Effects of novel news photographs
 
 
 
 
Effects of novel news photographs on readers' interest in
 and memory for newspaper content
 
 
 
 
by
 
Andrew Mendelson
Southern Illinois University at Edwardsville
 
April 1, 1998
 
 
 
 
 
 
Paper submitted to the Newspaper Division of AEJMC
 for the 1998 conference, Baltimore
 
 
 
 
 
 
 
Correspondence to:
 
Andrew Mendelson
Department of Mass Communications
Box 1775
Southern Illinois University at Edwardsville
Edwardsville, IL 62026-1775
tel.: (618) 692-2232
fax: (618) 692-3716
e-mail: [log in to unmask]
 Effects of novel news photographs on readers' interest in
 and memory for newspaper content
 
 
 
 
 
  The idea that news photography is highly conventional is not new.  Critics of
photojournalism suggest that most newspaper photography is repetitive both in
content and composition.  As Komenich said, "It's the same picture you took last
week with somebody else in it....Instead of the city council boss, it's the
mayor" (Komenich, 1994).  Or as Rosenblum (1978) put it,  "While the dramatis
personae may change over time as public officials, baseball players, and
murderers come and go, the basic news scenarios remain the same" (p. 427).  (see
also Hall, 1973)
  Apart from content, the style and composition of news photographs also are
conventional.  Composition has been defined as how the content elements of
photographs are arranged in the frame (Fox & Kerns, 1961; Kerns, 1980; Geraci,
1984; Hoy, 1986; Lewis, 1995). It is the style of news pictures, perhaps even
more than their content, that makes them recognizable.
  Photographers rely on a "formula" aesthetic, Rosenblum (1978) argued, that is
a highly conventionalized system for structuring cultural products.  This
produces "standard pictures for standard situations" (p. 428).  As Hagaman
(1996) saw it, the style of news photographs is not as simple as what the editor
says you can and cannot do.  Rather, it is built into what is considered an
appropriate subject for a newspaper (content) and what was considered the
correct way to photograph that subject (composition).
  Hagaman (1996) went on to state that news photographers
     solve the problem [of photographing a situation] by choosing from a
repertoire
     of standard types of pictures that used a limited number of visual
components and
     compositional devices (p. 11).
 
  This section has suggested that there is a limited range of content in news
photographs and a limited range of compositional devices used to convey this
content.  If news photographs are so conventional, then photographs may not
offer anything "new" to the viewer and so may not be looked at for long.  To
this end, the purpose of this study is to examine the effects of novelty in
content and composition of news photographs on newspaper readers' attention and
memory for the photographs and the accompanying stories.  This research will
attempt to isolate how atypical news photographs are processed.  Do people
remember these photographs and the accompanying stories better?  Further, the
studies proposed here hope to extend past research in a number of ways.  First,
it will use news photographs as the stimulus objects, instead of meaningless,
geometric shapes or illustrations of common scenes such as kitchens.  Second,
this research will attempt to demonstrate that readers do have expectations for
both the content and composition of news photographs.  Third, this research will
examine the effect of novelty on accompanying words in the naturalistic setting
of newspaper pages.
  This research has grown out of research suggesting the limited effects of news
photographs on readers.
 
Lack of effects for photographs in newspapers
  While it would seem obvious that photographs would have an effect on newspaper
readers, a review at past research indicates that effects only appear depending
on what methodology is used.  Certainly there is much anecdotal evidence of the
effects of various pictures on readers.  Goldberg (1991) cites numerous examples
through history of how pictures changed the way we see the world and ourselves.
For example, she argues that photographs have changed the way we see
politicians, social issues, and ourselves by their existence and use.  The
United States's recent experience in Somalia and the resulting images are cited
as causing pressure to remove the troops (Somalia, 1994).  But these examples
are photographs that are quite different from the everyday photographs one finds
in a newspaper, in terms of their distinctiveness.  Not every photograph that
runs in a newspaper stands out to readers.
  Most empirical studies compare a photograph-absent condition to a
photograph-present condition (see Levie & Lentz, 1982 for a review; Wanta, 1988;
Huh, 1996), rather than comparing one photograph condition to another.  These
studies show that stories with photographs are remembered better than stories
alone.  In a review of literature on picture-word interaction, Levie and Lentz
(1982) consistently found that pictures facilitate learning from text when
compared to no-picture conditions.  Glenberg and Langston (1992) found that
readers organized text better in memory with a picture than without.  Wanta
(1988) and Huh (1996) showed that stories with a photograph are ranked as more
important than the same story without a picture.
  The studies that do compare one photograph type to another either show few
effects or are highly confounded methodologically (Wanta & Roark, 1993).  Garcia
and Stark (1991), using Eye-Trac cameras, showed that photographs provide a
useful entry point to news stories, and readers "process" photographs more than
text.  They defined "processing" as just looking at a photograph.  These authors
did not suggest how long people actually studied the news photographs, or if any
effect on memory could be seen.  Wanta and Roark (1993) attempted to show what
impact different photographs have on readers' memories for news stories.
However, these authors did not use a systematic method for varying the
treatments.  They showed people different photographs in the pages in which they
appeared and measured what was remembered about the story.  But the story
content and size were not controlled for; thus, there are too many confounds to
allow thorough interpretation of the impact of the photographs.
  Other research has that photographs have little impact on newspaper readers'
processing of news stories.  Mendelson (1996) examined the effect of different
photographs on the attribution of responsibility for homelessness.  This
research was based on frame theory, which posits that changing the slant of a
news story (e.g., from being about an individual to being about society) will
affect how people attribute responsibility for social issues.  Two different
photographs were used, both of which were typical newspaper photographs.  One
showed a homeless person;  the other showed people rallying against
homelessness.  People were asked to read a story about homelessness that was
paired with one of the photographs (as part of a set of newspaper pages).  The
results showed that the photograph did not affect people's attributions about
responsibility for this issue; only the story content made a difference.  People
seldom mentioned the photographs in their recall of what they had seen on the
topic.  This was a first indication of minimal effects of typical newspaper
photographs.
  Thorson and Mendelson (1996) looked at the effects of photographs of Hillary
Rodham Clinton (HRC) on processing news stories.  In two experiments the author
varied the role in which HRC was portrayed.  In the first experiment (Thorson &
Mendelson, 1996), all the photographs used were typical newspaper photographs.
As in the previous study (Mendelson, 1996), no effects on recall and interest
were found for the different photographs, only for the stories.  Recall for the
photographs was much less than the recall of the stories, of which people could
cite many details.  This suggested that perhaps people don't look at typical
news photographs as sources of much information and that people do not spend
much time processing them.
  In the second experiment (Thorson & Mendelson, 1996), a new set of photographs
was used, in addition to the typical photographs used above.  These photographs
were award-winning photographs by P.F. Bentley (1993).  They were
behind-the-scenes images of Bill Clinton's 1992 Presidential campaign, and as
such, were very different from typical newspaper photographs.  Results showed
that these images did have an effect on the readers.  People found the
accompanying story more interesting in the presence of the atypical photographs
and seem to have noticed these photographs while normal news photographs were
ignored.  This study did not differentiate between the content and the
composition of the photographs, so it was not yet possible to say which aspect
affects people more.
  In this second experiment (Thorson & Mendelson, 1996), a no-photograph
condition was included to determine the impact of photograph presence in
comparison to absence.  Results showed that people found both photographic
conditions more interesting than the no-photograph condition.  So, the mere
presence of a photograph is enough to make a story more interesting regardless
of the picture's content.
  What can be concluded is that any photograph is better than no photograph, but
it is hard to say that one typical newspaper photograph has more impact than
another.  Perhaps people simply like stories to have photographs to break up the
grayness of the news pages, but do not look to the photographs for information.
As Levie and Lentz (1982) state, "Although most students do look at text
illustrations in most situations, they usually do not 'study' the pictures
unless prompted to do so.  Much research has indicated that people normally
notice only the global aspects of a picture" (p. 223).  Levie and Lentz (1982)
argued that people are not in the habit of learning from pictures, so
photographs may have to vary significantly from "normal" or familiar for people
to be drawn to study them and use them for information.
  From the limited effects of newspaper photographs in empirical research, it
appears that newness or atypicality is a prime factor for study.  I hope to show
this experiment that novelty is essential to getting people to spend time with
news photographs.
  To examine the impact of novelty of news photographs on readers' curiosity and
memory, a number of topics from psychology will need to be examined:  (1) schema
theory, to see how people organize their experiences with the world; (2) the
concept of novelty, in general, and more specifically how novelty can be
examined in the content and composition of news photographs; (3) the processes
of attention and exploration; (4) and the results of research showing effects of
novelty on interest and memory.
  In the next section, I will begin the literature review by examining how
people organize previously encountered information.  This draws on the concepts
of cognitive structures of schemata and categories.
 
Schema Theory
  People do not process and store every detail of events they have experienced.
This would require too much effort and too much memory.  Psychologists use the
concept of schema to represent how information is stored in long term memory.
Schemata are abstract knowledge structures that embody expectations
(abstractions or stereotypes) of the way things should be based on past
experience, each concept with its own schemata.  These expectations include the
attributes of the experience and the relations among these attributes (Abelson,
1981; Fiske & Taylor, 1991).
  Schema theory assumes that people use what they know about the world to
understand newly encountered events, whether real or mediated (Abbot, Black &
Smith, 1985).  Information that cannot be derived from existing memory
structures will be what we attend to.  "The schema prepares the person to see
certain kinds of things; consequently, little attention need be paid to those
things that match expectations, leaving attentional resources free to devote to
the more unusual, and therefore more informative, items" (Mandler, 1983, p.
105). Schemata are based on experience and, therefore, are dynamic structures
that change as we acquire more information.  With each experience we are testing
the schemata and updating those parts that are new.
  A schema is said to be activated as soon as the initial identification of one
or more objects in a scene is made (Mandler & Parker, 1976; Mandler & Johnson,
1977; Mandler & Ritchey, 1977).  We quickly process the overall gist of
situations, then move on to the unfamiliar details.  As will be seen, this fits
with how people process pictures.
  There probably exists a continuum of expectations; not an either/or.  For any
given event or scene, some attributes are always expected, others sometimes
expected, and still others seldom expected. A graded structure has especially
been shown in research on categories, memory structures of types of objects,
like animals or fruit.  Barsalou (1985) showed that members of categories vary
in how typical or expected they are.  For example, if given the category of
fruit, individuals are more likely to name "apple" or "orange" than a more
exotic fruit like "nectarine."  People are also very consistent and in agreement
as to these levels of typicality.
  We can see that schemata are abstractions from experience based on what is
most expected.  They are used to guide processing of incoming information, and
we focus on what is new or what is not known.  Schemata are dynamic,
ever-changing structures, and as such,  there is a continuum of expectations or
typicality.  As will be shown in the section on content novelty, there is a high
degree of agreement among people as to what is expected in many events or
scenes.  Given the large amount of time people spend with the media, people
should have strong expectations for attributes of types of news photographs.  In
the next section, I will look more closely at one type of schema: the scene
schema.
 
Scene Schema
  There are many different types of schemata referred to in past research, each
accounting for the organization of different types of knowledge (e.g., events,
roles, categories, etc.).  For the purpose of these studies, the concept of the
scene schema will be examined.  A scene schema is a structure that contains
information about what we expect different scenes to look like, an inventory of
typical objects and the expected locations relative to each other within the
scene (Mandler & Parker, 1976; Mandler & Johnson, 1977; Mandler & Ritchey, 1977;
Biederman, 1981
  It seems reasonable to expect that news photographs could be used by readers
to form scene schema.  We should have expectations for scenes commonly portrayed
in news photographs.  For example, we should expect to see flames, firemen,
charred buildings, etc. when we think of a "fire" scene.
  Besides the content of news photographs,  I predict that people also will have
expectations for the composition of news photograph types.  People have seen so
many news photographs that they should have developed a compositional component
of a scene schema, though it may not be as expressible as the content component.
Such a component would consist of the typical composition of the photograph.
  I have now examined how people organize expectations in memory.  The next
sections will examine the concept of novelty.  First, I will look at the concept
of novelty in general as used in psychological research.  Then, I will explicate
the concepts of content and composition novelty.
 
What is Novelty?
  People do not monitor everything in the world equally.  They especially want
to know what is new or what is changing "out there."  From schema theory, this
makes sense, as there is no need to focus on what we already known.  In essence,
novelty is anything that breaks with expectations.  This definition fits well
with the literature on schemata, which are abstractions of what is most
expected.  Moving on a continuum away from these expectations, the more novel
events will be those with lower probabilities of occurrence or those that have
been encountered less frequently (Berlyne, 1951; Greenberger, Woldman &
Yourshaw, 1967; Hutt, 1970; Barnard, Breeding & Cross, 1984;  Martindale, Moore,
& West, 1988).  This definition suggests that novelty is based on each person's
past experience (as schemata are based on one's experiences).  Something is only
informative based on what the viewer already knows (Wentworth & Witryol, 1986).
Even though novelty is based on experience, people show much agreement, as
people within a culture have many of the same experiences.
  Past literature refers mainly to two types of novelty for visual objects:
absolute and relative (Hutt, 1970; Berlyne, 1951, 1971).  Absolute novelty is
something that has never been seen before, while relative novelty refers to a
familiar object in a new place or previously experienced elements in
unprecedented combinations.  A stimulus could also be novel in relation to what
has immediately preceded it (Berlyne, 1951; Nunnally & Lemond, 1973).
  The next two sections will break the concept of novelty into two parts:
content and composition.  All of the past research on visual novelty has dealt
with content rather than composition.
 
Content Novelty
  As mentioned in the beginning of this paper, many of the events covered by
newspapers are quite routine.  Almost daily, newspaper photographers cover
meetings, speeches, rallies, fires, and crime scenes.  People should have
expectations for what types of objects and people appear in news scenes.
  Content is novel when people do not expect an object or person to appear in
the specific scene.  For example, a familiar person or object could be in a new
setting or role or an unfamiliar person or object could be in a common setting
(Durham, Nunnally and Lemond, 1971).
  The pictures most researchers have used are of common scenes with objects
varying in levels of novelty (unless only a single object is shown).  The entire
scene is usually not novel.  There have been two methods for selecting novel
objects for pictures.  The first method is to have the experimenter choose the
stimuli; the second is to have a preliminary group of people generate the
stimuli.  Regardless of which method is selected, a group of people then rate
the stimuli to confirm the assigned novelty.  As I used the first method of
selection for these studies, I will elaborate only on that one.
  In the first method, the experimenter either creates or finds stimuli that
theoretically fit his/her concepts of novel.  The participants in the study
confirm these choices by independently rating the stimuli on novelty scales
(Berlyne, 1957, experiment 1; Nunnally, Lemond & Wilson, 1977; Lewis, 1978).
These researchers created stimuli of varying levels of novelty by combining
aspects of various creatures to create increasingly bizarre images.  For
example, they combined parts of a cow, airplane, and elephant to make a new
creature.  Participant ratings showed that their stimuli did show increasing
levels of novelty.
  Goodman (1980) chose objects for standard scenes that were more or less
relevant to the scene schema.  She then had a group of students and non-students
rate these objects for relevancy to the scene.  Beltran and his colleagues
(Beltran, Herrando, & Pelegrina, 1992; Beltran & Duque, 1993; Beltran &
Herrando, 1994) have investigated the typicality of objects in a variety of
environmental scenes.  They defined content novelty as the suitability of a
given element in a contextual situation.  By having people rate the suitability
of objects the authors selected for scenes on a 0 to 10 scale, natural groupings
for high, medium, and low typicality were found.
  Thorson and Mendelson (1996) selected news photographs that showed Hillary
Rodham Clinton (HRC) in either typical or atypical settings.  The entire scene
(apart from HRC) was atypical or typical of political coverage in newspapers.
The photographs chosen as atypical (award-winning photographs by P.F. Bentley
(1993)) were behind-the-scenes glimpses of the 1992 Clinton presidential
campaign showing aspects not usually covered by most news.  Participants in the
study confirmed the greater novelty of these scenes.  (These scenes also were
chosen for compositional novelty, which will be discussed in the next section.)
It should be noted that this appears to be the only study to have examined the
issue of content novelty in news photographs.
  There are evident limitations in the past research on content novelty.  First,
for the most part, only objects within a scene have been examined for novelty;
the scenes themselves were held constant.  In terms of news photographs, not all
scenes may be as familiar to readers.  Readers may not have a developed schema
for all news photographs.
  Another problem seems to exist in many novelty studies is that what authors
called low probability stimuli seem to be "no probability" or bizarre stimuli.
For example, in one case subjects were shown a farm scene with a tractor (high
probability) in the foreground.  In the other version of this picture, the
tractor was replaced by an octopus sitting in front of the barn.  Because of
this, the experiment may not be testing attention to informative areas, but to
random items.  A better test would be to compare the tractor to a carriage.
Both items could be found on a farm, but a carriage is less expected.
  The above literature review led to the following operational definition for
rating the content in news photographs for expectedness or novelty:
     Regardless of composition, how often have you seen newspaper photographs of
     [news topic] having this content (i.e., subject matter, setting,
characters, props,
     action, gestures, etc.)?
 
  No known studies have established categories for content novelty of news
photographs.  In an initial pilot test, I will have people generate types of
news photographs they expect to see most often in newspapers.  This will be
elaborated upon in the stimulus materials design section.  In the next section,
the discussion of novelty will continue with an explication of
composition/stylistic novelty.
 
Compositional and stylistic novelty
  Earlier, I discussed the typical look of most news photographs.  This look,
based on common features of composition, went beyond the content of news
photographs.  Even so, the concept of novelty of style and composition has not
been explicated before.  This section will examine the various aspects of
photographs that could be varied to increase novelty.  It should be noted that
most of the studies reported below have not really measured people's
expectations for style and compositional features.
  Some authors have examined composition from a holistic level, as features of
composition tend to depend on and interact with other features.  Hagaman (1996)
discussed how the composition of her pictures evolved away from typical news
photograph style to a "new aesthetic."  Unlike her newspaper portraits, which
were tightly cropped, she began to frame her subjects more loosely, including
their feet and surroundings.  "I was making formal changes in the structure of
the photographs, including more information in the frame and consistently using
a greater depth of field in order to have more space from foreground to
background in focus" (p. 36).
  In an empirical study, Thorson and Mendelson (1996, experiment 2) varied the
composition of news photographs between traditional newspaper photographs and
the award winning, "behind-the-scenes" work of P.F. Bentley (1993).  As
mentioned in the last section, these photographs not only showed HRC in atypical
settings (content novelty), but they also were shot in atypical ways; for
example, including more of the setting in the frame.  Thorson and Mendelson did
not, however, specifically differentiate between content and composition of the
photographs, nor did this study break the images down into individual
compositional elements.
  Most authors of photojournalism textbooks break down composition into
individual elements.  However, no studies have completely examined readers'
expectations for these features, as was done for content.  Certainly all
variations of compositional features are not seen as often as every other.  The
rest of this section will examine the various elements that are used when
composing photographs.
  The only known study that looked at readers' preferences for compositional
features was an uncited report in a textbook by Fox and Kerns (1961).  This
study showed that for two compositional features (cropping and framing), people
preferred the tightly cropped and framed image.  This study, while minimal in
scope, at least shows that people can notice differences in compositional
elements when shown both versions.
  The first general area of composition addresses the point of view that the
viewer has toward the subject of a photograph.  Photography allows people to see
the world from many different and unusual vantage points (Szarkowski, 1966).
Subject-viewer point-of-view can be further broken down into vertical camera
angle, horizontal camera angle and proximity to the subject (Messaris, 1992).
  Vertical camera angle has been studied more than any other compositional
feature, not for typicality, but for effects on attitudes based on different
angles (high or low).  Most often the effects due to vertical camera angle are
based on "analogy with 'real-life' experience" (Messaris, 1992, p. 183).  We
feel superior to subjects we look down upon and inferior to subjects we look up
at.  Camera angle has been shown to influence viewers' perceptions of
politicians (Kepplinger & Donsbach, 1987), viewers' evaluations of characters in
a story (Kraft, 1987), and viewers' evaluation of advertised products, but only
when viewers had little motivated to process the ads (Meyers-Levy & Peracchio,
1992).
  Another feature of composition that a photographer can manipulate is
horizontal camera angle relative to the subject.  The photographer, in many
cases, can shoot from any position around the subject (360 degrees).  This
change in position gives viewers completely new ways of seeing familiar
subjects.
  The final way point-of-view can be manipulated is through the apparent
distance between the subject of a photograph and the viewer (or camera).  This
is shown through how much of the subject is visible in the frame and how much
space surrounds the subject.  The more space and the more of the subject that is
visible, the further away the subject is perceived to be.  This is affected by
the photographer moving closer to the subject or by the photographer choosing a
telephoto lens.
  Intraub and colleagues (Intraub, Renders, & Mangel, 1992; Intraub &
Richardson, 1989) showed that people have expectations (schematic ideas) about
how close objects in pictures should be (as indicated by how much of the picture
frame is filled), and this influences how they remember the pictures.  When an
object is shown too close to the subject, people recall the image as having more
space around the object.  The opposite is true for wide-angle pictures when the
object seems too far away.  People mentally crop in close to the subject.  These
authors argued that a person's general knowledge about pictures includes
prototypic viewing distance associated with depicted objects that is formed over
time.
  The second large group of features deals with the form, shape, or positioning
of the content as graphic elements in the photograph.  The photographer uses
graphics to create a sense of direction, which orients viewers and helps them
decide where the action starts, how it progresses, and where it ends
(Metallinos, 1996).
  The graphic use of shapes, lines, and curves link objects in the frame and
guide people around the frame.  One example is the use of leading lines, lines
that draw or lead the readers toward the focal point of the picture.  By placing
the subject in a spatial relationship to such forms, the photographer is able to
contain the viewer's eye.  Moreover, the dominant use of vertical lines can
suggest dignity, horizontal lines serenity, diagonal lines motion and depth, and
curves grace and beauty (Fox & Kerns, 1961; Kerns, 1980; Hoy, 1986; Lewis,
1995).
  The photographer also is able to control how the contents of a photograph are
distributed in a frame.  For example, the photographer can control where the
subject is placed horizontally in the frame.  Photojournalism textbooks instruct
that the major subject of interest should never be in the center of the frame,
and that photographers should avoid chopping a picture in half, either
vertically or horizontally.  The horizon line should always fall either above or
below the mathematical center of picture (Geraci, 1984; Hoy, 1986).  A
compositional guide to aid photographers in this is the "rule of thirds."  Lewis
(1995) stated that the photographer should visually divide the frame into three
equal parts vertically and horizontally.  The rule says that subjects should be
placed on one of the lines or at their junctions in order to create a more
dynamic photograph (see also Kerns, 1980).  Photographers also can frame a
subject with an object in the foreground (Lewis, 1995) and compose their picture
around repetitive or similar shapes or patterns (Fox & Kerns, 1961; Kerns, 1980;
Lewis 1995).
  Light also can be used as a graphic element in photographs.  Like camera
angles, light can be applied from different horizontal and vertical angles.
Also, light can be used in different intensities to create different moods (Fox
& Kerns, 1961; Feinberg, 1970; Kobre, 1980; Keene, 1995; Metallinos, 1996).  For
typical news photographs, the tendency is that the light source should not be
obvious and that the photographs should not be overly contrasty (Feinberg, 1970;
Kobre, 1980).
  Most examinations of composition describe only the elements of composition.
Only one known empirical study examined photographs stylistically in terms of
overall composition and specific features of composition.  Henry and Solano
(1983) developed a coding system for photographic variables, which included four
stylistic variables.  These authors were specifically interested in how
personality variables relate to the photographs people take.  Henry and Solano
first had 22 non-photographers take up to 36 pictures each (mean number of
photographs per person = 24).
  Two coders then rated all the pictures on four stylistic variables:  (1)
distance of photographer from subjects (close, medium, and far);  (2)  placement
of horizon line (not present, top, middle, bottom third of the frame);  (3)
motion (subject in motion, some other feature in motion, no motion);  and (4)
conventionality (a picture was unconventional if "either the subject of the
photograph or the viewpoint from which the picture was taken was highly
unusual...." (p. 82)).  The two coders had 76% agreement for distance, 55%
agreement for horizon, 85% agreement for motion, and 95% agreement for
conventionality.  Though it seems unbelievable that with such a vague definition
of conventionality that judges could get 95% agreement, in actuality, very few
photographs fell into the unconventional category (less than one picture per
photographer).  In the other categories, most of the photographers in this study
took photographs that were from the medium or far distance.  This would probably
not be true for professional photojournalists, who are taught to get close to
their subjects.  In terms of the horizon line, most pictures either had no
horizon or one that cut across the middle of the frame.  Finally, most of the
photographers used in this study showed no motion in their pictures.  This is
not surprising since shooting motion is a skill that photography students are
specifically taught.
  Although the results of Henry and Solano (1983) study might not be the same if
professional photographers were used, it does suggest that composition groupings
are possible to uncover.  The results of this study reveal it also is possible
to get coder agreement on stylistic variables and that these photographers were
fairly consistent in the levels of the stylistic variables they chose.
  Through this examination of stylistic and compositional novelty, it is
apparent that the current studies are starting nearly from scratch in terms of
how people expect news photographs to be composed.  From the above literature
review, the following operational definition for rating the composition in news
photographs for expectedness or novelty was derived:
     Regardless of content, how often have you seen newspaper photographs of
[news
     topic] having this composition (i.e., camera angle, lighting, proximity,
lens
     choice/distortion, how the content elements are combined, etc.)?
 
  Using the operational definitions for content and compositional novelty,
photographs were selected that were novel and typical for both content and
composition (two independent variables each with two levels for a total of four
different conditions).
 
  Typical Content
  Novel Content
  Typical Composition
 
 
  Novel Composition
 
 
 
The exact procedures will be elaborated in a later section.
  So far, I have explicated the expected and unexpected aspects of photographs
(the independent variables) that people have in mind.  The next section will
examine the concept of attention in terms of how people selectively explore
their environments, and, more specifically, how they process pictures (the
dependent variables).  Then, with this done, I will be able to see how novelty
affects attention, and formulate a set of hypotheses.
 
Attention, and Picture Processing
Attention
  Attention has been defined in many ways over the years.  Most of the
definitions of attention can be separated into two groups:  those about the
selective nature of processing (a spotlight notion) and those about the capacity
or intensive nature of attention (Kahneman, 1973).  For this study, I am
interested in the former; the idea of attention as a conscious, directed
processing, which is attracted by something new in the world.  To set a
framework, I begin with an examination of models of information processing.
  Cowan's (1988, 1995) model (see Figure 1) of attention and memory shows that a
stimulus first enters a sensory store, during which information in long-term
memory is automatically activated (schema).  Stimuli that are discrepant from
prior experience or especially significant to one's goals enter the focus of
attention.  His model uses a nested relation to illustrate that, structurally
speaking, the activated memory is a subset of long-term memory.  The focus of
attention is a subset of activated memory.  He postulated a central executive
that controls attention based on goals or external stimuli.  Thus, this model
shows that we focus our attention on material that does not fit with our prior
experiences. (see also Norman, 1968; Kihlstrom, 1984; Hoffman, 1987)
  Assumptions made by this model include a single memory store, automatic
semantic or meaning analysis, and attention positioned late rather than early in
the sequence of cognitive operations.  Perceptual analysis in this model is
unconscious and automatic and is not limited by attentional resources.  Which
stimuli are selected for attention or further conscious processing depends on
their novelty or pertinence to current cognitive activities.  Consciousness is
the end product of several unconscious memory processes.
  The model suggests the difference between automatic and controlled attentional
processes (Schneider, Dumais, & Shiffrin, 1984; Bargh, 1988).  Automatic
processing is a fast, parallel, fairly effortless process that is not limited by
attentional resources and is not under direct control of the viewer.  Controlled
processing is characterized as slow, generally serial, effortful, and
capacity-limited.  Controlled processes or attention is used to understand novel
or inconsistent information.  This is similar to the differentiation between
pre-attentive processing and focal attention (Neisser, 1967; Duncan, 1984) or
unconscious and conscious processing (Kihlstrom, 1984).
  We scan the world using automatic processing and when there is something
novel, we switch to controlled processing to incorporate the new information
(Kahneman, 1973).  The automatic matching of a stimulus with its stored
representation does not imply knowledge of everything related to that stimulus.
Unattended inputs are only partly analyzed.
  From these pre-attentive processes we focus our attention.  Attention, in this
light, is the process whereby some incoming stimuli are selected for more
complete processing.  Selective attention also has been used interchangeably
with the concepts of awareness, consciousness or working/primary memory (Matlin,
1983; Kihlstrom, 1984; Shiffrin, 1988; Cowan, 1995).  Attended inputs receive
more processing than unattended inputs and are remembered better (Norman, 1968;
Lachman, Lachman, & Butterfield, 1979).
 
Picture Processing:
  The above model of attention fits well with research on how people process
pictures.  Not all information in a picture is understood at once, so processing
happens over time.  As Kolers (1973) stated,  "The more familiar or simpler the
picture is the more readily its content can be identified or described" (p. 25).
Like words, we spend more time processing those parts of a picture that are new
to us.  When we come across a new word, we must stop and puzzle out its meaning
(if we can) by the context or its roots.
  Parker (1978) defined three steps of picture processing:  (1)  Extract
information from a wide area of the picture.  This identifies the global
information or "gist" of the picture (the main features or theme).  (2) Compare
this information with expectations or schemata.  (3) Take the information from
step two and use it to direct the eyes to the areas for direct fixation based on
information mismatch.  Similar models have been offered by other authors (Fisher
et al., 1983; Rabbit, 1984)  People systematically detect meaningful patterns
which they use to decide where to look and in what sequence to seek further
information.  Such a model is similar to the earlier cited attention models.
  The initial glance not only determines the global properties of the object, it
also effectively discriminates the properties such as novelty (Cupchik &
Berlyne, 1979).  In a single fixation (100-300 milliseconds) of an image, a
person can correctly identify the type of scene or the "gist" of a scene (Potter
& Levy, 1969; Biederman, Rabinowitz, Glass & Stacey, 1974;  Loftus, 1976;
Potter, 1976; Mandler & Johnson, 1977; Rayner & Pollatsek, 1992).
  So, attention is a process that begin with some sort of automatic featural
description of a stimulus, which leads to activation of memory structures
(schema), and then leads to further processing, when, for example, a stimulus is
novel.  Such models of attention also were shown to fit well with past work on
picture processing.
  Beyond conceptualizations of the attentional processes, there must be ways to
measure attention.  Since it is not possible to know the internal workings of
the brain, researchers use a variety of operational definitions of attention and
exploration.  The next sections will examine the effect of novelty on attention
using a variety of operationalizations.  I will look at the results of previous
studies that show how novelty affects attention and memory, leading to the
formulation of hypotheses.
 
Effects of Novelty on Interest
  People tend to report more interest in novel imagery using a self-report
measure for attention.  Berlyne (1963) had people rate a variety of patterns and
found that they rated the more irregular or novel as more interesting.  In a
later study (1974a), he was able to show that interestingness ratings declined
monotonically with each exposure (or a decline in novelty).  In both these
studies, pleasingness or liking was not shown to be correlated with
interestingness. (see also Lemond, 1978)
  Usually, interest is measured through a semantic differential scale:
interesting (Berlyne, 1963, 1974a; Thorson and Mendelson, 1996). The idea that
novel imagery will be found to be more interesting and not more pleasing leads
to two hypotheses and a research question.
  H1:  Photographs novel in content will be rated as more interesting than
photographs
      not novel in content.  Pleasingness will not be associated with content
novelty.
  H2:  Photographs novel in composition will be rated as more interesting than
      photographs not novel in composition.  Pleasingness will not be associated
with
      compositional novelty.
  RQ1:  How will content and compositional novelty interact for interest
ratings?
 
  The last section will address the effects of novelty on memory for
photographs themselves and for text accompanying the photographs.
 
Effects of novelty on memory for photographs
  Now that we have seen that novelty can get people to look at and spend more
time with an image, the next question is what is the effect that this increased
processing has on memory?  Attention, as conscious processing, is a precursor to
memory, therefore, it is not surprising that past research shows an effect of
novelty on memory.
  Although it is true that memory storage can take place automatically,
explicit, direct recall of stored material is possible only with the presence of
attention both at the time of encoding and at the time of recall.  Controlled
attention allows for a more complete encoding of a stimuli, and thus, better
memory (Bargh, 1984; Fisk & Schneider, 1984; Cowan, 1995).
  In a unique test of schema and memory, Brewers and Treyens (1981) used a real
office in which they placed objects of varying typicality and then brought
people into the office to wait for the "actual experiment."  Thus, the students
were free to examine whatever caught their interest.  The authors then
correlated the students' recall and recognition with schema expectancy.  These
students had better recall for atypical objects.
  Friedman (1979) found better memory for more novel items in a common scene.
People were able to notice changes to unexpected items better than expected
ones.  In fact, people seldom noticed changes of one expected object for a
different version of the expected object.  Memory for elements may have been
better because participants were told they would have to distinguish the old
pictures from new ones.
  These studies lend support to the idea that the typicality of a photograph's
content will affect memory for that photograph.  Memory in the these experiments
was measured using free recall only.  The research on novel images and memory
leads to two hypotheses and a research question.
  H3:  Content-novel photographs will be recalled better than non-novel
photographs.
  H4:  Composition-novel photographs will be recalled better than non-novel
photographs.
  RQ2:  How will content and compositional novelty interact with memory?
Effects of novelty on memory and interest for stories
  In the context of newspapers, it will be important to see if novel photographs
have an effect on the perceptions of and the memory for an accompanying story.
Novel photographs, being more interesting, may cause people to process the story
more or perhaps people will spend more time with the photograph to the detriment
of story memory.  It would seem logical that if someone is more curious about
the photograph, they would then become more interested in the story, and then
perhaps learn more from the story.
  The only known study to specifically examine the effect of novelty of
photographs on stories in a newspaper was by Thorson and Mendelson (1996),
albeit only for stories about Hillary Rodham Clinton.  When a photograph was
novel, interest in the accompanying story increased.  No effect of photograph
novelty was found for recall of story content.
  In an early examination of the effect of novel images on an accompanying text,
Paradowski (1967) paired common and novel animals with the same generic,
descriptive paragraph.  He found that there was a significant improvement in
learning for an incidental and intentional task.  The intentional task dealt
with learning from the accompanying paragraph, while the incidental learning
task dealt with learning about the background or setting in which the animal
appeared.  The participants were only told they were to be tested on memory for
the text.
  This leads to several hypotheses and research questions as to the effect of
novel photographs on story recall and interest.
  H5:  Stories paired with content-novel photographs will be found more
interesting than stories accompanied by non-novel photographs.
  H6:  Stories paired with composition-novel photographs will be found more
interesting than stories accompanied by non-novel photographs.
  RQ3:  How will content and compositional novelty interact for story
interestingness?
  H7:  Material from stories accompanied by content-novel photographs will be
recalled at higher rates than stories accompanied by non-novel photographs.
  H8:  Material from stories accompanied by composition-novel photographs will
be recalled at higher rates than stories accompanied by non-novel photographs.
  RQ4:  How will content and compositional novelty interact for story recall?
 
Having reviewed the literature and proposed a set of hypotheses, the next
chapter will present the designs for the three experiments used to test these
hypotheses.
 
Stimulus Materials Design
  The hypotheses proposed in the previous sections were tested through two
experiments.  Each of the experiments used the same independent variables,
photographs varying in content and compositional novelty.  This created four
different photographic conditions (typical content/typical composition; typical
content/novel composition; novel content/typical composition; and novel
content/novel composition).
  Within each condition, there were photographs of different news topics and
multiple examples of photographs within each topic.  Having multiple news topics
and multiple photographs for each topic prevents systematic bias due to one
message (Reeves & Geiger, 1994).
 
Pilot Test 1 -- General News Photograph Topics
  The first step was to determine what people thought were typical or common
photographic topics or events in the newspaper.  Students in two beginning
journalism courses (news writing and advertising copy-writing) were asked:
     "Please write down what topics or events you would expect to see
photographs
     of in a newspaper?"
 
  These students were told not to include sports topics, as this was beyond the
scope of the current research.  Thirty students gave a total of 219 different
photographic possibilities, which were grouped by the author into general news
categories.  From this list, topics that were mentioned by more than 10 people
were considered for the studies.  This resulted in 10 topics of news
photographs:  features, local/cultural events, politics, natural disaster, war,
celebrities/entertainment, accidents, crime, fire, and meetings.  The top ten
topics accounted for 76% of the 219 ideas.  To limit the scope of the current
research studies, photographic categories were limited to hard news items that
ran with a story.  This eliminated the categories of features, local/cultural
events, and celebrities/entertainment photographs.
  To select the specific photographs for these studies, slides from the 1996
National Press Photographers Association Pictures of the Year competition were
selected by the author fitting categories from Pilot Test 1.  This resulted in
photographs of war (an international news story), Bill Clinton (a national news
story/politics), and accidents (a local news story).  Then, a professor of
photojournalism and the author reduced the rough edit to a smaller number of
photographs that were perceived to be extremely typical or novel on the
appropriate dimensions.  This resulted in a total of 96 different photographs.
  In the final pilot test, two coders were selected.  One was a professor of
design and visual communication who has extensive experience as a picture editor
and recently served as a National Press Photographers Association Pictures of
the Year competition judge.  The other was a doctoral student (not the author)
who has worked extensively as a newspaper designer and editor and has much
experience with newspaper photographs.  The coders were trained and practice
trials were conducted using different news topics until both coders were
comfortable with the concepts.  The two coders independently rated 96 news
photographs on the following questions:
     Regardless of composition, how often have you seen newspaper photographs of
     [news topic] having this content (i.e., subject matter, setting,
characters, props,
     action, gestures, etc.)?
 
     Regardless of content, how often have you seen newspaper photographs of
[news
     topic] having this composition (i.e., camera angle, lighting, proximity,
lens
     choice/distortion, how the content elements are combined, etc.)?
 
  The questions were phrased to focus only on photographs that have been seen in
newspapers.  Also, eight-point scales were used to force the coders to make a
decision to either the typical or novel side of the spectrum
  Reliability results show that there was exact agreement for all photographs
coded.  Participants in each of the three experiments also rated the photographs
with these same scales to confirm these condition assignments.
Scott's Pi results:
Content
Composition
War
1.00
1.00
Accidents
1.00
1.00
Clinton
1.00
1.00
 
  The correlation between the two rating scores showed that these two concepts
were judged by the coders to be independent concepts (r = .10, p = .57).  From
these rated photographs, one photograph in each condition were selected for the
actual experiments.  This resulted in a set of 12 photographs (4 conditions x 1
photograph per condition x 3 news topics).
 
Experiment Design and Procedures
  In this study, photographs were used as part of mocked-up newspaper pages.
The main purpose was to see the effect that content novelty and composition
novelty in news photographs would have on accompanying stories when surrounded
by other material.  The main dependent measures for this study were story and
photograph interest ratings and story and photograph cued-recall.
  The experiment was a 2 (typical/novel content) x 2 (typical/novel composition)
x 3 (topic: Clinton, accidents and war) between-subject design.  One photograph
from each condition for each news topic (Clinton, war, and accidents) was
selected.  All photographs were horizontal shots for consistency in design.
Also, all the war photographs were of the conflict in Chechnya.  There were 12
different newspaper versions.  For each news topic, a story was found that was
general enough to go with a variety of different photographs.  The story on
Clinton was a recent news analysis piece on how he handled the stress of
political battles.  The story on war was an overview piece on the 400-year
history of the Russian-Chechan conflict.  The piece on traffic accidents was a
story about the increase in risky driving behavior.
  Each person was given three mocked-up pages from the Columbia Missourian, the
first and third pages were actual pages from the paper and were seen by all
respondents.  The second page contained one of the twelve story/photograph
packages in addition to other stories and advertisements.  Captions were written
that described the photographs and linked them with the story.  The pages were
created by a page designer for the Missourian and were printed on 11" x 17"
paper, so that the final product looked like a photocopy of the real newspaper.
  After filling out a consent form, the student participants were told to go
through the three Missourian news pages for 20 minutes.  They were told "to read
the stories, look at the photographs, whatever interested them."  They also were
told to be sure to go through all three pages in the 20 minutes.
  After the 20 minutes were up, participants were given a questionnaire.  As in
the earlier two experiments, the first questions investigated demographics and
media use.  This was followed by a set of questions examining preferences for
visual or written material.  Next, participants were asked to list the four most
interesting stories.  This was one measure of story interest (Thorson and
Mendelson, 1996).  The next page asked them to list everything they could
remember about the target story and photograph.  Recall was cued by the headline
of the story.  Under the recall questions were a set of lines, and participants
were instructed to write one recalled idea per line.  Thus, amount of recall
equaled the number of ideas listed by the respondent.
  After the recall questions, participants were given four semantic differential
scales probing how well the target photograph went with the target story
(fitting-not fitting; bad-good; congruent-incongruent;
appropriate-inappropriate).  All photographs should fit equally well with the
stories.
  A second story interest rating was gathered by three semantic differential
scales:  interesting/boring; surprising/unsurprising; and
informative/uninformative. Finally, all 36 photographs were rated on 10 semantic
differential scales (scales from 1 to 7) and the two typicality scales (scales
from 1 to 8) used by the coders.  The 10 semantic differential scales have been
used in previous research, and were combined to form five measures:  active/calm
and emotional/unemotional formed an activity measure; simple/complex and
organized/disorganized formed a complexity measure (Berlyne 1974b);
interesting/uninteresting was used by itself (Berlyne, 1963);
pleasing/displeasing and like/dislike  formed the pleasingness scale (Berlyne,
1974b); and surprising/unsurprising, informative/uninformative and
atypical/typical formed an informativeness scale (Loftus, 1976; Thorson and
Mendelson, 1996).  After completing this questionnaire, participants were
debriefed and thanked.
Results:
  Sixty-eight students participated in the study for class extra-credit or to
fulfill requirements for a course.
Manipulation checks:
  Each participant rated the photograph from their experimental condition for
content and compositional novelty with the same scales as the coders used to
confirm the photographs' condition assignments.   Though the overall model was
only marginally significant (F (3,64) = 2.10, p = .11), the analysis of the
content ratings showed a significant main effect only for content novelty
differences (F (1,64) = 5.02, p =.03).  As assigned, the typical content
photographs (mean = 2.8) were rated as less novel than the novel content
photographs (mean = 3.2).  (see Figure 2)  There was no significant effect of
the composition manipulation nor an interaction effect for the content novelty
ratings.
  Similarly, while the overall model for the composition ratings was only
marginally significant (F (3,64) = 2.10, p = .11), the main effect of
composition manipulation was significant (F (1,64) = 4.43, p = .04), with the
photographs with novel composition (mean = 4.4) being rated as more novel than
the photographs with typical composition (mean = 3.6).  As required for the
successful manipulation check, there was no significant effect of content nor
interaction effect on the composition score.  (see Figure 3).  Still, the two
ratings scales were significantly correlated (r = .49, p < .0001), suggesting
that the untrained participants did not completely distinguish the concepts of
content and composition novelty.  Therefore, even though the manipulation checks
were successful, this experiment was analyzed using a between-subjects
hierarchical regression, with the continuous novelty ratings of each participant
as the IVs.
  One other manipulation check for this experiment dealt with how well the
photograph matched the accompanying story.  The stories were chosen so that they
could run with a variety of photographs.  There should have been no difference
in how well the photographs went with the stories, so that this would not
account for increased attention to the story.  To ascertain this, four scales
were rated by each participant in terms of how well the photograph went with the
story:  good/bad, fitting/not fitting, congruent/incongruent, and
appropriate/inappropriate (Cronbach's  = .86).  There were no differences in how
well the photographic conditions fit with the stories (F (3, 50) = .18, p =
.91).
 
Hypotheses results: (see Table 1 for complete regression results)
  Interest ratings:  The first dependent measures examined were the
participants' ratings of interest in the photograph and story seen.  As in the
earlier experiments, interest in the photographs (H1 and H2) was measured by an
interesting/uninteresting semantic differential scale.  It was predicted that
participants would find the more novel photographs more interesting.  The
results of the regression analysis showed that there were no significant effects
for any of the independent variables.
  It has been shown that pleasingness is not associated with visual novelty.
The pleasingness scale had an acceptable reliability level (Cronbach's  = .76).
There was only a significant main effect of composition novelty, with
photographs with more composition novelty (mean = 4.72) being better liked than
photographs with less composition novelty (mean = 2.98). (see Figure 4)
  Story interest was measured in two ways.  First, participants were asked to
list the four most interesting stories.  The rank of the target story was
recorded, and if it was not listed, the rank was assigned a five.  Second,
participants rated the story on three semantic differentials:
interesting/boring, surprising/ unsurprising, and informative/uninformative
(Cronbach's  =.64).  It was predicted that stories accompanying more novel
photographs would be ranked as more interesting and would be rated as more
interesting on the interest index.
  How the participants ranked a story in terms of interest was not significantly
affected by the photograph's novelty level.  Likewise, story interest, as
measured by the three-variable index, was not affected by a photograph's novelty
level.  However, story interest was predicted by the topic of the story.  Using
Bonferroni's paired comparison tests (critical t' for three comparisons (df =
62) = 2.46).  The results show that the Clinton story (mean = 3.43) was rated as
significantly less interesting than either the war (mean = 4.64) or accident
story (mean = 4.85) (Clinton vs. accident t' = 6.11; Clinton vs. war t' = 5.21;
accident vs. war t' = .90).  (see Figure 5)
  Cued recall of photographs and stories:  It was predicted that participants
would recall more about more novel photographs and the stories accompanied by
them.  Recall for the photograph and the story was cued by the headline of the
package, and was measured by the number of ideas listed by the participant.
Looking at the regression results for photo recall, the only significant effect
was an interaction between the content and the composition novelty ratings.
Using Bonferroni's paired comparisons test (critical t' for six comparisons (df
= 56) = 2.75), we find that the photographs with matching novelty levels (high
content/high composition or low content/low composition) had more information
recalled about them than the photographs without matching novelty levels
(low/low vs. low/high  t' = 7.01; high/high vs. high/low t' = 10.04; low/low vs.
high/low t' = 9.45; high/high vs. low/high t' = 7.60).  There was no significant
difference between the amount recalled about the low content/low composition
photographs and the high content/high composition photographs (t' = .6), nor was
there a significant difference between the amount recalled about the low
content/high composition photographs and the high content/low composition
photographs (t' = 2.44). (see Figure 6)
  For story recall, neither type of novelty had a significant main effect on
recall.  However, the topic of the story did have a significant impact on the
amount recalled about the story.  The results of the Bonferroni test (critical
t' for three comparisons (df = 56) = 2.48) showed that more was recalled about
the accident story (mean = 4.70) than either the Clinton (mean = 3.30; t' =
2.85) or the war story (mean = 2.46; t' = 4.57).  The war and Clinton story were
not significantly different (t' = 1.71).  (see Figure 7)
  While there seems to be a significant three-way interaction effect, this
result is not interpretable due to missing cells.
  Semantic differentials ratings:   As in the other experiments, participants
rated the photographs on a number of semantic differentials that were then
combined into scales.  Three indices (complexity/disorganization,
activity/emotionality, and informativeness) were used to further examine the
effects of novelty in photographs.  The participants rated the photographs on
these indices while looking at the photographs unlike the recall and ranking
measures, which were made from memory.
  The complexity scale only had a Cronbach's  = .26, and the two component
scales (simple/complex and organized/unorganized) were not significantly
correlated (r = .16, p = .21).  The analyses showed that there was a significant
interaction between the two novelty scales (see Figure 8), as well as a
significant interaction between the topic and the composition novelty scale (see
Figure 9).
  In looking at the graph of the interaction between the two novelty scales, we
see that the participants viewed the photographs with matched novelty levels
(high/high and low/low) as less complex than the photographs with mismatched
novelty levels (low/high and high/low).  Through an analysis of paired
comparisons using Bonferroni's method (critical t' for six comparisons (df = 56)
= 2.75), we see that all levels are significantly different from each other (low
content/high composition vs. low content/low composition t' = 19.41; low/high
vs. high/high t' = 17.43; low/high vs. high/low t' = 6.14; low/high vs.
high/high t' = 11.29; and high/low vs. low/low t' = 13.29) except for low
content/low composition and high content/high composition (t' = 1.98).  The
photographs low in content and high in composition novelty were rated as the
most complex.
  Looking at the graph of the interaction between composition novelty and photo
topic, it is interesting to see that for low composition novelty photographs the
accident photographs are perceived as most complex (accident vs. war t' = 8.71;
accident vs. Clinton t' = 10.45; war vs. Clinton t' = 1.73; Bonferroni critical
t' (df = 56) = 2.75).  For high composition novelty photographs, accident
photographs are perceived as the least complex (accident vs. war t' = 6.93;
accident vs. Clinton t' = 6.48; war vs. Clinton t' = .45).
  The activity/emotionality scale had a Cronbach's  = .44, though the zero-order
correlation was significant (r = .25, p = .04).  There was a significant main
effect for content novelty, composition novelty and topic.  For content novelty
(see Figure 10), the low novelty photographs (mean = 2.49) were rated as more
active than the high novelty photographs (mean = 1.91).  For composition novelty
(see Figure 11), low novelty photographs (mean = 2.63) were rated as more active
than high novelty photographs (mean = 1.84).  Looking at the main effect for
topic (see Figure 12), we see that the war and accident photographs were rated
significantly more active than the Clinton photographs (accident  vs. Clinton t'
= 4.15; war vs. Clinton t' = 2.46; war vs. accidents t' = 1.69; Bonferroni
critical t' (df = 56) 2.46).
  Again, while there seems to be a significant three-way interaction effect,
this result is not interpretable due to missing cells.
  Last, the informative scale made from the informative/uninformative,
surprising/unsurprising, and atypical/typical scales, had a Cronbach's  = .80.
There was a significant main effect of content and composition novelty on the
informativeness of the photographs.  The more novel (high novel content mean =
4.42; low novel content mean = 3.38) the content the more informative the
photographs.  (see Figure 13)  Likewise, the more novel (high composition novel
mean = 4.73; low composition novel mean = 2.94) the composition the more
informative the photographs.  (see Figure 14)
  The informativeness scale correlated significantly with both the content
novelty scale (r = .23, p = .06) and the composition novelty scale (r = .37, p =
.002).
 
Discussion
  This experiment placed the photographs in newspaper pages surrounded by other
photographs, stories, and advertisements.  The results of revealed no
significant main effects of content or composition novelty on interest in either
the photographs or the accompanying stories.
  Further, there were no main effects of novelty on cued-recall about the
photographs or the stories.  The content of the photographs in a newspaper
context does not seem to matter.  All that seems to affect recall of and
interest in the story is the news topic.  Story interest and recall were
significantly predicted by the topic of the story.  Participants found the
Clinton story least interesting and recalled the most about the accident story.
There was, however, a significant interaction between the two types of novelty
on recall about the photograph.  Recall was better when the content and
composition novelty levels matched, either low or high for both.
  Participants also rated content novel photographs as more informative and less
active.  The same was true for photographs novel in composition.  There was also
a significant main effect of topic on activity levels, with the accident and war
photographs being rated as more active than the Clinton photographs.  The two
novelty variables interacted on the complexity scale.  The photographs with
mis-matched levels of novelty (e.g., high content and low composition novelty)
were found to be more complex than photographs with matched levels of novelty
(e.g., low content and low composition novelty).  It seems that these
participants felt that complexity was related to how well content and
composition match.  Further, pleasingness ratings were predicted by composition
novelty.
  Another interesting interaction was seen for the complexity variable between
composition novelty and topic.  For Clinton and war photographs, complexity
increased as the photographs went from low to high composition novelty.  On the
other hand, for the accident photographs, complexity decreased as composition
novelty went from low to high.  While it is hard to say exactly why this result
occurred, it does seem evident that people react differently to various topics.
  GENERAL DISCUSSION AND CONCLUSIONS
Review and Implications
  This research separated the two concepts of content and compositional novelty.
It was possible to come up with rating scales and operational definitions for
these concepts that could be reliably used by trained coders.  Past research on
visual novelty has never tried to conceptualize or operationalize content
novelty for news photographs, looking instead at what people expect in common
scenes, such as classrooms.  Likewise, although critics have discussed issues of
conventionality in news photographs, they have never stated what makes a
photograph new.  This study appears to be the first research to examine what is
typical and what is new in news photographs.
  Moreover,  no visual novelty research on any sort of stimuli has tried to
conceptualize or operationalize compositional novelty.  Many photojournalism
textbooks discuss composition, and photojournalism critics discuss the
conventional style of photojournalism, but, like content, nothing is ever said
about what is not conventional composition.  Also, until now the effects of
typical and novel compositional devices had never been studied.
  Even though the expert coders were able to differentiate between the concepts
of content and composition novelty, participants, who were not formally trained
in photography (specifically, photojournalism), were not able to cleanly rate
photographs on the concepts of content and compositional novelty.  Also,
although the coders' ratings cleanly placed the photographs in four different
groups, the experiment participants' ratings did not.  These results were
consistent across multiple examples of photographs for three very different, yet
common, news topics:  war, an international news topic, President Clinton, a
national news topic, and traffic accidents, a local news topic.  The concept of
composition was especially difficult for the participants.  The composition
manipulation was confounded with the content manipulation, as evidenced by the
significant main effect for content when the composition novelty scores were the
dependent variable.  The lack of sensitivity to composition novelty is also
evident in the lack of effects due to this independent variable.  More about the
differences between the trained coders and the untrained participants will be
discussed in the future research section.
  By using news photographs as stimuli, this research extends visual novelty
research beyond meaningless geometric shapes or illustrations of objects in
familiar scenes.  As suggested in the literature review, one of the main
problems with past visual novelty research was the use of illustrations of
creatures made from parts of many animals or common scenes with a bizarre object
included to provide novelty.  The novel content in the news photographs used in
the current studies was completely plausible and realistic, showing that novelty
does not have to be defined as bizarre to find effects.
  This research also shows that people have strong expectations for events
portrayed in news photographs, events not often directly experienced by people,
as opposed to scenes and events that people have direct experience with, such as
stores or kitchens.  This research shows that people have expectations for the
overall look of news photographs, and that experts seem to have expectations for
what type of people, settings, and objects typically appear for different news
topics and for the way photographers typically compose photographs of these
events.
  The results of this experiment also strengthens and extends schema theory,
which argues that people have expectations for commonly experienced events and
that they use these expectations to process new stimuli.  People were able to
judge news photographs on their newness and this newness affected variables like
attention and memory.
  The lack of effects found in this experiment was very consistent with past
research in this area (Mendelson, 1996; and Thorson & Mendelson, 1996).  The
different novelty conditions had no effect on story interest, story rank or
story recall, even when people judged the photographs as different from each
other in novelty.  Also, the participants did not feel any of the photographs
and stories were inappropriately paired.  Only specific news topics affected
story interest and recall.
  Garcia and Stark (1991) showed that photographs are the most effective way to
get people to read a story.  Perhaps photographs in a newspaper only serve as
entry-points or attention-getting devices and not as news content in their own
right.  Photographs may simply serve as an additional headline and
point-of-entry for a story, telling readers what the story is about so they can
decide whether to read further.  Stories with photographs have two points of
entry versus one for a story without a photograph (only the headline).  This may
be the advantage that photographs serve.
  In looking through a newspaper, we do not seem to stop to look at photographs
for very long.  This may relate to research that shows that people do not look
at photographs or illustrations as sources of information.  As Levie and Lentz
(1982) stated,  "Although most students do look at text illustrations in most
situations, they usually do not 'study' the pictures unless prompted to do so.
Much research has indicated that people normally notice only the global aspects
of a picture" (p. 223).  This fits with the idea of photographs simply being a
second headline to be briefly looked at once and then moved past.  If people are
not in the habit of learning from photographs, even novel photographs may not be
of much benefit.
  In most newspaper-reading situations, there may be no need to remember all the
details of news photographs.  Most people will not be tested to distinguish one
image from another.  If there is a photograph accompanying a political story,
the only "useful" information might be that a story is about President Clinton.
It probably does not matter what kind of suit he was wearing, or what the podium
looked like.  Processing of the image may stop after realizing Bill Clinton is
the subject, as people do not normally think they will need to make fine
distinctions on a recognition memory test.
  Past research on visual novelty has presented only one photograph at a time to
participants.  The newspapers in this experiment presented much more information
at one time, all competing for the viewer's attention.  Besides local
advertising, the newspaper pages contained many stories (some with photographs),
both hard news and features, that people found very interesting.  With so much
else to look at, effects due to the novelty of any one photograph with one story
may be hard to isolate.  Perhaps novelty's effects get washed out.
  Research on visual novelty needs to move beyond single stimuli at a time
studies.  Although important at a basic level, this type of research is
unrealistic with how people experience the world, with multiple sources of
stimulation at a time.  At this point, there has been enough basic level
research done.  The results of the newspaper experiment show that further
research must be done to understand how novelty operates in competition with
other sources of information.  This requires research designs of greater
complexity.  The present study was only a beginning, and apparently although
novelty works alone, it does not do so with other stimuli.  With so much
information presented, novelty may not be the only variable affecting readers.
  Cowan's (1995) attention model showed that long term goals or interests also
can drive attention -- the notion of pertinence.  When a person comes across
something in a newspaper that is novel at the same time encountering something
they have an interest in, does novelty or long-term interest win out in
attracting and maintaining attention?  What if something that is only novel is
competing with something novel and of long-term interest to the reader?  This
will be discussed further in the future research section.
Limitations
  In hindsight, there are a number of design factors from the experiment that
might have limited the effects of the novel photographs.  Putting the target
story on the second page may have been a problem.  Some of the stories, mainly
the ones people found most interesting, jumped from page one to page three, so
people spent less time on page two.  Page placement is a cue used by readers to
signal the importance of a story, and it may be that this subtle cue told the
readers that the target story was not that important.
  The target stories were placed on page two so that everyone would see the same
front page.  This prevented people who were sitting next to each other during
the experiment from noticing differences in the pages.  It might have been
better to have people read three front pages from different issues, rather than
three pages from one issue.  This would prevent any one page from having more
apparent importance;  all stories would have a "page one" importance cue.  Past
research has used both presentation methods.  Mendelson (1996) used three
different front pages from the Wisconsin State Journal, while Thorson and
Mendelson (1996) used pages from the same issue of the Columbia Missourian.
  Another design issue that might have limited novelty effects is that there
were other photographs, specifically feature photographs, that were run bigger
than the target photographs.  Photograph size is another important cue to
importance.  Huh (1994) showed that larger photographs lead people to rate the
accompanying stories as more important.  While the target photographs were not
run unusually small, nor were they run on a page with any other photographs,
other photographs did run larger.  If other photographs are larger, the reader
may be cued that the story with the relatively small photograph is not as
important.  Perhaps there is a minimum absolute size necessary for novel
photographs to have an effect, or perhaps the target photograph needs be
dominant in size.
  One other design limitation, though unlikely to have had much of an impact,
was that reproduction quality of the photographs in the newspapers could have
limited the photograph's effectiveness.  As all the photographs were printed
black-and-white and the entire newspaper page was photocopied, the poor quality
could have caused everyone to pay less attention to all photographs (not just
the target ones).  The photographs were all "readable," although they were
reproduced at a lesser quality than occurs in actual newspaper pages.
  The last possible limitation was that using news topics that people felt were
typical might have reduced differences between the groups, especially if there
were other stories that were less typical on the news pages.  It may be the case
that novel stories were working against a typical story (the target story) with
a novel photograph.  It would have been helpful to have readers rate other
photographs and stories that were in the pages for interest to compare to the
target ones.  Again, where does novelty's role in attracting attention end and a
person's long-term interest begin?  If people especially like human interest
stories, they may simply skip over war, Clinton or accident stories regardless
of the photographs paired with them.  Also, how would two novel photographs
compete with each other?  Does the size or placement on the page of the
photograph have an effect then?  These issues will be discussed in the future
research section.
 
Future Research
  There are many areas of photographic novelty that still need further research.
The main areas are:  research on novelty and design issues, research on the
make-up of photographic schemata, and research on the involuntary nature of
attention to novel photographs.
Novelty and Design Variables
  All of the newspaper design issues above could be examined in more detail.
Newspaper pages could be designed that compare target stories on page one to
inside pages.  The effect of this cue on the perceived importance of the news
story should not be discounted.  Simply putting a photograph and story on page
one increases the likelihood of a reader processing that story.  Another cue
would be whether the target story is placed above or below the fold on a page.
  The effect of the size of photographs should be examined.  As stated above,
target photographs may need to be equal or larger in size relative to other
photographs in the newspaper to signal that the story is worthy of the reader's
time and attention.  Garcia and Stark (1991) showed that increased size of a
photograph increased the likelihood of a reader "processing" it.
  Beyond size of photographs it may be that number of photographs are a factor.
It is hard to tell a complete story in one photograph, so portraying novel
situations (or even typical ones) through multiple photographs may increase the
effects on attention and memory.  This would not only increase the information
available to readers, but would increase the perceived importance of the story.
A story with many photographs is probably viewed as more important than a story
with just one.  Garcia and Stark (1991) showed that multiple photograph packages
led to more processing of the text, and this was especially true when the
photographs were in color rather than black-and-white.  However, color versus
black-and-white made no difference for single photographs.  All of these design
issues might interact with novelty by increasing the perceived importance of the
stories and photographs.
  It also needs to be examined what specific purpose a photograph serves on a
newspaper page.  Are photographs merely a "second headline?"  What source of
information do they serve for readers?  As a starting point, people could be
questioning during and after reading a newspaper as to how they perceive
themselves using the photographs.  Such think-aloud procedures have been used
successfully in many studies (Shapiro, 1994).  From there, other studies could
be designed to time how long people look at photographs.
 
Schema Research
  The second area of future research involves study about people's expectations
or schema for news photographs.  As this was a first attempt to examine people's
expectations for compositional devices, future research must isolate which
devices and what levels of these devices are most expected by people.  For
example, what camera angles do people most expect for different news photograph
types?  It is necessary to conduct a vast content analysis of a variety of
newspapers to determine the most typical types of photographs, in general, what
objects are most typical in them, and how photographs are most typically
composed.  It would be possible to devise scales to rate the levels for many
compositional devices in each photograph.
  Another important area to research is the differences in schema or
expectations between expert photographers and novices or non-photographers.
Since untrained viewers were not able to differentiate content and composition
novelty, future research needs to examine what these viewers are perceiving in
news photographs.  They do seem to be affected by variations in composition
novelty even if they can't specify these changes.  Training in visual expression
and its processing then should make a difference in the understanding of visual
stimuli.  Similar to the general differences between experts' and novices'
schemata, visual arts experts and novices exhibit differences in the processing
of visual stimuli.  Cupchik and Heinrichs (1981) argued that for those who are
more trained in art, there should be differences in processing of visual stimuli
that go beyond just faster processing.  They suggested the existence of
particular "aesthetic sets" that experts have constructed.  An aesthetic set is
tuned to the sensory and physical qualities of certain styles, for example,
impressionism.  Novices, on the other hand, are constrained by their lack of
knowledge in aesthetics to focus only on the surface content and the real
life-resemblance qualities of a painting.  Thus, there are differences in the
processing of pictures by experts and novices, at least, the "perceptual and
experiential levels" (Winston & Cupchik, 1992, p.  3).  Goude (1972) showed that
experienced art viewers were more focused on style and the quality of the color,
compared to naive viewers who concentrated on the paintings" content.
  Nodine, Locher and Krupinski (1993) found that art students had different
viewing styles (based on their eye movements) when compared to non-art students.
Art students preferred more ambiguous art work.  The untrained viewers were less
efficient in analyzing the relationships among pictorial elements.  Art-trained
viewers spent more time examining background features, while untrained viewers
focused more on individual objects rather than on relationships among the
elements.  This shows the untrained viewers' greater emphasis on
representational accuracy rather than on the aesthetics of the art piece.
Further, Nodine et al. (p. 224) demonstrated that "untrained viewers were less
prepared to separate issues of form and balance from issues of content in
judging aesthetic value than were art-trained viewers."
 
Physiological reactions to novelty
  The effects of novelty should be examined on two types of physiological
measures:  eye-movements and orienting responses (OR).  Using equipment that
tracks eye-movements will allow us to see what the eyes are actually doing when
people view a novel photograph in a newspaper compared to a typical one.  Do
people spend more time with the novel one?  And if so, what about the time spent
with the story?  This would help to illuminate what is producing the lack of
effects in newspapers that has been found.  Garcia and Stark (1991) only
reported the results for first looks at photographs, not successive looks.
Additional eye-tracking research also would reveal how photography experts
process the news pages differently than non-experts.
  Eye-track research would offer insight into what a viewer does with multiple
sources of novel information or how sources of novelty and long-term interest
interact.  Perhaps novelty cannot compete with long-term interest.  Given a
choice between something novel and something we are interested in, interest
probably wins out.  A study could have people rate what types of photos and
stories they prefer to read: for example, international news, politics, human
interest, sports, etc.  These data could be used to predict attention along with
novelty.
 
Conclusion
  This line of research has several implications for the practice of newspaper
journalism.  Eventually, it should be helpful in systematically isolating the
effects that photographs have in newspapers and isolating to which photographs
people respond better.  Any research that improves editorial decision-making is
important.  However, as was shown in the literature review, most of the previous
research on the effects of photographs in newspapers has been severely
confounded due to methodological issues.  The authors of these studies
attributed numerous effects to photographs, most of which were not tested in any
systematic way.  Visual stimuli, such as news photographs, are highly complex,
and this complexity is compounded when entire newspaper pages serve as stimuli.
For this reason, rigorous methodology is essential.
  The results of the experiment  do not give much hope to the newspaper
photographer doing daily assignments.  This research showed that what photograph
accompanies a story does not seem to matter.  Still, past research (Thorson &
Mendelson, 1996) has shown that having a photograph accompany a story was better
than not having one.   How photographers shoot a single photography assignment
does not seem to matter as long as they get something.  As stated above, future
research is needed to understand why this preference does not matter when a
photograph is part of a newspaper page.  Certainly, photographers should be
pushing themselves creatively, not just for the readers, but to keep themselves
inspired.
  Although it may be too soon to establish any clear guidelines for the
newspaper industry, it seems certain that what single photograph accompanies a
news story does not matter much, except, perhaps, as a second headline.
Newspaper editors may be better served having photographers spend more of their
time on long-term, in-depth photography projects than on daily assignments.
This will require that editors give more space for multiple-photograph projects.
With multiple photographs, a story can be more fully told.  Likewise, these
results should lead photographers to demand more time and space for such work.
But, such a change in newspaper photography usage would require a move away from
a reliance on the coverage of events like press conferences, meetings and
speeches, to more coverage of the issues behind the events.  In fact, Lewis
(1995), in his photojournalism textbook, argues that this is a better way to
approach photography assignments, as it produces photographs that tell a more
complete story.  As he states: "News conferences, meetings, awards ceremonies,
and the like are visual clich s that result in the same photos time after time.
The best approach is to show the cause or the consequence of the story" (p.
145).
  The present examination of novelty in news photographs is just the beginning
of a long line of research I have planned.  As is evident from the future
research section, there is much we do not know about what makes a photograph
novel and what effects this novelty has on people.
 REFERENCE LIST
 
Abbott, V., Black, J.B., & Smith, E.E. (1985).  The representation of scripts in
memory.
     Journal of Memory and Language, 24, 179-199.
 
Abelson, R.P. (1981).  Psychological status of the script concept.  American
Psychologist
     , 36, 715-729.
 
Bargh, J.A. (1984).  Automatic and conscious processing of social information.
in
     Handbook of Social Cognition, Vol. 3.  eds. R.S. Wyer, Jr. & T.K. Srull.
Hillsdale, NJ:
     Lawrence Earlbaum Associates.
 
Barnard, W.A., Breeding, M., & Cross, H.A. (1984)  Object recognition as a
function of
     stimulus characterisitcs.  Bulletin of the Psychonomic Society, 22, 15-18.
 
Barsalou, L.W. (1985).  Ideals, central tendency, and frequency of instantiation
as
     determinants of graded structure in category.  Journal of Experimental
Psychology:
     Learning, Memory, and Cognition, 11, 629-649.
 
Beltran, F.S., & Duque, Y. (1993).  Processing typical objects in scenes:
Effects of
     photographs versus line drawings.  Perceptual and Motor Skills, 76,
307-312.
 
Beltran, F.S., & Herrando, S. (1995).  Measuring the typicality of objects
included in
     environmental scenes:  A logisitic model for atypicality.  Perceptual and
Motor Skills,
     80, 1343-1352.
 
Beltran, F.S., Herrando, S., & Pelegrina, M. (1992).  Measuring the typicality
of objects
     included in environmental scenes:  A first scale.  Perceptual and Motor
Skills, 74,
     827-831.
 
Bentley, P.F. (1993).  Clinton: Portrait of Victory.  New York:  Warner Books.
 
Berlyne, D.E. (1951).  Attention to change.  British Journal of Psychology, 42,
269-278.
 
Berlyne, D.E. (1957).  Conflict and information-theory variables as determinants
of human
     perceptual curiosity.  Journal of Experimental Psychology, 53, 399-404.
 
Berlyne, D.E. (1963).  Complexity and incongruity variables as determinants of
     exploratory choice and evaluative ratings.  Canadian Journal of Psychology,
17, 274-290.
 
Berlyne, D.E. (1971).  Aesthetics and Psychobiology.  New York:
Appleton-Century-Crofts.
 
Berlyne, D.E. (1974a).  Novelty, complexity, and interestingness.  in Studies in
the New
     Experimental Aesthetics: Steps toward an Objective Psychology of Aesthetic
Appreciation.
     ed. D.E. Berlyne.  New York: John Wiley and Sons.
 
Berlyne, D.E. (1974b).  Verbal and exploratory responses to visual patterns
varying in
     uncertainty and in redundancy.  in Studies in the New Experimental
Aesthetics: Steps
     toward an Objective Psychology of Aesthetic Appreciation. ed. D.E. Berlyne.
New York:
     John Wiley and Sons.
 
Biederman, I. (1981).  On the semantics of a glance at a scene.  in Perceptual
     Organization. eds. M. Kubovy and J.R. Pomerantz.  Hillsdale, NJ: Lawrence
Erlbaum
     Associates.
 
Biederman, I., Rabinowitz, J.C., Glass, A.L., & Stacey, E.W., Jr. (1974).  On
the
     information extracted from a glance at a scene.  Journal of Experimental
Psychology, 103,
     597-600.
 
Brewer, W.F., & Treyens, J.C. (1981).  Role of schemata in memory for places.
Cognitive
     Psychology, 13, 207-230.
 
Cowan, N. (1988).  Evolving conceptions of memory storage, selective attention,
and their
     mutual constraints within the human information-processing system.
Psychological Bulletin
     , 104, 163-191.
 
Cowan, N. (1995).  Attention and Memory:  An Integrated Framework.  New York:
Oxford
     University Press.
 
Cupchik, G.C., & Berlyne, D.E. (1979).  The perception of collative properties
in visual
     stimuli.  Scandinavian Journal of Psychology, 20, 93-104.
 
Cupchik, G.C., & Heinrichs, R.W.  (1979).  The perception of collative
properties in
     visual stimuli.  Scandinavian Journal of Psychology, 20, 93-104.
 
Duncan, J. (1984).  Selective attention and the organization of visual
information.
     Journal of Experimental Psychology: General, 113, 501-517.
 
Durham, R.L., Nunnally, J.C., & Lemond, L.C. (1971).  The effects of levels of
     information conflict on visual selection.  Perception and Psychophysics,
10, 93-96.
 
Feinberg, M. (1970).  Techniques of Photojournalism.  New York:
Wiley-Interscience.
 
Fisher, D.F., Karsh, R., Breitenbach, F., & Barnette, B.D. (1983).  Eye
Movements and
     Picture Recognition: Contribution or Embellishment.  in Eye Movements and
Psychological
     Functions: International Views. eds. R. Groner, C. Menz, D.F. Fisher, and
R.A. Monty.
     Hillsdale, NJ: LEA.
 
Fisk, A.D., & Schneider, W. (1984).  Memory as a function of attention, level of
     processing, and automatization.  Journal of Experimental Psychology:
Learning, Memory,
     and Cognition, 10, 181-197.
 
Fiske, S.T., & Taylor, S.E. (1991).  Social Cognition. New York: McGraw-Hill,
Inc.
 
Fox, R., & Kerns, R. (1961).  Creative News Photography.  Ames, Iowa:  Iowa
State
     University Press.
 
Friedman, A. (1979).  Framing pictures:  The role of knowledge in automatized
encoding
     and memory for gist.  Journal of Experimental Psychology: General, 108,
316-355.
 
Garcia, M.R. & Stark, P. (1991).  Eyes on the News.  St. Petersburg, Florida:
The Poynter
     Institute for Media Studies.
 
Geraci, P.C. (1984).  Photojournalism:  New Images in Visual Communication.
Dubuque,
     Iowa:  Kendall/Hunt Publishing Company.
 
Glenberg, A.M., & Langston, W.E. (1992).  Comprehension of illustrated text:
Pictures
     help to build mental models.  Journal of Memory and Language, 31, 129-151.
 
Goldberg, V. (1991).  The Power of Photography:  How Photographs Changed Our
Lives.  New
     York: Abbeville Press.
 
Goodman, G.S. (1980).  Picture memory: How the action schema affects retention.
     Cognitive Psychology, 12, 473-495.
 
Goude, G. (1972).  A multidimensional scaling approach to the perception of art.
I.
     Scandinavian Journal of Psychology, 13, 258-271.
 
Greenberger, E., Woldman, J., & Yourshaw, S.W. (1967).  Components of curiosity:
Berlyne
     reconsidered.  British Journal of Psychology, 58, 375-386.
 
Hagaman, D. (1996).  How I Learned Not To Be A Photojournalist.  Lexington:
University
     Press of Kentucky.
 
Hall, S. (1973).  The determinations of news photographs.  in The Manufacture of
News:
     Social Problems, Deviance, and the Mass Media.  eds.  S. Cohen, and J.
Young.  London:
     Constable.
 
Henry, W.P., & Solano, C.H. (1983).  Photographic style and personality:
Developing a
     coding system for photographs.  The Journal of Psychology, 115, 79-87.
 
Hoffman, J. (1987).  Semantic control of selective attention.  Psychological
Research,
     49, 123-129.
 
Hoy, F.P. (1986).  Photojournalism: The Visual Approach.  Englewood Cliffs, New
Jersey:
     Prentice-Hall.
 
Huh, H.L. (1994).  The effect of newspaper picture size.  Visual Communication
Quarterly,
     1 (Spring), 14-15.
 
Hutt, C. (1970).  Specific and diversive exploration.  in Advances in Child
Development
     and Behavior, Volume 5.  eds. H.W. Reese and L.P. Lipsitt.  New York:
Academic Press.
 
Intraub, H., Bender, R.S., & Mangels, J.A. (1992).  Looking at pictures but
remembering
     scenes.  Journal of Experimental Psychology:  Learning, Memory, and
Cognition, 18,
     180-191.
 
Intraub, H., & Richardson, M. (1989).  Wide-angle memories of close-up scenes.
Journal
     of Experimental Psychology:  Learning, Memory, and Cognition, 15, 179-187.
 
Kahneman, D. (1973).  Attention and Effort.  Englewood Cliffs, NJ:
Prentice-Hall, Inc.
 
Keene, M. (1995).  Practical Photojournalism:  A Professional Guide.  Boston:
Focal Press.
 
Kepplinger, H.M., & Donsbach, W. (1987).  The influence of camera perspectives
on the
     perception of a politician by supporters, opponents, and neutral viewers.
in Political
     Communication Research.  ed. D.L. Paletz.  Norwood, NJ: Ablex.
 
Kerns, R.L. (1980).  Photojournalism:  Photography with a Purpose.  Englewood
Cliffs, NJ:
     Prentice-Hall, Inc.
 
Kihlstrom, J.F. (1984).  Conscious, subconscious, unconscious: A cognitive
perspective.
     in The Unconscious Reconsidered.  eds. K.S. Bowers & D. Meichenbaum.  New
York:  John
     Wiley & Sons.
 
Kobre, K. (1980).  Photojournalism: the Professional's Approach.  Sommerville,
MA: Curtin
     and London.
 
Kolers, P.A. (1973).  Some Modes of Representation.  in Communication and
Affect:
     Language and Thought.  eds. P. Pliner, L. Krames, and T. Alloway.  New
York:  Academic
     Press.
 
Komenich, K.  Telephone interview with the author.  7 May 1994.
 
Kraft, R.N. (1987).  The influence of camera angle on comprehension and
retention of
     pictorial events.  Memory & Cognition, 15, 291-307.
 
Lachman, R., Lachman, J.L., & Butterfield, E.C. (1979).  Cognitive Psychology
and
     Information Processes: An Introduction.  Hillsdale, NJ: Lawrence Erlbaum
Associates.
 
Lemond, L.C. (1978).  Complexity, incongruity, pre-exposure and the familiarity
effect in
     visual selection.  Perceptual and Motor Skills, 46, 99-106.
 
Levie, W.H., & Lentz, R. (1982).  Effects of illustrations:  A review of
research.
     Education and Communication Technology Journal, 30, 195-232.
 
Lewis, G. (1995).  Photojournalism: Content & Technique.  Madison, WI: Brown &
Benchmark.
 
Lewis, M. (1978).  Attention and verbal labeling behavior in preschool children:
A study
     in the measurement of internal representations.  The Journal of Genetic
Psychology, 133,
     191-202.
 
Loftus, G.R. (1976).  A framework for a theory of picture recognition.  in Eye
Movements
     and Psychological Processes.  eds. R.A. Monty and J.W. Senders.  Hillsdale,
NJ: Lawrence
     Erlbaum Associates.
 
Mandler, J.M. (1983).  Stories, Scripts, Schemas: Aspects of Schema Theory.  New
York:
     LEA.
 
Mandler, J.M., & Johnson, N.S. (1976).  Some of the thousand words a picture is
worth.
     Journal of Experimental Psychology: Human Learning and Memory, 2, 529-540.
 
Mandler, J.M., & Parker, R.E. (1976).  Memory for descriptive and spatial
information in
     complex pictures.  Journal of Experimental Psychology: Human Memory and
Learning, 2,
     38-48.
 
Mandler, J.M., & Ritchey, G.H. (1977).  Long-term memory for pictures.  Journal
of
     Experimental Psychology: Human Learning and Memory, 3, 386-396.
 
Martindale, C., Moore, K., & West, A. (1988).  Relationship of preference
judgments to
     typicality, novelty, and mere exposure.  Empirical Studies of the Arts, 6,
79-96.
 
Matlin, M.W. (1983).  Cognition.  New York:  Holt, Rinehart, and Winston.
 
Mendelson, A.L. (1996).  The effect of pictures on the attribution of
homelessness.
     Paper presented at the Association for Education in Journalism and Mass
Communication
     Annual Conference, August, Anaheim.
 
Messaris, P. (1992).  Visual "manipulation:" Visual means if affecting responses
to
     images.  Communication, 13, 181-195.
 
Metallinos, N. (1996).  Television Aesthetics:  Perceptual, Cognitive, and
Compositional
     Bases.  Hillsdale, NJ:  Lawrence Erlbaum Associates.
 
Meyers-Levy, J., & Peracchio, L.A. (1992).  Getting an angle in advertising:
The effect
     of camera angle on product evaluations.  Journal of Marketing Research, 29,
454-461.
 
Neisser, U. (1967).  Cognitive Psychology.  New York: Appleton-Century-Crofts.
 
Nodine, C.F., Locher, P.J., & Krupinski, E.A. (1993).  The role of formal art
training on
     perception and aesthetic judgment of art compositions.  Leonardo, 26,
219-227.
 
Norman, D.A. (1968).  Toward a theory of memory and attention.  Psychological
Review, 75,
     522-536.
 
Nunnally, J.C., & Lemond, L.C. (1973).  Exploratory behavior and human
development.  in
     Advances in Child Development and Behavior. ed. H.W. Reese.  New York:
Academic Press.
 
Nunnally, J.C., Lemond, L.C., & Wilson, W.H. (1977).  Studies of voluntary
visual
     attention -- Theory, methods, and psychometric issues.  Applied
Psychological Measurement
     , 1, 203-218.
 
Paradowski, W. (1967).  Effect of curiosity on incidental learning.  Journal of
     Educational Psychology, 58, 50-55.
 
Parker, R.E. (1978).  Picture processing during recognition.  Journal of
Experimental
     Psychology: Human Perception and Performance, 4, 284-293.
 
Potter, M.C. (1976).  Short-term conceptual memory for pictures. Journal of
Experimental
     Psychology: Human Learning and Memory, 2, 509-522.
 
Potter, M.C., & Levy, E.I. (1969).  Recognition memory for a rapid sequence of
pictures.
     Journal of Experimental Psychology, 81, 10-15.
 
Rabbitt, P. (1984)  The control of attention in visual search.  in Varieties of
Attention
     . eds. R. Parasuraman and D.R. Davies.  New York: Academic Press.
 
Rayner, K., & Pollatsek, A. (1992).  Eye Movements and Scene Perception."
Canadian
     Journal of Psychology, 46, 342-376.
 
Reeves, B., & Geiger, S. (1994).  Designing Experiments that Assess
Psychological
     Responses to Media Messages.  in Measuring Psychological Responses to
Media.  ed. Annie
     Lang.  Hillsdale, New Jersey: Lawrence Erlbaum Associates.
 
Rosenblum, B. (1978).  Style as social process.  American Sociological Review,
43,
     422-438.
 
Schneider, W., Dumais, S.T., & Shiffrin, R.M. (1994).  Automatic and control
processing
     and attention.  in Varieties of Attention. eds. R. Parasuraman and D.R.
Davies.  New
     York: Academic Press.
 
Somalia: We saw the horror. (1995).  News Photographer, 49 (1), 26-31.
 
Szarkowski, J. (1966).  The Photographer's Eye.  New York: The Museum of Modern
Art.
 
Thorson, E.L., & Mendelson, A.L. (1996).  The impact of news stories and photos
on the
     processing of information about Hillary Rodham Clinton.  Paper presented at
the
     International Communication Association Annual Conference, May, Chicago,
IL.
 
Wanta, W. (1988).  The effect of dominant photographs:  An agenda-setting
experiment.
     Journalism Quarterly, Spring, 107-111.
 
Wanta, W., & Roark, V. (1993).  Cognitive and affective responses to newspaper
     photographs.  Paper presented at the Association for Education in
Journalism and Mass
     Communication Annual Conference, August, Kansas City, MO.
 
Wentworth, N., & Witryol, S.L. (1986).  What's new? Three dimensions for
defining
     novelty.  The Journal of Genetic Psychology, 147, 209-218.
 
Winston, A.S., & Cupchik, G.C. (1992).  The evaluation of high art and popular
art by
     naive and experienced viewers.  Visual Arts Research, 18, 1-14.