Guided by The Grid Raising Attention With The Rule

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 21

Journalism Practice

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/rjop20

Guided by the Grid: Raising Attention with the Rule


of Thirds

Michael Koliska & Klive (Soo-Kwang) Oh

To cite this article: Michael Koliska & Klive (Soo-Kwang) Oh (2021): Guided by the Grid: Raising
Attention with the Rule of Thirds, Journalism Practice, DOI: 10.1080/17512786.2021.1916402

To link to this article: https://doi.org/10.1080/17512786.2021.1916402

Published online: 24 Apr 2021.

Submit your article to this journal

Article views: 35

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at


https://www.tandfonline.com/action/journalInformation?journalCode=rjop20
JOURNALISM PRACTICE
https://doi.org/10.1080/17512786.2021.1916402

Guided by the Grid: Raising Attention with the Rule of Thirds


a
Michael Koliska and Klive (Soo-Kwang) Ohb
a
Communication, Culture and Technology, Georgetown University, Washington, DC, USA; bCommunication
Division, Frank R Seaver College of Letters Arts and Sciences, Pepperdine University, Malibu, CA, USA

ABSTRACT KEYWORDS
The rule of thirds is a longstanding principle that helps artists and Rule of thirds; news
photographers to compose aesthetically pleasing images. Despite photography; main subject
the rule’s wide use by photographers, little is known about its recognition; visual cognition;
impact on viewers in particular to raise attention to the main visual information
processing; compositional
subject and narrative of an image but also to possibly limit principle; attention; vision
cognitive load. The findings of this exploratory, experimental
study of undergraduate students at a U.S. university suggest that
the rule of thirds serves as a geometric guiding structure that
directs participants’ vision. The rule of thirds appears to enhance
participants’ ease of information processing by raising attention
to the predominant visual content of an image, making it more
salient to the observer. Overall, the rule of thirds enabled a more
correct recognition of the primary valence and the main subject
or focus of interest of an image, indicating that participants may
glean a better understanding of the intended message of an
image. But findings also indicate significant differences between
the rule of thirds composition and the placement of the main
subject along the rule of thirds grid, suggesting the latter as
being more effective.

Introduction
The rule of thirds (RoT),1 in art and photography, refers to making use of the intersections
of four imaginary lines that divide an image into thirds, horizontally and vertically, to alter
the composition of an image (Smith 1797; Stern 2011). Although many people may not be
aware of the term “rule of thirds,” they may very well apply it on a daily basis when taking
pictures with their mobile devices. Digital cameras and camera apps normally have a RoT
grid option to assist the user in taking a photograph. Most photo editing software also
provide a RoT layer to guide the cropping process of a picture. Virtually any book on learn-
ing photography will discuss the RoT. One of the goals of employing the RoT is to
compose an aesthetically pleasing and balanced image composition2 but also to position
the most important information within the image in such a way as to communicate it sali-
ently (Smith 1797; Stern 2011; Ensenberger 2011; Amirshahi et al. 2014), thereby increas-
ing the probability of a more accurate identification of the most important information of
an image.

CONTACT Michael Koliska michael.koliska@georgetown.edu


© 2021 Informa UK Limited, trading as Taylor & Francis Group
2 M. KOLISKA AND K.-K. OH

The RoT as a guiding principle of image composition and image perception is


especially relevant in the digital news media environment, with its increasing emphasis
on visual culture (Russell 2011). Visual news has been predicted to be “replacing text-
based news as the main source of information” (Van der Haak, Parks, and Castells 2012,
2931). Photos and videos are and continue to be important in the online communication
economy on social media (Kane and Pear 2016) as well as in political communication and
participation (Geise, Heck, and Panke 2020). In fact, news consumers increasingly share
and consume visual information (Newman 2020). For instance, over the past four years,
the consumption of online news videos alone doubled across several countries in the
Americas and Europe (Newman 2016, 2020).
The proliferation of visual information has a number of implications for the production
and perception of visual news content. On the one hand, the demand for images as an
essential tool for journalistic storytelling, created a demand for multimedia skills (McLellan
and Newton 2012; Huang et al. 2006) to produce professional high-quality visual content
(Barnhurst 1993). Students in video and photography classes, including in the U.S., where
this study was conducted, learn using RoT in order to improve the composition of an
image (while abiding by journalistic principles and ethics). Possessing such a skill has
been deemed to be one of the differences between visual content produced by pro-
fessional journalists and amateurs (Seelig 2005).3
On the other hand, a shift toward more visual content may impact the way users, readers,
and audiences perceive information and news. Scholars speak of a visual primacy effect that
highlights the salience of visual cues over verbal ones (DePaulo et al. 1978). Empirical
research shows that visual cues can overshadow textual information or carry biases such
as in political news coverage (Grabe and Bucy 2009), environmental news (Lester and
Cottle 2009), conflict coverage (Greenwood and Jenkins 2015), but also on social media.
Van der Heide, D’Angelo, and Schumaker (2012) pointed out that people have the ten-
dency—at least on Facebook—to form impressions about others primarily based on their
profile photos, rather than on the verbal self-disclosures that accompany these images.
The proliferation of visual information and the dominance of visual over textual infor-
mation, as described by the visual primacy effect, indicate the need to improve the cre-
ation of visual information such as photographs. The RoT maybe able to assist in this
process by making the focus point of an image more salient and thus reducing the prob-
ability of information processing errors (Ensenberger 2011; Stern 2011). Yet, despite the
ubiquity of visual content and wide use of the RoT as a decision-making tool for compo-
sition (Seelig 2005), research has not expansively explored the effectiveness of the RoT as
a way to guide attention within visual communication. This exploratory study attempts to
fill this gap in research by theorizing and testing the RoT as a geometrical structure that
assists the processing of visual information (Lockhead 1989; Marr 1982, 2000; Palmer
1989; Uttal 1988; Pinker 1985).

Literature Review
The Rule of Thirds: Practice and Theory
Since the fifteenth century, artists have been using the RoT to make certain crucial parts of
an image more salient (Stern 2011). Specifically, the RoT “is used to determine those areas
JOURNALISM PRACTICE 3

Figure 1. Positioning of the main subject along the rule of thirds.

of interest which lie at the four intersections of four imaginary lines drawn to divide
the composition into thirds horizontally and again vertically” (Stern 2011, 353) (see
Figures 1 and 2).
Many visual artists in the past and photo professionals today maintain the belief that
composition rules will “warrant your attention for good reasons—[leading] you in the

Figure 2. Image composition complying with the rule of thirds.


4 M. KOLISKA AND K.-K. OH

right direction and produce desired effects in most circumstances” (Ensenberger 2011,
65). In particular, the main subject will often be put off-center, in one of the intersections
of the thirds, to avoid static effects (Figure 1). The RoT helps to balance and structure the
different elements of an image by “establishing visual priorities among the collection of
lines, shapes colors and textures” and “it helps orchestrate viewer eye movement that
encourages exploration of the entire photograph” (Ensenberger 2011, 67) (Figure 2).
Users of the RoT frequently accept these suppositions behind the rule as taken for
granted principles even though little research has tested them. Kaufman (1974) and
Zinchenko (1975) suggest that artistic practices have a lot of merit and can serve as a
good starting point for exploratory research. Zinchenko argues that artists’ theory often
precedes research because artists “recognized before psychologists did that the visual
image is the result of the observer’s active perceptual activity” (49).
Di Dio, Macaluso, and Rizzolatti (2007) found evidence that artistic rules can produce
measurable effects. Using functional Magnetic Resonance Imaging (fMRI) they were
able to show that the golden ratio (a + b is to a as a is to b) did in fact produce the
“desired effects.” Subjects exposed to a Greek statuary in its original canonical form
experienced strong feelings and identified the object as beautiful whereas subjects
that saw the same sculpture, but whose proportions had been altered so that it violated
the golden ratio, evaluated it as ugly. The fMRI results supported these subjective evalu-
ations as physical manifestations within the brain.
Despite its widespread use, relatively little research has attempted to theorize, measure
or explore the effects of the RoT. Yegiyan and Lang (2010) tested the centrality-focused
and periphery-focused aspects of visual content based on the RoT. They were able to
show that the centrality of an image will produce a higher emotional intensity and
valence. Yet their study only used the RoT to define a frame within an image, which
they called the periphery, while the centrality of an image was a square that was based
on the four intersectional lines of the RoT grid. In other words, Yegiyan and Lang didn’t
test if the RoT composition or the geometrical alignment of visual content will produce
any effects. In fact, most research focusing on the RoT in photography appears to take
the effectiveness of the rule for granted. Several studies examined computer vision i.e.,
the detection of RoT presence as marker of image aesthetics and quality but also compu-
tational object detection of information salience (Mai et al. 2011; Maleš, Heđi, and Grgić
2012; Singh, Arya, and Agrawal 2017). Other studies focused on how well computer
models compare to human perceptions of RoT images, indicating that computer
models can approximate human perceptions (Amirshahi et al. 2014; Osman et al. 2019).
Amirshahi et al. (2014) also correlated the RoT presence with aesthetical evaluations of
artistic and professional photos and paintings, suggesting that RoT may only play a
minor role in such assessments. Yet, little to no studies (across various fields including
journalism studies) have convincingly theorized why RoT may be an effective compo-
sitional rule by showing connections to human perceptions nor did much research test
whether RoT assists in human information processing by making the central information
of an image salient.
In this study, we argue that the RoT composition will guide attention and ease the pro-
cessing of the primary visual content of images. Specifically, we argue the RoT will sali-
ently raise attention to the main subject matter, easily convey crucial information of a
news story through the positioning of the main subject along the RoT grid, and
JOURNALISM PRACTICE 5

enhance engagement through a higher perceived valence. In other words, the RoT helps
to reduce the complexity of visual information by placing visual content within a geo-
metrical structure, akin to the human visual system.

Vision: A System to Reduce Complexity


The human visual system allows for one to see, comprehend and make sense of their sur-
roundings, and the human brain is considered to be more effective and faster in proces-
sing audio-visual than verbal information (Pike, Edgar, and Edgar 2012). Williams (2005)
for instance suggests the visual system can stimulate behavior before the conscious
mind is even aware of it. Similarly, Graber (1996) argues that the visual system is superior
in processing information. “One quick glance at complex visual scenes suffices to identify
situations that words would describe with far less fidelity in information transmission”
(Graber 1996, 86). Research shows that the visual system developed as an integral part
in the management of information and social complexity allowing for higher order think-
ing and processing of nuanced social information (Joffe and Dunbar 1997). Yet, while the
visual system is equipped to manage and process complex sets of information, the lack of
explicit guidelines such as syntactical and lexical rules in visual communication problema-
tizes the transfer of exact information through images and videos (Grabe and Bucy 2009).
Neuroscientists and psychologists have shown that human vision operates as a system
reducing complexity. Barry (2005) for instance combines neurological and psychological
principles with the study of visual communication. Her concept of perception theory
builds on the premise that “the eyes are, in fact, a direct extension of the brain into the
environment” (48). Perception is a very complex process and brain research shows that:
sight may indeed begin with light hitting the retina, vision occurs deep within the brain; and
that perception, the process by which we derive meaning from what we see, is an elaborate
symphony played first and foremost through the unconscious emotional system. (Barry 2005,
46)

In contrast, Pinker (1985) and later Williams (2005) speak of visual cognition describing the
idea that people behave, change and learn as a direct reaction to the input of visual infor-
mation. Information processing, behavior and knowledge are then responses to visual
recognition, which “allows us to determine on the basis of retinal input that particular
shapes, configurations of shapes, objects, scenes, and their properties are before us”
(Pinker 1985, 3).
Much of the research about visual cognition focuses on two aspects: first, on the rec-
ognition of shapes and surfaces with a concurrent frame of reference to establish spatial
relations in the visual cognition process; and second, on how these shapes and surfaces
may be remembered or recalled (Pinker 1985; Jacob and Jeannerod 2003). Overall the lit-
erature indicates that geometrical shapes, surfaces and redundancy will ease and speed
up information processing of visual input (Zusne 1970; Marr 1982, 2000; Lockhead 1989;
Palmer 1989; Uttal 1988). At the same time aesthetics may also play a role. Gestalt psy-
chology for instance predicts “that ‘good’ figures will be better remembered than
‘poor’ ones” (Zusne 1970, 63).
Even though shapes and redundancy help to process visual information more efficien-
tly there are several factors that can complicate this process. Newell (1990) points out that
6 M. KOLISKA AND K.-K. OH

“vision is built around eye movements of about 200 milliseconds” (358), which suggests
that visual information cannot be stored for a very long time due to the limited data
storage capabilities of the visual system. Moreover, since eye movements follow in
quick succession, perception can be influenced by forward or backward masking
(Turvey 1973). Masking describes an interference from one visual stimulus to another
that can either affect the following or the preceding stimulus (hence forward or backward
masking). At the same time emotions can also create a bias in visual perception (Stefa-
nucci 2010). Such cognitive limitations indicate that in order for vision to be effective, a
selection process must occur that focuses on crucial parts of the visual information acces-
sible to the visual cortex (Pylyshyn 1998). As indicated above, research shows that geo-
metrical structures can assist in the selection process. This idea of selecting and
reducing complexity in vision through geometrical structures is echoed in the RoT grid,
where visual information is placed to geometrically align content and guide attention.

Attention, a Guiding Principle for Vision


Researchers describe attention as a type of selection process distinguishing relevant from
irrelevant stimuli, which is required to process visual information. Tipper and Weaver
(1998) suggest that the visual system relies on attention; without it, objects cannot be
processed, nor recognized. Pylyshyn (2003) stresses the importance of the “focal attention
as an interface between vision and cognition” (88). Attention, in other words, enhances
specific spatial points or areas, which define or constitute particular objects. Specifically,
attention heightens but also diminishes aspects of perception. Thus, attention operates
like a filtering system that includes and excludes visual categories from the focus of
awareness (Kaufman 1974). Pylyshyn (2003) argues that the filtering process exists to
prevent overloading of the visual system and to reduce interference on perception and
cognition.
Several researchers have suggested that attention is guided by specific patterns and
symmetries of geometrical shapes that facilitate information processing (Pylyshyn 1998;
Lockhead 1989; Palmer 1989). Uttal (1988) posits that “human visual systems are sym-
metry information-processing systems” (145). In other words, specific layouts of a scene
or surface representations, that highlight the edges of surfaces, induce attention,
making perception possible. Pylyshyn (2003) describes visual analysis or vision as “a
highly complex information processing system [that] individuates, or picks out, objects
in a scene and computes the spatial layout of visible surfaces and the 3D shapes of the
objects in the scene” (50).
Marr (2000) conceptualizes the complexity of the visual process and cognition by
defining vision from the user-centered perspective. His model, which converges the con-
cepts of vision, attention and an even-spaced grid, specifically evokes the idea that the
RoT may mimic the function of the human visual system on a rudimentary level (see
Figure 3). Marr describes vision as “a process that produces from images of the external
world a description that is useful to the viewer and not cluttered with irrelevant infor-
mation” (76). Marr’s model stresses that the filtering of information is achieved through
geometrical structures that raise attention and accelerate the processing of the filtered
visual information. Marr (1982) speaks of three different levels of analysis within the
visual process, which all utilize geometrical structures akin to a coordinate system or a
JOURNALISM PRACTICE 7

Figure 3. Marr’s 2½ sketch (1982, 278) highlights how the RoT appears to mimic the visual system.

grid. The “primal sketch” provides important information about the two-dimensional
sphere by detecting intensity changes and a geometrical distribution. The “2 ½-D
sketch” highlights spatial relationships within a viewer centered coordinate frame. The
last stage is the “3-D model representation” which changes the frame to an object cen-
tered coordinate frame to “describe shapes and their spatial organization” (Marr 2000, 79).
While scholars point out that perception is not governed by specific laws (Uttal 1998),
they agree on an underlying mechanism of attention, i.e., filtering visual information on
the basis of geometrical structures that are spatially aligned. Attention can then be
described as a process, which is guided by symmetry and geometrical structures that
assist in the filtering of information and contribute to the ease of information processing
within the complexity of visual cognition. The idea of the RoT as a geometrical structure
that guides attention may then have merit. As a compositional tool RoT assists human
vision by placing visual content on a grid, guiding the observer’s attention toward the
most crucial visual information of an image. Thus, allowing for an easier, faster and less
error prone processing of visual content. Following the rationale of these theoretical
underpinnings, images that follow the RoT should make the focus of interest more acces-
sible and easier to be understood. Thus, also making, among other aspects, the message
and valence of an image4 more salient.
Research shows that attention is closely connected to valence (Lane, Chua, and Dolan
1999; Morris, Leclerc, and Kensinger 2014), which has been described as a form of cogni-
tion (Duncan and Barrett 2007). Valence perceptions can thus be considered as an
expression of information processing, reflecting a level of attention or how salient the
main message of an image may be. We posit that measuring and comparing valence per-
ceptions of images that are compliant and noncompliant with the RoT can thus indicate
how the RoT aides image information processing.
8 M. KOLISKA AND K.-K. OH

In summary, on the basis of these theoretical propositions, we posit that the RoT
manipulates the visual structure of an image to raise attention, facilitating an easier
visual analysis and understanding of a photograph. In order to examine this proposition,
this study explores two research questions, which were deemed appropriate for this first-
time research inquiry synthesizing and exploring these concepts (Salkind 2010). First,
research has often argued that one of the critical aspects of understanding the narrative
of an image is to perceive the valence (positive or negative) of the image (Seelig 2005;
Yegiyan and Lang 2010) as a form of information processing (Barry 2005). Therefore,
we explore whether ratings of the perceived valence of images are significantly
affected by the compliance with the RoT.
RQ1. Does the rule of thirds have a significant effect on the perceived valence of images?

Another important aspect of image production is to place the main subject or focus of
interest in such a way that it draws attention to the central information conveyed by
the image (Ensenberger 2011). The literature on vision suggests that the RoT would
help raise attention to the main subject of an image, which is arguably the starting
point for perceiving the main information or narrative of an image. Thus, we sought to
explore the relationship between the perceived main subject of the image and compli-
ance with the RoT.
RQ2. Does the rule of thirds have a significant effect on the perceived main subject of images?

Methodology
Design
In this experimental exploratory study, we employed a within-subjects 2 (valence type) ×
2 (rule of thirds composition) × 2 (rule of thirds main subject) repeated-measures design.
“Valence type” refers to the positive or negative (two levels) types of emotions of an
image; “rule of thirds composition” indicates whether an image composition was in com-
pliance with the RoT (two levels: “yes” and “no”); and the “rule of thirds main subject”
referred to whether the main subject of the image was situated on one of the four RoT
grid intersections (two levels: “yes” and “no”).
Employing the repeated-measures design to test “perceived valence” and “perceived
main subject of image” ensured stronger statistical power while also being economical,
requiring fewer participants (Murphy and Myors 2008). The repeated measures design
also made sense as the level of variability between participants across all images (that
contain different characteristics of the independent variables) was low and the design
allowed for a higher number of respondents to produce a given number of observations
in contrast to other experimental designs that only match participants to one image or
independent variable (Lomax 2007). Thus, due to the nature of the within-subjects exper-
iment design the sample size was deemed appropriate. The 41 participants completing
the study were each exposed to 24 different news images in random order (see below),
making the total n = 984 for the experiment. For this research design, power analysis indi-
cated that a total sample of 34 participants would be needed to detect large effects (d
= .8) with 90% power using a t-test (with alpha at .05).
JOURNALISM PRACTICE 9

Sample Population
A total of 52 undergraduate students majoring in journalism were recruited from a large
mid-Atlantic university in the United States via in-class announcements. Participants were
told they would view images and answer a short set of questions for each image. In the
recruitment phase, participants were also informed of the potentially graphic nature of
some images based on the topics being portrayed and were given the option to
decline participation. Under the instructors’ discretion, students who participated in the
study received extra credit, or if choosing not to participate in the study, students were
provided with an opportunity to complete a separate assignment for the same credit.
Of the 52 students surveyed, 41, aged 19–25, completed the survey. Twenty-four partici-
pants were female (58.6%) and seventeen were male (41.4%).

Material
Twenty-four news images across six news topics were selected for this study. The six news
topics were determined by surveying 25 university students in a pretest where they dis-
cussed perceived controversy levels of 20 news topics. The perceived controversial nature
of the images was an important element for this study because we sought to understand
how the respondents would perceive the valence (positive or negative) of the image
differently based on various RoT conditions. In other words, we postulated that depicting
controversial topics (pro vs con perspectives) would result in a distinct valence structure
(positive or negative emotions). The following six topics received the most responses in
the pretest, which we then selected for this study: drinking (healthy lifestyle vs binge
drinking), gay marriage (pro vs con), Greek life (brother/sisterhood vs hazing), football
(glory of sport vs injuries), U.S. troops abroad (fighting for our country vs mistreating pris-
oners), and President (pro vs con).
Four news images for each news topic were selected using a Google image search of
each topic followed by each word in parentheses above (e.g., “Greek life brotherhood/sis-
terhood” and “Greek life hazing”). Two images per story and per aspect (i.e., two positive/
two negative) were selected randomly from the search results. The image orientation was
not a factor in selection of photos for this study but the majority of images were horizon-
tal (22) while only two were vertical. The general Google image search was used deliber-
ately to reflect the generic exposure to news images online. For the purpose of this
experiment, we aimed to control perceived valence type by selecting images for all
topics that tended in one valence direction (either positive or negative). For instance,
negative images about gay marriage depicted only anti-gay marriage protesters
holding disparaging signs and positive images depicted only happy (smiling) just-
married gay couples. By the same token positive military images only showed smiling sol-
diers and negative images depicted scenes of a wounded soldier or abuse of a prisoner by
a soldier. Using a wide range of topics allowed us to control for possible individual biases
regarding a specific topic.
In order to further ensure that our selection of images portray a valence direction (posi-
tive or negative) we conducted a manipulation check of the stimuli material. Ten photo-
journalism experts with expertise in both still photography and videography (professional
photographers—including photojournalists at CNN, the White House Press Corps and the
10 M. KOLISKA AND K.-K. OH

Table 1. Compliance with the rule of thirds composition (RoT C).


Compliant (RoT C) Non-compliant (No-RoT C)
Positive Negative Positive Negative
246 369 246 123
Sub-totals 615 369
Total 984

Table 2. Compliance with the rule of thirds main subject (RoT MS).
Compliant (RoT MS) Non-compliant (No-RoT MS)
Positive Negative Positive Negative
287 287 205 205
Sub-totals 574 410
Total 984

Baltimore Sun,—university educators or both) examined the selected images for compli-
ance with the RoT. The experts also rated and determined the valence (as either positive
or negative) of each image. Their comments about each image determined the main
subject of each image, and the levels for each independent variable (“rule of thirds com-
position,” “rule of thirds main subject”), which was essential for coding the experiment
survey data. In sum, we operationalized the experts’ opinion as the accurate understand-
ing of the main subject, composition and the intended valence of each image (positive/
negative).
Respondents’ answers were then coded as either “congruent with experts” opinion or
“not congruent with experts” opinion. As a result, 15 out of the 24 images were identified
as complying with the RoT in terms of composition, and 14 images as having main sub-
jects complying with the RoT. In summary, out of 984 total cases, there were 492 per
valence type (positive, negative); 615 following the RoT in terms of image composition;
and 574 with the main subject following the RoT (see Tables 1 and 2).

Procedure
Participants received a link to an online survey hosted by Qualtrics software. Upon activat-
ing the link, participants received brief instructions concerning the procedures and the
option to withdraw at any time from the experiment. Then each of the 41 participants
was shown the stimulus material (24 images) one image at a time, equaling 984 image
views in total. In order to avoid potential carry over effects arising from the order of
images shown, we randomized the image sequence for each participant (Lomax 2007).
After the exposure to each image participants had to fill out a three questions survey.5
The first question asked about the perceived valence of the image or how it made partici-
pants feel, using 1–9 Likert scale (1 being “extremely negative” and 9 being “extremely
positive”). The second and third questions asked participants to write down what they
perceived as the main subject of the image and what they thought the story was as
could be told from the image. In the online survey, participants were not able to navigate
back and forth from image to image. Participants took on average 42 min to complete the
experiment. The completion time ranged from 23 to 57 min.
JOURNALISM PRACTICE 11

Results
Rule of Thirds and Perceived Valence
In order to address the first research question—Does the rule of thirds have a significant
effect on the perceived valence of images?—we tested whether participants perceived
the direction of valence for an image to each image’s predetermined valence (positive
or negative). A one-way ANOVA with mean values for perceived valence ratings for
each image valence type (positive n1 = 492; negative n2 = 492) found that valence type
of the image was a significant factor for the respondents’ perceived valence and its direc-
tion. Participants’ responses were highly skewed toward each end of the scale (1 and 9) in
the direction of the predetermined (positive and negative) valence (M1 = 6.82, SD = 1.38,
M2 = 3.09, SD = 1.66, F(1, 982) = 1473.80, p < .001).
As indicated in Tables 1 and 2 the distribution of images with a positive and negative
valence in respect to RoT composition and main subject is not fully balanced. This imbal-
ance is due to the Google search that we conducted, which aimed to reflect a wide variety
of images, as they are found online. We took the conservative approach and tested if the
unequal distribution may skew the valence perception. We ran several independent
sample t-tests to examine if equality of variances is met. The Levene’s tests for equality
of variances were not statistically significant indicating that the different groups’ variances
are similar. Thus, the group size differences did not impact the analysis. Perceived valence
scores were more negative when a negative valence image complied with the “rule of
thirds composition” (M = 2.69, SD = 1.54) then if it did not (M = 4.27. SD = 1.45), t(490) =
9.98, p < .001, d = 1.05. A t-test between the positive valence groups also showed that par-
ticipants perceived valence stronger when images complied with the rule n = 246 (M =
7.32, SD = 1.26) than if they did not n = 246 (M = 6.33, SD = 1.31), t(490) = 8.59, p < .001,
d = .75.
A t-test was also conducted to test the unequal group sizes for main subject RoT com-
pliance and valence. Participants perceived a higher positive valence when images com-
plied with the RoT in respect to the main subject position n = 287 (M = 7.13, SD = 1.36)
than when images did not have a main subject placed in the RoT grid n = 205 (M = 6.39,
SD = 1.29), t(490) = 6.1, p < .001, d = .56. Another t-test confirmed the overall tendency
that compliance with the RoT evokes stronger directional valence perceptions. Participants
perceived images more negatively when the main subject was positioned in the RoT grid
(M = 2.61, SD = 1.44) than if it wasn’t (M = 3.75, SD = 1.73), t(389) = 7.67, p < .001, d = .71.
We also ran a number of ANOVAs to compare valence perceptions (positive and nega-
tive) in respect to the presence of the rules of thirds (i.e., RoT compliance/ non-compliance)
and whether the main subject was positioned in the RoT grid or not. All of the ANOVAs sup-
ported the t-test results. First, we compared positive valence perceptions of images compli-
ant with the RoT with images that were non-compliant. Again, the presence of RoT
composition showed that participants perceived an image as more positive (M1) than
without RoT (M2) composition (M1 = 7.32, SD = 1.26, M2 = 6.33, SD = 1.31, F(1, 491) = 73.84,
p < .001). We then repeated this step by comparing whether negative valence perceptions
followed a similar pattern. Again, when images were compliant with the RoT, negative
valence perception was more pronounced (M1) than without RoT (M2) compliance (M1 =
2.69, SD = 1.54, M2 = 4.27, SD = 1.45, F(1, 491) = 99.61, p < .001) (Figure 4 and Table 3).
12 M. KOLISKA AND K.-K. OH

Figure 4. Interactions for composition (RoT C and No-RoT C) and perceived valence (positive and
negative).

Next, we examined if the valence perceptions were affected by the presence of the main
subject (MS) in the RoT grid. Both the positive valence (M1 = 7.13, SD = 1.36, M2 = 6.39, SD =
1.29, F(1, 491) = 37.28, p < .001) and the negative valence (MS on RoT grid M1 = 2.61, SD =
1.44, MS not on RoT grid M2 = 3.75, SD = 1.73, F(1, 491) = 62.44, p < .001) were significantly
more pronounced when the main subject was positioned on RoT grid than not (Figure 5).
In order to see if the previous tendencies are robust, we further examined whether
there were significant differences in the overall perceived valence ratings of the images
(positive and negative combined) across the two RoT characteristics (composition,
main subject). A two-way ANOVA was conducted to compare means of perceived
valence for the two types of RoT characteristics. When RoT was followed for composition
(n = 615), the average perceived valence rating was 4.59 (SD = 2.71), and 4.34 (SD = 2.57)

Table 3. Valence, main subject and composition compliance with the rule of thirds.
Rule of thirds No-rule of thirds
Mean SD Mean SD F
Valence (positive) 7.32 1.26 6.33 1.31 73.84**
Valence (negative) 2.69 1.54 4.27 1.45 99.61**
Valence (positive) RoT MS 7.13 1.36 6.39 1.29 37.28**
Valence (negative) RoT MS 2.61 1.44 3.75 1.73 62.44**
Valence (positive and negative) RoT C 4.59 2.71 4.34 2.57 64.87**
Valence (positive and negative) RoT MS 6.59 1.36 5.37 1.65 15.39**
*p < .01, **p < .0016.
JOURNALISM PRACTICE 13

Figure 5. Interactions for main subject (RoT MS and No-RoT MS) and perceived valence (positive and
negative).

when it was not (n = 369), showing a significantly different rating when the RoT for com-
position was followed, F(1, 983) = 64.87, p < .001. Similarly, RoT for the main subject was
also a significant factor for differences in the overall perceived valence ratings. Compli-
ance to RoT for main subject (n = 574) showed a mean of 6.59 (SD = 1.36) and non-com-
pliance a mean of 5.37 (SD = 1.65). Differences were significant, F(1, 983) = 15.39, p < .001.
We further examined the valence and RoT compliance by analyzing each type of RoT
compliance. Differences in mean perceived valence was significantly higher when the RoT
was not followed in terms of composition, but followed in terms of main subject (F(1, 983)
= 6.55, p < .05). In other words, placing the main subject on the RoT grid proved to be a
more significant indicator for differences in valence perception (Figure 6).

Rule of Thirds and Identification of Main Subject


To explore the second research question—Does the rule of thirds have a significant effect
on the perceived main subject of images?—we conducted a two-way ANCOVA to test the
differences of mean scores recorded in each case. An ANCOVA with valence as a covariate
was used in order to test only the most important aspects of an image that may pertain to
congruency in identifying the main subject, thus ensuring lower error probability and
raising statistical power. We wanted to see whether the average participant could
locate the main subject in an image for any valence type and whether this task would
be affected by an image’s compliance with the RoT.
14 M. KOLISKA AND K.-K. OH

Figure 6. Interactions between rule of thirds (main subject) and rule of thirds (composition).

ANCOVA showed that only the RoT in terms of main subject was a significant factor for
congruency of main subject identification (see Table 4). When the RoT was present for
main subject (n = 573), the mean score for congruency was .59 (SD = .49) and when it
was not present (n = 411), the mean score was significantly lower at .50 (SD = .50), F(1,
983) = 11.43, p < .01.
Also, significant interactions were found for each type of RoT compliance. Differences
of congruency between predetermined (expert based) main subject identification and
respondents were significantly higher when the RoT was not followed in terms of compo-
sition, but followed in terms of main subject (F(1, 983) = 54.70, p < .001). Thus, the RoT
compliance with regard to main subject yielded higher scores, making it easier for partici-
pants to identify the main subject (Figure 7).

Conclusion and Discussion


This exploratory study took a first step to test possible effects of the RoT on participants’
perceptions of news images. We tested whether the RoT would make the main subject
and the valence of a news photograph more salient. Our data analysis provides a tentative

Table 4. ANCOVA of rule of thirds characteristics and identification of main subject.


RoT No-RoT
Mean SD Mean SD F SS df MS
Main subject .59 .49 .50 .50 11.43** 2.64 1 2.64
Thirds Composition .57 .50 .52 .50 .015 .003 1 .003
a
R Squared = .071 (Adjusted R Squared = .067); **p < .01.
JOURNALISM PRACTICE 15

Figure 7. Interactions between rule of thirds (main subject) and rule of thirds (composition).

but promising indication that the RoT may be more than simply an aesthetical compo-
sitional tool; it appears to also assist in the information processing of visual stimuli. The
results for both of our research questions–that explored whether the RoT has a significant
effect on valence and main subject perception—suggest that participants’ perceptions
were indeed guided by the RoT grid as they processed the visual information more accu-
rately. The perceived valence of images was significantly higher when those images fol-
lowed the RoT in their composition than when they didn’t. The differences for perceived
valence was also higher when the main subject was positioned in the intersections of the
RoT grid, than if it was not. This means that valence perception was stronger (more vivid)
for images with the main subject complying with the RoT. In other words, participants
were better able to correctly recognize the most important part of an image and as
such an integral part of the narrative of the image (Barrett 2006; Ensenberger 2011).
The RoT also aided the identification of the main subject. Our analysis of covariance
showed that participants were more likely than not to identify the main subject when
it was located in the RoT grid. The interaction effects that we found also pointed
towards the importance of placing the main subject on the RoT grid. Valence was
higher when the main subject was located on the RoT grid than when an image
merely followed a “rule of thirds composition.” This indicates that the RoT could have sig-
nificant effects in conveying the story of a picture more forcefully, making it more salient
and easier to process.
These findings are promising indicators that the RoT may ease information processing
and as such has the potential to minimize possible errors in visual communication (Turvey
1973; Pinker 1985). Specifically, our findings suggest, that the RoT could help avoid mis-
identifications of the main subject. The geometrical structures guide the process of vision
16 M. KOLISKA AND K.-K. OH

Figure 8. Military photo with positive valence and without RoT grid subject placement.

by raising attention toward the focus of interest of an image hereby filtering out irrelevant
information. The process of visual cognition is guided by attention, which in turn is pri-
marily triggered by the recognition of geometrical shapes and surfaces, symmetry and
spatial relationships (Marr 1982, 2000; Lockhead 1989; Palmer 1989; Uttal 1988; Pinker
1985; Tipper and Weaver 1998; Kaufman 1974; Pylyshyn 1998). We argue that a similar
process appears to facilitate the effectiveness of the RoT in establishing “visual priorities.”
The significance of our results emphasizes the importance of the RoT for journalistic
storytelling—not merely to raise attention, but also to effectively convey the focus
point of an image. Images complying with the RoT can be considered more efficient in
pointing toward relevant news content. But more specifically, when photojournalists
have to choose between overall image composition or placing the main subject on to
the RoT grid, they should favor the latter over the former to most effectively communicate
visual content (see Figures 8 and 9 below, for example, photos used in the study). The
need to clearly communicate visual content is getting increasingly important as the

Figure 9. Military photo with positive valence placing main subject on RoT grid.
JOURNALISM PRACTICE 17

consumption of visual news content continuous to grow (Newman 2020). Findings also
indicate that in particular controversial news concerning social, political, or environmental
issues (Grabe and Bucy 2009; Lester and Cottle 2009; Greenwood and Jenkins 2015) would
benefit from employing the RoT to emphasize the valence of an image, as increasing the
emotional and/or affective salience of an image can facilitate greater engagement with
the visual content (Broersma et al. 2019).

Limitations and Prospects for Future Research


Even though our study shows significant effects, there are a number of limitations that need
to be addressed to move research on the RoT forward. First of all, there are no clear guide-
lines about the use and implementation of the RoT (Ensenberger 2011). This makes it very
challenging to assess images, and even experienced photojournalists and experts often dis-
agree on whether a particular photo follows the rule or not. This also created some pro-
blems in our study since we designed our project to reflect the everyday news photo
consumption experience online. The Google Image search did not provide us with an
equal number of images that followed the RoT and those that did not comply with the com-
positional guideline. In a follow up study, it would seem prudent to implement a series of
tests by using images that unmistakably abide by the RoT or to possibly use a machine
vision analysis to help detect the presence of the RoT (Mai et al. 2011). At the same time,
it would be greatly beneficial to control for other aspects of an image such as color, con-
trast, brightness, depth of field, object size, image orientation (vertical vs horizontal) and
so on to clearly parse out the impact of the RoT on information processing and main infor-
mation detection of an image. Such examinations would also allow to investigate the per-
ception of overall news photography quality in relationship to the presence and absence of
the RoT in order to tease out possible aesthetical effects.
Moreover, our conceptualization of valence has some limitations because not all indi-
viduals may follow normative valence perceptions. In other words, some may evaluate
distraught faces, abuse, or anger as positive rather than negative. Thus, our study
cannot claim to universally describe, measure or predict valence. In future studies it
would also be worthwhile to expand the research, using a larger population sample,
test and control for political partisanship, attitudes or other biases that may be influencing
participants perceptions of controversial topics. The fact that we used an online exper-
imental setting may also have influenced our findings and it may be worthwhile to
conduct a similar study in a more controlled environment.
This study is just a first step in looking more closely at the significance of the RoT in
conveying relevant information within journalistic storytelling. Much more needs to be
done to understand this compositional rule and its effects on visual cognition especially
as images are becoming increasingly pivotal in the selection of news content and influ-
ential in the understanding of news.

Notes
1. While the RoT is a widely used concept in the visual arts and image creation including video,
painting, drawing, animation, rendering, etc., this research focuses on photos (still images)
and specifically on the use of RoT in news photography.
18 M. KOLISKA AND K.-K. OH

2. Ensenberger (2011) points out that the RoT is an important tool to create balance in a photo-
graph. “Balance implies that the elements within the frame have a sense of visual weight,
larger objects obviously outweigh smaller objects, and dark or saturated objects weigh
more than lighter colored ones. The positioning of each element and its proportional
relationships to other elements is critical to the visual hierarchy within the composition. It
communicates to the viewer what’s most important in your photograph” (68).
3. Having a command of the RoT is just one skill among several others that can distinguish pro-
fessionals from amateurs. Other critical skills include lighting, image clarity/quality, angle
usage, distance selection, and editing skills (see Greenwood and Thomas 2015; Nilsson and
Wadbring 2015; Mortensen 2014; Barnhurst 1993).
4. By valence we refer to the affective and/or emotional quality of an image, which assumes a
bipolar dimension ranging from negative to positive. Valence is an important factor in per-
ception and has been described as a basic building block of narratives and human emotional
life (Barrett 2006).
5. Given the number of images each participant had to view and comment on, we limited the
survey questions to avoid fatigue effects.
6. This has to be right underneath the table

Acknowledgement
We want to thank Dr. John E. Newhagen for his inspiration, guidance and support.

Disclosure Statement
No potential conflict of interest was reported by the author(s).

ORCID
Michael Koliska http://orcid.org/0000-0003-2098-2630

References
Amirshahi, S. A., G. U. Hayn-Leichsenring, J. Denzler, and C. Redies. 2014. “Evaluating the Rule of
Thirds in Photographs and Paintings.” Art & Perception 2: 163–182.
Barnhurst, K. 1993. “The First Exercise: What It Teaches About Photojournalism Practice.” The
Journalism Educator 48 (2): 56–66.
Barrett, L. F. 2006. “Valence is a Basic Building Block of Emotional Life.” Journal of Research in
Personality 40 (1): 35–55.
Barry, A. M. 2005. “Perception Theory.” In Handbook of Visual Communication: Theory, Methods, and
Media, edited by K. Smith, S. Moriarty, G. Barbatsis, and K. Kenney, 45–62. Mahwah: Lawrence
Erlbaum Associcates.
Broersma, M. 2019. “Audience Engagement.” In The International Encyclopedia of Journalism Studies,
edited by T. P. Vos, F. Hanusch, A. Sehl, D. Dimitrakopoulou, and M. Geertsema-Sligh,, 1–6. Wiley.
DePaulo, B. M., R. Rosenthal, R. A. Eisenstat, P. L. Rogers, and S. Finkelstein. 1978. “Decoding
Discrepant Nonverbal Cues.” Journal of Personality and Social Psychology 36 (3): 313–323.
Di Dio, C., E. Macaluso, and G. Rizzolatti. 2007. “The Golden Beauty: Brain Response to Classical and
Renaissance Sculptures.” Plos One 2 (11), e1201: 1–9.
Duncan, S., and L. F. Barrett. 2007. “Affect Is a Form of Cognition: A Neurobiological Analysis.”
Cognition & Emotion 21 (6): 1184–1211. doi:10.1080/02699930701437931.
Ensenberger, P. 2011. Focus on Composing Photos. Burlington: Focal Press.
JOURNALISM PRACTICE 19

Geise, S., A. Heck, and D. Panke. 2020. “The Effects of Digital Media Images on Political Participation
Online: Results of an Eye-Tracking Experiment Integrating Individual Perceptions of ‘Photo News
Factors’.” Policy & Internet 13 (1): 54–85.
Grabe, M. E., and E. P. Bucy. 2009. Image Bite Politics: News and the Visual Framing of Elections. New
York: Oxford University Press.
Graber, D. 1996. “Say It with Pictures.” The ANNALS of the American Academy of Political and Social
Science 546: 85–96.
Greenwood, K., and J. Jenkins. 2015. “Visual Framing of the Syrian Conflict in News and Public Affairs
Magazines.” Journalism Studies 16 (2): 207–227.
Greenwood, K., and R. J. Thomas. 2015. “Locating the Journalism in Citizen Photojournalism: The use
and Content of Citizen-Generated Imagery.” Digital Journalism 3 (4): 615–633. doi:10.1080/
21670811.2015.1034528.
Huang, E., K. Davison, S. Shreve, T. Davis, E. Bettendorf, and A. Nair. 2006. “Bridging Newsrooms and
Classrooms: Preparing the Next Generation of Journalists for Converged Media.” Journalism &
Communication Monographs 8: 221–262.
Jacob, P., and M. Jeannerod. 2003. Ways of Seeing: The Scope and Limits of Visual Cognition.
New York: Oxford University Press.
Joffe, T. H., and R. I. Dunbar. 1997. “Visual and Socio–Cognitive Information Processing in Primate Brain
Evolution.” Proceedings of the Royal Society of London. Series B: Biological Sciences 264: 1303–1307.
Kane, G. C., and A. Pear. 2016. “The Rise of Visual Content Online.” MIT Sloan Management Review,
January 4. https://sloanreview.mit.edu/article/the-rise-of-visual-content-online/.
Kaufman, L. 1974. Sight and Mind: An Introduction to Visual Perception. New York: Oxford University
Press.
Lane, R. D., P. M.-L. Chua, and R. J. Dolan. 1999. “Common Effects of Emotional Valence, Arousal and
Attention on Neural Activation During Visual Processing of Pictures.” Neuropsychologia 37 (9):
989–997. doi:10.1016/S0028-3932(99)00017-2.
Lester, L., and S. Cottle. 2009. “Visualizing Climate Change: Television News and Ecological
Citizenship.” International Journal of Communication 3: 920–936.
Lockhead, G. 1989. “Category Bounds and Stimulus Variability.” In Object Perception: Structure and
Process, edited by B. E. Shepp and S. Ballesteros, 267–295. Hillsdale: Lawrence Erlbaum Associates.
Lomax, R. G. 2007. Statistical Concepts: A Second Course of Education and the Behavioral Sciences. 3rd
ed. New York: Longman.
Mai, L., H. Le, Y. Niu, and F. Liu. 2011. Rule of Thirds Detection from Photograph. In 2011 IEEE
International Symposium on Multimedia (pp. 91–96). IEEE.
Maleš, M., A. Heđi, and M. Grgić. 2012. Compositional Rule of Thirds Detection. In Proceedings
ELMAR-2012, 41–44, September. IEEE.
Marr, D. 1982. Vision: A Computational Investigation into the Human Representation and Processing of
Visual Information. San Francisco: W.H. Freeman.
Marr, D. 2000. “Vision.” In Minds, Brains, and Computers: The Foundations of Cognitive Science: An
Anthology, edited by R. Cummins and D. D. Cummins, 69–83. Malden: Blackwell Publishers.
McLellan, M., and E. Newton. 2012. Digital Training Comes of Age: How Knight Journalism Fellows and
Trainees benefit – and Why They Want Even More. http://www.knightfoundation.org/media/
uploads/publication_pdfs/KFTrainingFieldReportWEB.pdf.
Morris, J. A., C. M. Leclerc, and E. A. Kensinger. 2014. “Effects of Valence and Divided Attention on
Cognitive Reappraisal Processes.” Social Cognitive and Affective Neuroscience 9 (12): 1952–1961.
doi:10.1093/scan/nsu004.
Mortensen, T. M. 2014. “Blurry and Centered or Clear and Balanced?” Journalism Practice 8 (6): 704–
725. doi:10.1080/17512786.2014.892703.
Murphy, K. R., and B. Myors. 2008. Statistical Power Analysis: A Simple and General Model for
Traditional and Modern Hypothesis Tests. 2nd ed. Mahwah: Lawrence Erlbaum Associates.
Newell, A. 1990. Unified Theories of Cognition. Cambridge, MA: Harvard University Press.
Newman, N. 2016. Reuters Institute Digital News Report. Oxford: Reuters Institute for the Study of
Journalism.
20 M. KOLISKA AND K.-K. OH

Newman, N. 2020. Reuters Institute Digital News Report. Oxford: Reuters Institute for the Study of
Journalism.
Nilsson, M., and I. Wadbring. 2015. “Not Good Enough?” Journalism Practice 9 (4): 484–501. doi:10.
1080/17512786.2015.1030135.
Osman, T., S. S. Psyche, T. Deb, A. Firoze, and R. M. Rahman. 2019. “An Algorithmic Approach to
Estimate Cognitive Aesthetics of Images Relative to Ground Truth of Human Psychology
Through a Large User Study.” Journal of Information and Telecommunication 3 (2): 156–179.
doi:10.1080/24751839.2018.1542574.
Palmer, S. 1989. “Reference Frames in the Perception of Shape and Orientation.” In Object Perception:
Structure and Process, edited by B. E. Shepp and S. Ballestero, 121–163. Hillsdale: Lawrence
Erlbaum Associates.
Pike, G., G. Edgar, and H. Edgar. 2012. “Perception.” In Cognitive Psychology, edited by N. Braisby and
A. Gellatly, 65–99. Oxford: Oxford University Press.
Pinker, S. 1985. Visual Cognition. Cambridge, MA: MIT Press.
Pylyshyn, Z. 1998. “Visual Indexes in Spatial Vision and Imagery.” In Visual Attention, edited by R. D.
Wright, 215–231. New York: Oxford University Press.
Pylyshyn, Z. W. 2003. Seeing and Visualizing: It’s not What you Think. Cambridge, MA:: MIT Press.
Russell, A. 2011. Networked: A Contemporary History of News in Transition. Cambridge: Polity.
Salkind, N. J. 2010. Encyclopedia of Research Design. Thousand Oaks: Sage.
Seelig, M. 2005. “A Case for the Visual Elite.” Visual Communication Quarterly 12 (3/4): 164–181.
Singh, N., R. Arya, and R. K. Agrawal. 2017. “A Novel Position Prior Using Fusion of Rule of Thirds and
Image Center for Salient Object Detection.” Multimedia Tools and Applications 76 (8): 10521–
10538.
Smith, J. T. 1797. Remarks on Rural Scenery. London, UK: Nathaniel Smith.
Stefanucci, J. K. 2010. “Emotional High: Emotions and the Perception of Spatial Layout.” In Social
Psychology of Visual Perception, edited by E. Balcetis and G. D. Lassiter, 273–297. New York:
Psychology Press.
Stern, K. 2011. Photo 1: An Introduction to the Art of Photography. Clifton Park: Delmar.
Tipper, S. P., and B. Weaver. 1998. “The Medium of Attention: Location-Based, Object-Centered,
Scene-Based?” In Visual Attention, edited by R. D. Wright, 77–107. New York: Oxford University
Press.
Turvey, M. T. 1973. “On Peripheral and Central Processes in Vision: Inferences from an Information-
Processing Analysis of Masking with Patterned Stimuli.” Psychological Review 80 (1): 1–52.
Uttal, W. R. 1988. On Seeing Forms. Hillsdale: L. Erlbaum Associates.
Uttal, W. R. 1998. Toward a New Behaviorism: The Case Against Perceptual Reductionism. Lawrence
Erlbaum Associates Publishers.
Van der Haak, B., M. Parks, and M. Castells. 2012. “The Future of Journalism: Networked Journalism.”
International Journal of Communication 6: 2923–2938.
Van der Heide, B., J. D. D’Angelo, and E. M. Schumaker. 2012. “The Effects of Verbal Versus
Photographic Self-Presentation on Impression Formation in Facebook.” Journal of
Communication 62 (1): 98–116.
Williams, R. 2005. “Cognitive Theory.” In Handbook of Visual Communication Research: Theory,
Methods, and Media, edited by K. Smith, S. Moriarity, G. Barbatis, and K. Kenney, 193–210.
Mahwah: L. Erlbaum.
Yegiyan, N. S., and A. Lang. 2010. “Processing Central and Peripheral Detail: How Content Arousal
and Emotional Tone Influence Encoding.” Media Psychology 13 (1): 77–99.
Zinchenko, V. P. 1975. “Problems of Visual Culture in the Light of Modern Investigations of Visual
Perception and Visual Thinking.” Russian Social Science Review 16 (2): 47–58.
Zusne, L. 1970. Visual Perception of Form. New York: Academic Press.

You might also like