Submitted for Publication in Reading Psychology. Prepared with the T&F Journal Template.
The Influence of Illustrations on Children’s Book Preferences and Comprehension
Suggested short title: The Influence of Illustrations
Jamye B. Brookshire
Stephen F. Austin State University
Lauren F. V. Scharff, Ph.D
Stephen F. Austin State University
Laurie E. Moses
Stephen F. Austin State University
Address Correspondence to: Lauren Scharff, Stephen F. Austin State University, Box 13046 SFA Station,
Nacogdoches, TX 75962; email@example.com
Brookshire, J., Scharff, L. & Moses, L. (2002). The influence of illustrations on
children's book preferences and comprehension Reading Psychology, Vol 23(4), 323 -
The influence of illustrations on children’s book preferences and comprehension were studied. Seventy-one first
and third graders were shown one of nine books varying in illustration style (realistic or abstract), illustration
brightness (bright or somber), and content (illustrations with no text, illustrations with text, or text with no
illustrations). Participants were asked 15 comprehension questions and several illustration preference questions.
While third graders scored higher than first graders, both groups showed highest comprehension for the text-plus-
illustrations book content and the lowest for the illustrations-only book content. Participants preferred the bright-
realistic book stimuli significantly more than any other book stimuli.
“Vividness, compactness, and inexpressibility” are three ways by which pictures can aid the reader in
attaining the message the author is portraying (Schallert, 1980). A picture can be used in a book to literally replace a
large number of words. For example, consider the amount of space it takes to write directions to get to someone’s
house versus drawing a map. Illustrations can also draw attention to a book by making it colorful and interesting.
Fang (1996) points out several additional functions of illustrations in picture books for children. Pictures can
establish the setting and mood of a story. They can also define and develop characters. This is especially effective
in short stories that do not usually allow for more fully-developed characters. Finally, in many picture storybooks,
illustrations extend or develop the plot.
Several authors and artists have written books about general guidelines for illustrating children’s books. One
such author, Cianciolo (1970), believed that the purpose of children’s books is to be read, and that illustrations
should not defeat this purpose by being too novel or eccentric. Rather, the illustrations must legitimately help deliver
the writer’s message. She subscribed to the notion that illustrations do facilitate reading, and that surrounding
children with attractive books about things of interest is one of the most effective ways to stimulate learning.
Further, she maintained that illustrations in books could facilitate the reader’s comprehension by building children’s
vocabulary and allowing them to form various concepts about the world around them. Finally, according to
Cianciolo, illustrated books provide a literary and cultural heritage and help in the development of appreciation and
understanding of the graphic arts.
Despite such favorable claims by artists, many researchers have argued that illustrations in children’s literature
can impair the reader’s comprehension and recall of the information presented in the text. Elster and Simons (1985)
believe that the way in which the picture and the text interact affects the reader’s ability to understand the story.
Sometimes a text uses words that refer to information that must be retrieved from the pictures (e.g., Look at him,
Look at that, I want that one over there). Such a text is known as a “picture-dependent” text (Simons, 1990). In
contrast, picture-independent text (Elster & Simons, 1985) can be understood by the reader without having to refer
to the picture (e.g., I see a boy. That boy is Nick.). In this case, the picture could be used as an additional context
cue. Elster and Simons (1985) believe picture-dependent stories distract both low ability readers and those readers
just beginning to read. They support the idea that in a picture-dependent story, the reader must divide his/her
attention between the words and the pictures because they have to process both visually. This slows down
processing and makes learning to read more difficult. Elster and Simons (1985) also mention that children learning
to read need to learn how to decode and look for meaning in the words, not in the pictures. They concluded that
pictures should only be used as a supplement to reading (picture-independent books), allowing children to detect
meaning by focusing on the words, not the pictures.
Beck (1984) cited evidence (Peeck, 1974) that pictures can sometimes hinder comprehension in first and second
grade readers. She concluded that story comprehension was only aided by illustrations if the pictures agree with the
text. If the picture carries concepts that are contrary to the text then children’s performance will be hindered. Beck
also determined that two styles of pictures (abstract and elaborate) could interfere with comprehension. She
proposed that elaborate and abstract illustrations both increase the demands of the reading task and slow the process
of comprehension. Further, she cited evidence (Poulsen, Kintsch, Kintsch & Premack, 1979) that comprehension
could be delayed by large shifts in perspective (for example, illustrating an animal as larger than human characters
on one page and on the next page picturing the same animal much smaller).
While some believe that pictures can hinder children learning to read, several researchers have found evidence
supporting their use. A study by Read and Barnsley (1977) found that pictures in basal reading materials can have
significant positive effects when long-term memory is tested. Fang (1996) holds that illustrations are important to a
child learning to read. He believes that the usefulness of illustrations in the development of children’s reading
behaviors far outweighs its potential dangers. He supports this by mentioning that pictures lure children to read and
interact with the text. Therefore, pictures stimulate and promote children’s interest in books. Consequently, this
enhanced involvement with the book motivates them to predict what is going to happen next or find objects in the
pictures. Fang notes that pictures, as first-order symbols, exhibit familiar experiences with which children are likely
to identify more easily. As second-order symbols, words are abstract and do not show experience. This implies that
a child can use the pictures in order to figure out unknown words in the text because the scene in the picture is
familiar to the child. Perhaps the most important use of illustrations in books, according to Fang, is to provide
mental images for child readers so that they can understand the written text easily and remember it longer.
Considering how limited a child’s vocabulary, syntax, and world knowledge are, pictures lessen the complexity of
the written text by helping develop the plot and characters. Hence, with pictures, the text can use fewer words and
less difficult syntax. This is especially important when considering basal materials that have vocabulary control
constraints for beginner readers. In this case, characters, objects, and situations essential to the story can be
portrayed in the pictures and not included in the text (i.e. the result would be a form of picture dependent text).
A study by Peeck (1974) supports that pictures facilitate the learning of information in the text. Peeck showed
nine- and ten-year-old children cartoon strips with reading instructions. Some children received the story with
pictures and the others just received the story. Afterwards, Peeck gave the children an unannounced retention test.
Questions referred to information in the text alone, in the pictures alone, or in both the text and pictures. Also
included were questions testing content that was either congruently or incongruently represented in both the text and
pictures. Peeck found upon immediate testing that the group that received the story with illustrations remembered
more information than the no-picture group. This effect was attributed to the picture group performing significantly
better on the picture questions. When retention was measured one day or one week after reading, the picture group
again correctly answered more picture-only questions as well as more congruent picture-and-text questions than the
no-picture group. When the text and pictures were incongruent, the picture group picked the alternative congruent
with the picture but incongruent with the text more often than the no-picture group. Peeck concluded from these
results that retention measures designed to separately measure comprehension of illustrated and unillustrated
information can be valuable in determining the specific effects of illustrations.
At this point, it is still debated whether illustrations are useful additions to children’s books. However, the
theorists that believe illustrations are beneficial have proposed several hypotheses that explain how and why pictures
facilitate comprehension of text. The motivation hypothesis states that children may better enjoy stories with
pictures and, thus, they may try harder to understand the text (Glenberg & Langston, 1992). This hypothesis
predicts that pictures will improve performance on tests pertaining to the illustrations and other aspects of the text as
well. The repetition hypothesis states that the integration of both verbal information from the text and the
information from the illustrations leads to the information being presented twice, therefore enhancing
comprehension and memory (Gyselinck & Tardieu, 1999). This hypothesis predicts that pictures that explicitly
portray information in the text (basically, pictures that repeat exactly what the text says) will facilitate performance
on tests of that information (Glenberg & Langston, 1992). Another proposition is the dual-code theory. This theory
is comparable to the repetition hypothesis, and states that information repeated or double-coded will facilitate
performance on tests of the information. However, this theory also proposes that text and pictures produce two
different kinds of conceptual representations, one in a verbal form and one in a nonverbal form. By having two
representations, comprehension may be improved (Glenberg & Langston, 1992). Both the repetition hypothesis and
the dual-code theory state that a picture should have little effect on information not represented in the text.
As summarized above, numerous studies have proposed theories regarding illustrations and their effects on
reading comprehension. However, the few studies that have collected data show some inconsistent results. Further,
there are few studies (especially recent ones) regarding children’s preferences for the style of illustrations in books.
Comparing several studies researching children’s preferences for illustrations in books, King (1967) made the
following conclusions. First, children preferred colored pictures to black and white ones. She also discovered that
realism tended to be more important in illustrations than color to children. Finally, young children preferred
uncluttered pages and larger type.
For those who subscribe to the opinion that illustrations facilitate reading, it would be informative to know
which types of illustrations currently interest children and hold their attention long enough to measurably increase
their learning and understanding of the text. Book sales of children’s books could be a starting point to assess
children’s preferences, but adults often pick out and purchase the books, especially for younger readers. Therefore,
book sales will most likely reflect adults’ preferences of children’s books instead of what the children truly find
interesting and appealing.
Because of the above reasons, the present study investigated children’s preferences for different illustration
styles and measured comprehension using three combinations of text and/or illustrations. More specifically, we
created a children’s book with four versions that used bright-abstract, bright-real, somber-abstract, or somber-real
illustrations. All pictures were drawn to scale with the pictures in the real conditions looking more like photographs
and those in the abstract condition looking more cartoon-like. The pictures were uniformly colored with the
exception that the somber pictures had darker shades of the same colors used in the bright pictures. These
systematic variations in the pictures allowed us to control for differences in color and style of the picture, in order to
gather information regarding the type of illustrations children prefer.
In order to examine the influence of illustrations on comprehension, an approach similar to the one in Peeck’s
(1974) study was used. Like Peeck, some of the participants in our study received the book with just the text and no
illustrations, while others received the book with both the text and illustrations. However, for our study, we added a
third group of participants who received the book with just the illustrations and no text. All groups received
questions covering information presented only in the text, only in the pictures, and both in the text and the
illustrations. We added the illustrations-only group in order to assess how well basic story comprehension can be
obtained solely through the pictures.
All of the theoretical approaches outlined above led to the prediction that the books with both the text and
illustrations would lead to better comprehension of the text than the books with just the text or illustrations alone.
Based on a preference pilot study using popular children’s books, we also hypothesized that the bright and abstract
illustrations would be preferred, better engage the reader, and thus lead to better comprehension. Finally, we
hypothesized that third grade participants would have better comprehension than first grade participants, primarily
because they were more experienced readers.
The researchers collected their data from 71 first and third graders at two local elementary schools. One of the
schools was rural while the other one was urban, therefore allowing for a wider variety of students from different
backgrounds. These two grades were chosen because they are ones in which children are learning to master basic
reading skills, and illustrations are often an integral part of their reading material. Although rather large differences
in basic reading ability were expected, since this was an exploratory study, the two-grade difference was chosen so
that developmental changes in comprehension and preferences might be more apparent. In order to better generalize
to all first and third grade students, participants were not selected based on any reading level criteria.
The distribution of ethnicity among the students was as follows: 58 white, 7 Hispanic, 5 black, and 1 Asian
Indian. Forty-five females and 26 males participated in this study. The researchers collaborated with the schools’
principals, counselors, and teachers as to when the best times would be to come get the students out of class. All
teachers involved sent parental consent forms home with the students, and only those students who returned them
were allowed to participate.
Design and Materials
One of the researchers wrote an original, second-grade-level story to use for the experiment. Readability
grade level was determined using the Flesch-Kincaid Grade Level statistic available in Microsoft Word. Original
illustrations for the book were created by a local artist. We created a new book in order to eliminate any previous
exposure effects the participants might have had to the text and/or illustrations. Also, using original illustrations
allowed the illustration conditions to be systematically manipulated.
A 2 (grade) x 2 (illustration style) x 2 (illustration brightness) x 3 (book content) x 3 (question source) design
was conducted. All variables were between-subject variables except question source. The between variable of grade
indicates whether the student was enrolled in the first or third grade. The second between-subjects variable was
illustration style: realistic (R) or abstract (A). An artist drew two sets of pictures with the same content and color
assignments in each (e.g. the house was yellow in both sets). The realistic pictures were drawn to scale similar to
photographs, while the abstract pictures were similar to cartoons.
The third between-subjects variable was illustration brightness: bright (B) or somber (S). The original pictures
that the artist drew were painted with brightly-colored watercolors. They were then scanned into a computer and
Adobe PhotoShop was used to uniformly change the color brightness of each picture so that the color printout
matched the original watercolor pictures. These matched color printouts were used for the bright conditions. Then
the brightness of the images in the computer was turned down 35% to make the somber conditions. Examples of the
different illustration styles can be viewed on Lauren Scharff’s web page, currently at
The between-subjects variable of book content refers to whether the student received a "book" that had the
illustrations only, the text only, or the illustrations with the text. The text was presented in 16 point, Arial Black
font. In the combined condition, the text was generally placed across the bottom of the page under the picture. On
three pages, however, the text was placed on the right-hand side of the page.
The within-subjects variable of question source indicates from what source (text, illustration, or both) the
answer to the question can be found. (See Table 1 for the list of questions.) The codes following each question
indicate the source: I (illustration only), T (text only), or TI (text and illustration). In addition to these questions,
following the book content questions, all participants received the same set of illustration preference questions
regarding a set of small-scale illustration examples. The preference questions are also shown in Table 1.
<Insert Table 1 here>
All students were tested individually. Each student that had returned the signed parental consent form was
dismissed from class and taken to a quiet room by one of the researchers. The researcher began every experiment
session by saying, “This book tells a story about a boy named Tommy, his dog named Oro, and his friend named
Justine. The story will describe how different people and animals see things differently.” For the book versions
with text, participants in the third grade read the story, while participants in the first grade had the story read to
them. For the illustrations-only version, the participants were given the names of the characters and then instructed
to make up a story to go with the pictures. The instructions for these different conditions are as follows. If the child
was assigned to the text-only condition and was in the third grade, the researcher said verbatim to the student, “This
is like a big kid’s book. It has only words and no pictures. I want you to read the story out loud. If you get stuck on
a word, I will help you.” For participants in this condition that were in the first grade, the researcher said, “This is
like a big kid’s book. It has only words and no pictures. I am going to read the story out loud while you follow
along.” If the student was assigned to the illustrations-only condition, the researcher said verbatim, “This story has
pictures but no words. I want you to look at the pictures and make up a story about Tommy, his dog Oro, and his
friend Justine.” Finally, if the third-grade student was assigned to the text-plus-illustrations condition, the researcher
said verbatim, “I want you to read the story out loud. If you get stuck on a word, I will help you” or “I will read the
story out loud while you follow along” for first grade participants. After these instructions, every student was told,
“When you are done with each page, I want you to flip it over and look at the next page. After we are done with the
book, I will be asking you questions about the story.” After the student was finished with the book, the researcher
said, “Now I am going to ask you some questions about the story. If you don’t know the answer, say ‘I don’t know’
or give me your best guess.” Then the researcher proceeded to ask the 15 comprehension questions and the other
preference questions listed in Table 1. When asking the last two preference questions, a set of small-scale
illustrations showing examples of each illustration style and brightness level was shown to each student. It took
approximately 20 minutes to complete the experiment with each student. After all of the questions were answered,
the student received a university sticker and cup for their participation.
For each student, ethnicity, gender, school, grade, age, and testing condition were recorded. For the
comprehension questions, correct answers were coded as a one and incorrect answers as a zero. The questions were
grouped so that the number correct could be totaled separately for each source condition (T, I, or T+I). Because
there were five questions per source, the range for correct scores was 0 - 5. An overall total correct was also
calculated for each student. For the preference scores, a tally was made for the most- and least-preferred illustration
styles, along with a list of the children's reasons for their choices. A simple grouping and tally was performed for
the other subjective questions. An analysis of variance (ANOVA) was performed on the comprehension data, and
chi-squares were calculated for the illustration preference data.
The preference data will be discussed first because it seems that book preference may influence
comprehension. Please see Table 2 for the simple tallies of the book style preferences. Students preferred the bright-
realistic book illustrations significantly more (χ2 (3, N = 71) = 48.61, p < .01) than any other book illustration style
and preferred the somber-abstract book illustrations significantly less (χ2 (3, N=71) = 37.00, p < .01) than any other
book illustration style. Overall, students preferred bright book illustrations significantly more than somber book
illustrations (χ2 (1, N=71) = 45.76, p < .01). There was not a significant difference between the preferences for
abstract and realistic illustrations although there was a trend that students preferred the realistic illustrations better
than the abstract illustrations.
<Insert Table 2 here>
Because the design was uneven, (i.e. there were 4 brightness and illustration style combinations for the
illustrations-only and text-plus-illustrations books, but only one level of the text-only book), the variables of
illustration style, brightness and book content were all placed into one, 9-level independent variable. Thus, data
were analyzed using a 9 (book stimulus) x 3 (question source) x 2 (grade) mixed ANOVA. Given the small N and
the exploratory nature of this project, the post-hoc comparisons were performed using the LSD test with a critical
cut-off of p < 0.05 (although many were significant at the 0.01 level).
There was a main effect for book stimulus, (F (8, 53) = 3.02; p < .01). All the illustrations-only books yielded
significantly worse comprehension (mean score of 1.88 for all four style conditions), than all the text-plus-
illustrations books, except the one with the somber-abstract picture style (mean scores: TIBA 2.76, TIBR 2.88, TISA
2.46, and TISR 2.75).. The text-only book (mean score 2.42) was not significantly different than any of the other
There was also a main effect for grade (F (1, 53) = 4.08; p < .05), with third graders showing better
comprehension than first graders (mean scores of 2.48 and 2.14, respectively). Finally, question source led to
significant differences (F (2, 106) = 63.66; p < .01), with questions based on illustrations-only content (mean of
1.41) leading to significantly lower scores than those from text-only content (mean of 2.01), and the text-only
content questions leading to significantly lower scores than those based on information that was in both the text and
illustrations (mean of 3.5).
Some of the above main effects can be explained by the significant two-way interaction found between book
stimulus and question source (F (16, 106) = 5.89; p < .01). Please see Figure 1 for a graph of this interaction. Not
surprisingly, the higher scores overall for the text-plus-illustrations books and the text-plus-illustrations questions
resulted from other conditions where participants were asked about information that they had not received. For
example, post hoc analyses indicated that when given the text-only book, comprehension was significantly worse
when tested on illustrations-only questions than any questions that had text as a source. Conversely, when given the
illustrations-only book, students had lower comprehension when tested on text-only material versus any questions
that had illustrations as a source.
However, the interaction also reveals some more meaningful trends. Post-hoc analyses show that students
given the text-only book performed better on text-only questions than students given the illustrations-only book did
on illustrations-only questions. There also seem to be some additional asymmetric influences of text and
illustrations on comprehension. Text-plus-illustrations books led to lower illustrations-only question scores than
illustrations-only books. However, for the text-only questions, there was no difference between the text-only
books and the text-plus-illustrations books. Along the same line, students given the text-plus-illustrations book had
significantly better comprehension for the text-only questions than for the illustrations-only questions. Their best
performance was for the text-plus-illustrations questions, although this was only significant for the TIBA and TIBR
Although there were no overall comprehension differences between the book illustration styles (bright v.
somber and realistic v. abstract), there were some trends for the text-plus-illustrations books when participants were
asked text-plus-illustrations questions. The TIBA and TIBR books led to increases in comprehension compared to
the text-only book, while both the TISA and TISR books led to essentially equal comprehension as the text-only
book. When comparing comprehension for just the different text-plus-illustrations styles, the TIBR book led to
significantly better comprehension than the TISA book, and nearly significantly better comprehension than the TISR
<Insert Figure 1 here>
In conclusion, we found effects of illustrations on story comprehension and preferences. Illustration style
significantly affected preferences, with the bright and realistic illustrations liked best. An effect of illustrations is
important because it suggests that illustration style may affect the likelihood that children will self-select a particular
book. Further, book type (text-only, illustrations-only, or text-plus-illustrations) significantly affected
comprehension. Illustration style alone didn’t significantly affect comprehension, although there were interactive
trends with book type such that the most preferred styles led to better comprehension for the text-plus-illustrations
books when asked text-plus-illustrations questions.
The interaction between book type and question type reveals some interesting trends that should be further
explored in future research. The fact that students given the text-only book performed better on text–only questions
than students given the illustrations-only book did on the illustrations-only questions suggests that, while
illustrations alone can lead to some story comprehension (since the scores on the illustrations-only questions are
well above zero for any books containing illustrations), text information is more influential in comprehension.
Possibly this is due to the presence of a more coherent story than the one made up by the students as they viewed the
pictures. However, the potential of illustrations to aid comprehension should not be overlooked. Although the
illustrations-only group participants weren’t actually exposed to text words, the fact that they achieved measurable
story comprehension suggests that the illustrations can serve as a contextual aid to help readers determine the
meaning of ambiguous text or new words, as proposed by Fang (1996).
Other interaction trends help address the concern that illustrations may distract from text comprehension. As
pointed out above, text-plus-illustrations books led to lower illustrations-only question scores than illustrations-only
books. In contrast, there was no difference between text-only books and text-plus-illustrations books for the text-
only questions. This suggests that students (perhaps through experience) focus more on the text if it is present, to
the detriment of processing the illustration information. An additional finding extends the conclusion that
illustrations are not detrimental to comprehension. The effect of the preferred illustration styles on comprehension of
the text-plus-illustrations questions for the text-plus-illustrations readers was to enhance comprehension above the
level seen for the students receiving the text-only book. This finding further suggests that, if the pictures are
thoroughly processed (because they are liked), the redundancy may increase comprehension and long-term memory
for the items.
However, there was a potential downside to the pictures. For the illustrations that were most liked, there was a
nonsignificant trend for text-plus-illustrations groups to answer the text-only questions less accurately than the
groups that did not receive the illustrations. Because they may contain more attention-capturing information,
pictures may act as a distraction. This supports Elster and Simons (1985) theory about the negative potential of
illustrations and lends some support to their belief that books should use only picture-independent text.
However, we also have some data that suggest that even incongruent text and pictures are not always a big
distraction. The last page of the book contained a contradiction between the text and the illustration. The text
indicated that Oro, the dog, won the race back to the house, while the illustration showed Justine, the girl, winning
the race. Twenty-six of the participants who received the text-plus-illustrations version of the book correctly
answered the question of who won the race based on the text (Oro won the race) versus only five that answered
incorrectly based on the illustration. These comprehension results from the one “contrary” picture strongly suggest
that, when given both text and illustrations, most children attend to the text information more closely than the
With respect to overall story comprehension, third graders performed better than first graders, but the influence
of illustrations and the other variables was similar for both groups. It should be explicitly pointed out that
comprehension for the two groups was measured differently. The first grade students were measured on listening
comprehension, while the third grade students were measured on reading comprehension. Thus, the difference seen
between first and third grade students could be due to differences in the development of general comprehension
skills, or to the different means by which the story was presented. A larger grade difference might be expected if
both grades were tested on reading comprehension. Importantly, however, the lack of any interactions with grade
suggests that, overall, reading and listening comprehension were not differently affected by the other variables.
There are a few additional issues that could be addressed in future studies of this nature. First, there may have
been many individual differences across students that would affect comprehension and the influence of illustrations.
An obvious possible difference is in reading ability, which we did not measure. Better readers might show less
reliance on illustrations than do poor readers, and they might find different illustrations more effective or distracting.
The idea that there are individual differences in the strategies that participants might use when processing
illustrations was highlighted by Tierney and Cunningham (1984). Future studies might benefit from correlating
reading ability with comprehension for the different text and/or illustrations conditions.
Second, many of the children reported that the somber conditions were too dark in some pictures, making it
hard to see some of the detail. Therefore, when the participants were asked a question involving details of the
illustrations that were dark, they occasionally answered the question incorrectly (i.e., #7 asks where Tommy sat to
wait for Justine, but in the somber pictures it was hard for some participants to see that Tommy and Oro are waiting
on the front porch). This problem is due to the methods used to systematically make the somber condition (i.e., the
lightness of all parts of all pictures was equally adjusted).
Finally, this study’s “abstract” illustrations were not very abstract. The pictures were still drawn to scale and
colored realistically. This could explain the lack of significant difference between the preferences for realistic
versus abstract. Because this study was the first of its kind, we felt that it was more important to control picture
content and placement of the items in the pictures. We wanted the pictures to be as similar as possible while still
manipulating the style. However, a review of children’s books already in publication shows that many of them use
much more abstract images. Given that we found no difference, and King (1967) concluded realistic pictures were
more preferred, it would be of interest to see how a wider range of abstractness would influence both comprehension
Despite the limitations of the current study, it is the first to systematically test the dual influence of illustrations
on preferences and reading comprehension. Several implications result from our findings. Most importantly,
comprehension can be significantly influenced by the presence of illustrations. Further, illustration styles do
significantly influence preferences, which have not been studied recently. We suggest that adults should be
sensitive to children’s responses to illustrated books, since illustrations can influence a child’s motivation to
approach a book and, in turn, the child’s resultant comprehension of the book. Publishers may want to create focus
groups to determine children’s responses to proposed illustrations in new textbooks.
We thank Nancy Reams Coon, who created the original illustrations for the books used in this study. We also
appreciate all the coordination efforts by the teachers and administrators at the SFA Childhood Lab and Charter
School, Central Heights Elementary, Raguet Elementary, and Cushing Elementary. This work was supported by
Stephen F. Austin State University faculty research grant 1-14051.
Beck, I. (1984). Developing Comprehension: The impact of the directed reading lesson. In R. C. Anderson, J. Osborn, & R. J.
Tierney (Eds.). Learning to read in American schools: Basal readers and content texts. (pp. 3-20). Mahwah, NJ:
Lawrence Erlbaum Associates.
Cianciolo, P. (1970). Illustrations in children’s books (2nd ed.). Dubuque, IA: William C. Brown Company Publishers.
Elster, C., & Simons, H.D. (1985, November). How important are illustrations in children’s readers? The Reading Teacher, 148-
Fang, Z. (1996). Illustrations, text, and the child reader: what are pictures in children’s storybooks for? Reading Horizons, 37,
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.
Gyselinck, V., & Tardieu, H. (1999). The role of illustrations in text comprehension: what, when, for whom, and why? In H.
van Oostendorp & S.R. Goldman (Eds.), The construction of mental representations during reading (pp. 195-218).
Mahwah, NJ: Lawrence Erlbaum.
King, E. (1967). Critical appraisal of research on children’s reading interests, preferences, and habits. Canadian Education and
Research Digest, 7, 312-326.
Peeck, J. (1974). Retention of pictorial and verbal content of a text with illustrations. Journal of Educational Psychology, 66,
Read, J.D., & Barnsley, R.H. (1977). Remember Dick and Jane? Memory for elementary school readers. Canadian Journal of
Behavioral Science, 9, 361-370.
Schallert, D.L. (1980). The role of illustrations in reading comprehension. In R.J. Spiro, B.C. Bruce, & W.F. Brewer (Eds.),
Theoretical issues in comprehension instruction (pp. 503-525). Hillsdale, NJ: Lawrence Erlbaum.
Simons, H. (1990). Picture dependence in first-grade basal texts. Journal of Educational Research, 84(2), 86-92.
Tierney, R. J. & Cunningham, J. W. (1984). Research on teaching reading comprehension. In P. D. Pearson (Ed.), Handbook of
Reading Research (pp.609-655). New York: Longman Inc.
Table 1: The 15 comprehension questions and the preference questions asked of each participant. ‘I’ indicates questions whose
answers were only found in the illustrations, ‘T’ indicates those questions whose answers were only found in the text,
and ‘TI’ indicates questions whose answers were redundantly available in both the text and illustrations. Participants
were shown small examples of all illustration styles when asked the questions comparing the four groups of pictures.
How many balls were in the grass? TI
Why do Tommy and Oro like playing with the tennis ball? T
How does a rat see the world differently from people? I
What color was the tractor to Tommy? TI
What color was the tractor to Justine? TI
How many horses can the barn hold? T
Where did Tommy sit to wait for his friend Justine? TI
What color is Tommy’s house? I
Why can’t Tommy see some colors? T
What animals are in the yard in front of the barn? I
What did the truck pass on the way down the driveway? I
Oro means what in Spanish? T
In what color truck did Justine arrive? TI
What color is Justine’s shirt? I
Who won the race to the barn and to the house? T
What did you like most about the book?
What did you like least about the book?
*What did you like most about the pictures?
*What did you like least about the pictures?
Of the four groups [of pictures], which one do you like the most? Why?
Of the four groups [of pictures], which one do you like the least? Why?
*only asked to those participating whose conditions contained illustrations
Table 2: The numbers of participants who reported liking each of the four book styles the best and least.
Bright - Realistic
Bright - Abstract
Somber - Realistic
Somber - Abstract
Chosen as Best Liked
Chosen as Least Liked
Figure 1: Shows for each of the Question Source categories, the mean number (and standard error) of comprehension questions
correct for each of the nine books (N = 8 per book condition, except the bright-abstract text-plus-illustrations condition in which
N= 7). When the book included only text, participants did equally well as long as the questions asked for information that was
included in the text (with or without illustrations). Performance was poor on questions based on illustrations only. For
participant’s who received only the illustrations, performance was very poor for text-only questions, better for illustrations-only
questions, and even better for text-plus-illustrations questions. Book style (bright versus somber and realistic versus abstract) did
not significantly affect these scores. Participants who received books that included both text and illustrations, performed
similarly to participants receiving the text-only book, although they did score higher on the illustrations-only questions. Book
style showed a trend that the bright illustrations (which were the most preferred illustration styles) led to significantly enhanced
comprehension for the text-plus-illustrationss questions, while they led to slightly lower comprehension scores for the text-only
Text Illustrations Text & Illustrations
I-BA I-BR I-SA I-SR Text Only TI-BA TI-BR TI-SA TI-SR