkenauer et al., 1998). Novelty. Participants rated how (a) common, (b) unusual, (c) special, (d) strange, and (e) ordinary the event was for them. Surprise. Participants rated to what extent (a) they felt surprised, (b) the news was expected and (c) unbelievable. Importance/consequentiality. Participants rated the extent to which the event was important (a) for themselves, (b) for their family and friends, (c) for their country, and (d) the importance of the event at the international level. Additionally, they rated to extent to which the original event had consequences for them. Emotional feeling state. Participants rated the extent to which they were upset by the news. Additionally, they rated the impact of

Hearing the News of the 9/11 Attacks the news by assessing how much they felt (a) shaken, (b) affected, and (c) indifferent. Social sharing. Respondents described (a) how many times they had discussed the events, (b) how long after hearing the news they first spoke about it, (c) with how many people they had discussed the news. Following the mass media. Participants rated how often in the past three days they had followed the news via (a) TV, (b) radio, (c) newspapers, and (d) internet. Rumination. Participants rated the frequency over the last three days of thoughts, memories or images related to the event. Background knowledge. Participants were assessed on 11 questions. Questions were for instance about (a) the number of floors in the Twin Towers or (b) the name of the administrations located at the Pentagon. Attitudes. Non-US respondents rated to what extent they liked (a) the United States as a country, and (b) United States citizens. Memory for event-related facts. Memory for the event was assessed by nine questions concerning for instance (a) the number of planes hijacked or (b) the time at which the Pentagon was hit.

were run on each set of indicators, and, for each set, the one-dimensionalsolution wasconsidered. In CatPCA, a respondent’s score on a given indicator variable is weighted with respect to the distribution of frequencies for that indicator variable (Greenacre, 1993). Optimal scores are computed. Optimal scores are standardized coefficients, and, comparable to ordinary correlation coefficients, they may be positive or negative in conformity with the direction of the correlation of the corresponding observed scores with the dimensional solution. Optimal scores indicate the position of respondents along a continuum represented by the latent construct. A positive score indicates that that respondent scored above the mean for the whole sample on a given construct, a negative score indicates that the respondent scores below the mean. For example, a positive optimal score on memory for event-related facts means that this country had a better memory for the event than a country who had a negative score. Given the high number of statistical tests, for both main effects and planned contrasts, and in order to reduce the possibility of type I errors, comparisons significant at p > .005 level were not considered.

Procedure Questionnaires were distributed among university students, university faculty and staff, and experimenters’ acquaintances one to six weeks after the attacks RESULTS Measurement Issues and Structural Analyses In order to aggregate each set of indicators (memory for the reception context, novelty, surprise, importance, emotional feeling states, rehearsal, background knowledge, attitudes, memory for eventrelated facts) in single composite scores, the SPSS CatPCA (Categorical Principal Component Analysis) procedure was used. CatPCA is an exploratory analysis aimed at identifying the latent dimensions underlying a set of categorical variables. Separate analyses

Structural Analyses by Citizenship First, the sample of US respondents was split into three groups (West coast, Central regions, East coast), to test whether location within the US affected the results. One-way ANOVAs were run on the CatPCA Optimal Scores for US respondents, with location (West coast vs. Central regions vs. East coast) as the between subjects factor. Results showed that the US sample was very homogeneous for the vast majority of the variables investigated and was considered as a whole . Then, in order to compare responses from US and non-US participants, a one-way ANOVA was also run on the CatPCA optimal scores for each set of indicators, with citizenship (US vs. non-US respondents) as the between-subjects factor, and elapsed time between 9/11/01 and questionnaire completion 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 Citizenship Composite variables

Effect M(US) M (non-US) F (df) (SD) size (ES) (SD) Memory for the 12.72* .004 .080 -.017 reception context (1) (1, 3268) (1.120) (.988) Novelty 36.17* .010 .191 -.054 (1, 3432) (.922) (1.014) Surprise 47.16* .013 .219 -.060 (1, 3541) (.973) (.999) Importance 527.42* .128 .688 -.185 (1, 3605) (.812) (.965) Emotional feeling state 134.88* .043 .375 -.143 (1) (1, 3027) (.982) (.968) Social sharing (1) 110.63* .037 .114 -.047 (1, 3203) (.999) (.994) Following the mass 47.87* .015 .115 -.038 media (1) (1, 3182) (1.116) (.962) Background 199.29* .057 .323 -.087 knowledge about WTC (1, 3268) (.924) (1.012) and Pentagon (1) Background 239.35* .068 .444 -.119 knowledge about (1, 3268) (1.070) (.943) terrorist acts (1) Rumination (1) (2) 61.54* .019 2.98 2.90 (1, 3221) (1.23) (1.17) Memory for event517.01* .137 .498 -.165 related facts (1) (1, 3268) (.601) (1.056) * p < .001 (1) Results were controlled for delay from the event, significant at p < .001. (2) Raw scores, with range 1-7.

Raw scores M (SD) (range) 10.46 (1.03) (0-11) 5.91 (1.03) (1-7) 5.98 (1.06) (1-7) 5.48 (1.03) (1-7) 5.37 (1.29) (1-7) 3.48 (.68) (1-6) 2.10 (.73) (1-5) 4.62 (3.68) (0-19) .60 (.84) (0-3) 2.93 (1.19) (1-5) 11.73 (4.02) (10-30)

Table 1. ANOVAs by citizenship (US vs. non-US) on the CatPCA Optimal Scores

were significant at .001 alpha level. Thus, US respondents scored higher than non-US respondents on all considered variables. The largest effect sizes (> .10) were found for the ratings of importance and memory for eventrelated facts. Moderate effect sizes (< .10 but above > .05) were found for the comparisons on background knowledge about the WTC and Pentagon and terrorist acts. Finally, low effect sizes (< .05) were observed for all the other variables (memory for the reception context, novelty, surprise, emotional feeling state, social sharing, following the mass media, and rumination). The effect of the covariate 216

(elapsed time) was significant for all variables except for novelty, surprise, and importance. This means that the ratings of novelty, surprise, and importance were not affected by when the questionnaire was completed. The next step in the analysis was to consider differences among the non-US respondents. One-way ANOVAs were run on the CatPCA optimal scores, with citizenship (French-speaking countries, Italy, Japan, Romania, The Netherlands, Turkey) as the between-subjects factor, and covaring the time elapsed between 9/11/01 and the completion of the questionnaire. Table 2 displays the results

Hearing the News of the 9/11 Attacks Citizenship Composite variables

F (df)

Effect M (Frenchsize speaking ES countries)(SD)

M (Italy) (SD)

M (Japan) (SD)

M (Romania) (SD)

M (The Ne M dherlands) (Turkey) (SD) (SD)

Memory for the Reception Context

3.394* (5, 2891)

.006

.006 (.980)

.002 (1.262)

-.041 (.944)

-.119a (.924)

.108 a (.987)

.090 (.512)

Novelty

19.86** (5, 2663)

.036

-.089a (.960)

.044b (.843)

.127a,c (1.102)

.002d (.856)

-.212c,e (1.159)

-.644a,b,c,d,e (1.008)

Surprise (1)

16.69** (5, 2437)

.039

.076a (1.056)

.014b (1.003)

-.170a,c (.944)

.150c,d (.986)

-.216a,b,d (.959)

-.516a,b,c,d (.758)

Importance (2)

10.36** (5, 2484)

.020

-.293a (.906)

.141a,b (.890)

-.138a,b (.973)

-.148b (1.075)

-.309b (.822)

-.232b (.920)

Emotional feeling state Social sharing (2)

114.70** (5, 2611) 23.40** (5, 2473) 11.89** (5, 2450) 24.56** (5, 2520)

.180

.128a (.956) .230a (.828) -.128a (.915) 2.930a (1.192)

.506a,b (.875) .488a,b (.880) .225a,b (1.045) 3.402a,b (1.040)

-.694a,b,c (.677) -.254a,b,c (.957) .066a,c (.997) 2.766b,c (1.081)

.172b,c,d (1.024) -.367a,b,d (1.085) -.285b,c,d (.870) 2.606a,b,d (1.133)

-.110b,c,d,e (.897) .371c,d,e (.799) .276a,c,d,e (1.056) 3.587a,c,d,e (1.195)

-.653a,b,d,e (.693) -.560a,b,c,e (.890) -.250b,c,e (.645) 2.410a,b,c,e (.935)

Following the mass media (2) Rumination (2) (3) Background knowledge about WTC and Pentagon Background knowledge on terrorist acts (2) Attitude Memory for eventrelated facts (2)

.047 .024 .046

111.15** (5, 2891)

.161

.530a (.884)

-.410a,b (.808)

-.286a,c (1.092)

-.475a,c,d (.842)

.293a,b,c,d,e (.870)

-.210a,e (.917)

50.85** (5, 2535) 43.86** (5, 2841) 24.84** (5, 2535)

.091

.264a (1.088) -.175a (.793) .384a (.720)

-.471a,b (.792) .104a,b (1.006) -.088a,b (.980)

-.335a,c (.831) .132a,c (.954) -.385a,b,c (1. 230)

-.351a,d (.823) .310a,c (1.212) -.630a,b,c,d (1.024)

.115b,c,d (1.020) -.492a,b,c (.686) .365b,c,d,e (.693)

.427a,b,c,d (.503) -.326b,c (.824) -.075a,c,d,e (.688)

.072 .047

* 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. Table 2. ANOVAs by citizenship (non-US countries) on the CatPCA Optimal Scores

of these comparisons. All comparisons were significant at .001 alpha level, except that memory for the reception context was significant at .005 alpha level. Overall, the effect sizes for the comparisons were not high, except for the ANOVAs on emotional feeling states (.108), and background knowledge about WTC and Pentagon (.161), thus indicating that the differences among countries were strong only for those variables. The effect of the covariate (elapsed time) was found to be significant for all variables except for memory for the reception context, novelty, emotional feeling states, background knowledge about

WTC and Pentagon, and attitudes. This means that the ratings corresponding to these constructs did not seem to depend on when the questionnaire was completed. A ceiling effect was expected for memory for the reception context, novelty, and surprise. On memory for the reception context, the only difference was that Romanians scored significantly lower than Dutch respondents, but the effect size was extremely low (.006). Large differences were only observed, as expected, for Turkish people, who had the lowest coefficients for novelty and surprise. 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 countries. The lowest ratings of importance were found for Dutch, French-speaking, and Turkish respondents. As predicted, participants from individualistic countries (Frenchspeaking countries, Italy, and The Netherlands) scored higher than participants from collectivistic countries (Japan and Turkey) on emotional feeling states, social sharing, and rumination – planned comparisons were significant at .001 alpha level. As for rehearsal through the mass media, planned contrasts were consistent with the expectation that Romanian and Turkish respondents would have had the lowest access to the media. As a consequence, Romanian and Turkish respondents were also expected to exhibit very low levels of background knowledge about the WTC and Pentagon. This prediction was confirmed through a planned contrast, significant at .001 alpha level. However, Italians also scored significantly low on this variable. On the other hand, Turkish participants showed the highest background knowledge about terrorist acts compared to participants from the other involved countries. As expected, Turkish respondents exhibited very low scores on attitudes towards the US. However, the lowest scores were found for Dutch respondents. Finally, the planned contrasts confirmed the prediction of low memory for event-related facts for respondents from Turkey and Romania. However, Japanese people also showed significantly lower memory for eventrelated facts compared to respondents from the other involved countries. DISCUSSION Comparison between US and non-US Countries Confirming our predictions, only very small differences emerged for the memory for the reception context (the effect size was below 1%). There was a ceiling effect across the two samples, with response rate above 95% in the US group and above 90% in the non-US one. Both groups reported remembering most 218

of the details of how they heard the news. This ceiling effect is similar to other studies assessing immediate memory for the reception context (e.g., Christianson & Engelberg, 1999; Conway et al., 1994; Curci et al., 2001; Pezdek, 2003; Schmolk et al., 2000; Wright, 1993). We will now turn to the major aim of the present study, i.e., examining group differences in predictors of flashbulb memories. Moderate and large differences will be examined first, followed by small ones. Three findings fall into the category of moderate to large differences. First, Americans rated the events as much more important to them. More specifically, Americans rated the events as having greater national and personal importance than did nonUS respondents. US respondents were more likely to have known someone killed or injured, and the attacks involved more changes in daily activities in the US than in any other country. Finally, financial and political consequences of the events were more prevalent in the US (e.g., bankruptcies of airline companies, the new priorities of the US government towards military exercises and expenses). US respondents showed greater background knowledge, on both relevant composite scores. These differences can be explained by different factors. Even prior to the attacks, US respondents likely knew more basic facts about the WTC and the Pentagon. Such facts were also salient in media coverage, as they were crucial for understanding event implications (e.g., information about the number of floors in the WTC was key to understanding missing peoples’ fates). Item effects support the hypothesis that people learned some of these facts as events unfolded. For example, people were much better at listing the number of floors in the WTC (59% correct) than its height (30-32%). These two facts refer to the same thing: the size of the building. However, media coverage framed the planes’ points of impact in terms of floors, not feet or meters. Interestingly, US vs. nonUS differences were larger for number of

Hearing the News of the 9/11 Attacks floors (72 vs. 55.5%) than for height of the towers (33 vs. 23.7%). Again, Americans were more likely to know people who worked in the buildings (and knew their location in terms of floors) and thus needed to rapidly assess where these people were located in relation to the critical floors. Finally, some knowledge was likely learned after the events. In particular, knowledge about terrorist acts likely increased after the attacks. US citizens were probably more motivated than non-US ones to learn about such issues in the aftermath of their national tragedy. The largest difference between US and non-US respondents was in memory for eventrelated facts. US respondents found the events more important, and had more background knowledge - , two variables that the literature suggests are antecedents of memory for eventrelated facts (e.g., Finkenauer et al., 1998). Higher event-memory could thus be the result of a cumulative advantage of US respondents for these antecedents. Both groups assessed the events as very novel and surprising. These results are consistent with media labels of the 9/11 attacks as distinctive and unpredictable. The terrorists sent no explicit warnings, and no events in particular suggested such an attack was in the works. US and non-US groups also rehearsed the event to a similar extent, again pointing to the unique nature of the events. Both groups also had similar emotional reactions to the attacks. Although raw scores for emotional feeling states were high, they did not reach a ceiling level, as novelty and surprise did. This can be explained by variations within the items assessing the emotional feeling states. While overall means were slightly higher for US than for non- US citizens, it is interesting to note that large variations were observed across countries on specific items. Comparison between non-US Countries Major differences were observed across non-US countries for three variables that have a direct or an indirect effect on the formation of flashbulb memories: emotional

feeling states, attitudes and background knowledge. Only very small differences were observed, however, for memory for the reception context, novelty, surprise, and importance. For each of these variables, ceiling levels were observed -- indicating that the 9/11 attacks had a major international impact that was unaffected by economic and cultural differences. For each country or group of countries, we will now examine the relation of the results to predictions. As predicted, the group of Western European countries (French-speaking countries, Italy and The Netherlands) had high scores on almost all the variables investigated. For each of these groups, only one or two predictions out of 12 were not confirmed (i.e., French-speaking respondents rated importance as low, Italians showed low levels of background knowledge and citizens of The Netherlands found the events less important than other Western Europeans and also had less positive attitudes towards US). Thus, we can conclude that, as expected, Western European citizens were similar to Americans in their reactions, although the intensity of their reactions was sometimes lower. Due to their lack of experience with terrorism, and their cultural similarity, Romanians was expected to react like Western Europeans in their ratings of novelty, surprise and emotional feeling states. These results were confirmed. However, Romanians were expected to differ from Western Europeans on variables more directly related to their economic status, such as following the media, background knowledge and for the two types of memories (event-related facts and reception context). Again, these predictions were fully supported. The only unexpected finding was related to low levels of social sharing and rumination. This finding may be due to the delayed data collection in Romania ; Romanians responded on average 30 days after the attacks whereas Western Europeans responded after 15 days. As these questions specifically required assessment of behavior in the past three days , it is not surprising that 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 Japanese group. They responded similarly to Western Europeans on variables related to economic development (following the media and attitudes). In contrast, Japanese respondents reported lower levels of emotional feeling states, social sharing, and rumination. These results are consistent with findings showing that collectivistic countries report less intense immediate and long-term emotional reactions (e.g., Basabe et al., 2000, 2002; Fernandez et al., 2000). Unexpectedly, however, Japanese respondents reported less background knowledge and memory for eventrelated facts. One explanation is that the Japanese may have been less interested in the 9/11 events than citizens of other countries. This hypothesis is supported by differences in the mean level of “interest” in the news, which respondents rated on a 7-point scale (from 1, “not at all”, to 7, “very much”). The mean level for Japanese respondents was only 4.77, while all other respondents scored close to or above 6. Finally, as expected, Turkish respondents displayed the most differences from the other respondents Turkish respondents had particularly low scores on novelty, surprise, emotional feeling states, social sharing, rumination, following the media, attitudes, and background knowledge about the WTC and the Pentagon. Smaller differences were observed for memory for the reception context and event-related facts. Overall, the pattern of results can be explained by considering the Turkish experience. That is, Turkey had a number of terrorist attacks in the last few years, and also a serious earthquake. In both cases, many thousands were killed, likely habituating Turkish citizens to such catastrophes (and thus reducing their responses to the 9/11 attacks). Turkey’s history of terrorist attacks (and its geographical proximity to many terrorist bases) likely played a role in their high level of knowledge about terrorist groups (M = 83.8% correct, as compared to a sample mean of 31.6%). Turkish respondents’ knowledge about terrorism may have help them to encode and 220

assimilate the news about the attacks (see Conway et al., 1994; Finkenauer et al., 1998), leading to average levels observed for memory for the reception-context and for memory for event-related facts. Cultural and economic differences help us to understand the other differences. For example, the collectivistic nature of the country might explain low scores on emotional feeling states, social sharing and rumination (e.g., Basabe et al., 2000, 2002; Fernandez et al., 2000). Additionally, this country shares a different religion and different values as compared to Western countries, which might have exacerbated differences. Turkey is also less economically developed, likely reducing media coverage of the events. Turkey has also suffered economically from US imposed prohibitions on trade with countries accused of helping terrorist acts immediately after the 9/11 attacks ; this likely plays into the less positive attitudes held by the Turkish respondents. Limitations and Strengths of the Study Studies of natural events are difficult to do; the data must be collected soon after the surprising event and thus there is little time to plan what data are of interest and how they should be obtained. Methodologically, the biggest weakness in the current study was the occurrence of a lag between when the event occurred and when the data were collected, and that this lag varied across countries. However, by introducing elapsed time between 9/11/01 and questionnaire as a covariate, we controlled for this difference in the analyses. Methodologically, the strength of the study was its use of a questionnaire validated by its use in previous studies (Curci et al., 2001; Finkenauer et al., 1998). One advantage of this questionnaire is that it assesses each predictor with multiple items. It thus allows for computing composite scores, which are more reliable than single items. In addition, in comparison to previous studies investigating memory for the reception context, the present one has the advantage of assessing a larger set of predictors. To our knowledge, no previous study investigated 12 predictors

Hearing the News of the 9/11 Attacks simultaneously as we did in the present study. Even fewer studies have examined group differences on the different predictor variables. For example, only few have examined group differences in memory for event-related facts (see Curci et al., 2001 for an exception). Such differences can be critical as memory for event-related facts is a direct predictor for flashbulb memories (e.g., Finkenauer et al., 1998). The other major methodological strength is of course the number and nature of the comparison groups. While previous studies usually involved only one or two comparison groups (Brown & Kulik, 1977; Conway et al., 1994; Curci et al., 2001), eight non-US national groups were involved in the present study. Also previous comparison studies involved other national groups that were close economically and culturally (e.g., UK vs. US and Denmark for Conway et al., 1994; France and Belgium for Curci et al., 2001). In the present study, we had a first group of countries economically and culturally close to the US (French-speaking countries, Italy and The Netherlands), one country culturally close but economically different (Romania), one country economically close but culturally different (Japan), and one country with both large cultural and economic differences (Turkey). Another major improvement relates to the nature of the event investigated. Previous studies did not use events with high scores on all the major predictors of the formation of flashbulb memories. In contrast, our data confirm that all predictors were high for Americans remembering 9/11 . Almost all American respondents had a vivid memory for the circumstances in which they learned about the news . Similarly, they were at ceiling level in ratings of novelty and surprise, which is in sharp contrast with previous studies investigating the formation of flashbulb memories (e.g., Christianson & Engleberg, 1999; Conway et al., 1994; Curci et al., 2001; Pillemer, 1984). Third, many studies that have examined the formation of flashbulb memories were criticized for their lack of personal relevance, with only national or international

importance reaching high levels (e.g., Conway, 1995), while personal importance has been defined as the key feature (Brown and Kulik, 1977). In contrast to many other studies in which personal importance was either low (e.g., Schmolk et al., 2000; Wright, 1993) or not assessed (e.g., Bohannon, 1988; Christianson, 1989; McCloskey et al., 1988; Neisser & Harsch, 1992; Neisser et al., 1996), our subjects clearly rated the events of 9/11 as personally important (Ms above 5 on scales ranging from 1 to 7). Fourth, mean levels for emotional reactions and rehearsal were remarkably high, while some previous studies failed to find high levels on these variables (e.g., Conway et al., 1994; Pillemer, 1984). Fifth, memory for event-related facts reached high levels in the US sample. Out of 10 variables investigated, four had a correct answer rate above 90% and seven were above 70%. In the US sample, background knowledge was the only predictor for which some ratings were low-- but the overall mean level was still in the middle range. The present event is thus unique as it is characterized by high ratings on all the required characteristics for the formation and maintenance of flashbulb memories. Finally, we wish to note an empirical strength of the current study. A statistical procedure called CatPCA was employed that allowed a scoring system analogous to the standard WAS procedure used by Neisser & Harsch (1992) and Conway and his colleagues (1994; 1995). This procedure assigns different weights to the different attributes of the reception context and, in general, to all indicator variables in the questionnaire. The procedure is aimed at creating composite scores corresponding to the considered theoretical constructs. Unlike the WAS procedure, the weights assigned to the indicator variables by CatPCA were not decided a priori by the researcher, but came from the empirical distributions of the response categories for each indicator variable.

221

O. Luminet, A. Curci, E. J. Marsh, I.Wessel, T. Constantin, F. Gencoz, and M. Yogo REFERENCES Basabe, N., Paez, D., Valencia, J., Rimé, B., Pennebaker, J.W., Diener, E., & Gonzalez, J. L. (2000). Sociocultural factors predicting subjective experience of emotion: A collective level analysis. Psicothema, 12, 55-69. Basabe, N., Paez, D., Valencia, J., Gonzalez, J. L., Rimé, B., & Diener, E. (2002). Cultural dimensions, socioeconomic development, climate, and emotional hedonic level. Cognition and Emotion, 16, 103-125. Bohannon, J. N., III (1988). Flashbulb memories for the space shuttle disaster: A tale of two theories. Cognition, 29, 179196. Bohannon, J. N. & Symons, V. L. (1992). Flashbulb memories: Confidence, consistency, and quality. In E. Winograd & U. Neisser (Eds.), Affect and Accuracy in Recall. Studies of “flashbulb memories” (pp. 65-91). New York: Cambridge University Press. Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 73-99. Christianson, S-Å. (1989). Flashbulb memories: Special, but not so special. Memory and Cognition, 17, 435-443. Christianson, S- Å, & Engelberg, E. (1999). Memory and emotional consistency: The MS Estonia ferry disaster. Memory, 7, 471-482. Conway, M. A. (1995). Flashbulb memories. Hove: Erlbaum. Conway, M. A. (2002). Testing the flashbulb memory hypothesis. Unpublished manuscript, University of Bristol, UK. Conway, M. A., Anderson, S. J., Larsen, S. F., Donnelly, C. M., McDaniel, M. A., McClelland, A. G. R., Rawles, R. E., & Logie, R. H. (1994). The formation of flashbulb memories. Memory and Cognition, 22, 326-343. Curci, A., Luminet, O., Finkenauer, C., & Gisle, L. (2001). Flashbulb memories in social groups: A comparative test-retest study of the memory of French President 222

Mitterand’s death in a French and a Belgian group. Memory, 9, 81-101. Fernandez, I., Carrera, P., Sanchez, F., Paez, D., & Candia, L. (2000). Differences between cultures in emotional verbal and non-verbal reactions. Psicothema, 12, 8392. Finkenauer, C., Luminet, O., Gisle, L., ElAhmadi, A., Van Der Linden, M., & Philippot, P. (1998). Flashbulb memories and the underlying mechanisms of their formation: Toward an emotionalintegrative model. Memory and Cognition, 26, 516-531. Greenacre, M. J. (1993). Correspondence analysis in practice. London: Academic Press. McCloskey, M., Wible, C. G. & Cohen, N. J. (1988). Is the a special flashbulb-memory mechanism? Journal of Experimental Psychology, 117, 171-181. Neisser, U. (1982). Snapshots or benchmaks? In U. Neisser (Ed.), Memory observed (pp. 43-48). San Francisco: Freeman. Neisser, U., & Harsch, N. (1992). Phantom flashbulbs: False recollections of hearing the news about Challenger. In E. Winograd & U. Neisser (Eds.) Affect and accuracy in recall: Studies of ‘flashbulb’ memories (pp. 9-31). New York: Cambridge University Press. Neisser, U., Winograd, E., Bergman, E. T., Schreiber, C. A., Palmer, S. E., & Weldon, M. S. (1996). Remembering the earthquake: Direct experience vs. hearing the news. Memory, 4, 337-357. Oatley, K., & Johnson-Laird, P. N. (1987). Towards a cognitive theory of emotion. Cognition and Emotion, 1, 29-50. Pezdek, K. (2003). Memory for the events of September 11, 2001. Manuscript submitted for publication. Pillemer, D. B. (1984). Flashbulb memories of the assassination attempt on President Reagan. Cognition, 16, 63-80. Rubin, D. C. & Kozin, M. (1984). Vivid memories. Cognition, 16, 81-95. Schmolck, H., Buffalo, E. A., & Squire, L. R. (2000). Memory distortions develop over

Hearing the News of the 9/11 Attacks time: Recollections of the O. J. Simpson trial verdict after 15 and 32 months. Psychological Science, 11, 39-45. Tulving, E., & Kroll, N. (1995). Novelty assessment in the brain and long-term memory encoding. Psychonomic Bulletin and Review, 2, 387-390. Wright, D. B. (1993). Recall of the Hillsborough disaster over time: Systematic biases of ‘flashbulb’ memories. Applied Cognitive Psychology, 7, 129-138.

223