Eyewitness Memory

R E S E A R CH A R T I C L E

Cross-cultural differences in eyewitness memory reports

Nkansah Anakwah1,2 | Robert Horselenberg1 | Lorraine Hope2 |

Margaret Amankwah-Poku3 | Peter J. van Koppen1,4

1Department of Criminal Law and Criminology,

Maastricht University, Maastricht, The

Netherlands

2Department of Psychology, University of

Portsmouth, Portsmouth, UK

3Department of Psychology, University of

Ghana, Legon, Ghana

4Department of Criminal Law and Criminology,

VU University, Amsterdam, The Netherlands

Correspondence

Nkansah Anakwah, Faculty of Law, Maastricht

University, Bouillonstraat 3, Maastricht 6211

LH, The Netherlands.

Email: nkansah.anakwah@

maastrichtuniversity.nl

Funding information

Erasmus Mundus Joint Doctorate Program The

House of Legal Psychology (EMJD-LP), Grant/

Award Numbers: 2013-0036, 532473-EM-

5-2017-1-NL-ERA MUNDUS-EPJD

Summary

Increasingly, investigators conduct interviews with eyewitnesses from different cul-

tures. The culture in which people have been socialised can impact the way they

encode, remember, and report information about their experiences. We examined

whether eyewitness memory reports of mock witnesses from collectivistic (sub-

Saharan Africa) and individualistic (Northern Europe) cultures differed regarding quan-

tity and quality of central and background details reported. Mock witnesses (total

N = 200) from rural Ghana, urban Ghana, and the Netherlands were shown stimuli

scenes of crimes in Dutch and Ghanaian settings and provided free and cued recalls.

Individualistic culture mock witnesses reported the most details, irrespective of detail

type. For each cultural group, mock witnesses reported more correct central details

when crime was witnessed in their own native setting than a non-native setting,

though for different recall domains. The findings provide insight for legal and investiga-

tive professionals as well as immigration officials eliciting memory reports in cross-

cultural contexts.

K E YWORD S

cultural differences, eyewitness memory reports, individualism–collectivism, interview

1 | INTRODUCTION

An international criminal tribunal, The Special Court for Sierra

Leone, put Charles Taylor, a former president of the West African

state of Liberia, on trial in The Hague. Taylor was accused of war

crimes, crimes against humanity, and violations of international

human rights law during the civil war in Sierra Leone. He was alleged

to have supplied arms to rebel groups in Sierra Leone in exchange of

diamonds and also to have been involved in the massacre of many

innocent people. In the legal proceedings that ensued in his trial,

eyewitness evidence from sub-Saharan African witnesses was

instrumental (Keith, 2012). As in the trial of Taylor, eyewitness

memory reports in international criminal settings are crucial in pros-

ecuting alleged atrocities. However, due to the cross-cultural con-

text of international criminal settings, investigators who interview

witnesses in such settings may find it challenging, particularly if

insight into culturally determined reporting norms of the witnesses

is limited.

Aside from international criminal settings, the increase in interna-

tional migration has made it more likely that legal and investigative

professionals in different countries will need to obtain eyewitness

memory reports in cross-cultural contexts. For instance, police detec-

tives are increasingly likely to interview eyewitnesses from cultural

backgrounds different to their own. In other contexts, immigration

officials typically interview asylum seekers from different cultures

about their recollections of events and locations in order to verify

their claims (van Veldhuizen, Maas, Horselenberg, & van Koppen,

2018). Irrespective of the case type, such interviewees will have been

socialised into their respective cultures, and embedded in these cul-

tures are norms (Hofstede, 2001). Various cultural norms may have

Received: 8 May 2019 Revised: 20 December 2019 Accepted: 8 January 2020

DOI: 10.1002/acp.3637

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium,

provided the original work is properly cited.

© 2020 The Authors. Applied Cognitive Psychology published by John Wiley & Sons Ltd.

504 Appl Cognit Psychol. 2020;34:504–515.wileyonlinelibrary.com/journal/acp

implications for how people view, remember, and report about their

experiences and how they behave in the course of cross-cultural

interactions (Kastanakis & Voyer, 2014; Wang, Song, & Kim Koh,

2017). Hence, it is entirely possible that witnesses, victims, and other

interviewees reflect culturally determined reporting norms when

being questioned in legal and forensic contexts. Therefore, an

increased understanding of the impact of cross-cultural differences on

interviews in forensic settings is vital (see Hope & Gabbert, 2019).

The culture in which people have been socialised has been shown

to impact both behaviour and psychological processes (Schwartz,

Boduroglu, & Gutchess, 2014; Wang, 2004). The individualism–

collectivism cultural dimension has been particularly influential in

research exploring cross-cultural differences across various social phe-

nomena (Triandis, 2001; Triandis, Bontempo, Villareal, Asai, & Lucca,

1988). Individualism refers to a cultural orientation where the ties

between individuals in a society are relatively loose, whereas collectiv-

ism refers to a cultural orientation where a person is embedded in a

complex web of social relationships (Hofstede, 1983). The

individualism–collectivism cultural dimension may lead to biases in what

is considered worthy and informative to report when people from these

cultures are exposed to similar scenes (Boduroglu, Priti, & Nisbett,

2009). For example, drawing on the individualism–collectivism dimen-

sion, Markus and Kitayama (1991, 2003) proposed independent–

interdependent cognitive styles. According to Markus and Kitayama

(1991, 2003), an independent construal of the self is a characteristic of

individualistic societies and features the self as having significant dispo-

sitional attributes, and as being more autonomous and independent. For

that reason, individuals with an independent self-construal become

more perceptually oriented towards the properties of an object than

the context (analytic perception). Accordingly, they become more prone

to attend to the properties and characteristics of an object and as a

result, narrow their attentional resources to focal objects at a visual field

(Boduroglu et al., 2009). In contrast, an interdependent construal of the

self, whereby individuals view the self as integrated with (i.e., not sepa-

rate from) the social context, is proposed as a characteristic of collectiv-

istic cultures. Markus and Kitayama (1991) argued that due to the

interdependent self-construal, people from collectivistic cultures

become more perceptually oriented towards a broader visual field

(holistic perception) and, as a result, are more likely to allocate their

attentional resources broadly. Applying Markus and Kitayama's (1991)

framework, it might be predicted that reports about events by people

from individualistic and collectivistic cultures may differ as their cultural

background biases them to be either analytically or holistically oriented.

Aside from an individual's cultural background, it has also been

suggested that the characteristics of a cultural setting could direct atten-

tion (Masuda & Nisbett, 2006; Miyamoto, Nisbett, & Masuda, 2006).

Proponents of that perspective have argued that irrespective of their cul-

tural background, individuals are likely to detect changes to focal objects

of scenes from individualistic cultures than scenes from collectivistic cul-

tures. Conversely, they argue that individuals, regardless of their cultural

background, are more likely to detect changes to contextual objects for

scenes from collectivistic cultures than scenes from individualistic cul-

tures (Masuda & Nisbett, 2006). Other researchers have observed a

tendency for familiar environments to modulate the processing of visual

stimuli (Epstein, Higgins, Jablonski, & Feiler, 2007). According to Epstein

et al. (2007), people activate long-term representations of spatial struc-

tures of familiar environments to aid recall. Therefore, it is plausible that

eyewitnesses are likely to have superior performance when attending to

scenes in their native cultural environment (own-setting effect) than

scenes located in a different cultural environment.

Consistent with these perspectives, research suggests that individ-

uals' cultural orientation can bias their perceptual processing and content

of their reports (Boduroglu et al., 2009; Istomin, Panáková, & Heady,

2014; Masuda & Nisbett, 2006). For example, in a study comparing chil-

dren from three Siberian cultures, Istomin et al. (2014) found children

from the two cultures with holistic perception included more contextual

information in their drawings than those from the culture with analytic

perception. They also found that children from a collectivistic cultural ori-

entation tend to draw background objects before drawing focal objects,

whereas the reverse was true for those from individualistic cultural ori-

entation. Istomin et al. (2014) attributed these findings to differences in

attention that the different cultures accord to contextual information.

However, other results have been inconclusive with respect to

cultural differences in memory reporting. For example, Wong, Yin,

Yang, Li, and Spaniol (2017) compared Canadian and Chinese partici-

pants with respect to memory for individual and background objects

of picture scenes. Participants were exposed to picture drawings con-

taining focal and background scenes and later reported whether they

attended to the focal or background scene. Irrespective of partici-

pants' cultural background, participants reported attending more to

focal details than background details, and there was no difference in

memory for focal objects between cultures. However, Canadian par-

ticipants reported attending more to background scenes than Chinese

participants did. Thus, there seem to be mixed findings on research on

the influence of culture on memory.

1.1 | The current research

Increasingly, investigators interview witnesses from diverse cultural

backgrounds, and given that cultural norms may influence the nature or

content of the information reported in such interviews, this may have

implications for the criminal justice system. Criminal justice profes-

sionals can be confronted with challenges when they lack the relevant

awareness, knowledge, and training about cultural differences in eye-

witness memory reports. To date, research in this area has largely been

conducted using Western, Educated, Industrialised, Rich and Demo-

cratic (WEIRD; Henrich, Heine, & Norenzayan, 2010) samples, with little

consideration of cross-cultural factors or comparisons. There have been

calls for cross-cultural research to go beyond Western borders to

enhance our understanding of cultural variations in behaviour (Brady,

Fryberg, & Shoda, 2018; Gelfand, Harrington, & Jackson, 2017) and,

more specifically, to appreciate cultural differences relevant for the field

of investigative interviewing (Hope & Gabbert, 2019).

Consequently, efforts are being made in psychological science to

explore other non-WEIRD samples. However, a recent meta-analytic

ANAKWAH ET AL. 505

review revealed that, even for the small proportion of non-WEIRD

populations studied in cross-cultural research, the majority of these

non-WEIRD populations were from East Asia (collectivistic culture),

with only 0.63% of the non-WEIRD sample populations from Africa

(Veillard, 2017). Hence, in the current study, we sampled participants

from sub-Saharan Africa (typifying collectivistic culture) and Western

Europe (typifying individualistic culture). Within the collectivistic cul-

ture, we were also interested in comparing rural and urban cultures,

as the latter tends to be less collectivistic than the former (Rooks,

Klyver, & Sserwanga, 2016). This difference is likely due to the fact

that urban centres are prone to cultural infiltration, and there is

greater exposure to western cultural values in urban areas than in

rural areas (Ma, Pei, Jin, & De Wit, 2015). To date, the literature on

cross-cultural cognition has rarely made the distinction between rural

and urban dwellers in collectivistic societies. To address this issue in

the current research, we compared eyewitnesses from Western

European culture with eyewitnesses from urban and rural sub-Saharan

African cultures.

Mock witnesses from sub-Saharan Africa and Western Europe

viewed stimuli scenes presenting African and European settings and

reported what they saw in a free recall test. Afterwards, they were

asked cued recall questions that focused on both central and back-

ground details of the scenes. Drawing on theory and previous findings,

we predicted cultural differences in the types of details reported by the

cultural groups. Specifically, we expected Western European mock wit-

nesses to report more focal details about the crime scene than sub-

Saharan African mock witnesses. Conversely, we expected sub-Saharan

African mock witnesses to report more contextual details than Western

European mock witnesses. Among sub-Saharan African mock wit-

nesses, we expected differences between mock witnesses from rural

and urban areas in the type of details reported. Specifically, we

predicted that those from rural sub-Saharan Africa would report more

contextual details than those from urban sub-Saharan Africa. Mock wit-

nesses from urban sub-Saharan Africa were expected to report more

focal details than those from rural sub-Saharan Africa. We also

expected that cultural setting would play a role in the memory reports

of mock witnesses of all cultural groups. Specifically, we predicted that

mock witnesses across cultures would report more central details about

Western European cultural settings than sub-Saharan African cultural

settings. Mock witnesses across cultures were also predicted to report

more background details for sub-Saharan African settings than Western

European settings. Finally, we expected mock witnesses from sub-

Saharan Africa to report more central and background details about

sub-Saharan African settings than Western European settings, whereas

we expected the reverse for mock witnesses fromWestern Europe.

2 | METHOD

2.1 | Participants and design

A total of 207 participants were sampled from Ghana (nrural Ghana = 78;

n urban Ghana

= 73) and the Netherlands (n = 56). The selection of countries

for inclusion is consistent with previous research (Hofstede, 1983,

2001).1 Out of the 207 participants recruited, seven were excluded.

These participants were excluded because they did not follow instruc-

tions (n = 2), viewed only three out of the four scenes (n = 4), and had

East Asian parents although born in the Netherlands (n = 1). Our final

sample comprised 200 participants (103 males and 97 females;

Mage = 28.44, SD = 12.43). The urban sample (n = 70; Mage = 26.39,

SD = 10.79) in Ghana was recruited in the capital city, Accra, whereas

the rural sample (n = 75; Mage = 31.61, SD = 14.29) was recruited in

Akim Aduasa, a farming community in the Eastern Region of Ghana.

Participants from the Netherlands (n = 55; Mage = 26.78, SD = 10.96)

were recruited in Maastricht, a provincial capital in the south of the

country. Student participants in the Netherlands were awarded course

credits, whereas nonstudent participants received a €5 shopping

voucher. Student and nonstudent participants from Ghana received a

GHС| 5 voucher for phone credit.

The design for the study was a 3 (cultural group: rural Ghana,

urban Ghana, The Netherlands) × 2 (crime setting: Ghanaian setting,

Dutch setting) mixed factorial design. The between-group variable

was cultural group, and the within-group variable was crime setting.

The dependent variables were correct, incorrect, and withheld (Don't

know) details, for both central and background information.

2.2 | Materials

2.2.1 | Stimuli

The stimuli used were eight photographs rich in central and back-

ground details. The photographs depicted four crime scenarios (theft,

assault, accident, and robbery). Each of these crime scenarios was

photographed in a Ghanaian setting as well as in a Dutch setting. For

example, for a crime depicting, a theft in a Ghanaian setting, the same

crime was depicted in a Dutch setting. Each participant viewed four

of these stimuli (two stimuli each for Dutch and Ghanaian settings).

The stimuli were prepared in the Netherlands and Ghana. Scenarios

were prepared with a very clear central event that was distinct from

the background.

Two of the stimuli (one Ghanaian setting and one Dutch setting)

were piloted in the respective countries. A total of 14 participants

(nine males and five females, M = 24.07, SD = 3.20) from Ghana and

15 participants (four males and 11 females, M = 30.40, SD = 13.12)

from the Netherlands provided ratings, using a 5-point Likert scale.

They rated the extent the stimulus (a) represented their native setting

and (b) represented a crime scene. Consistent with Paz-Alonso, Good-

man, and Ibabe (2013), the mid-rating score was used in deciding

whether a stimulus received sufficient rating. The stimuli settings

were rated by participants to adequately represent settings in their

respective countries (Ghanaian stimuli—M = 3.79, SD = .97; Dutch

stimuli—M = 3.33, SD = .62) and reflect plausible crime scenes

(Ghanaian stimuli—M = 3.43, SD = 1.28; Dutch stimuli—M = 3.47,

SD = .83). The pilot study also determined which details participants

regarded as central and background details in each scene. To establish

ANAKWAH ET AL.506

stimulus centrality, the participants were asked two open-ended ques-

tions: “What do you regard as the central event in the picture?” and

“What do you regard as background event(s) in the picture?.” All par-

ticipants identified the central and contextual events in a manner con-

sistent with our intended central and contextual elements when

constructing the stimuli (with the exception of one participant who

did not identify central event for the Ghanaian stimuli as such).

Results from this pilot informed the development of the remaining

stimuli with Ghanaian and Dutch settings, which were developed to

have a clear central event distinct from the background. The stimuli

are available on Open Science Framework at https://osf.io/t89hu/?

view_only=59e038117b2d4d5588e00c804de3539a

2.2.2 | Cultural orientation scale

We used the cultural orientation scale (Triandis & Gelfand, 1998) to

measure self-reported individualism and collectivism of participants.

That scale has 16 items with a 9-point Likert scale (1 = never or defi-

nitely no and 9 = always or definitely yes). It has four subscales: vertical

individualism (VI), horizontal individualism (HI), vertical collectivism

(VC), and horizontal collectivism (HC).2 Sample items on the scale

include VI—“winning is everything”; HI—“I often do my own thing”;

VC—“Parents and children must stay together as much as possible”;

and HC—“If a co-worker gets a prize, I would feel proud.” The coeffi-

cient alphas of the subscales range from .62 to .75 (Soh &

Leong, 2002).

2.3 | Procedure

All participants in the study were tested individually. After consenting

to participate, participants completed the cultural orientation scale

and a short demographic questionnaire. Participants then viewed the

stimulus scenes, one at a time. Consistent with previous research

(e.g., Wang & Pomplun, 2012), participants viewed each scene for

5 seconds. After viewing a scene, participants worked on a distractor

task (mathematical problems) for 5 minutes. Participants were then

instructed to provide a verbal free recall describing what they could

remember about the scene they viewed. Participants were asked to

be as detailed and accurate as possible in their reports about the

scene. Participants had up to 6 minutes to provide that account.

After the free recall task, participants answered 20 cued recall

questions about central and background events or items in the stimu-

lus (e.g., “How was the attacker dressed” and “Can you describe the

colour of the building?”). The order of questions alternated between

questions on central and background details. The instructions and

questions for some participants in rural Ghana were given in the local

language (Twi) as these participants had a low level of English

comprehension.3

After completing both recall tasks, participants saw the next

scene, and the procedure was repeated until they had viewed all four

scenes. The presentation of the scenes was counterbalanced.

Participants received the same instructions for all tasks. Participants'

responses were audio recorded. After completing the procedures,

they were thanked and debriefed. The test session took approxi-

mately 60 min per participant. The study received ethical approval

from the Ethics Review Committee Inner City faculties, Maastricht

University, and the Ethics Committee for the Humanities, University

of Ghana.

2.4 | Coding

Verbal responses were transcribed. The interviews conducted in Twi

in rural Ghana were translated into English during the transcription by

one of the research assistants indigenous to the region. A detailed

coding template for each of the stimulus scenes was developed by the

first author and was adapted from previous research (Gabbert,

Hope, & Fisher, 2009; Wright & Holliday, 2007). For the purposes of

our study, details provided by participants were classified as either a

background detail or central detail, in both free and cued recall, adher-

ing a coding manual prepared in advance.4 An item was coded as cor-

rect if it was present in the stimuli scene and given a correct

description. Incorrect items were also coded and scored accordingly.

Vague responses (e.g., it was a red or green bag) or subjective infer-

ences (e.g., the car belonged to the woman lying on the floor) were

not coded. “Don't know” responses were coded as withheld details. A

second coder coded 20% of the transcripts, which were randomly

selected to check for coding consistency. We found high intercoder

reliability (intraclass correlation coefficient) for free recall with regard

to correct central details (r = .97) and correct background details

(r = .95). The details provided by participants were collated across all

stimuli, and analysis was based on data for all scenes.

3 | RESULTS

Analyses were conducted using a mixed factorial analysis of variance

(ANOVA), except analysis on type of detail that dominated in the

memory reports of the cultural groups, where repeated measures

ANOVA was used. Where significance difference existed, we used

Games–Howell multiple comparisons test as this post-hoc test is suit-

able for comparison groups of unequal size (Lee & Lee, 2018). We

applied a Bonferronni correction (.017) to control for increased error

rates arising from multiple tests.

3.1 | Free recall

3.1.1 | Central details

Cultural group had a significant main effect on the number of correct

central details reported, F(2, 197) = 43.02, p < .001, ηp 2 = .30. Partici-

pants from the Netherlands reported significantly more correct central

details than participants from urban Ghana (p = .003), who also reported

ANAKWAH ET AL. 507

significantly more correct central details than participants from rural

Ghana (p < .001; see Table 1). We also found a significant main effect

for crime setting on correct central details, F(1, 197) = 8.78, p = .003,

ηp 2 = .04. Participants reported more correct central details when the

crime scene was a Ghanaian setting (M = 15.91, SD = 7.50) than when it

was a Dutch setting (M = 14.54, SD = 7.35). There was no significant

interaction effect between cultural group and crime setting, F

(2, 197) = 3.28, p = .04, η2p = .03. In order to test evidence in favour of

the null, we proceeded with a Bayesian ANOVA analysis using JASP

(Wagenmakers, 2007). The analysis yielded a Bayes Factor of

BF10 = 2.35 × 1014. According to Raftery (1995), Bayes factor of

150 and above is indicative of very strong evidence in favour of the

alternate hypothesis. A planned comparison revealed both participants

from rural Ghana (p = .019) and urban Ghana (p = .001) significantly

reported more correct central details for Ghanaian crime settings than

Dutch crime settings. Participants from the Netherlands, however, did

not significantly differ in correct central details reported for Ghanaian

and Dutch crime settings (p = .770). Results are shown in Figure 1.

There was a significant main effect of cultural group on the num-

ber of incorrect central details reported, F(2, 197) = 9.27, p < .001,

ηp 2 = .09. Participants from rural Ghana reported significantly fewer

incorrect central details than participants from the Netherlands

(p = .001). Participants from urban Ghana and the Netherlands did not

significantly differ in incorrect central details reported (p = .055). Par-

ticipants from rural Ghana and urban Ghana also did not significantly

differ in incorrect central details reported (p = .146; see Table 1).

Crime setting did not have a significant effect on incorrect central

details F(1, 197) = 3.80, p = .05, ηp 2 = .02. The interaction effect for

cultural group and crime setting for incorrect central details was not

significant, F(2, 197) = 2.05, p = .13, ηp 2 = .02.

3.1.2 | Background details

There was a significant main effect of cultural group on the reporting of

correct background details F(2, 197) = 45.35, p < .001, ηp 2 = .32. Partici-

pants from the Netherlands reported more correct background details

than participants from urban Ghana (p = .002). Participants from urban

Ghana also reported more correct background details than participants

from rural Ghana (p < .001; see Table 1). There was also a significant

main effect for crime setting, F(1, 197) = 38.03, p < .001, ηp 2 = .16. Par-

ticipants reported more correct background details for crime scenes

with Dutch settings (M = 9.22, SD = 5.94) than Ghanaian settings

(M = 6.93, SD = 4.81). However, the interaction between cultural group

and crime setting was not significant, F(2, 197) = .94, p = .39, ηp 2 = .01.

Cultural group had no significant main effect on incorrect back-

ground details reported, F(2, 197) = .47, p = .62, ηp 2 = .01. Crime set-

ting also had no significant main effect on incorrect background

details reported, F(1, 197) = .33, p = .57, ηp 2 = .00. The interaction

between cultural group and crime setting on incorrect background

details was also not significant F(2, 197) = 1.13, p = .33, ηp 2 = .01.

3.1.3 | Type of detail reported

We examined the total (correct and incorrect) amount of central and

background details reported by each group. Participants from the

Netherlands reported more central details than background details, F

(1, 54) = 93.25, p < .001, ηp 2 = .63. A similar pattern was found for

participants from urban Ghana who also reported more central details

than background details, F(1, 69) = 100.85, p < .001, ηp 2 = .59. Partici-

pants from rural Ghana also reported more central details than

TABLE 1 Mean (standard deviation) correct, incorrect, and withheld central and background details reported in free and cued recall by cultural groups

Rural Ghana Urban Ghana The Netherlands

Free recall Correct Central 9.71 (6.58) 15.57 (6.53) 20.39 (5.78)

Background 3.99 (4.76) 8.36 (4.69) 11.87 (4.75)

Incorrect Central 1.13 (1.39) 1.51 (1.34) 2.17 (1.33)

Background .78 (1.13) .91 (1.09) .96 (1.11)

Cued recall Correct Central 13. 47 (5.72) 18.06 (5.69) 22.77 (5.71)

Background 4.59 (3.38) 6.26 (3.35) 10.73 (3.34)

Incorrect Central 5.59 (2.25) 5.15 (2.26) 6.10 (2.30)

Background 4.27 (2.51) 4.27 (2.51) 5.73 (2.52)

Withheld Central 6.25 (4.16) 5.54 (4.18) 3.75 (4.15)

Background 10.54 (3.64) 10.13 (3.68) 7.18 (3.63)

0

5

10

15

20

25

Rural Ghana Urban Ghana The Netherlands

M e a n c

o rr

e c t

d e ta

ils

Cultural Group

Ghanaian Setting Dutch Setting

F IGURE 1 Mean correct details for different crime settings reported across cultural groups under free recall. Error bars represent 95% confidence intervals

ANAKWAH ET AL.508

background details, F(1, 74) = 156.35, p < .001, ηp 2 = .68 (see Table 3).

Although central details dominated in the memory reports of all cul-

tural groups, there was a significant difference in the total amount of

central details reported across cultural groups, F(2, 197) = 43.09,

p < .001, η2p = .30. Participants from the Netherlands significantly

reported more central details than participants from urban Ghana

(p = .002), who also reported more central details than participants

from rural Ghana (p < .001; see Table 3).

3.2 | Cued recall

3.2.1 | Central details

There was a significant main effect of cultural group on correct central

details reported in response to cued recall questions focused on central

details, F(2, 197) = 42.66, p < .001, ηp 2 = .30. Participants from the

Netherlands reported more correct central details than participants

from urban Ghana (p < .001), who also reported more correct central

details than participants from rural Ghana (p < .001; see Table 1). There

was also a significant main effect of crime setting on correct central

details reported, F(1, 197) = 5.82, p = .017, ηp 2 = .03. Participants

reported more correct central details when the crime scene was a

Dutch setting (M = 18.66, SD = 7.21) than when it was a Ghanaian set-

ting (M = 17.55, SD = 6.08). The interaction between cultural group and

crime setting was not significant, F(2, 197) = 2.85, p = .06, ηp 2 = .02. We

proceeded with a Bayesian ANOVA to test for evidence for the null.

We found the Bayes Factor to be BF10 = 7.964 × 1012, indicative of

very strong evidence (Raftery, 1995) in favour of the alternate hypothe-

sis. A planned comparison revealed participants from rural Ghana did

not significantly differ on correct central details reported for Ghanaian

and Dutch crime settings (p = .91). Participants from urban Ghana also

did not significantly differ on correct central details reported for the

two cultural settings (p = .36). However, participants from the Nether-

lands reported more correct central details for Dutch crime settings

than they did for Ghanaian crime settings (p = .01; see Figure 2).

The main effect of cultural group on incorrect central details

reported, in response to questions focused on central details, was not

significant, F(2, 197) = 2.66, p = .07, ηp 2 = .02. There was, however, a

significant main effect of crime setting on incorrect central details

reported, F(1, 197) = 10.16, p = .002, ηp 2 = .05. Participants reported

more incorrect central details when crime setting was a Ghanaian set-

ting (M = 6.02, SD = 2.97) than when it was a Dutch setting (M = 5.21,

SD = 2.83). The interaction between cultural group and crime setting

was not significant, F(2, 197) = .36, p = .699, ηp 2 = .00.

Cultural group had a significant main effect on the central details

withheld by participants, F(2, 197) = 5.97, p = .003, ηp 2 = .06. Partici-

pants from rural Ghana withheld more responses for questions about

central details than participants from the Netherlands (p = .004). Par-

ticipants from urban Ghana also withheld more central details than

participants from the Netherlands (p < .00). Participants from rural

Ghana and urban Ghana did not significantly differ in central details

withheld (p = .619; see Table 1). Crime setting did not have a

significant main effect on the central details withheld by participants,

F(1, 197) = .90, p = .34, ηp 2 = .01. The interaction between cultural

group and crime setting on withheld central details was also not sig-

nificant, F(2, 197) = 1.29, p = .28, η2p = .01.

3.2.2 | Background details

There was a significant main effect of cultural group on correct back-

ground details reported in response to questions about background

details, F(2, 197) = 55.59, p < .001, ηp 2 = .36. Participants from the

Netherlands reported more correct background details than participants

from urban Ghana (p < .001) and rural Ghana (p < .001). Participants

from urban Ghana also reported more correct background details than

participants from rural Ghana (p = .004; see Table 1). The main effect of

crime setting on correct background details reported was significant, F

(1, 197) = 130.51, p < .001, ηp 2 = .40. Participants reported more cor-

rect background details when crime setting was a Dutch setting

(M = 8.95, SD = 4.38) than when it was a Ghanaian setting (M = 5.44,

SD = 3.68). The interaction between cultural group and crime setting

was also significant, F(2, 197) = 15.23, p < .001, ηp 2 = .13. A planned

comparison revealed participants from rural Ghana reported more cor-

rect background details for Dutch settings than Ghanaian settings

(p < .001). Participants from urban Ghana also reported more correct

background details when crime scene was a Dutch setting than Ghana-

ian setting (p < .001). We found a similar pattern for participants from

the Netherlands, who reported more correct background details when

crime setting was a Dutch setting than when it was a Ghanaian setting

(p < .001). The interaction effect for correct background details could

be accounted for by the magnitude of the simple main effect. This is

because, for all cultural groups, the slopes of the simple main effect of

crime setting have the same direction. See Table 2 for descriptive statis-

tics on interaction between cultural group and crime setting.

There was also a significant main effect of cultural group on incor-

rect background details, F(2, 197) = 6.81, p = .001, ηp 2 = .07. Partici-

pants from urban Ghana reported few incorrect background details

than participants from the Netherlands (p = .009). Participants from

rural Ghana also reported few incorrect background details than partici-

pants from the Netherlands (p = .005). Participants from urban Ghana

and rural Ghana did not differ in incorrect background details reported

(p = 1.00; see Table 1). Setting of crime had a significant main effect on

incorrect background details reported, F(1, 197) = 15.29, p < .001,

ηp 2 = .07. Participants reported more incorrect background details for

Dutch crime settings (M = 5.22, SD = 3.25) than Ghanaian crime settings

(M = 4.30, SD = 2.83). The interaction effect between cultural group

and crime setting on incorrect background details reported was not sig-

nificant, F(2, 197) = 1.50, p = .23, ηp 2 = .02.

The analysis also revealed that the main effect of cultural group

on background details withheld by participants was significant, F

(2, 197) = 15.06, p < .001, ηp 2 = .13. Participants from urban Ghana

withheld significantly more responses for questions on background

details than participants from the Netherlands (p < .001). We also

found a similar pattern for participants from rural Ghana, who

ANAKWAH ET AL. 509

withheld significantly more responses to questions on background

details, than participants from the Netherlands (p < .001). No signifi-

cant difference was observed for withheld responses for participants

from rural Ghana and urban Ghana (p = .781; see Table 1). The setting

of crime also had a significant main effect on background details with-

held by participants, F(1, 197) = 54.54, p < .001, ηp 2 = .22. Participants

withheld more background details for Ghanaian crime settings

(M = 10.33, SD = 4.53) than Dutch crime settings (M = 8.24, SD = 3.81).

The interaction effect between cultural group and crime setting for

background details withheld by participants was not significant, F

(2, 197) = 2.47, p = .09, ηp 2 = .02.

3.2.3 | Type of detail reported

The total (correct and incorrect) amount of details reported for central

and background details for each group was compared to find out the

type of detail that dominated in their reports. Participants from rural

Ghana significantly reported more central details than background

details, F(1, 74) = 304.58, p < .001, ηp 2 = .81. Participants from urban

Ghana also significantly reported more central details than background

details, F(1, 69) = 370.02, p < .001, ηp 2 = .84. We found the same pat-

tern for participants from the Netherlands who also significantly

reported more central details than background details, F

(1, 54) = 334.83, p < .001, ηp 2 = .86 (see Table 3). Notwithstanding the

observation that in all cultural groups central details dominated in the

memory reports, the cultural groups significantly differed in amount of

central details reported, F(2, 197) = 44.11, p < .001, ηp 2 = .31. Partici-

pants from the Netherlands significantly reported more central details

than participants from urban Ghana (p < 001), who also reported more

central details than participants from rural Ghana (p < .001).

3.3 | Self-reported cultural orientation

We conducted an exploratory analysis on the self-reported cultural

orientation of participants from the cultural groups. The analysis rev-

ealed that the cultural groups did not differ on horizontal collectivism,

F(2, 197) = .69, p = .50, ηp 2 = .01, but did differ on vertical collectivism,

0

5

10

15

20

25

30

Rural Ghana Urban Ghana The Netherlands

M e a n c

o rr

e c t d e ta

ils

Ghanaian setting Dutch setting

Cultural Group

F IGURE 2 Mean correct details for different crime settings reported across cultural groups under cued recall. Error bars represent 95% confidence intervals

TABLE 2 Mean (standard deviation) correct and incorrect details reported in free and cued recall for cultural groups by crime setting

Rural Ghana Urban Ghana The Netherlands

Ghanaian setting

Dutch setting

Ghanaian setting

Dutch setting

Ghanaian setting

Dutch setting

Free recall Correct Central 10.56 (5.45) 8.87 (5.02) 16.93 (7.80) 14.21 (7.38) 20.25 (9.08) 20.53 (9.29)

Background 3.20 (2.68) 4.79 (4.52) 7.06 (4.86) 9.66 (6.76) 10.55 (6.63) 13.20 (6.40)

Incorrect Central 1.29 (1.55) .97 (1.00) 1.46 (1.73) 1.56 (1.97) 2.53 (2.40) 1.82 (1.86)

Background .63 (1.17) .93 (1.26) .97 (2.02) .86 (1.12) .95 (1.15) .96 (1.41)

Cued

recall

Correct Central 13.51 (5.70) 13.44 (6.70) 17.69 (6.30) 18.44 (7.72) 21.45 (5.85) 24.09 (6.70)

Background 3.72 (3.10) 5.45 (3.86) 4.81 (3.17) 7.71 (4.06) 7.78 (4.54) 13.67 (5.17)

Incorrect Central 5.85 (3.28) 5.33 (2.96) 5.61 (2.47) 4.69 (2.61) 6.60 (3.20) 5.60 (2.86)

Background 3.59 (2.60) 4.96 (3.02) 4.06 (2.76) 4.49 (3.15) 5.25 (3.06) 6.20 (3.37)

Withheld Central 6.45 (11.26) 6.05 (3.05) 4.81 (2.52) 6.27 (2.47) 3.56 (2.49) 3.93 (2.36)

Background 11.96 (4.64) 9.39 (4.16) 10.74 (4.17) 9.51 (3.69) 8.55 (4.61) 5.81 (3.53)

TABLE 3 Mean (standard deviation) of amount of central vs background details for cultural groups under free and cued recall

The Netherlands Urban Ghana Rural Ghana

Central Background Central Background Central Background

M SD M SD M SD M SD M SD M SD

Free recall 45.13 19.21 25.65 11.70 34.16 14.27 18.54 12.34 21.69 9.41 9.55 6.66

Cued recall 57.75 10.97 32.91 9.89 46.43 11.83 20.97 8.02 38.13 12.24 17.64 8.59

ANAKWAH ET AL.510

F(2, 197) = 8.30, p < .001, ηp 2 = .08. Participants from rural Ghana

(M = 29.72, SD = 6.87) significantly scored higher on vertical collectiv-

ism than participants from the Netherlands (M = 26.20, SD = 4.67;

p = .002). Participants from urban Ghana (M = 30.01, SD = 4.99) also

scored higher on self-reported vertical collectivism than participants

from the Netherlands (p < .001). There was no significant difference

between participants from rural Ghana and urban Ghana on vertical

collectivism (p = .95).

There was also a (marginally) significant difference between the

cultural groups on horizontal individualism, F(2, 197) = 3.05, p = .05,

ηp 2 = .03. Participants from rural Ghana (M = 26.83, SD = 6.45) and

the Netherlands (M = 25.84, SD = 4.78) did not differ on scores on

horizontal individualism (p = .57). There was also no significant differ-

ence between participants from rural Ghana and urban Ghana

(M = 28.29, SD = 5.21) on self-reported horizontal individualism

(p = .29). However, there was a significant difference in self-reported

horizontal individualism between participants from urban Ghana and

the Netherlands (p = .02). Participants from urban Ghana gave higher

ratings than participants from the Netherlands on horizontal individu-

alism. The cultural groups significantly differed on self-reported verti-

cal individualism F(2, 197) = 14.86, p < .001, ηp 2 = .13. Participants

from rural Ghana (M = 24.52, SD = 6.80) reported higher scores on

vertical individualism than participants from the Netherlands

(M = 17.98, SD = 7.26; p < .001). Participants from urban Ghana

(M = 23.04, SD = 6.83) also significantly gave higher ratings on vertical

individualism than participants from the Netherlands (p < .001). There

was no significant difference between participants from rural Ghana

and urban Ghana on self-reported vertical individualism (p = .40).

4 | DISCUSSION

We examined eyewitness memory reports of individuals from differ-

ent cultural groups thought to typify individualistic (Western Europe)

and collectivistic (sub-Saharan Africa) cultures. The results appear to

reveal a tendency towards the underreporting of details by sub-

Saharan African mock witnesses. In addition, central details dominated

in the eyewitness memory reports provided across cultures. The

results also showed that in free recall, sub-Saharan African mock wit-

nesses reported more correct central details when the crime scenario

was witnessed in their own native setting than when it was witnessed

in a non-native setting. Western European mock witnesses also

reported more correct central details in cued recall when the crime

scenario was witnessed in their own native setting than a non-native

setting. Mock witnesses from sub-Saharan Africa reported more back-

ground details about a non-native setting than they did for their own

setting under cued recall. Crime context did not appear to affect the

nature of correct background details that Western European Mock

witnesses reported in free recall. However, they reported more cor-

rect background details when crime was witnessed in their own native

setting than a non-native setting in cued recall.

The differences between cultural groups with respect to the

amount of reported details is noteworthy. One possible explanation

for this finding could be elaboration differences due to socialisation

affordances (Peterson, Sales, Rees, & Fivush, 2007). Such a differ-

ence is conspicuous in childrearing practices, where it has been

observed that parents from individualistic cultures provide much

more feedback to their children in conversations than those from

collectivistic cultures (Wang, 2004). It may be the case that differ-

ences in linguistic elaboration are transmitted to children and persist

to later adulthood. Consequently, although eyewitnesses from col-

lectivistic cultures report details about a crime scene, they may not

spontaneously provide a detailed elaboration in their memory narra-

tives. This speculation fits with assertions that individuals from col-

lectivistic cultures report less specific and more generic details than

individuals from individualistic cultures (Millar, Serbun, Vadalia, &

Gutchess, 2013; Wang & Ross, 2005). Similar results have been

observed in research on deception detection, showing interviewees

in individualistic cultures typically report more explicit details than

interviewees from collectivistic cultures (Leal et al., 2018). Leal et al.

(2018) argued that interviewees from collectivistic cultures tend to

leave many things unsaid, allowing the context to communicate

what is implied, whereas in individualistic cultures, the communica-

tion style tends to be more explicit. Therefore, during investigative

interviews, it may be necessary to prompt and encourage eyewit-

nesses from collectivistic cultures to elaborate further on the initial

information they provide.

Apart from the possibility of elaborative differences, it may be the

case that individuals from collectivistic cultures have a tendency to be

more modest or restrained when providing their memorial accounts

than those from individualistic cultures. Cultural differences in self-

effacement and self-enhancement have been documented, with self-

effacement attributed to collectivistic cultures and self-enhancement

attributed to individualistic cultures (Takata, 2003; Yamagishi et al.,

2012). Such differences may reflect cultural disparities in the

independent–interdependent construal of the self (Markus &

Kitayama, 1991). Individuals from cultures with independent construal

of the self are more likely to emphasise the unique attributes of a per-

son. This tendency may be reflected in their self-presentation in

regard to expressing themselves, as they may be inclined to empha-

sise their positive attributes (self-enhancement; Takata, 2003). In con-

trast, individuals from collectivistic cultures, in comparison with

individuals from individualistic cultures, have a tendency to be self-

critical and modest about emphasising their unique attributes (self-

effacement; Heine, Lehman, & Takata, 2000). Therefore, individuals

from collectivistic cultures are more likely to be modest in terms of

self-presentation and expression (Wise, Gong, Safer, & Lee, 2010).

These concepts have been identified as powerful determinants of

behaviour, especially within a social context (Brown & Gallagher,

1992). It is possible for a witness from a collectivistic culture to self-

efface when being interviewed, by being modest in terms of the

extent of the personal memory narrative provided (i.e., providing a

less elaborative or detailed account spontaneously). However, it is

worth noting that this tendency to self-efface may attenuate when

the implications or stakes of self-effacing are high (Yamagishi et al.,

2012). Future research should explore whether this tendency is

ANAKWAH ET AL. 511

attenuated when investigators emphasise the importance of providing

details to pursue an investigation.

In the current study, mock witnesses from the collectivistic cul-

tural groups provided more “Don't Know” responses than those from

the individualistic cultural group. Thus, in this study at least, partici-

pants from collectivistic cultures might have applied a relatively strict

criterion for reporting, and withheld details they remembered but

were not confident about (Cai, Brown, Deng, & Oakes, 2007). This

pattern aligns well with the self-effacing tendency of collectivistic cul-

tures. In a study on self-effacement and self-enhancement among

Canadians and Japanese participants, Heine et al. (2000) found that

although the former were confident they performed well on a test,

the latter were reluctant to admit that they had performed better. It

may be that when sub-Saharan mock witnesses were not confident

about memory for certain details, they simply decided not to report

them. Consistent with this notion is the observation that participants

from Western Europe, who tend to be more assertive and expressive

than people from collectivistic cultures (Matsumoto et al., 2008), pro-

vided more inaccurate responses than participants from sub-Saharan

Africa, which suggests Western European mock witnesses had a

looser threshold for reporting accurate details. Future research should

examine the extent to which there are cultural differences in the

reporting of low-confidence memories.

The social dynamics during the interview may have also played a

role in the amount of information mock witnesses reported, particu-

larly those from sub-Saharan Africa. Individuals from sub-Saharan

Africa have been shown to be high on the cultural dimension of power

distance (Hofstede, 1983). Power distance, another dimension in

which cultures differ, is the extent to which a society endorses hierar-

chy in social relationships (Oyserman, 2006). High power distance

(endorsement of hierarchy in social relationships) may inhibit free and

spontaneous communication when an individual is in a social interac-

tion with an authority figure (Ghosh, 2011). Consistent with this spec-

ulation, in the present study, sub-Saharan African mock witnesses

endorsed more hierarchy in social relationships (vertical collectivism)

than Western European mock witnesses. Therefore, there is a possi-

bility that the mere fact of reporting to an authority or expert (i.e., a

researcher) may have produced cultural differences in the amount of

details provided. Future research should explore the impact of this

dimension further to (a) determine whether in an interview context,

the presence of an authority figure plays a culture-related role in the

amount of information reported by witnesses and (b) explore how

such differences might be attenuated.

None of the cultural groups appear to have processed back-

ground information deeply (cf. central details; Wong et al., 2017) as,

regardless of cultural background, central details dominated in the

memory reports provided. This finding does not align with previous

research suggesting collectivistic cultures attend holistically to a visual

field (Istomin et al., 2014). However, it is worth noting that the stimuli

used in our study were crime scenes and quite different to the stimuli

used in previous research. Previous studies used stimuli such as pic-

tures from the physical environment and artistic representations

(Boduroglu et al., 2009; Miyamoto et al., 2006). The focus of attention

when a crime occurs is likely not the same as any ordinary or neutral

everyday scene. For example, in a robbery, the threatening and

unusual nature of the scene will make it more likely for people at the

scene to attend to this focal event than other activities that may be

going on at the background. The tendency to attend more to notice-

able details at a visual field is well-documented (Loftus & Mackworth,

1978; Masuda & Nisbett, 2006, Experiment 3; Wang &

Pomplun, 2012).

It is also worth noting that past research on culture and visual

attention focused mostly on comparing East Asian and other Western

cultures. Hence, even though African cultures are regarded as collec-

tivistic, the findings for East Asian cultures may not be generalisable

to sub-Saharan Africa. Studies in cross-cultural cognition have largely

studied East Asian cultures, and it may be that the collectivistic self

(interdependent self-construal) may not be a one-size-fits-all phenom-

enon for all collectivistic cultures. This conclusion is consistent with

the notion that collectivism is not a context-free construct (Triandis,

2001). As such, the self-construal for collectivistic cultures may be

context-specific. For example, it has been argued that the

interdependent self-construal among Africans does not suggest a total

loss of the independent self in the collective (Adams & Dzokoto,

2003), and there may be different variations of the interdependent

self-construal among collectivistic cultures. In that vein, the holistic–

analytic categorisation of visual attention across cultures may be rela-

tive. Future research should explore differences between and within

different collectivist cultures.

The current results suggest that the cultural setting in which a

crime is witnessed may also be important when considering eyewit-

ness reports. Mock witnesses reported more correct central details

for Ghanaian crime settings than for Dutch crime settings for free

recall. When cued recall questions were asked, mock witnesses

reported more correct central details for Dutch settings than Ghana-

ian settings. That finding partially aligns with the results of previous

research. For example, Masuda and Nisbett (2006) found that both

participants from individualistic (North America) and collectivistic

(Japan) cultural groups detected focal changes to North American

stimuli scenes quicker than they did for Japanese stimuli scenes. In

the current research, sub-Saharan African mock witnesses reported

more correct central details in free recall, when reporting about crime

witnessed in their own native setting than when it was witnessed in a

non-native setting. This superior performance for crime witnessed in a

native setting was not observed when cued recall questions were

asked. However, Western European mock witnesses reported more

correct central details when the witnessed crime was in their own

native setting than a non-native setting in cued recall, but not for free

recall. The own-setting effect for central details observed for the cul-

tural groups is consistent with work that shows familiar environments

have the tendency to modulate the processing of visual details

(Epstein et al., 2007). However, that explanation does not fit for cor-

rect background details witnessed by sub-Saharan African mock wit-

nesses when crime setting was considered, as sub-Saharan African

mock-witnesses reported more contextual information about a non-

native setting than they did for their own setting in cued recall. We

ANAKWAH ET AL.512

suspect that because the non-native setting was an unfamiliar setting,

participants from sub-Saharan Africa may have attended more to con-

textual information in that setting than they did for their own setting.

Future work should pursue the issue of crime context and how this

relates to reporting in cross-cultural contexts.

There are some limitations associated with the current research.

The first limitation relates to some unavoidable differences in the edu-

cation levels for one of the cultural group samples. Although the

Dutch and urban Ghanaian samples comprised mainly university-level

students with a similar age range and were, as such, well-matched

with respect to education level, this was not the case for the rural

Ghanaian sample. Participants from rural Ghana had a minimal level of

education and were relatively older. Both of these factors may have

affected the performance of this group relative to the other experi-

mental groups—although it is also worth noting that it would likely be

impossible to recruit university-level educated sample in rural Ghana.

Similar issues relating to the difficulty of matching samples across dif-

ferent cultures are common in the cultural literature (Buil, De

Chernatony, & Martínez, 2012). A second possible methodological

concern relates to the test language. As the study instructions were

translated for participants in rural Ghana who lacked adequate com-

prehension of the English language, we do not rule out the possibility

that the translation into a different language may have in some way

affected the outcomes for the rural sample. Finally, we acknowledge

that the static nature of the stimuli used limit generalisability to the

eyewitness context. Typically, crime events involve dynamic move-

ment and action, and the reporting of such information may also vary

culturally. Although static images might be a useful starting point to

examine reporting from memory, future research should adopt the

more typical mock witness paradigm using recorded or live events.

5 | CONCLUSION

In this research, we sought to take the first steps in addressing an

important gap in the eyewitness literature. Specifically, drawing on

samples from sub-Saharan Africa and Western Europe, we examined

eyewitness memory reports for differences predicted by theory in the

cross-cultural literature. Our results show that individuals from indi-

vidualistic cultures provide more details in their account of crime

scene information, irrespective of type of detail. We also found evi-

dence that regardless of the cultural background of eyewitnesses,

central details dominate in their reports of crime scene information.

Finally, we found evidence that the cultural setting in which a crime is

witnessed may play a role in eyewitness memory reports. These find-

ings not only identify important routes for future research in this area

but also highlight the importance of considering the cultural back-

ground of the witness when eliciting memory reports. As such, these

findings should be informative for legal and investigative professionals

working in international criminal justice settings, border and security

practitioners interviewing in asylum, migration and intelligence-

gathering contexts, and law enforcement personnel who regularly

interview witnesses from different cultural backgrounds.

CONFLICT OF INTEREST

The authors declare no conflict of interest

ACKNOWLEDGEMENTS

This research is supported by a fellowship awarded from the Erasmus

Mundus Joint Doctorate Program The House of Legal Psychology

(EMJD-LP) with framework partnership agreement (FPA) 2013-0036

and specific grant agreement (SGA) 532473-EM-5-2017-1-NL-ERA

MUNDUS-EPJD to Nkansah Anakwah. We thank Manouk Vrouch,

George Ofori, and Melody Konadu Frempong for their assistance with

data collection. We also thank Michael Obiri Yeboah for his assistance

in translating the study instructions into Twi.

ORCID

Nkansah Anakwah https://orcid.org/0000-0002-4271-4728

ENDNOTES 1 Hofstede's individualism–collectivism index indicates the extent to

which countries are individualistic and collectivistic. On Hofstede's index

(ranging from 0 to 100), the Netherlands is associated with an individual-

ism index of 80, whereas Ghana is associated with an index of 14, where

a higher score reflects greater individualism. 2 Vertical individualism refers to individualistic cultures where hierarchy is

emphasised in social relationships; horizontal individualism refers to indi-

vidualistic cultures where equality is emphasised in social relationships;

vertical collectivism refers to collectivistic cultures where hierarchy is

emphasised in social relationship; and horizontal collectivism refers to

collectivistic cultures where equality is emphasise in social relationships

(Triandis & Gelfand, 1998). 3 A PhD student in Linguistics with expertise in the Ghanaian language

translated the protocol. The interviewer who also had a good command

of the local language explained the study instructions to these partici-

pants thoroughly and also read the questions out to such participants in

the Twi language. 4 Classification of central and background details in this coding manual

was based on stimulus centrality established in the pilot study earlier

reported.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are openly

available at https://osf.io/9gs78/?view_only=135f537f7bf8437998

906da24a4184c5

ORCID

Nkansah Anakwah https://orcid.org/0000-0002-4271-4728

REFERENCES

Adams, G., & Dzokoto, V. A. (2003). Self and identity in African

studies. Self and Identity, 2, 345–359. https://doi.org/10.1080/

714050252

Boduroglu, A., Priti, S., & Nisbett, R. E. (2009). Cultural differences in allo-

cation of attention in visual information processing. Journal of Cross-

Cultural Psychology, 40, 349–360. https://doi.org/10.1177/

0022022108331005

Brady, L. M., Fryberg, S. A., & Shoda, Y. (2018). Expanding the interpretive

power of psychological science by attending to culture. Proceedings of

ANAKWAH ET AL. 513

the National Academy of Sciences of the United States of America, 115,

11406–11413. https://doi.org/10.1073/pnas.1803526115 Brown, J. D., & Gallagher, F. M. (1992). Coming to terms with failure: Pri-

vate self-enhancement and public self-effacement. Journal of Experi-

mental Social Psychology, 28, 3–22. https://doi.org/https://doi.org/10. 1016/0022-1031(92)90029-J

Buil, I., De Chernatony, L., & Martínez, E. (2012). Methodological issues in

cross-cultural research: An overview and recommendations. Journal of

Targeting, Measurement and Analysis for Marketing, 20, 223–234. https://doi.org/10.1057/jt.2012.18

Cai, H., Brown, J. D., Deng, C., & Oakes, M. A. (2007). Self-esteem and cul-

ture: Differences in cognitive self-evaluations or affective self-regard?

Asian Journal of Social Psychology, 10, 162–170. https://doi.org/10. 1111/j.1467-839X.2007.00222.x

Epstein, R. A., Higgins, J. S., Jablonski, K., & Feiler, A. M. (2007). Visual

scene processing in familiar and unfamiliar environments. Journal of

Neurophysiology, 97, 3670–3683. https://doi.org/10.1152/jn.00003.

2007

Gabbert, F., Hope, L., & Fisher, R. P. (2009). Protecting eyewitness evi-

dence: Examining the efficacy of a self-administered interview tool.

Law and Human Behavior, 33, 298–307. https://doi.org/10.1007/

s10979-008-9146-8

Gelfand, M. J., Harrington, J. R., & Jackson, J. C. (2017). The strength of

social norms across human groups. Perspectives on Psychological Sci-

ence, 12, 800–809. https://doi.org/10.1177/1745691617708631 Ghosh, A. (2011). Power distance in organizational contexts—A review of

collectivist cultures. Indian Journal of Industrial Relations, 47, 89–101. Retrieved from. http://0-search.ebscohost.com.library.regent.edu/

login.aspx?direct=true&db=bth&AN=69815239&site=eds-live

Heine, S. J., Lehman, D. R., & Takata, T. (2000). Beyond self-presentation:

Evidence for self-criticism among Japanese. Personality and Social Psy-

chology Bulltin, 26, 71–78. https://doi.org/doi.org/10.1177%

2F0146167200261007

Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in

the world? Behavioral and Brain Sciences, 33, 61–83. https://doi.org/ 10.1017/S0140525X0999152X

Hofstede, G. (1983). Dimensions of national cultures in fifty countries and

three region. In J. Deregowski, S. Dzuirawiec, & R. Annis (Eds.), Explo-

rations in cross-cultural psychology (pp. 335–355). Lisse: Swets and

Zeitlinger.

Hofstede, G. (2001). Culture's consequences: Comparing values, behaviors,

institutions and organisations (2nd ed.). Thousand Oaks, CA: Sage.

Hope, L., & Gabbert, F. (2019). Interviewing witnesses and victims. In

N. Brewer & A. B. Douglass (Eds.), Psychological science and the law

(pp. 56–74). New York: The Guildford Press.

Istomin, K. V., Panáková, J., & Heady, P. (2014). Culture, perception, and

artistic visualization: A comparative study of children's drawings in

three siberian cultural groups. Cognitive Science, 38, 76–100. https:// doi.org/10.1111/cogs.12051

Kastanakis, M. N., & Voyer, B. G. (2014). The effect of culture on perception

and cognition: A conceptual framework. Journal of Business Research, 67,

425–433. https://doi.org/10.1016/j.jbusres.2013.03.028 Keith, K. (2012). Blood diamonds and war crimes: The case against Charles

Taylor. Southern Cross University Law Review, 15, 99–110. Retrieved from. http://classic.austlii.edu.au/au/journals/SCULawRw/2012/7.pdf

Leal, S., Vrij, A., Vernham, Z., Dalton, G., Jupe, L., Harvey, A., & Nahari, G.

(2018). Cross-cultural verbal deception. Legal and Criminological Psy-

chology, 23, 192–213. https://doi.org/10.1111/lcrp.12131 Lee, S., & Lee, D. K. (2018). What is the proper way to apply the multiple

comparison test? Korean Journal of Anesthesiology, 71, 353–360. https://doi.org/doi.org/10.4097%2Fkja.d.18.00242

Loftus, G. R., & Mackworth, N. H. (1978). Cognitive determinants of fixa-

tion location during picture viewing. Journal of Experimental Psychol-

ogy: Human Perception and Performance, 4, 565–572. https://doi.org/ doi/10.1037/0096-1523.4.4.565

Ma, Q., Pei, G., Jin, J., & De Wit, H. De. (2015). What makes you generous?

The influence of rural and urban rearing on social discounting in China.

PLoS One, 10, 1–11. https://doi.org/10.1371/journal.pone.0133078 Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for

cognition, emotion, and motivation. Psychological Review, 98,

224–253. https://doi.org/10.1037/0033-295X.98.2.224 Markus, H. R., & Kitayama, S. (2003). Culture, self, and the reality of the

social. Psychological Inquiry, 14, 277–283. https://doi.org/10.1080/

1047840X.2003.9682893

Masuda, T., & Nisbett, R. E. (2006). Culture and change blindness. Cognitive

Science, 30, 381–399. https://doi.org/10.1207/s15516709cog0000_63 Millar, P. R., Serbun, S. J., Vadalia, A., & Gutchess, A. H. (2013). Cross-

cultural differences in memory specificity. Culture and Brain, 1(2–4), 138–157. https://doi.org/10.1007/s40167-013-0011-3

Miyamoto, Y., Nisbett, R., & Masuda, T. (2006). Culture and the physical

environment. Psychological Science, 17, 113–119. https://doi.org/10. 1111/j.1467-9280.2006.01673.x

Oyserman, D. (2006). High power, low power, and equality: Culture

beyond individualism and collectivism. Journal of Consumer Psychology,

16, 352–356. https://doi.org/10.1207/s15327663jcp1604_6 Paz-Alonso, P. M., Goodman, G. S., & Ibabe, I. (2013). Adult eyewitness

memory and compliance: Effects of post-event misinformation on

memory for a negative event. Behavioral Sciences & the Law, 31,

541–558. https://doi.org/10.1002/bsl Peterson, C., Sales, J. M., Rees, M., & Fivush, R. (2007). Parent-child talk

and children's memory for stressful events. Applied Cognitive Psychol-

ogy, 21, 1057–1075. https://doi.org/10.1002/acp Raftery, A. E. (1995). Bayesian model selection in social research. Sociologi-

cal Methodology, 25, 111–163. https://doi.org/10.2307/271063 Rooks, G., Klyver, K., & Sserwanga, A. (2016). The context of social capital:

A comparison of rural and urban entrepreneurs in Uganda. Entrepre-

neurship: Theory and Practice, 40, 111–130. https://doi.org/10.1111/ etap.12107

Schwartz, A. J., Boduroglu, A., & Gutchess, A. H. (2014). Cross-cultural dif-

ferences in categorical memory errors. Cognitive Science, 38(5),

997–1007. https://doi.org/10.1111/cogs.12109 Soh, S., & Leong, F. T. L. (2002). Validity of vertical and horizontal individualism

and collectivism in Singapore. Journal of Cross-Cultural Psychology, 33,

3–15. https://doi.org/https://doi.org/10.1177/0022022102033001001 Takata, T. (2003). Self-enhancement and self-criticism in Japanese culture.

Journal of Cross-Cultural Psychology, 34, 542–551. https://doi.org/10. 1177/0022022103256477

Triandis, H. C. (2001). Individualism-collectivism and personality. Journal of

Personality, 69, 907–924. https://doi.org/10.1111/1467-6494.696169 Triandis, H. C., Bontempo, R., Villareal, M. J., Asai, M., & Lucca, N. (1988).

Individualism and collectivism: Cross-cultural perspectives on self-in-

group relations. Journal of Personality and Social Psychology, 54,

323–338. https://doi.org/10.1037/0022-3514.54.2.323 Triandis, H. C., & Gelfand, M. J. (1998). Converging measurement of hori-

zontal and vertical individualism and collectivism. Journal of Personality

and Social Psychology, 74, 118–128. https://doi.org/10.1037/0022-

3514.74.1.118

van Veldhuizen, T. S., Maas, R. P. A. E., Horselenberg, R., & van

Koppen, P. J. (2018). Establishing origin: Analysing the questions asked

in asylum interviews. Psychiatry, Psychology and Law, 25, 283–302. https://doi.org/10.1080/13218719.2017.1376607

Veillard, N. (2017). WEIRD sampling in cross-cultural psychology, should it

not be less WEIRD and more representative?: The over-representation

of individuals from Western, educated, industrialized, and democratic

countries as sample populations in cross-cultural psychology research

(master's thesis). Retrieved from https://openaccess.leidenuniv.nl/

handle/1887/56361

Wagenmakers, E.-J. (2007). A practical solution to the pervasive problems

of p values. Psychonomic Bulletin & Review, 14, 779–804. https://doi. org/https://doi.org/10.3758/BF03194105

ANAKWAH ET AL.514

Wang, H., & Pomplun, M. (2012). The attraction of visual attention to texts

in real-world scenes, 12, 1–17. https://doi.org/10.1167/12.6.26 Wang, Q. (2004). The emergence of cultural self-constructs: Autobio-

graphical memory and self-description in European American and Chi-

nese children. Developmental Psychology, 40, 3–15. https://doi.org/10. 1037/0012-1649.40.1.3

Wang, Q., & Ross, M. (2005).What we remember and what we tell: The effects

of culture and self-priming on memory representations and narratives.

Memory, 13, 594–606. https://doi.org/10.1080/09658210444000223 Wang, Q., Song, Q., & Kim Koh, J. B. (2017). Culture, memory, and narra-

tive self-making. Imagination, Cognition and Personality, 37, 199–223. https://doi.org/10.1177/0276236617733827

Wise, R. A., Gong, X., Safer, M. A., & Lee, Y. T. (2010). A comparison of

Chinese judges' and US judges' knowledge and beliefs about eyewit-

ness testimony. Psychology, Crime and Law, 16, 695–713. https://doi. org/10.1080/10683160903153893

Wong, B. I., Yin, S., Yang, L., Li, J., & Spaniol, J. (2017). Cultural differences

in memory for objects and backgrounds in pictures. Journal of

Cross-Cultural Psychology, 49, 1–14. https://doi.org/10.1177/

0022022117748763

Wright, A. M., & Holliday, R. E. (2007). Enhancing the recall of young,

young–old and old–old adults with cognitive interviews. Applied Cogni-

tive Psychology, 21, 19–43. https://doi.org/10.1002/acp.1260 Yamagishi, T., Hashimoto, H., Cook, K. S., Kiyonari, T., Shinada, M.,

Mifune, N., … Li, Y. (2012). Modesty in self-presentation: A compari-

son between the USA and Japan. Asian Journal of Social Psychology, 15,

60–68. https://doi.org/10.1111/j.1467-839x.2011.01362.x

How to cite this article: Anakwah N, Horselenberg R, Hope L,

Amankwah-Poku M, van Koppen PJ. Cross-cultural

differences in eyewitness memory reports. Appl Cognit

Psychol. 2020;34:504–515. https://doi.org/10.1002/acp.3637

ANAKWAH ET AL. 515