Content uploaded by Patti M. Valkenburg
Author content
All content in this area was uploaded by Patti M. Valkenburg on Dec 08, 2014
Content may be subject to copyright.
Children’s Creative Imagination
21
Children’s Creative Imagination
in Response to Radio and
Television Stories
by Patti M. Valkenburg and Johannes W. J. Beentjes
The prevailing explanation for the experimental finding that radio
stories elicit more novel responses than do television stories is that
viewers have difficulty dissociating themselves from ready-made
television images (visualization hypothesis). In this experiment, we
investigated a rival hypothesis that radio stories elicit more novel
responses than do television stories because they are less well
remembered (faulty-memory hypothesis). We presented 64 children
at two age levels (grades 1 to 2 and 3 to 4) with one radio story
and one television story, and exposed half the children in both age
groups to the radio story twice to stimulate their memory. Contrary
to the faulty-memory hypothesis, double presentation of a radio
story did not result in fewer novel ideas than did a single presenta-
tion. In the older age group, radio stories elicited more novel
responses than did television stories. We found no medium differ-
ence in the younger age group.
Television viewing, it is often claimed, evokes less creative imagination than
does radio listening (Greenfield & Beagles-Roos, 1988; Mander, 1978; Meline,
1976; Webb, 1980; Winn, 1985). According to the visualization hypothesis,
television, unlike radio, presents viewers with ready-made visual images and
leaves them little opportunity to form their own images. When engaged in
Patti M. Valkenburg (PhD, Leiden University, 1995) is an associate professor and Dutch Royal
Academy fellow at the University of Amsterdam, the Netherlands. Her research interests include
children and media, methodology, and political communication. Johannes W. J. Beentjes (PhD,
Leiden University, 1990) is an associate professor of Communication at the University of Nijmegen,
the Netherlands. His research interests include media and children’s culture. An earlier version of
this article was presented at the 1996 annual meeting of the International Communication Associa-
tion, Chicago, IL. The data were collected when the first author was working at Leiden University,
the Netherlands. We thank Tom H. A. van der Voort and Leo J. Th. van der Kamp for their useful
suggestions, Patricia Greenfield for the stimulus materials and scoring manuals used in the study by
Greenfield, Farrar-Hartmann, and Beagles-Roos (1986), and three anonymous reviewers for their
valuable comments.
CopCopyright
Copyright © 1997 Journal of Communication 47(2), Spring. 0021-9916/97/$5.00
Journal of Communication, Spring 1997
22
creative thinking, viewers find it hard to dissociate themselves from the images
supplied by television, so that they have difficulty generating novel ideas. This
hypothesis attributes the reductive effect of television on creative imagination to
the medium’s visual nature (Valkenburg & van der Voort, 1994).
Six media comparison experiments, conducted with children between the
ages of 7 and 13, have investigated the effect of radio and television on creative
imagination. In one study, researchers presented children with problems to be
solved (Meline, 1976). In the remaining five studies they presented the children
with stories (Greenfield & Beagles-Roos, 1988; Greenfield, Farrar, & Beagles-
Roos, 1986; Kerns, 1981; Runco & Pezdek, 1984; Vibbert & Meringoff, 1981).
The problems and stories were presented in either radio or television format.
Because the sound tracks were the same for the radio and television versions,
the television versions differed only in that they included images. In all studies,
researchers defined creative imagination as the number of (novel) ideas
produced in response to a creative assignment. Five different types of creative
assignments were used:
1. Problem solutions (Meline, 1976). Researchers presented children, via
radio or television, with four problems (e.g., getting people to stop smoking),
then asked them to invent one creative solution for each problem. Children’s
problem solutions were scored on stimulus freedom, that is, the number of
novel elements not present in the given stimulus information.
2. Drawings (Vibbert & Meringoff, 1981). After being exposed to a radio or
television story, children were asked to create drawings depicting various se-
lected story scenes. The researchers scored the drawings on stimulus freedom.
3. Questions about a story (Kerns, 1981). Following exposure to either a
radio or a television story, children answered questions designed to elicit
creative responses to the story. The responses were scored on stimulus free-
dom, fluency (the number of ideas), flexibility (the number of distinct and
different conceptual categories), and originality (the number of ideas statistically
infrequent relative to the ideas of the other children in the sample).
4. A just suppose test (Runco & Pezdek, 1984). Following a radio or televi-
sion story, researchers gave children an adaptation of Torrance’s (1974) Just
Suppose Test, which required them to generate ideas about what might have
happened if the story had turned out differently. Responses were scored on
fluency, flexibility, and originality.
5. Story completions (Greenfield & Beagles-Roos, 1988; Greenfield et al.,
1986). Researchers presented, via radio or television, a story that was stopped
just before the end. Then, the children were asked to think up an ending for
the incomplete story. The story completions were scored on stimulus freedom.
With four of the five creative assignments, radio led to more novel responses
than did television. Only the adapted Just Suppose Test used by Runco and
Pezdek (1984) failed to yield a significant medium effect on creative imagina-
tion. However, this absence of an effect might be attributable to the low
sensitivity of the test, which, compared to the other operationalizations of
creative imagination, elicited remarkably few creative responses in both the
radio and television conditions (Valkenburg & van der Voort, 1994). In the
Children’s Creative Imagination
23
above-mentioned experiments, the medium differences in favor of radio were
explained with the visualization hypothesis. According to the authors, the
television presentations led to fewer novel ideas than did the radio presenta-
tions because children in the television condition had difficulty dissociating
themselves from television images during creative thinking.
However, the results of the previous media comparison experiments can also
be explained in a different way. According to a rival hypothesis, discussed by
Greenfield and Beagles-Roos (1988), and Valkenburg and van der Voort (1994),
radio presentations elicited more novel responses than television presentations,
not because radio presentations were more stimulating for creative imagination,
but because they were remembered less well. According to this faulty-memory
hypothesis, radio listeners were less able to incorporate repetitive elements in
their creative responses than were television viewers, because the radio presen-
tation was less clear in their minds. The novel ideas produced by children in
the radio condition might not have been creative responses but irrelevant
fabrications invented to compensate for faulty memory.
The visualization and the faulty-memory hypotheses both provide an expla-
nation for the finding in previous media comparison experiments that children
produce more novel ideas following a radio presentation than following a
television presentation. However, the two hypotheses fundamentally differ in
the interpretation of the cognitive processes that underlie the production of
novel ideas. The visualization hypothesis states that the novel ideas after a radio
story are expressions of children’s creativity. According to this hypothesis, radio
leaves children more cognitive room to generate creative ideas than does
television, because the production of novel ideas in the radio condition is not
hindered by ready-made images. The faulty-memory hypothesis, however,
disputes that the superior production of novel ideas after a radio presentation
should be attributed to creativity. This hypothesis claims that the higher number
of novel ideas after a radio presentation results from inferior story recall. The
novel ideas produced after radio listening are not creative ideas, but merely
inventions to fill in holes in a faulty memory.
A part of the faulty-memory hypothesis is that radio information is remem-
bered less well than television information. This is supported by experimental
evidence (Beagles-Roos & Gat, 1983; Gibbons, Anderson, Smith, Field, & Fisher,
1986; Greenfield & Beagles-Roos, 1988; Hayes, Kelly, & Mandel, 1986). How-
ever, to date, none of the media comparison experiments has investigated
whether the relatively poor recall of radio information is responsible for the
incorporation of fewer repetitive and more novel ideas in children’s creative
products. Therefore, the faulty-memory hypothesis cannot be excluded as a
rival explanation for the differences in creative imagination elicited by radio
and television found in earlier media experiments.
The purpose of the present experiment was to test whether the faulty-
memory hypothesis should be maintained or excluded as a rival explanation for
the relatively high number of novel ideas in creative products following a radio
presentation. In addition, we investigated whether the differences found in
creative imagination in favor of radio could be replicated. To enhance compara-
Journal of Communication, Spring 1997
24
bility, we used the same stimulus materials employed in three of the six previ-
ous media comparison experiments. In these studies, children’s creative imagi-
nation was operationalized by means of story completions (Greenfield &
Beagles-Roos, 1988; Greenfield et al., 1986) or a just suppose test (Runco &
Pezdek, 1984). We decided to use story completions, because Runco and
Pezdek’s (1984) test might have lacked sensitivity (Valkenburg & van der Voort,
1994). Following Greenfield and Beagles-Roos and Greenfield et al., we pre-
sented children from two different age groups (grades 1 to 2 and 3 to 4) with
one radio story and one television story. Both stories were stopped just before
the end. After each story, the children were asked to retell the story as well as
they could. We assessed children’s story recall on the basis of their retellings.
Subsequently, we asked the children to think of an ending for the incomplete
story. To determine children’s creative imagination from their story endings, we
followed the content analytic procedure developed by Greenfield and col-
leagues. We investigated how many ideas in the story completions were repro-
duced from the stimulus story (i.e., the number of repetitive ideas) and how
many ideas in the story completions were not present in the given stimulus
story (i.e., the number of novel ideas).
We investigated the faulty-memory hypothesis in two ways. First, we intro-
duced an extra radio condition in addition to the usual radio and television
conditions, in which children were exposed twice to the same radio story.
Because there is ample evidence that repetitive stimulus presentation improves
recall (e.g., Crowder, 1976; Richardson-Klavehn & Bjork, 1988), we expected
that double presentation of a radio story would stimulate children’s recall. If this
expectation is verified, the validity of the faulty-memory hypothesis can be
investigated by comparing children’s creative story completions following a
double radio presentation with those following a single radio presentation.
After all, if the faulty-memory hypothesis is valid, improved recall of a radio
story should result in fewer novel and more repetitive responses in the story
com-pletions.
A second way in which we investigated the faulty-memory hypothesis was
by using an alternative scoring method for assessing children’s creative imagina-
tion that takes the quality of the novel ideas into account. Earlier media com-
parison experiments focused only on the number of novel ideas produced and
failed to pay attention to the quality of these novel ideas. Therefore, it was
possible that the novel ideas produced by children who listened to the radio
story were irrelevant fabrications invented to compensate for faulty memory.
To assess both the novelty and the quality of children’s story completions,
we used Amabile’s (1982) consensus method for assessing creative products.
This method assumes that the creative value of a product (i.e., story comple-
tions) cannot be assessed through objective analysis. Instead, the most suitable
method is to rely on subjective judgments of creative imagination made by
observers familiar with the domain in question. If these observers agree on the
creative value of the individual products, then their judgments can be accepted
as a theoretically valid measure of creative imagination. Similarly, in everyday
life, a product is labeled as creative because appropriate judges collectively
Children’s Creative Imagination
25
agree on this evaluation (Amabile, 1982; Sternberg, 1988; Sternberg & Lubart,
1991). The consensus method has been applied to creative products of both
adults (e.g., poems) and children (e.g., collages, stories made up to accompany
a picture book). In all these studies, there was consensus on the creative value
of the individual creative products, evidenced by high interjudge reliabilities
(Amabile, 1982; Amabile & Gittomer, 1984; Amabile, Hennessey, & Grossman,
1986; Hennessy & Amabile, 1988; Lubart & Sternberg, 1991)
Hypotheses
The faulty-memory hypothesis postulates that inferior story recall results in story
completions with fewer repetitive and more novel ideas. To stimulate recall, we
presented children with a radio story twice instead of once. If these children
remember the radio story better than do children presented with the radio story
once, the faulty-memory hypothesis predicts that:
H1a: The story completions of children presented with the radio story twice
will contain fewer novel ideas than those of the children who listen to the
radio story once.
H1b: The story completions of children presented with the radio story twice
will contain more repetitive ideas than those of the children who listen to the
radio story once.
The faulty-memory hypothesis can be investigated experimentally, and
correlationally. The hypothesis predicts that the level of story recall is related
negatively to the number of novel ideas and positively to the number of
repetitive ideas in the story completions. The correlational prediction of the
faulty-memory hypothesis results in the following hypothesis.
H2: As children remember the stimulus story less well, they produce more
novel and fewer repetitive ideas in their story completions.
The faulty-memory hypothesis also predicts that faulty memory results in a
higher number of irrelevant fabrications in the story completions. We investi-
gated this prediction by comparing the quality of the story completions of
children who heard the radio story twice with that of the children who heard
the story once. We assessed the quality of the story completions through
Amabile’s (1982) consensus method. After having checked whether the double
radio presentation caused superior story recall, we investigated the following
hypothesis.
H3: The story completions of children presented with the radio story twice are
of a better quality than those of the children who listen to the radio story once.
Because various media comparison experiments found that radio stories
elicited more novel (Greenfield & Beagles-Roos, 1988; Greenfield et al., 1986;
Journal of Communication, Spring 1997
26
Kerns, 1981; Meline, 1976; Vibbert & Meringoff, 1981) and fewer repetitive
(Greenfield & Beagles-Roos; Greenfield et al.) responses than did television
stories, we investigated the following hypothesis.
H4a: In response to a radio story, children generate more novel elements in
their story completions than in response to a television story.
H4b: In response to a radio story, children generate fewer repetitive elements
in their story completions than in response to a television story.
In the present study, the number of novel ideas in children’s story comple-
tions was scored in two different ways: by means of the content analytic
method of Greenfield et al. (1986) and by means of the consensus method of
Amabile (1982). The use of two scoring methods offered the opportunity to
investigate the extent to which the content analytic and the consensus methods
produced the same results. The content analytic method might yield biased
results in two ways. First, there is a risk that a story completion that contains a
high number of irrelevant fabrications receives an undeservedly high creative-
imagination score. Second, through the content analytic method, a short story
completion gets a low creative-imagination score by definition, although the
story completion might actually be very creative. Addition of the consensus
method offers the opportunity to investigate the possible scoring artifacts of the
content analytic method. Therefore, in this study we investigated the following
research question.
RQ: To what extent do the results of the content analytic scoring method agree
with those of the consensus method?
Method
Sample
We recruited the subjects, 32 first and second graders (6.5 to 8 years old) and 32
third and fourth graders (8.5 to 10 years old), from eight primary schools in an
urban district in the Netherlands. The sampling method used insured that the
sample of children was representative of Dutch society in terms of socioeco-
nomic status. The children were all native speakers of Dutch without learning,
visual, speech, or hearing disabilities. Boys and girls were equally represented
in the two age groups.
Story Materials
Three of the earlier media comparison experiments (Greenfield & Beagles-Roos,
1988; Greenfield et al., 1986; Runco & Pezdek, 1984) used two animated stories
produced by Weston Woods Studios. They were Strega Nonna (De Paola, 1975),
a story about a kind-hearted old witch with a magic pasta pot, and A Story, A
Story, an African folktale. To enhance comparability with earlier media compari-
son experiments, we asked Patricia Greenfield to send us these stories. How-
Children’s Creative Imagination
27
ever, after inspection of the two stories, we agreed with Greenfield et al.’s
(1986) conclusion that A Story, A Story is too difficult for younger children to
comprehend. Accordingly, we used only Strega Nonna in the present study.
Instead of A Story, A Story, we selected Doctor De Soto (Steig, 1983), another
animated story produced by Weston Woods Studios. Doctor De Soto has an
easier story line, about a mouse dentist and his wife who get a hungry fox as
a patient.
A female Dutch narrator translated and recorded the text of the U.S. sound
track of Strega Nonna. To keep the voice and style of narration similar in the
two stories, the same narrator also recorded Doctor De Soto. The sound tracks
were identical for the radio and television versions of each story. As in the
studies by Greenfield and colleagues, each story was stopped approximately
1 minute before the end. The incomplete versions of both Strega Nonna and
Doctor De Soto each had running times of approximately 7 minutes.
Design
The experiment used a 2 (age group: grades 1 to 2 and 3 to 4) x 2 (radio
presentation: single vs. double presentation of radio story) x 2 (medium: radio
vs. television story) mixed design. Medium was varied within subjects; age
group and radio presentation were between-subjects variables. We presented
each child in each age group with two stories, one in radio format and another
in television format. Half the subjects in each age group listened to the radio
story just once; the other half heard it twice. Within each age group and radio
condition, half the subjects were exposed to the television version of Strega
Nonna and the radio version of Doctor De Soto. The other half listened to the
radio version of Strega Nonna and watched the television version of Doctor De
Soto. To control for possible order effects, half the children in each age group
and radio condition received the television story first, and half received the
radio story first.
Procedure
A female experimenter in an empty room in the child’s primary school individu-
ally tested each child. To prevent the children from discussing the stories with
classmates, we selected only two children from each classroom. We chose the
first child randomly, then asked the teacher to select a second child of the
opposite sex with the same level of linguistic competence as the first. Finally,
we assigned the two children randomly, one to the double and one to the
single radio story. This matching method was applied to reduce the error
variance of the between-subjects variable radio condition (single vs. double
radio story), and hereby to improve the power.
During the experimental session, the child sat to the right of the experi-
menter at a table. A television monitor and a radio cassette player sat on the
table in front of the subject. Each session began with an informal conversation
of approximately 5 minutes to help the child feel at ease, and to make certain
that he or she understood that the experiment had nothing to do with formal
grades or testing. The first story was introduced as follows: “We are going to
Journal of Communication, Spring 1997
28
1 For a detailed description of the conventions used in converting verbal texts into propositions, see
Kintsch (1974) and Kintsch and Van Dijk (1978).
listen to/watch a story. The story is called Strega Nonna /Doctor De Soto. After
we have heard/seen this story, we are going to talk about it.” Then the incom-
plete story was presented.
At the conclusion of the story presentation, the experimenter asked if the
child had ever heard or seen the story before. Children who were familiar with
either story were replaced by a classmate with the same level of linguistic
competence. Subsequently, we gave the following recall instruction: “Now I
would like you to tell me what the story was about. It was a long story and
nobody can remember everything about a story, but I would like you to tell me
everything that happened in the story from the beginning to the end.”
After the child finished his or her retelling of the story, we gave the creative
imagination instruction. With respect to Strega Nonna, the instruction was as
follows: “At the end of the story Strega Nonna blew the three kisses, and with a
sputter the pasta pot stopped boiling. ‘Oh grazie, thank you,’ the people cried.
And then the story ended. But in reality the story still continues. And now I
would like you to tell me a story about what you think is going to happen
next.” With respect to Doctor De Soto, we gave the same instruction, but with
this introduction: “At the end of the story the dentist said to the fox: ‘You will
never have to visit us again.’ And the fox thought: ‘he he, no one will ever visit
you again.’ And then the story ended. . . .”
When the child had finished the first creative imagination assignment, the ex-
perimenter introduced the second story. The procedure for the second story was
identical to that for the first story. However, a child who had listened to a radio
story first received the second story in television format. A child who first had seen
a television story received the second story in radio format. Children’s retellings
and story completions were audiotaped for later transcription.
Measures
Free recall. To score the children’s retellings, we conducted a propositional
analysis as developed by Kintsch (1974). In a first step, we converted the
transcribed sound track of each stimulus story into an ordered list of verbal
propositions. A verbal proposition consists of a predicate (e.g., verb, adjectives,
adverbs) and one or more arguments (e.g., subject, object, or another proposi-
tion).1 The verbal propositional analysis yielded 284 verbal propositions for
Strega Nonna and 312 for Doctor De Soto.
To measure the memory of visual story information, in a second step, the
visual part of each stimulus story was converted into an ordered list of visual
propositions (i.e., stimulus propositions conveyed visually). However, because
Kintsch (1974) focused only on verbal texts, we had to develop an additional
procedure. First, we developed a visual observation instruction in which we
described clear criteria concerning how to score visual propositions about
characters, objects, time and location settings, and actions. Subsequently, the
first author and another rater analyzed the visual parts of each stimulus story in
Children’s Creative Imagination
29
accordance with the criteria established in the visual observation instruction.
The interrater reliability for this procedure, based on the ratio of the number of
commonly scored propositions to the total number of scored propositions, was
.87. The visual content analysis yielded 542 visual propositions for Strega
Nonna (57 shots) and 694 for Doctor De Soto (67 shots).
In a final step, we scored children’s transcribed retellings. Each retelling was
converted into a list of propositions in accordance with the conventions of
Kintsch (1974). The interrater reliability for the number of propositions counted
in each free-recall protocol, conducted on a sample of 16 retellings, was .95. We
scored each recall proposition in a child’s retelling on whether it was semanti-
cally equivalent to a verbal or visual stimulus proposition. The interrater reliabil-
ity of this scoring procedure, conducted on a sample of 40 free-recall protocols,
was .94. A child’s total recall score was the sum of the verbal and visual stimu-
lus propositions that were recalled. This scoring procedure implied that
children’s recall after the television presentation was made up of verbal as well
as visual propositions, whereas children’s recall after the radio presentation was
based on only verbal propositions. One could argue that this scoring procedure
unjustly privileged children in the television condition and that this bias should
be controlled. However, the point of media comparison experiments is to
investigate differences between media. Therefore, one should not control for
differences between the two media. After all, dual information provision is
inherent to television.
During data screening, we located one outlier with an extreme value on the
recall score. Because outliers may exert a negative influence on statistical power
(Stevens, 1992), we assigned the outlier a score that was one unit higher than
the next highest score in the cell in which it was detected (Tabachnick & Fidell,
1989, p. 70).
Creativity. We coded the children’s story completions according to the
protocol in the scoring manual of Farrar-Hartman (1984), which was sent to us
by Patricia Greenfield. Greenfield et al.’s (1986) basic coding unit was the
proposition. A proposition in the story completions was considered novel if it
was not found in the stimulus story, and repetitive if it was semantically equiva-
lent to a verbal or visual stimulus proposition. In our study, we replicated the
propositional analysis of Greenfield and colleagues. The story completions were
converted into a list of propositions by two independent raters. The interrater
reliability for this procedure was .96. Then, each proposition was coded as
either novel or repetitive. The interrater reliability for this procedure was .90.
Data screening revealed four outliers, cases with an extremely high number of
novel propositions in response to both the television and the radio story. We
assigned each outlier a score that was one unit higher than the next highest
score in the cell in which it was detected (Tabachnick & Fidell, 1989).
In addition to using the content analytic method of Greenfield and col-
leagues, we evaluated the story completions using Amabile’s (1982) consensus
method. The story completions were rated on five dimensions, two concerning
the novelty of the story completions and three concerning the quality of the
story completions. The two novelty dimensions (creativity and novelty) were
Journal of Communication, Spring 1997
30
adopted from Amabile (1982) and Hennessey and Amabile (1988). We had to
develop the three quality dimensions ourselves, because story completions
require different quality criteria than do the creative products that were used in
the Amabile studies (e.g., poems, collages). We asked 16 developmental
psychologists to read three story completions and to indicate which quality
criteria a creative story completion has to meet to be deemed high in quality.
According to these psychologists, a story completion was high in quality if it
satisfied the following three criteria: (a) It maintained continuity with the
preceding stimulus story, (b) the events within the story completion were
coherent, and (c) the story completion provided an ending to the previous,
incomplete stimulus story.
We recruited five developmental psychologists to judge the story comple-
tions. The judges were unaware of the research hypotheses and made their
assessments independently. Before judging the story completions, we familiar-
ized the judges with the stimulus stories and the dimensions of judgment. Then,
we asked them to read the story completions once. They were then instructed
to go through the story completions a second time and to judge each story
completion on a discrete 7-point rating scale on the following dimensions:
• Creativity: Using your own, subjective definition of creativity, assess how
creative the story completion is.
• Novelty: How novel is the story completion? Does the story completion
have an original twist, or is it only a variation on the preceding story theme?
• Continuity: Does the story completion maintain continuity with the preced-
ing stimulus story or is it irrelevant?
• Coherence: To what extent are the events within the story completion
coherent? Do the events within the story completion fit together well, or is there
no clear connection between them?
• Completion: Does the story completion round off well or does it stop
abruptly?
The judgment procedure took 2 days. On the first day, judges evaluated the
64 completions of Strega Nonna, and on the second day, the 64 completions of
Doctor De Soto, with story completions and judgment dimensions presented to
the judges in random order. The judgments were made per dimension. There-
fore, a judge evaluated all story completions on each dimension before going to
the next.
Following Amabile (1982), we calculated a Spearman-Brown interjudge
reliability for each judgment dimension (Nunnally, 1967). For Strega Nonna, the
reliability coefficients of the dimensions were as follows: creativity .87, novelty
.90, coherence .80, continuity .79, and completion .81. For Doctor De Soto, they
were as follows: creativity .92, novelty .86, coherence .80, continuity .79, and
completion .81.
To investigate the extent to which the scores on the five judgment dimen-
sions reflected underlying factors, a principal component analysis with varimax
rotation was carried out. A clear two-factor pattern emerged. The creativity and
novelty dimensions loaded high on the first factor, which could by labeled
novelty. The continuity, coherence, and completion dimensions loaded high on
Children’s Creative Imagination
31
the second factor, which could be labeled quality. Together, the two factors
accounted for 86.5% of the variance in the five judgment dimensions (see
Table 1).
Because of the high amount of variance accounted for by the factor solution,
we created a novelty scale and a quality scale on the basis of the five judgment
dimensions. We formed the novelty scale by totaling the child’s unweighted
scores on the creativity and novelty dimensions. We constructed the quality
scale by summing the child’s unweighted scores on the continuity, coherence,
and completion dimensions. The reliabilities of the novelty and quality scales
were satisfactory. Cronbach’s alpha values were .86 for the novelty scale and
.81 for the quality scale. The correlation between the novelty and quality scale
was, r = .19, p > .05.
Results
Correlation Between the Two Creative-Imagination Measures
Our research question asked to what extent the results of the two scoring
methods for creative imagination, the content analytic method (i.e., the number
of novel propositions) and the consensus method (i.e., novelty judgments),
would agree. To investigate this research question, we calculated a Pearson
correlation between the number of novel propositions and the scores on the
novelty scale found through principal component analysis. The number of
novel propositions correlated strongly with the scores on the novelty scale (r =
.76, p < .01), showing that the results of the two scoring methods agreed to a
large extent. Because of the high correlation between the content analytic and
the consensus method, the two scoring methods cannot be considered as
independent from each other. Therefore, in the analyses below, the results from
the two scoring methods were simultaneously included in a multivariate analysis of
variance (Tabachnick & Fidell, 1989).
Table 1. Varimax Rotated Factor Solution for the Five Judgment Dimensions of Creative
Imagination
Factor 1 Factor 2
Judgment Dimension Novelty Quality
Factor 1: Novelty
Creativity .94 .19
Novelty .97 -.01
Factor 2: Quality
Continuity -.31 .87
Coherence .23 .89
Completion .28 .80
Eigenvalues 2.22 2.08
Journal of Communication, Spring 1997
32
Manipulation Check
To investigate the faulty-memory hypothesis, we introduced an extra radio
condition in which the radio story was presented twice to stimulate the child’s
story recall. A major premise of this study was that a double radio presentation
would be remembered better than a single one. To check if this occurred, we
performed an analysis of variance on the number of correctly recalled proposi-
tions in children’s retellings following a radio story, with age group (grades 1 to
2 and 3 to 4) and radio condition (single presentation and double presentation)
as between-subjects variables. Table 2 presents the mean number of proposi-
tions recalled from a single or double presentation of the radio story in the two
age groups.
The ANOVA revealed a significant main effect for radio story, F(1,60) =
10.43, p < .01, η2 = .15, and age group, F(1,60) = 4.65, p < .05, η2 = .07. The
interaction between radio story and age group was not significant, F(1,60) =
3.59, p > .05. The main effect of radio story implied that a double presentation
of a radio story caused better story recall than a single presentation. The main
effect of age group meant that the older children recalled more from both the
single and double presentations than did the younger children.
To compare children’s story recall following a single or double presentation
of the radio story with story recall in response to a television story, two separate
ANOVAs were conducted. Both analyses included age group as a between-
subjects variable and medium (radio vs. television) as a within-subjects variable.
In the first ANOVA we investigated the differences in story recall following a
radio story presented once and a television story. The difference in recall
between a television story (M = 43.47, SD = 22.10) and a single radio story (M =
40.53, SD = 20.69) was not statistically significant, F(1,30) = .83, p = .37. In the
second ANOVA we investigated the differences in story recall in response to a
radio story presented twice and a television story. The analysis revealed that, in
both age groups, children remembered a radio story presented twice (M =
55.94, SD = 19.31) better than a television story presented once (M = 40.59, SD =
18.41), F(1,30) = 11.54, p < .01, η2 = .28.
Number of Novel and Repetitive Ideas in the Story Completions
Two research hypotheses made predictions about the number of novel ideas in
the story completions. H1a predicted that the story completions of children
presented with a radio story twice would contain fewer novel ideas than those
Table 2. Mean Number of Recalled Stimulus Propositions Following a Single or Double
Radio Story by Age Group
Single Radio Story Double Radio Story
Grades 1–2 Grades 3–4 Grades 1–2 Grades 3–4
M
30.88 50.19 55.31 56.56
SD
18.98 18.03 21.74 17.22
Children’s Creative Imagination
33
of children who listened to a radio story once. H4a predicted that the story
completions in response to a radio presentation would contain more novel
ideas than those in response to a television presentation. To test H1a and H4a,
we conducted a MANOVA on the number of novel propositions and the scores
on the novelty scale with age group (grades 1 to 2 and 3 to 4) and radio
condition (single and double radio story) as between-subjects variables and
medium (radio vs. television) as a within-subjects variable. The MANOVA did
not reveal a significant interaction between medium and radio condition,
F(1,60) = .03, p = .86. This implied that, contrary to H1a, a radio story pre-
sented twice elicited as many novel ideas as a radio story presented once. The
MANOVA did reveal a significant main effect of age group, F(1,60) = 19.21, p <
.001, η2 =.24, and a significant interaction between medium and age group,
F(1,60) = 4.23, p < .05, η2 =.07. Table 3 presents the mean numbers of novel
propositions and the mean scores on the novelty scale for the two age groups
and the two media.
The main effect of age group meant that the story completions of the older
children contained more novel ideas than those of the younger children. To
explain the interaction between medium and age group, we conducted a simple
effects analysis (Tabachnick & Fidell, 1989). This analysis showed that for older
children, in agreement with H4a, the story completions following a radio story
contained more novel ideas than those following a television story, F(1,60) =
4.67, p < .05, η2 =.07. For the younger children, however, there was no signifi-
cant difference in the number of novel ideas in response to a radio or television
story, F(1,60) = .63, p = .43.
Two hypotheses pertained to the number of repetitive ideas in the story
completions. H1b predicted that the story completions of children who heard a
radio story twice would contain more repetitive ideas than those of children
who listened to a radio story once. H4b predicted that the story completions in
response to a radio presentation would contain fewer repetitive ideas than
those in response to a television presentation. H1b and H4b were tested by
means of an ANOVA on the number of repetitive propositions in the story
Table 3. Mean Number of Novel Propositions and Novelty Judgments by Age Group and
Medium
Grades 1–2 Grades 3–4
(n
= 32) (
n
= 32)
Creative-Imagination Measures Radio TV Radio TV
Number of Novel Propositions
M
12.56 13.22 25.94 22.38
SD
9.26 7.56 15.85 13.89
Novelty Judgments
M
34.59 36.72 49.63 45.59
SD
13.01 13.34 13.61 13.59
Journal of Communication, Spring 1997
34
completions, with age group and radio condition as between-subjects variables
and medium as a within-subjects variable. The ANOVA did not yield a signifi-
cant interaction between medium and radio condition, F(1,60) = .44, p = .51.
This result implied that, contrary to H1b, the story completions after a double
radio story (M = 3.45, SD = 3.10) did not contain more repetitive ideas than
those after a single radio story (M = 5.39, SD = 4.77). The ANOVA showed only
a significant interaction between medium and age group, F(1,60) = 5.89, p <
.05, η2 = .09. A simple effects analysis indicated that H4b held only for the older
age group. The story completions of the older children contained more repeti-
tive propositions following a television story (M = 5.46, SD = 4.31) than follow-
ing a radio story (M = 3.81, SD = 3.25), F(1,60) = 4.73, p < .05, η2 =.07, regard-
less of whether the children heard the radio story once or twice. For the younger
children, we found no significant difference in the number of repetitive propo-
sitions between the radio and television presentations, F(1,60) = 1.72, p = .19.
Correlations of Story Recall with the Number of Novel and Repetitive Ideas in the
Story Completions
The faulty-memory hypothesis contained in H2 predicted that level of story
recall would be related negatively to the number of novel ideas and positively
to the number of repetitive ideas in the story completions. To investigate this
prediction, we correlated story recall with the number of novel propositions,
the scores on the novelty scale, and the number of repetitive propositions.
Contrary to H2, level of story recall was related positively to the number of
novel propositions in the story completions (r = .47, p < .01) as well as to the
scores on the novelty scale (r = .52, p < .01). Also contrary to H2, level of story
recall was not significantly related to the number of repetitive propositions
(r = .17, p > .05).
Quality Differences in Story Completions Following a Single or Double Radio
Presentation
The faulty-memory hypothesis contained in H3 predicted that the quality of the
story completions following a double presentation of a radio story would be
superior to that of story completions following a single radio presentation. To
test H3, an ANOVA was conducted on the quality judgments of the story
completions, again with age group and radio condition as between-subject
variables and medium as a within-subject variable. Contrary to the faulty-
memory hypothesis, the quality of the story completions following a double
radio story (M = 64.44, SD = 16.96) was not superior to that of the story
completions following a single radio story (M = 61.56, SD = 15.86), F(1,60) =
.84, p = .36. The quality of a radio story completion after a single presentation
also was not inferior to that of a television story completion (M = 64.20, SD =
14.23). The ANOVA revealed only a main effect of age group, F(1,60) = 4.72, p
< .05, η2 =.08. Older children generated story completions of higher quality than
did the younger children, regardless of whether they were presented with a
television story or a (single or double) radio story.
Children’s Creative Imagination
35
Discussion
The basic purpose of this study was to test the faulty-memory hypothesis, which
postulates that radio presentations elicit more novel ideas than do television
presentations, because radio information is less well remembered. In this
hypothesis, novel ideas are seen as fabrications to compensate for a faulty
memory. The faulty-memory hypothesis challenges the prevailing visualization
hypothesis, which states that children generate relatively few novel ideas in
response to a television presentation because they have difficulty dissociating
themselves from television’s ready-made images. In the visualization hypothesis,
novel ideas are considered expressions of children’s creativity.
In this study, we derived three predictions from the faulty-memory hypoth-
esis. The first prediction, that double presentation of a radio story leads to fewer
novel (H1a) and more repetitive ideas (H1b) in the story completions than does
a single radio presentation, was not confirmed. Two radio presentations elicited
as many novel and repetitive elements in the story completions as a single radio
presentation. The second prediction, that superior story recall is associated with
fewer novel and more repetitive elements in the story completions (H2), also
was not supported. Superior story recall was related to significantly more novel
ideas in the story completions and was not significantly related to the number
of repetitive ideas in the story completions. The third prediction, that the story
completions in response to a double presentation of a radio story are of a
higher quality than those in response to a single radio presentation (H3), was
not confirmed either. Our results showed that the quality of story completions
in response to a single presentation of a radio story was not inferior to the
quality of story endings after two presentations. Because none of the predic-
tions derived from the faulty-memory hypothesis was confirmed, there is no
reason to assume that the hypothesis provides a plausible rival explanation for
the finding in previous experiments that radio presentations elicit more novel
ideas than do television presentations.
Because the faulty-memory hypothesis proved to be implausible, the visual-
ization hypothesis has lost its rival. The visualization hypothesis, which states
that children have difficulty generating novel ideas because they cannot easily
dissociate themselves from television images, has never been directly investi-
gated. The visualization hypothesis did, however, receive indirect support. In
studies by Greenfield and colleagues, and in our study, children incorporated
more repetitive ideas in their television story completions than in their radio
story completions. The relatively high number of repetitive ideas in television
story completions cannot be attributed to a better story recall, because the
number of repetitive elements following a television story was also higher than
that following a double presentation of a radio story. Because the television
story differed from the radio story only in that it included images, it is not
illogical to ascribe the relatively high number of repetitive ideas following a
television story to an inability to dissociate oneself from ready-made television
images.
Journal of Communication, Spring 1997
36
A second purpose of our study was to investigate the RQ about the extent to
which the content analytic method (i.e., the number of novel propositions)
agreed with the novelty judgments obtained through the consensus method. As
discussed in the introduction, the content analytic method seems susceptible to
scoring artifacts because it does not evaluate the quality of story endings.
Despite this shortcoming, the content analytic method proved to correlate
highly with the consensus method. In addition, both scoring methods yielded a
similar pattern of results for (a) the older and younger children and (b) radio
and television. The agreement between the results of the content analytic and
consensus methods is an indication of the convergent validity of both methods
and increases confidence in the results obtained by the content analytic
method.
A final purpose of this study was to investigate whether the differences in
creative imagination in favor of the radio that were found in earlier experiments
could be replicated. On the basis of previous research, we predicted that a
radio story would elicit more novel (H4a) and fewer repetitive ideas (H4b) in
children’s story completions than a television story would. This hypothesis held
only for the older age group. In the younger age group, we found no significant
differences between the numbers of novel and repetitive ideas following a radio
or television story.
The different findings for the younger and the older children might be
explained in terms of children’s cognitive development. According to Piaget and
Inhelder (1969), two kinds of mental images develop during childhood:
reproductory images and anticipatory images. Reproductory images are images
in which scenes that have been observed earlier are brought into mind again.
Anticipatory images are images of movements and transformations, as well as
their possible outcomes, that were not previously observed. According to Piaget
and Inhelder, it is not until the level of concrete operations (7 to 8 years old)
that children are capable of creating anticipatory images. The ability to form
anticipatory images could be important for the creative-imagination assignment
in the present study. After all, we asked children to complete an unfinished
story, a task that calls for the ability to imagine scenes that have not been
observed before. Because children in grades 1 and 2 may have difficulty
forming anticipatory images, completing an unfinished story is a difficult
assignment, regardless of whether the story is conveyed via radio or television.
This may explain why the medium used did not affect the number of novel
elements produced by the younger children. Older children, who are already
capable of forming anticipatory images, apparently experience greater difficulty
generating a novel story completion after a television story than after a radio
story.
Our finding about the difference between younger and older children can be
compared only with the media comparison experiments by Greenfield and
colleagues (Greenfield & Beagles-Roos, 1988; Greenfield et al., 1986), because
the children in the other studies (Kerns, 1981; Meline, 1976; Runco & Pezdek,
1984; Vibbert & Meringoff, 1981) were older than those in our youngest age
group. The interaction between medium and age group found in the present
Children’s Creative Imagination
37
study does not agree with the results of the studies by Greenfield and col-
leagues. They found that radio stories elicited more novel ideas than did
television stories in both age groups. This discrepancy between the findings of
Greenfield and colleagues and the present results might also be explained in
terms of the Piagetian theory. The younger children in both our study and the
two studies by Greenfield and colleagues were between 6.5 and 8 years old
and, therefore, on the threshold of the level of concrete operations (7 to 8
years). This stage is, as mentioned before, important for the creative-imagina-
tion assignment employed in the studies involved. Maybe the younger children
in the studies by Greenfield and colleagues were just a little further in their
cognitive development than the younger children in our study, so that the
younger children in our study might have experienced greater difficulty com-
pleting the creative-imagination assignment.
As in earlier media comparison experiments, the results of the present study
are consistent with the visualization hypothesis. However, because the question
of whether children have difficulty dissociating themselves from television
images has never been investigated, the evidence for the visualization hypoth-
esis is not conclusive. In future media comparison studies into the effect of
radio and television on children’s creative imagination, the validity of the
visualization hypothesis can be investigated by distinguishing between elements
of the stimulus story that have been visualized (audiovisual elements) and
elements that have been conveyed only via audio track (audio-only elements).
The visualization hypothesis predicts that children’s creative imagination is
hindered only if creative assignments concern audiovisual elements of the
stimulus story, but not if assignments pertain to audio-only elements.
References
Amabile, T. M. (1982). Social psychology of creativity: A consensual assessment technique. Journal
of Personality and Social Psychology, 43, 997–1013.
Amabile, T. M., & Gitomer, J. (1984). Children’s artistic creativity: Effects of choice in task materials.
Personality and Social Psychology Bulletin, 10, 209–215.
Amabile, T. M., Hennesey, B. A., & Grossman, B. S. (1986). Social influences on creativity: The
effects of contracted-for reward. Journal of Personality and Social Psychology, 50, 14–23.
Beagles-Roos, J., & Gat, I. (1983). Specific impact of radio and television on children’s story
comprehension. Journal of Educational Psychology, 75, 128–137.
Crowder, R. G. (1976). Principles of learning and memory. Hillsdale, NJ: Erlbaum.
de Paola, T. (1975). Strega nonna. Englewood Cliffs, NJ: Prentice-Hall.
Farrar-Hartman, D. (1984). The differential effects of radio and television on the imaginativeness of
children’s story completions. Dissertation Abstracts International, 45(09), 3092B. (University
Microfilms No. DA84-28511)
Gibbons, J., Anderson, D. R., Smith, R., Field, D. E., & Fischer, C. (1986). Young children’s recall and
reconstruction of audio and audiovisual narratives. Child Development, 57, 1014–1023.
Greenfield, P. M., & Beagles-Roos, J. (1988). Radio vs. television: Their cognitive impact on children
of different socioeconomic and ethnic groups. Journal of Communication, 38(2), 71–92.
Journal of Communication, Spring 1997
38
Greenfield, P. M., Farrar, D., & Beagles-Roos, J. (1986). Is the medium the message? An experimental
comparison of the effects of radio and television on imagination. Journal of Applied Developmen-
tal Psychology, 7, 201–218.
Hayes, D. S., Kelly, S. B., & Mandel, M. (1986). Media differences in children’s story synopses: Radio
and television contrasted. Journal of Educational Psychology, 78, 341–346.
Hennessey, B. A., & Amabile, T. M. (1988). Story-telling: A method for assessing children’s creativity.
Journal of Creative Behavior, 22, 235–246.
Kerns, T. Y. (1981). Television: A bisensory bombardment that stifles children’s creativity. Phi Delta
Kappan, 62, 456–457.
Kintsch, W. (1974). The representation of meaning in memory. Hillsdale, NJ: Erlbaum.
Kintsch, W., & van Dijk, T. A. (1978). Toward a model of text comprehension and production.
Psychological Review, 85, 363–394.
Lubart, T. I., & Sternberg, R. J. (1991). Creative performance: Testing an investment theory.
Unpublished Manuscript, Yale University.
Mander, J. (1978). Four arguments for the elimination of television. New York: Quill.
Meline, C. W. (1976). Does the medium matter? Journal of Communication, 26(3), 81–89.
Nunnally, J. C. (1967). Psychometric theory. New York: McGraw-Hill.
Piaget, J., & Inhelder, B. (1969). The psychology of the child. London: Routledge & Kegan Paul.
Richardson-Klavehn, A., & Bjork, R. A. (1988). Measures of memory. Annual Review of Psychology,
39, 475–543.
Runco, M. A., & Pezdek, K. (1984). The effect of television and radio on children’s creativity.
Human Communication Research, 11, 109–120.
Steig, W. (1983). Doctor De Soto. New York: Farrar, Straus & Giroux.
Sternberg, R. J. (1988). The nature of creativity. New York: Cambridge University Press.
Sternberg, R. J., & Lubart, T. I. (1991). An investment theory of creativity and its development.
Human Development, 34, 1–31.
Stevens, J. (1992). Applied multivariate statistics for the social sciences (2nd ed.). Hillsdale, NJ:
Erlbaum.
Tabachnick, B. G., & Fidell, L. S. (1989). Using multivariate statistics (2nd ed.). New York: Harper
Collins.
Torrance, E. P. (1974). Torrance tests of creative thinking. Lexington, MA: Personal Press.
Valkenburg, P. M., & van der Voort, T. H. A. (1994). Influence of television on daydreaming and
creative imagination: A review of research. Psychological Bulletin, 116, 316–339.
Vibbert, M. M., & Meringoff, L. K. (1981). Children’s production and application of story imagery: A
cross-medium investigation (Tech. Rep. No. 23). Cambridge, MA: Project Zero, Harvard University.
(ERIC Document Reproduction Service No. ED 210 682).
Webb, E. (1980). Images of television: Some reflections. Use of English, 32(1), 10–17.
Winn, M. (1985). The plug-in drug: Television, children, and the family. New York: Penguin Books.