Вы находитесь на странице: 1из 11

Escaping the corner of death?

An eye-tracking
study of reading direction influence on
attention and memory
Monica D. Hernandez
Department of Management and Marketing, Texas A&M University-Corpus Christi, Texas, USA
Yong Wang
Department of Marketing, Ohio University, Athens, Ohio, USA
Hong Sheng
Missouri University of Science and Technology, Rolla, Missouri, USA
Morris Kalliny
Department of Marketing, Eastern Washington University, Spokane, Washington, USA, and
Michael Minor
Department of Marketing, Robert C. Vackar College of Business & Entrepreneurship, University of Texas Rio Grande Valley,
Edinburg, Texas, USA

Abstract
Purpose – The authors aim to examine the effect of location-driven logo placement on attention and memory on the web addressing differences
between individuals that read unidirectionally (left-to-right [LTR]) versus bidirectionally (both right-to-left and LTR).
Design/methodology/approach – Using an eye-tracking approach combined with traditional verbal measures, the authors compared attention
and memory measures from a sample composed of bidirectional (Arab/English) readers and unidirectional readers.
Findings – The findings reveal that unidirectional and bidirectional readers differ in attention patterns. Compared to bidirectional readers,
unidirectional readers pay less attention to the logo on the bottom right corner of the webpage based on verbal measures. The eye-tracking data
of the two groups further identify differences based on total hits and duration time. Unidirectional LTR readers demonstrate higher fluency in
feature-based attention whereas bidirectional readers show higher fluency in spatial attention.
Originality/value – The authors expand on scarce research on reading direction bias effect on location-driven stimuli placement in online settings.
They contribute to the understanding of the differences between unidirectional and bidirectional readers in their cognitive responses (attention and
memory) to organization of marketing stimuli.
Keywords Online marketing, Attention, International consumer, Location-driven placement, Reading direction
Paper type Research paper

Internet users are constantly exposed to marketing stimuli Outing and Ruel’s (2006) exploratory study with American
such as banners, sponsored search, logos, pop-up ads and participants. Considering the diverse internet audience, the
rectangle ads in multiple locations of a webpage. However, question of whether this scanning pattern also applies to any
Agarwal et al. (2011) highlighted the matter that the specific internet reader remains unanswered. Despite being one of the
location on a webpage for placing the marketing stimuli typical places where marketers and advertisers place their logo,
determines different levels of revenue generation by the according to Outing and Ruel (2006), the corner of death is
stimuli. Along this vein, research about the most effective and typically the second to last place an audience looks on a page.
memorable location of such elements on the webpage is On the other hand, contrary to the advice of avoiding the
fragmented and inconclusive. Hill (2010) warns advertisers to corner of death or lower right-hand corner, Goodrich (2010)
avoid the “corner of death”, a term coined to indicate the
found that online ads placed on the lower visual field of the
lower right-hand corner of any layout, whether it is a web site
screen received more attention than those placed on the upper
page, a TV screen, a print ad, direct mail piece or billboard. It
visual field. Therefore, decisions about marketing stimuli
is important to note that Hill’s (2010) advice is based on
placement in online settings become critical if we consider the
average ad viewing time per reader usually is less than 2 s
The current issue and full text archive of this journal is available on (Hill, 2010).
Emerald Insight at: www.emeraldinsight.com/0736-3761.htm

Journal of Consumer Marketing


Received 16 February 2016
34/1 (2017) 1–10 Revised 30 June 2016
© Emerald Publishing Limited [ISSN 0736-3761] 21 August 2016
[DOI 10.1108/JCM-02-2016-1710] Accepted 23 August 2016

1
Attention and memory Journal of Consumer Marketing
Monica D. Hernandez et al. Volume 34 · Number 1 · 2017 · 1–10

Outing and Ruel (2006) noted this pattern on a webpage – Body specificity theory (Casasanto, 2009) supports the
Eyes fixate first in the upper left then hover in that area before claim that right- and left-handers’ patterns of motor
going left to right, and, after perusing the top, eyes explore experience influence the way of thinking. Relevant to our
down. Correspondingly, it is common to see online advertisers study, Casasanto (2011) contended that there are differences
striving to secure the top left screen position under the premise in the functional role of language processing, providing
that the higher the ad placement, the more likely to receive evidence that different areas of the brain are activated to both
extended exposure and, therefore, the more effective its communicate and represent the meaning of action verbs and
economic performance. abstract ideas. Furthermore, language writing and reading are
The inconclusive literature calls for a closer investigation of also habitual. For example, the literature has documented
the effects of marketing stimuli location-driven placement significant habitual differences in information processing
and, particularly, logo position on the visual attention and among individuals speaking dissimilar languages, based on
memory of a diverse internet audience. Given that internet is different alphabets (Schmitt et al., 1994). In particular, there
borderless, it is critical to address consumer responses to are significant differences in the reading direction of written
location-driven placement in an international context. Our languages (Smith and Elias, 2013). The focus of our study is
study attends to this issue by investigating the differences the difference in reading direction. Toward this end, we
between unidirectional and bidirectional readers. The study classify languages as follows. The left-to-right (hereafter,
addresses the question: What is the differed outcome of LTR) language family includes most alphabetic languages
attention and memory for unidirectional and bidirectional (e.g. English, Spanish), and the modern versions of Far East
readers when they respond to a logo placed within the “corner languages (e.g. Chinese, Korean, Japanese). The right-to-left
of death” of a webpage? In efforts to consider these (hereafter, RTL) language family includes Arabic, Farsi,
differences, we analyze the impact of the position on visual Hebrew, Urdu, Pashto and a few African languages.
information search pattern and memory relative to the natural Addressing this distinction, Smith and Elias (2013)
direction of reading. Given the scarce use of physiological considered the native reading direction and its effect on
measures in addressing the effect of written language, we preferences for lit images. They concluded that LTR natives
attempt to use eye-tracking device to uncover the exhibit a larger number of fixations on the left side as well as
physiologically different patterns across natives of different preference for lit-images in the same side, whereas RTL
languages. We attempt to advance our knowledge in this area natives exhibit the mirror pattern and preference. While
by making contributions to the literature, particularly on extensive work has been conducted with LTR natives, there
visual attention and memory of logos in the online setting that are only a handful of studies addressing the eye movement of
can be generalized to online advertising and communication in RTL readers. The Arabic alphabet – second only to the Latin
the global marketplace. alphabet in usage around the world – consists of 28 letters (32
in Farsi), all representing consonants, and is written from
RTL (Britannica, 2016). Among the studies including Arabic
Conceptual background natives, Maass and Russo (2003) advanced the hypothesis that
Should we place a logo to the left or to the right of the content hemispheric brain specialization and scanning habit affect
on a webpage? Previous research has offered insights about the visual imaging. Specifically, their study addressed cultural
directional/lateral organization of marketing stimuli. The influences on the semantic representation of actions among
directional or lateral display influences consumers’ cognitive Italian (LTR) and Arabic (RTL) natives. Italians showed a
(e.g. information processing and value judgment) as well as clear LTR preference, whereas Arabs showed the reverse
affective (e.g. attitude) evaluations of stimuli (Janiszewski, pattern. In the same vein, Spalek and Hammad (2005) found
1990, 1993; Valenzuela and Raghubir, 2009; Chae and a LTR bias in a group of English (LTR) speaking natives.
Hoegg, 2013; Romero and Biswas, 2016). Among them, Conversely, a RTL bias was found among Arabic (RTL)
Atalay et al. (2012) examined the effect of product layout and speaking natives. The Spalek and Hammad (2005) results
found how items are laterally displayed has a prominent were attributed to the direction of text reading. More recent
influence on consumers’ visual attention in a shopping scholarly work has also found significant differences in
environment. information processing, which presumably are acquired
Hoyer et al. (1997) defined attention as the process by because of language constraints. Dobel et al. (2007) compared
which a person apportions his or her cognitive resources to a Germans (LTR) and Israelis (RTL), and preschool-age
visual stimulus. The importance of attention in marketing has children in both countries showed no directional bias.
been highlighted in various major advertising response Nevertheless, adults manifested a bias that was consistent with
models, such as AIDA and the hierarchy of effects. The the writing system of their language.
importance of attention stems from its position as a Despite having the same direction of reading, it is important
prerequisite for the effectiveness of visual information to note there are diverse structural differences between
processing. In other words, there is little chance that any alphabetic and non-alphabetic based writing systems. Among
information provided to the consumers will produce desired them, sound structure, meaning and reading direction are
result without consumers devoting appropriate attention to fundamental to classifying writing systems. For instance,
the information. Thus, understanding attention patterns may Akmajian et al. (1990) distinguished between two major types
help advance academic marketing knowledge and guide of writing systems based on the graphic symbols associated
managerial efforts toward more successful advertising with sound or meaning. These graphic symbols or “letters”
outcomes, particularly in the online setting. represent the sound structure of the spoken work in alphabetic

2
Attention and memory Journal of Consumer Marketing
Monica D. Hernandez et al. Volume 34 · Number 1 · 2017 · 1–10

languages (such as English). In contrast, the graphic symbols processing of logo in the bottom right corner of a webpage, the
or “logographs” represent a complete word in the Chinese “corner of death”, depends on the bidirectionality of readers.
language. Another prominent structural difference was that The unidirectional LTR readers are found to pay less
ancient Chinese and Japanese were written from top to bottom attention on the corner of death (Hill, 2010). This is because
(with the columns printed from right to left). Currently, of the LTR language constraints, which influences the
Chinese is written from LTR. Moreover, at a difference from attention mechanism of unidirectional LTR readers. Hill’s
alphabetic writing systems, Chinese is written without spaces (2010) study is in a North American context. For RTL
between successive characters and words. readers, such as Arab people, the “corner of death” will likely
Despite these and other differences, few scholarly studies not to be in the same location as for LTR readers. It means
have addressed the underlying mechanisms of eye movement that RTL readers will likely to avoid the specific “corner of
taking into consideration language differences. Interestingly, death” the LTR readers often experience. Putting together, we
some differences have been found between natives of think that the bidirectional (RTL/LTR) readers will have
alphabetic and non-alphabetic writing systems. Among them, higher attention at this spot because of their RTL reading
Chinese readers make shorter saccades than do English ability. With the underlying fluency of reading from either
readers, and the perceptual span of Chinese readers is smaller side, bidirectional readers have a higher tendency for spatial
than that of English readers (Rayner et al., 2005). It is central attention than their unidirectional counterparts because of
to point out that, according to Rayner et al. (2005), their bidirectional bias. On the contrary, unidirectional LTR
informational density is much higher in Chinese than in readers have developed the directional habitat in visual
English. information processing. The reading fluency because of
Given the increasing number of bilinguals, particularly in directional reading habitat eases the processing of other
Asia (e.g. Arab/English), the number of users reading related features and enhances the accuracy of such
bidirectionally is also frequent. In the era of globalization, the information processing (Chen, 2000). When the directional
behavior of consumers who are bidirectional readers should be habitat does not function, time spent on visual information
well examined to offer guidance for international marketing process may be higher. Thus, unidirectional LTR readers are
managers. As this tendency indicates, there is a need to more fluent in feature caption. The processing, feedback and
address the various effects of bidirectional reading. Román differentiating time period used to distinguish selected visual
et al. (2013) noted that developed contradictory directional features in the webpage is shorter for unidirectional readers.
habits of bidirectional readers (e.g. Arabic and English This results in their higher tendency for fluent feature-based
bilinguals) will enter in competition, and resulting bias would attention than the bidirectional readers.
be different from that of the unidirectional readers. However, Because of the reasoning, unidirectional LTR readers are
there are still a number of key questions on the role of more likely to ignore the logo placed in the corner of death
bidirectionality in reader’s cognitive responses. In the than the bidirectional readers are. We argue that, when we
following section, we attempt to address the gap by examining measure attention through eye fixation hits, bidirectional
their attention and memory patterns as compared to readers will be able to scan a larger number of spots including
unidirectional readers. the corner of death, an indicator for higher fluency of spatial
attention. When we measure attention through eye fixation
Hypotheses duration, however, bidirectional readers will use a larger
amount of time in visual information processing on the
Relevant to the study of two-dimensional space, there are two
webpage, indicating lower fluency for feature-based attention.
main mechanisms of visual attention: spatial attention and
Hence, we hypothesize:
feature-based attention (Liu et al., 2007; Carrasco, 2011).
Spatial attention is the attention process onto a specific spot H1. Indicated by verbal measure, unidirectional (LTR)
from a visual field. It allows individuals to selectively process readers have lower amount of visual attention to logo
all visual information of a visual area. It is overt and is placed in the corner of death than do bidirectional
indicated by eye movement to the focus of attention. On the (RTL/LTR) readers.
other side, feature-based attention can be deployed regardless
of the focus of attention. Feature-based attention is deployed H2. Indicated by smaller total eye fixation hits,
to differentiate and highlight particular features in a visual unidirectional (LTR) readers have a lower tendency for
field. It operates both within and outside the spatial focus of fluent spatial attention to logo placed in the corner of
attention. For example, individuals follow two mechanisms of death than do bidirectional (RTL/LTR) readers.
attention when browsing products on an e-tailing website:
1 where to look on the webpage (spatial attention); and H3. Indicated by shorter eye fixation duration,
2 what product to look for (feature-based attention). unidirectional (LTR) readers have a higher tendency
for fluent feature-based attention placed in the corner
Individuals are likely to locate the items after directing of death than do bidirectional (RTL/LTR) readers.
attention to a specific spot and/or to distinctive visual features
of a target product being sought. Memory is often measured by recall and/or recognition. While
A number of studies linked the directions of text reading recall refers to the ability to reproduce previously presented
and patterns of information processing, presumably based on items, with or without aid, recognition refers to the ability to
language constraints (Rayner et al., 2005; Román et al., 2013). identify an item that has been recently encountered (Wixted
By extension, we think that the amount of visual information and Squire, 2004). The literature addressing the relationship

3
Attention and memory Journal of Consumer Marketing
Monica D. Hernandez et al. Volume 34 · Number 1 · 2017 · 1–10

of attention and memory measured with an eye-tracker is week, 27 per cent spent between 1 and 10 h online a week and
inconclusive. Agarwal et al. (2011) did not find a correlation 64 per cent spent more than 10 h a week online. As for
between eye-tracking measures and memory-based attention shopping, 32 per cent stated that they rarely shop online,
measures. However, Yeh et al. (2011) noticed that trained whereas 45 per cent shop several times a year, and 23 per cent
readers looked back more often and spent a higher percentage stated that they shopped online several times a week. The
of time on a sentence. They concluded that a change in eye Chinese-speaking sample consisted of 70 per cent males and
movement and increased recall occurred together on second 30 per cent females with educational level ranging from
language English readers, supporting the efficacy of a reading 19 per cent having an undergraduate degree and 81 per cent
training program. We expect that, as a consequence of the having or pursing a graduate degree; 19 per cent of the sample
native reading directional bias, bidirectional readers will spent 1 to 10 h online a week, and 81 per cent spent more than
allocate stronger cognitive effort to remember information. By 10 h online a week, whereas 4 per cent rarely shop online,
extension, we hypothesize that: 41 per cent shop several times a year and 55 per cent shop
several times a month. Table I contains the demographics of
H4. Indicated by verbal measure, unidirectional (LTR) the participants.
readers have lower amount of logo memory than do
bidirectional (RTL/LTR) readers when it is placed in
Stimuli
the corner of death.
We used one webpage for all groups in the experiment. A
complete experimental webpage was created for this purpose.
Method The webpage was designed as the homepage for a fictitious
In this section, we discuss how we carried out the non-profit organization. Rayner’s (2009) review indicated that
experimental procedures. This process includes recruiting average fixation durations and the mean saccade length in
participants who are specifically bidirectional and silent reading, oral reading, scene perception and visual search
unidirectional readers, creating an experimental webpage as may be influenced by text difficulty, reading skill and
the marketing stimuli and instructing the participants to characteristics of the writing system. Therefore, no text was
conduct the experiment in lab settings. used on the webpage to void the intervention of written
language. A total of 25 200 ⫻ 150 pixel pictures were used as
Participants the content of the webpage. Figure 1 shows the experimental
Normal (and corrected-to-normal) vision adult participants webpage. All the items in the pictures were tested for their
were recruited to participate in the experiment. Participants cultural meanings using subjects from each of the three
received a $10 gift card in return for their participation. groups. The meanings reported by the three groups were
Participants were unaware of the purpose of the experiment identical across the three cultures. In addition, meanings of
until they had completed all tests. Rayner et al. (2006) found the items in the pictures are culturally neutral. The pictures
differences in reading between old and young participants require significant processing time from viewers. Moreover,
(e.g. older participants read more slowly and their perceptual reading LTR or RTL will lead to similar interpretations and
span is smaller). To avoid this issue, we restricted our an almost identical amount of processing time because
participants to those older than 18 years old. Our convenience pictures and design were symmetrical. To eliminate brand
sample was composed of 25 English-speaking natives, 27 familiarity, fictitious brand logos were created. Four different
Chinese-speaking natives and 22 Arabic-speaking natives. It is logos were placed in the corners, respectively, and
important to note that all participants in both Chinese and aesthetically incorporated to the webpage. The four logos have
Arab groups are bilingual, with English as second language. the same number of color elements and follow the same color
Thus, the 22 Arab participants were treated as the schemes. The meanings of the four logos are all linked to the
bidirectional readers (both LTR and RTL) because of their theme of the pictures. Based on our focus group test, the
bilingual reading capacity. On the other hand, the American meanings of the logos are also culturally neutral.
and Chinese participants were regarded as unidirectional
readers (LTR). The Chinese bilinguals are unidirectional Measures – self-report
because both Chinese and English languages are LTR. We Our self-report verbal measures of attention and memory
compare American and Chinese participants to see if there is (recall) were adopted from Moorman et al. (2012) who also
within-group difference in the unidirectional family. intended to capture attention and memory of participants.
Our English-speaking native sample consisted of 60 per cent Attention was measured by asking participants to report in a
males and 40 per cent females with educational level ranging scale of one-seven how much attention was paid to the brand
from high school (4 per cent), undergraduate (84 per cent) logos displayed in the four corners. Unaided recall was
and graduate (12 per cent). The English sample also varied in obtained by asking the participants to check from a list of
terms of time spent online and frequency of shopping online, verbal descriptions as many logos, as they might have seen
where 32 per cent of our sample spent 1-10 h online a week when the experimental stimuli were presented. In addition, we
and 68 per cent spent more than 10 h a week. The Arabic added a recognition measure, which was obtained by
speaking sample consisted of 59 per cent male and 41 per cent displaying sets of four logos and asking the participants to
female with educational level of 32 per cent having completed mark the one seen in each corner of the screen. A structured
high school, 23 per cent having an undergraduate degree and print questionnaire was used, and all the questions were
45 per cent having completed or currently pursuing a graduate presented in English. The questionnaire also addressed native
degree: 9 per cent of the sample spent less than 1 h online a language and second language proficiency. Six additional

4
Attention and memory Journal of Consumer Marketing
Monica D. Hernandez et al. Volume 34 · Number 1 · 2017 · 1–10

Table I Participant demographicsa


American (N ⴝ 25) Chinese (N ⴝ 27) Arab (N ⴝ 22)
Demographic variable Unidirectional (LTR) Unidirectional (LTR) Bidirectional (LTR/RTL)
Gender distribution
Male 60% (N ⫽ 15) 70% (N ⫽ 19) 59% (N ⫽ 13)
Female 40% (N ⫽ 10) 30% (N ⫽ 8) 41% (N ⫽ 9)
Age (average) 24 25 30

Education level
High school 4% (N ⫽ 1) 0% 32% (N ⫽ 7)
Undergraduate 84% (N ⫽ 22) 19% (N ⫽ 5) 23% (N ⫽ 5)
Graduate 12% (N ⫽ 3) 81% (N ⫽ 22) 45% (N ⫽ 10)

Currently a student
Yes 76% (N ⫽ 19) 33% (N ⫽ 9) 50% (N ⫽ 11)
No 24% (N ⫽ 6) 67% (N ⫽ 18) 50% (N ⫽ 11)

Time spent online (per week)


Less than 1 h 0% 0% 9% (N ⫽ 2)
From 1 to 10 h 32% (N ⫽ 8) 19% (N ⫽ 5) 27% (N ⫽ 6)
More than 10 h 68% (N ⫽ 17) 81% (N ⫽ 22) 64% (N ⫽ 14)

Frequency of shopping online


Rarely 24% (N ⫽ 6) 4% (N ⫽ 1) 32% (N ⫽ 7)
Several times a year 44% (N ⫽ 11) 41% (N ⫽ 11) 45% (N ⫽ 10)
Several times a month 32% (N ⫽ 8) 55% (N ⫽ 15) 23% (N ⫽ 5)
Note: a All the Chinese and Arabic participants are bilingual based on their English proficiency test

Figure 1 Experimental webpage

5
Attention and memory Journal of Consumer Marketing
Monica D. Hernandez et al. Volume 34 · Number 1 · 2017 · 1–10

items addressing demographic background and Internet usage case, left bias). Results of a t-test revealed a significant effect of
were collected at the end of the questionnaire. frequency of hits on the top right corner of the screen as
captured by the eye-tracking measures (t ⫽ ⫺2.641, p ⬍
Measures – eye-tracker 0.01). As expected, the bidirectional readers exhibited a larger
Eye movements were recorded with a Tobii Studio T60 eye number of hits on the right side. However, the t-test revealed
tracker system. Before starting the session, we calibrated the no significant effect of frequency of hits on the bottom right
eye tracker. A check calibration pattern occurred with five corner (t ⫽ 0.793, p ⬎ 0.05). H2 was not supported by the
fixed targets and a calibration cross that moved in eye-tracking data. According to Atalay et al. (2012), orienting
synchronization with the eye. If there was a discrepancy oneself may be the reason for the initial fixations. So,
between the calibration cross and the fixed target, the system additional analyses were conducted after the first fixation, and
was recalibrated for the participant. The experimental website the results were similar.
designed for the present study was then launched. Although there is no difference in the spatial attention
between unidirectional and bidirectional readers on the
Procedure bottom right corner, statistical evidence was found on the top
Initially, informed consent was obtained from each right corner. Further, as demonstrated in Figure 2 (heat
participant. Each participant was individually led to the maps), the likelihood of fixating on the right side logos,
eye-tracking lab and asked to sit in front of the experimental compared to the logos on the left side, was higher by the
computer. Participants were seated approximately 60-70 cm bidirectional readers. This effect was particularly prevalent in
in front of the computer screen. Rayner (2009) indicated that the first hit. Therefore, there is a spatial attention difference
average fixation durations and the mean saccade length in between the two groups.
silent reading and oral reading tasks differ. Specifically, mean
Attention – percentage of time
fixation duration for silent reading is shorter than for oral
We hypothesized that unidirectional readers would allocate a
reading whereas mean saccade length for silent reading is
greater percentage of time to the direction of their native
larger than for oral reading. To avoid these disparities between
language bias (in our case, left side). Because the time of
participants, all participants were asked to complete the tasks
completion of the experiment varied by participant, the
in silence.
percentage gaze duration of fixation was deemed appropriate
The instruction we gave the participants for the experiment
to measure the attention allocation of the participants who hit
is:
the AOIs. A t-test was conducted to compare the gaze
You will see a webpage that has 25 pictures presenting a thematic story. You duration for bi- and unidirectional readers. The t-test revealed
have to scan the webpage in 10 seconds and tell the story based on all the
pictures.
significant differences between bidirectional and
unidirectional readers [t(72) ⫽ ⫺2.549, p ⬍ 0.05]. The
Their eye movement patterns were tracked. After the difference was not significant between Chinese and American
experimental procedure was completed, they were asked to unidirectional readers [t(47) ⫽ 0.281, p ⬎ 0.05].
briefly write down a filler task consisting of the thematic story Unambiguously, we found that the bidirectional readers
based on their own interpretation. Lastly, participants were allocated a greater percentage of time to fixations when they
asked to complete the paper-and-pencil questionnaire process the logo information. The results indicated that
described above. The total experimental procedure for every unidirectional readers spend less time in visual information
participant was completed in 6-12 min. processing when their eyes are focused on the logo in the
bottom right corner. Thus, H3 was supported.
Results
Eye-tracking data Self-reported verbal measures
Eye-movement data, including both gaze duration and Self-reported attention
fixation, are informative of the visual attention process. An independent samples t-test was performed, with
Analyses were conducted on both time allocation (total self-reported attention as the dependent variable and direction
fixation duration) and location of fixation (AOI hits). Overall, of reading as the independent variable. Results showed a
the logos placed on the right side of the screen received less significant difference between the two groups [t(72) ⫽
frequent eye fixations [frequency of hits on the area of interest ⫺2.921, p ⬍ 0.01], with bidirectional readers reporting a
(hereafter AOI)], and they were looked at more briefly (total greater degree of attention than the unidirectional
fixation duration) over the data collection process. Table II counterparts. The two unidirectional groups, American and
contains a summary of statistical results. Chinese, did not exhibit significant difference (p ⬎ 0.10). H1
To ensure information acquisition per each fixation, was supported by the result of the analysis on self-reported
fixations that lasted less than 100 milliseconds were eliminated attention to the logo in the corner of death.
(Atalay et al., 2012; Pieters and Wedel, 2012). Consistent with
Memory
Atalay et al. (2012) and Pieters and Wedel (2012), the cases of
Memory for the logo placed on the bottom right corner shows
200 milliseconds and above did not alter the results.
a similar level between unidirectional and bidirectional readers
Therefore, these cases were retained.
[t(72) ⫽ ⫺0.968, p ⬎ 0.05]. In addition, results of an ANOVA
Attention – hits testing each corner of the screen revealed no significant effect
It was argued that unidirectional readers would have more of logo memory between groups via either recall or
frequent eye fixations on their native language bias (in our recognition. The rate of correct recall of each logo (one logo

6
Attention and memory Journal of Consumer Marketing
Monica D. Hernandez et al. Volume 34 · Number 1 · 2017 · 1–10

Table II Summary of statistical analysis


Dependent variable Means (SD) t-value df p-value
Eye-tracker data (on corner of death)
Attention – hits to AOI 0.793 3612 0.428
Unidirectional 0.00 (0.044)
Bidirectional 0.00 (0.031)
Attention – percentage of time ⫺2.549 3612 0.011
Unidirectional 1.92% (1.21%)
Bidirectional 2.05% (1.86%)

Self-reported verbal measures (on corner of death)


Attention (self-report) ⫺2.787 72 0.008
Unidirectional 2.17 (1.31)
Bidirectional 3.18 (1.47)
Memory (self-report) ⫺0.968 72 0.339
Unidirectional 2.62 (1.05)
Bidirectional 2.86 (0.99)

Figure 2 Representative heat maps


per screen corner) was used as the dependent variables. The
classified group (unidirectional vs. bidirectional) was the
independent variable. The interaction effect was
nonsignificant [F(2,71) ⫽ 0.211, MSE ⫽ 0.814,
p ⫽ 0.811]. The results indicate that there is no overall
difference in recall or recognition between the two groups
because of reading direction. Thus, H4 was not supported.
Interestingly, attention between the two groups differ but
not for memory as measured by recognition. In other words,
the bidirectional group’s self-reported data indicated higher
attention to stimuli. However, this time allocation did not
improve their logo memory.

Discussion
The corner of death has been identified as the least desired
place for a logo because of the lack of attention (Hill, 2010).
Such notion was based on experiments conducted with North
American consumers. Previously, how the different locations
on a webpage are processed or remembered across countries
and cultural groups has not been fully understood. Among the
earliest to shed light on this research topic, our findings
contribute to the understanding of the differences between
unidirectional and bidirectional readers in their responses to
directional organization of marketing stimuli. The results from
the verbal measures revealed that consumers who are
bidirectional readers pay more visual attention on the corner
of death than the unidirectional (LTR) readers. Further, from
the eye-movement analysis, it was found that unidirectional
(LTR) readers and bidirectional readers differ in attention as
measured by total fixations and duration time. In particular,
we found that bidirectional readers spent a higher percentage
of time fixed on the bottom right corner (the corner of death).
They also have a larger amount of eye fixations on the right
side of the webpage and, in particular, the top right corner.
These results are consistent with previous findings reported by
Román et al. (2013) and Dobel et al. (2007). Interestingly, we
found non-significant effect in logo memory. This finding
converges with Agarwal et al. (2011), who did not find a
correlation between attention and memory.

7
Attention and memory Journal of Consumer Marketing
Monica D. Hernandez et al. Volume 34 · Number 1 · 2017 · 1–10

In summary, we found that bidirectional readers are not as Arabic and English, it suggests that directional bias needs to
vulnerable as unidirectional LTR readers when encountering be taken into consideration when bidirectional readers are
the corner of death phenomenon because of the differences in targeted.
their visual attention mechanisms. More illustrative of the
phenomenon, compared to unidirectional (LTR) readers, Managerial implications
bidirectional readers are less fluent in deploying feature-based
From a managerial standpoint, the findings can help global
attention but are more fluent in using spatial attention when
marketing managers understand how to improve the
presented with the same webpage. This offers insights into
effectiveness of advertisements and webpage design. The
understanding the habitual nature of certain international
findings render clear managerial expertise in regard to logo
consumers who are accustomed to reading from both
placement. The placement of a logo on the corner of death is
directions. Meanwhile, the findings suggest that the corner of
not as detrimental in an advertising platform targeting
death phenomenon is at least partially caused by
bidirectional (LTR/RTL) readers. This is especially important
unidirectional LTR readers’ habitual tendency in higher
when the time of fixation is abundant. For ads or web pages
feature-based attention and lower spatial attention. When intended for certain international populations that are
consumers are able to read from both directions, they show bidirectional readers, the corner of death might be strategically
higher fluency in spatial attention, which reduced the used for logo placement to achieve good results in visual
possibility of not visually capturing the corner of death. attention.
There was no difference in memory found in the results, in The implications can also be extended to online visual
a general sense suggesting that the attention pattern search patterns. Search engines such as Google and Bing
differences do not result in differences in memory. Given that follow the same patterns displaying organic results and
the means for recall were low (Table II), logos placed in the sponsored ads. The rationale to position the relevant
corner of death are not effective in terms of memory, information closer to the top left corner (diagonally opposite
regardless of viewers’ reading directions. The advantage of the to the corner of death) is the visibility and number of
bottom right corner for bidirectional readers is not further click-through on the listings in this area. A similar pattern
transferred to the long-term memory outcomes for a logo. occurs in social media outlets such as Twitter. However, the
Thus, we should be aware that, in a content-rich web page, location of information seems to follow the natural LTR
logo placed in the corner of death could achieve no more than pattern and both the RTL and bidirectional pattern have been
a marginal memory effect, regardless of reading directions of overlooked. Our findings shed light when international
the consumer group. consumers are of concern.
The findings may help to generate specific insights to the Given the popularity and prominence of social media, the
marketing and advertising industry. The discovered patterns position of marketing stimuli in any given screen is important.
of difference help marketers delineate better directional/lateral As social media differs from traditional advertising in that the
layouts when physically placing marketing stimuli, such as former motivates users to engage and share, the elements with
brand logos, banner ads, remarks, sponsored search results, which the user interacts with the most should be easily
pop-up ads and information ads on a webpage, with accessible and within rapid reach. Decisions about the
anticipated international impact. Location-driven ad location of display and/or banner ads can also be targeted to
placement in view of our findings has implications for Internet the specific type of readers. We commonly find this type of
managers managing certain international audience. The advertising at the top or at the left side of the screen. Again,
numerous marketing activities on the web, such as social the decision to place these ads in these specific locations
networking and advertising copy, require an integrated perhaps follows the natural LTR pattern. However, ads
platform. In an international marketing context, marketers are targeting bidirectional readers may benefit from a right-side
also aware of the combined use of the native language in the placement instead.
home country and the language of English as the universal
business language. Limitations and future research
Specifically, the effect of logo position on visual attention Regarding sample selection, Arabic and Chinese speaking
and memory by a diverse Internet audience helps global participants are recruited to represent bidirectional and
marketers to understand the adaptation of webpage design unidirectional reader groups. It seems as though individuals
based on reading directions and English language fluency. from those cultures would have a considerably different
The results also offer important suggestions for online mindset regarding exposure to pictorial materials in general.
interface design. In a North American consumer environment, Essentially the question becomes the comparability on many
the corner of death is said to be the worse place for important variables. Future research that replicates or extends the
visual information. However, the corner of death may be a current study needs to ensure comparability via cultural-
corner of opportunity based on the characteristics of the web psychological analysis and demographic analysis.
browsers. Previous thoughts on location-driven ad placement Future research also needs to devote additional effort to
have not fully considered reading directions and its impact on external validity. Another limiting factor in our study was that
the information processing patterns. Our findings highlight the time of completion of the experiment slightly varied by
the attention pattern differences between North American participants. We attempted to correct this issue by considering
consumers who are by default unidirectional LTR readers and the percentage of fixation instead of the raw data. For future
many Arab consumers who are bidirectional (RTL/LTR) work, one can further explore other memory measures, such as
readers. With the example of Arab natives who are bilingual in aided recall, to see if the results still hold. Future work can

8
Attention and memory Journal of Consumer Marketing
Monica D. Hernandez et al. Volume 34 · Number 1 · 2017 · 1–10

consider the relevance of position given the purpose of a Liu, T., Stevens, S.T. and Carrasco, M. (2007), “Comparing
search. For example, does the attention shift to different areas the time course and efficacy of spatial and feature-based
of the screen whether the activity of the reader is leisure or attention”, Vision Research, Vol. 47 No. 1, pp. 108-113.
purchase oriented? In addition, future work can explore the Maass, A. and Russo, A.(2003), “Directional bias in the
attention, memory and click-through behavior of display/ mental representation of spatial events: nature or culture?”,
banner ads between groups of unidirectional versus Psychological Science (Wiley-Blackwell), Vol. 14 No. 4,
bidirectional readers. For social media, future research can pp. 296-301.
explore the effect of marketing stimuli position on consumer Moorman, M., Willemsen, L.M., Neijens, P.C. and Smit,
engagement and/or interaction. As social media networks E.G. (2012), “Program-involvement effects on commercial
evolve, more questions about the relevance of the position of attention and recall of successive and embedded
marketing stimuli may arise. advertising”, Journal of Advertising, Vol. 41 No. 2,
pp. 25-37.
References Outing, S. and Ruel, L. (2006), “The best of eyetrack III:
what we saw when we looked through their eyes”, Eyetrack
Agarwal A., Hosanagar, K. and Smith, M.D. (2011),
III, available at: www.uvsc.edu/disted/decourses/dgm/2740/
“Location, location, location: an analysis of profitability of
position in online advertising markets”, Journal of Marketing IN/steinja/lessons/05/docs/eyetrack_iii.pdf (accessed 1 June
Research, Vol. 48 No. 6, pp. 1057-1073. 2016).
Akmajian, A., Demers, R.A., Farmer, A.K. and Harnish, Pieters, R. and Wedel, M. (2012), “Ad gist: ad
R.M. (1990), Linguistics: An Introduction to Language and communication in a single eye fixation”, Marketing Science,
Communication, MIT Press, Cambridge, MA. Vol. 31 No. 1, pp. 59-73.
Atalay, A. Selin, Bodur H.O. and Rasolofoarison, D. (2012), Rayner, K. (2009), “Eye movements and attention in
“Shining in the center: central gaze cascade effect on reading”, Scene Perception, and Visual Search, The Quarterly
product choice”, Journal of Consumer Research, Vol. 39 Journal of Experimental Psychology, Vol. 62 No. 8,
No. 4, pp. 848-866. pp. 1457-1506.
Britannica, Encyclopedia (2016), “Arabic alphabet”, available Rayner, K., Li, X., Juhasz, B.J. and Yan, G.(2005), “The
at: www.britannica.com/EBchecked/topic/31666/Arabic- effect of word predictability on the eye movements of
alphabet (accessed 1 June 2016). Chinese readers”, Psychonomic Bulletin & Review, Vol. 12
Carrasco, M. 2011), “Visual attention: the past 25 years”, No. 6, pp. 1089-1093.
Vision Research, Vol. 51 No. 13, pp. 1484-1525. Rayner, K., Reichle, E.D., Stroud, M.J., Williams, C.C. and
Casasanto, D. (2009), “Embodiment of abstract concepts: Pollatsek, A. (2006), “The effect of word frequency, word
good and bad in right- and left-handers”, Journal of predictability, and font difficulty on the eye movements of
Experimental Psychology: General, Vol. 138 No. 3, young and older readers”, Psychology and Aging, Vol. 21
pp. 351-367. No. 3, pp. 448-465.
Casasanto, D. (2011), “Different bodies, different minds: the Román, A., Fathi, A. and Santiago, J. (2013), “Spatial biases
body specificity of language and thought”, Current Directions in understanding descriptions of static scenes: the role of
in Psychological Science, Vol. 20 No. 6, pp. 378-383. reading and writing direction”, Memory & Cognition,
Chae, B.G. and Hoegg, J.A. (2013), “The future looks ‘right’: Vol. 41 No. 4, pp. 588-599.
effects of the horizontal location of advertising images on Romero, M. and Biswas, D. (2016), “Healthy-left,
product attitude”, Journal of Consumer Research, Vol. 40 unhealthy-right: can displaying healthy items to the left
No. 2, pp. 223-238. (versus right) of unhealthy items nudge healthier choices?”
Chen, Z. (2000), “An object-based cost of visual filtering”, Journal of Consumer Research, Vol. 43 No. 1, pp. 103-112.
Perception and Psychophysics, Vol. 62 No. 3, pp. 482-495. Schmitt, B.H., Pan, Y. and Tavassoli, N.T. (1994),
Dobel, C., Diesendruck, G. and Bolte, J. (2007), “How “Language and consumer memory: the impact of linguistic
writing system and age influence spatial representations differences between Chinese and English”, Journal of
of actions: a developmental, cross-linguistic study”, Consumer Research, Vol. 21 No. 3, pp. 419-431.
Psychological Science, Vol. 18 No. 6, pp. 487-491. Smith, A.K. and Elias, L.J. (2013), “Native reading direction
Goodrich, K. (2010), “What’s up? exploring upper and lower and corresponding preferences for left- or right-lit images”,
visual field advertising effects”, Journal of Advertising Perceptual & Motor Skills: Learning & Memory, Vol. 116
Research, Vol. 50 No. 1, pp. 91-106. No. 2, pp. 355-367.
Hill, D. (2010), “The 6 secrets of eye-tracking”, Marketing Spalek, T.M. and Hammad, S. (2005), “The left-to-right bias
Daily, 15 October, available at: www.mediapost.com/ in inhibition of return is due to the direction of reading”,
publications/?art_aid⫽137686&fa⫽Articles.showArticle Psychological Science (Wiley-Blackwell), Vol. 16 No. 1,
(accessed 1 June 2016). pp. 15-18.
Hoyer, W.D., MacInnis, D.J. and Pieters, R.(1997), Consumer Valenzuela, A. and Raghubir, P. (2009), “Position-based
Behavior, Cengage, Mason, OH. beliefs: the center-stage effect”, Journal of Consumer
Janiszewski, C. (1990), “The influence of print advertisement Psychology, Vol. 19 No. 2, pp. 185-196.
organization on affect toward a brand name”, Journal of Wixted, J.T. and Squire, L.R. (2004), “Recall and recognition
Consumer Research, Vol. 17 No. 1, pp. 53-65. are equally impaired in patients with selective hippocampal
Janiszewski, C. (1993), “Preattentive mere exposure effects”, damage”, Cognitive, Affective, and Behavioral Neuroscience,
Journal of Consumer Research, Vol. 20 No. 3, pp. 376-392. Vol. 4 No. 1, pp. 58-66.

9
Attention and memory Journal of Consumer Marketing
Monica D. Hernandez et al. Volume 34 · Number 1 · 2017 · 1–10

Yeh, L.H., Schwartz, A.I. and Baule, A.L. (2011), “The Krueger, L.E. (1970), “Visual comparison in a redundant
impact of text-structure strategy instruction on the text display”, Cognitive Psychology, Vol. 7, pp. 314-357.
recall and eye-movement patterns of second language Luna, D. and Peracchio, L.A. (2001), “Moderators of
English readers”, Reading Psychology, Vol. 32, pp. 495-519. language effects in advertising to bilinguals: a
psycholinguistic approach”, Journal of Consumer Research,
Vol. 28 No. 2, pp. 284-295.
Further reading
Rayner, K., Li, X., Williams, C.C., Cave, K.R. and Well,
Dooley, R.(2010), “Avoid the corner of death! A.D. (2007), “Eye movements during information
neuromarketing”, available at: www.neuroscience processing tasks: individual differences and cultural
marketing.com/blog/articles/corner-of-death.htm (accessed effects”, Vision Research, Vol. 47 No. 21, pp. 2714-2726.
1 June 2016). Siok, W.T., Perfetti, C.A., Jin, Z. and Tan, L.H. (2004),
Green, D.W. and Meara, P. (1987), “The effects of script on “Biological abnormality of impaired reading is constrained
visual search”, Second Language Research, Vol. 3 No. 2, by culture”, Nature, Vol. 431 No. 7004, p. 71.
pp. 102-117.
Green, D.W., Liow, S.J.R., Tng, S.K. and Zielinski, S.
(1996), “Are visual search procedures adapted to the nature
of the script?”, British Journal of Psychology, Vol. 87 No. 2, Corresponding author
pp. 311-326. Yong Wang can be contacted at: wangy@ohio.edu

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com

10
Reproduced with permission of the copyright owner. Further reproduction prohibited without
permission.

Вам также может понравиться