Content uploaded by Daniel R Anderson
Author content
All content in this area was uploaded by Daniel R Anderson on Feb 24, 2014
Content may be subject to copyright.
Journal of Educational Psychology
1999,
Vol.
91,
No. 4, 630-637Copyright 1999 by the American Psychological Association, Inc.
0022-0663/99/$3.00
Effects of Repeated Exposures to a Single Episode of the Television
Program Blue's Clues on the Viewing Behaviors and Comprehension
of Preschool Children
Alisha M. Crawley and Daniel R. Anderson
University of Massachusetts at AmherstAlice Wilder, Marsha Williams,
and Angela Santomero
Nickelodeon
A single episode of the preschool educational television program Blue's Clues was shown once
or repeated on 5 consecutive days for
3-
to 5-year-old viewers. A comparison group watched a
different preschool program one time. Viewer behavior was videotaped, and comprehension
and learning of Blue's Clues content was tested. With repetition, looking at the television
screen remained at a high level. Only 5-year-old boys' looking decreased. Verbal and
nonverbal interactions with the program (such as answering questions and pointing at the
screen) increased, especially during educational content. Comprehension improved with
repetition. Episode repetition is an effective strategy for enhancing learning and program
involvement for a preschool audience.
Preschool children often ask for repetition of storybooks
and videos. Mares (1998) surveyed parents of children aged
2 to
17
years about frequency of repeated videotape viewing.
Sixty-nine percent of parents of children aged 2 to 4 reported
that their children "almost always" watched videos repeat-
edly. The parents of less than 10% of the children in this age
group reported that their children "very rarely" or "never"
watched videos repeatedly. The percentage of children who
"almost always" watched repeatedly substantially de-
creased with age beyond the preschool years.
Almost any theory of comprehension would predict that
children's comprehension of a television program improves
with repetition. In the small amount of research concerning
television that has examined this issue, the prediction has
been supported. Peracchio (1992) studied the effects of
repetition on the acquisition of the event knowledge of a
product-exchange story presented aurally or audiovisually.
Five-
and 7-year-old children were shown either once or
three consecutive times a televised version of a child
returning a birthday present to a store. Repetition benefited
Alisha M. Crawley and Daniel R. Anderson, Department of
Psychology, University of Massachusetts at Amherst; Alice Wilder,
Marsha Williams, and Angela Santomero, Nickelodeon, New York.
Aspects of this research were reported in Alisha M. Crawley's
master's thesis at the University of Massachusetts at Amherst and
in a presentation at the annual meeting of the International
Communication Association in Montreal, May 1997. Funding for
this research was provided by Nickelodeon.
We thank Scott Gamer, Jennifer Shulman, and Phyllis Schwartz
for assistance with data collection. We also thank Allyssa Brandt,
Feme Friedman-Berg, Hillary Knowlton, Rachel Revell, Erica
Rutt, and Kate Tevlia for coding the videotape data.
Correspondence concerning this article should be addressed to
Daniel R. Anderson, Department of Psychology, University of
Massachusetts, Amherst, Massachusetts 01003. Electronic mail
may be sent to anderson@psych.umass.edu.
comprehension in both versions, but the benefit for the
audiovisual version was primarily for younger children.
Mares (1997) had 4- and 9-year-old children watch a
televised story that had been misunderstood in previous
research. Experimental versions of the story concerned an
old woman who was pretty and mean or ugly and kind;
viewers watched either once or four times, a week apart.
Mares found that comprehension of the younger children
was worse than comprehension of the older children for the
group that saw the story only once, but there was no age
difference for comprehension following repetition. The
younger children who watched the story four times reached
the ceiling comprehension score that the older children
achieved with one viewing. Children who saw the story only
once rated their enjoyment as slightly more than those who
saw it four times.
Mares (1997) conducted a second study in which 6- to
9-year-olds watched an edited version of a children's feature
film and compared children who were watching the film for
the first time with those who had seen it before. She
measured comprehension by asking the children to name
characters, make inferences about the movie, and
state the moral. She found that children who had seen the
film previously did better on character identification and
inferences.
Sell, Ray, and Lovelace (1995) had 4-year-olds watch an
edited version of a Sesame Street tape depicting a muppet
game show ("Alphabet Treasure Hunt") three times over a
3-week period. After each viewing, the children were asked,
"How do you play 'Alphabet Treasure Hunt'?" With
repetition, children increased their knowledge of the game
show script, focusing more on relevant details. This study
confounded repeated questioning with repeated viewing, so
some of the increase may have come from alerting the
children to the relevant features of the program.
In addition to the work on repetition of television
programs, there has been a small amount of research on
630
REPEATED VIEWING OF BLUE'S CLUES 631
repeated reading of stories to young children, which appears
to occur quite commonly. With repeated readings, young
children asked more questions and made more comments;
they also shifted the focus of their questions from relatively
superficial aspects of the books to the meaning of the text.
The more the children understood the books that were
repeatedly read, the more they appreciated them (Beaver,
1982;
Martinez, 1983; Martinez & Roser, 1985; Morrow,
1988;
Snow, 1983; Snow & Goldfield, 1983; Sulzby, 1985;
also see Samuels, 1979).
Despite the apparent benefits of repetition to learning,
curriculum-based television programs are not systematically
repeated. We assume that broadcasters avoid repetition
because they believe that attention to the program would
habituate and that audiences would thereby be diminished. If
so,
this assumption flies in the face of observations that
children repeatedly watch videotapes. The primary question
addressed in this article is whether preschool viewer atten-
tion and behaviors indicating program involvement necessar-
ily diminish with moderate amounts of program repetition.
Theories of attention to television provide conflicting
perspectives. Singer (1980) argued that young children's
attention to television is primarily attracted and held by
movement, scene changes, and other formal features that
elicit orienting reactions. From its simplicity, this theory
provides the most straightforward prediction for the present
study. Because orienting reactions readily habituate with
stimulus repetition, attention should decrease as the viewer
habituates. This should happen equivalently for each age
group.
Anderson and Lorch (1983) proposed that young chil-
dren's attention to television is primarily driven by their
comprehension of the content of television. Because televi-
sion programs are frequently difficult for preschoolers to
understand, this theory emphasizes comprehensibility of the
content. Anderson and Lorch argued that to the degree to
which the television program is optimally comprehensible, it
holds attention. If a program is unfamiliar but somewhat
understandable, attention should increase with repetition
until the program becomes optimally comprehensible. Be-
yond that point, when the content becomes overly familiar
and predictable, attention should decrease. Huston and
Wright
(1983,
1989) provided a more comprehensive theory
along similar lines, explicitly arguing that attention may
follow an inverted U-shaped function related to, among
other things, familiarity with the content.
Based on unpublished formative research done by the
producers, the episode of Blue's Clues used in the present
study was designed to be optimally comprehensible but
intellectually challenging to 4-year-olds. From this consider-
ation, our predictions, based on the Anderson and Lorch
(1983) and the Huston and Wright (1983,1989) theories, are
that 4-year-olds' attention should either be sustained or
increase with repetition. Because repetition should make the
program more comprehensible to 3-year-olds, their attention
should increase. Five-year-olds, on the other hand, would be
the most likely to show some decrease in attention because
the program may begin to become overly predictable.
There is another theory that is relevant to the present
work. Salomon (1983) argued that children invest more or
less mental effort in processing media. The amount of
mental effort that is invested depends on the degree to which
the medium is perceived as demanding such effort. For
example, elementary school children who are told that they
will be tested on the contents of a television program invest
more effort than children who do not receive such instruc-
tions (Salomon & Leigh, 1984; van der Molen & van der
Voort, 1997). Field and Anderson (1985) replicated this
result for younger children and found that visual attention
was correlated with the children's reports of invested mental
effort.
Blue's Clues differs from other curriculum-based televi-
sion programs insofar as the audience is frequently asked to
"help"
solve various problems presented during the pro-
gram. After delays designed to allow the audience time to
overtly or covertly provide answers, feedback on solution
strategies as well as correct answers are provided by the
program. In designing Blue's Clues, the creators hoped that
the opportunity to "participate" would increase the mental
effort the viewers invest in the program. This, combined
with the sense of mastery instilled by solving the problems
or otherwise knowing the answers, would help children
sustain interest in and enjoyment of the program. Salomon's
(1983) theory does not directly predict the effects of episode
repetition, but it is reasonable to suppose that as children
become familiar with the Blue's Clues format, they increase
the effort they invest in solving the problems and helping
provide answers. On this basis, it might be predicted that
visual attention should be maintained and that overt interac-
tions with the program should increase.
Only two prior studies examined visual attention to
television in relationship to repetition. These studies found
no changes in visual attention to several repetitions of very
brief television segments (Anderson & Levin, 1976; Wright,
Kotler, Hughes, and Donley, 1997). Otherwise, there has
been no prior research on viewing behavior during repeated
presentations of a television program. Given the likelihood
that repeated experience with a television program extends
and deepens comprehension by young children, the relative
lack of research is surprising. Especially for curriculum-
based television and videotapes, repetition provides a poten-
tially inexpensive means of increasing learning.
Method
Design
One group of children
(3-,
4-, and 5-year-olds) viewed the Blue's
Clues episode once, with comprehension testing immediately
following. A second group viewed the same Blue's Clues episode
on 5 consecutive days, with comprehension testing occurring
immediately following the fifth exposure. A comparison group
watched an episode of a different preschool television program one
time.
The comparison group was included in order to provide an
indication of baseline knowledge of the curriculum items included
in the Blue's Clues episode. The design of the analyses for
comprehension was Condition (3) X Age (3) X Gender
(2),
with all
factors between subjects. The design of the analyses for viewer
behavior focused on the repeated exposure group and is a mixed
632CRAWLEY, ANDERSON, WILDER, WILLIAMS, AND SANTOMERO
design of Age (3) X Gender (2) X Repetition (5) X Content Type
(2),
with Repetition and Content Type within-subjects factors.
Content Type refers to content that is primarily either educational
or entertaining in intent.
Participants
Participants were recruited from three day-care centers in New
York and Connecticut. The 108 participants included 36 three-year-
olds (range was 2 years 11 months to 3 years 11 months; mean age
was 3 years 7 months; 13 girls), 38 four-year-olds (range was 4
years 0 months to 4 years 11 months; mean age was 4 years 6
months; 18 girls), and 34 five-year-olds, (range was 5 years 0
months to 5 years 10 months; mean age was 5 years 3 months; 16
girls).
The broad ethnicities of the participants were
41%
African
American, 36% White,
21%
Hispanic, and 2% Asian American.
The participants were assigned to one of the three experimental
groups based on their age, gender, and ethnicity. For example, 3
children of the same age, gender, and ethnicity were put in three
different groups. Unfortunately, there were not enough participants
to match perfectly by these criteria, so 6 of the children were
assigned randomly, consequently producing unequal numbers of
participants in the Conditions X Age X Gender design. In the
comparison group, there were 12 three-year-olds (4 girls), 12
four-year-olds (6 girls), and 12 five-year-olds (6 girls). In the
one-exposure group, there were 12 three-year-olds (5 girls), 14
four-year-olds (8 girls), and 10 five-year-olds (4 girls). In the
repeated-exposure group, there were
11
three-year-olds (3 girls), 13
four-year-olds (6 girls), and 12 five-year-olds (6 girls).
Of the 52 participants who watched on the 1st day, there were 36
who watched all 5 days and were tested. Fourteen were dropped
because of absence, and 2 were not tested although they did watch
from the 1st to the 4th day and were present on the 5th. In addition,
there were 13 children who were in the comparison or one-
exposure group who completed the session but were not used in the
final sample for reasons of equipment failure or other violation of
the experimental procedure.
Setting
Participants were observed at their day-care centers. One, in
Stamford, Connecticut, provided 60 participants (19 in the repeated-
exposure group, 21 in the one-exposure group, and 20 in the
comparison group). Another, in Queens, New York City, provided
20 participants (7 in the repeated-exposure group, 6 in the
one-exposure group, and 7 in the comparison group). The third,
from Harlem, New York City, provided 28 participants (10 in the
repeated-exposure group, 9 in the one-exposure group, and 9 in the
comparison group). The research was conducted in rooms that were
quiet and away from most distractions. The television was set up
near an electrical outlet, and the video camera was on a tripod next
to the television. The experimenter sat behind or next to the camera,
in order to move it if the child moved out of camera range.
Stimulus Programs
The experimental tape was Blue's Clues, episode 101, "Snack-
time,"
(Santomero, 1996) with a running time of 23.8 min,
excluding closing credits. The series is designed to teach preschool
children cognitive skills and to exercise new and already acquired
cognitive skills and knowledge. A human host, Steve, guides the
child viewers through his animated world, solving problems left by
his puppy Blue. In this episode, Blue gives Steve and the viewers
three clues to try to figure out what Blue wants with her snack. On
the way, Steve and Blue invite viewers to help elephants paint their
family (using color identification), help characters put food of
different shapes away (using shape recognition), and help chicks
find their friends with matching hats (using matching skills). The
series is designed to elicit overt verbal and behavioral reactions
from the preschool audience watching. This is accomplished by
having Steve or other characters ask questions of the audience that
can be answered verbally or through nonverbal behaviors such as
pointing. The series had not yet been telecast at the time of the
study and was therefore new to the participants in this study.
The comparison program was episode 41 of The Busy World of
Richard
Scarry,
(LaPierre, 1996) with a running time of 22.8 min
without closing credits. This episode features the animated resi-
dents of Busytown. This series was chosen as a comparison
because it is popular with preschool children but represents a more
traditional approach insofar as it presents stories without attempts
to elicit overt behaviors from the audience. This episode, too, had
not been telecast at the time of the study.
Procedure
Children were brought to the viewing room individually. The
child was seated on the floor in front of the television, with paper,
crayons, and blocks as distractors within reaching distance. The
experimenter talked to the child for a few minutes and then played
the videotape. Children were not told that they would subsequently
be tested on their knowledge of the program. During the program,
interaction with the child was kept to a minimum. If the child asked
a question or directed a comment to the experimenter, she would
indicate that she was doing work and that the child should go back
to playing or watching television. On the 5th
day,
the children in the
repeated-exposure group were tested after viewing. The children of
the comparison and one-exposure groups were tested after their
first and only viewing session.
The two comprehension tests for Blue's Clues and Busy World
viewers were almost identical. During the test, the participant was
shown pictures and asked questions about them. The test consisted
of five types of
items.
Educational items contained questions about
the educational content, primarily the games in which the audience
was invited to provide the answers (e.g., using the clues cup, straw,
and cow to infer what Blue wanted with her snack). Entertainment
items contained questions about the entertainment content, includ-
ing names of characters and questions such as what happened when
the telephone rang (Steve fell down). The third type we called far
transfer. In these items, the concepts of the thinking games (color
identification, shape recognition, and matching) were tested using
stimuli different from those shown in the episode. The fourth type
examined a strategy that was modeled in the program for solving
the matching problems (placing the standard next to the compari-
son stimuli). How often the child used this strategy was assessed
for the items using the same stimuli shown in the episode, as well as
for far transfer problems using different stimuli. For
example,
in the
matching part of the comprehension test, children matched chicks
with matching hats, as shown in the show. In the far transfer
matching game, children matched kangaroos with matching shoes.
These same items were given to the comparison group to
establish baseline knowledge. There were also questions regarding
the specific content of the episodes. These questions were different
depending on which show the child watched. All children were also
asked to answer the question "What did you just see?"
Immediately after the video, the experimenter moved to the floor
near the participant. After a brief warm-up, the experimenter
administered the comprehension tests without feedback. The
repeated-exposure children watched with the same experimenter
for the first 4 days, but on the 5th day, they watched with and were
REPEATED VIEWING OF BLUE'S CLUES 633
interviewed by one of the other two experimenters, randomly
chosen. This latter procedure was designed to reduce the possible
effect of a familiar researcher eliciting more answers from the
repeated-exposure children.
Videotape Coding
The videotapes that were used to record the children included
standard vertical interval and serial time code, allowing for frame
accurate computer-assisted coding. Looking at the television screen
was coded on a continuous basis. A coder advanced the tape to the
exact frame where she judged the child to be visually oriented to
the television screen. The coder then pressed a button on the
computer and held it until she judged the look
ended.
The computer
recorded the video frame numbers concurrent with the button push
and release. Because of the clarity of these tapes, there was little
difficulty in deciding where the child was looking. However, there
were some parts of the session when the coder was uncertain about
where the child was looking, as when the child moved out of
camera view. Here, the coder pressed a button on the computer at
the exact frame where the uncertainty began, and released it at the
exact frame where uncertainty ended.
Program-related verbal and nonverbal behaviors other than
looking were also coded. The verbal behaviors were answers
directed to the television host or other characters, verbalizations
about the television that were not answers, and other (including
laughing and singing). The coded nonverbal behaviors were
nonverbal direct responses to a query by a television character
(such as nodding and pointing), imitation, and other. There were
two uncertainty codes, one for verbal uncertainty and one for
nonverbal uncertainty. An example of an answer directed to the
television is "Steve, the clue is on the cow!" A verbalization about
the television that is not an answer would be "I have a puppy, too."
Behaviors not related to the program, such as talking in general and
playing, were not coded.
Interobserver reliability for looking at the television screen was
assessed by having each of the five coders code a tape in common.
Phi correlations (based on judgments of looking vs. not looking)
between pairs of coders ranged from .97 to .99, consistent with
corresponding reliabilities in previous research (e.g., Anderson &
Levin, 1976).
Interobserver reliability for interactions with the television was
also assessed by having each of the coders code a tape in common.
Phi correlations were based on judgments of whether or not an
interaction occurred during each second of the show. They ranged
from .76 to .92.
Results
Viewing
Behavior
Combining the single session and first exposure session of
the repeated-exposure Blue's Clues groups, we compared
percentage looking, verbal interactions per minute, and
nonverbal interactions per minute with the same for The
Busy
World
of Richard
Scarry.
The Groups (2) X Age (3) X
Gender (2) analyses of variance (ANOVAs) revealed greater
looking at Blue's Clues, F(l, 96) = 10.09, p < .01 (77.4%
looking vs. 62.2%), and more nonverbal interactions, F(l,
96) =
7.85,
p < .01 (.17 per minute vs. .06). The difference
in verbal interactions (.41 per minute for Blue's Clues vs.
.30) was not significant.
Blue's Clues is designed to elicit active audience participa-
tion in problem solving and other educational activities. The
program, nevertheless, is designed to entertain as well as
educate. Although the entertainment portions of
the
program
invite audience participation, these require less intense
cognitive involvement than do the educational portions of
the program. Accordingly, we parsed the show into constitu-
ent content units on the basis of whether the units were
strongly educational in nature or whether the intent was
primarily entertainment (consensus decisions by two of the
investigators). More of the episode involved educational
content (14.5 minutes or 61%) than entertainment content
(9.3 minutes or 39%). An example of educational content is
a pattern matching game, and an example of entertainment
occurs when Blue, hiding from Steve, is sitting on top of his
head.
The primary analyses of the viewer behaviors were Age
(3) X Gender (2) X Content Type (2) X Repetition (5)
multivariate analyses of variance (MANOVAs) with polyno-
mial trend analysis (up to the quadratic component) of the
repetition factor. In these analyses, scores were estimated for
1 participant whose session was missing due to a lost
videotape. This session was from the fourth exposure of a
3-year-old
boy.
The missing data point was estimated using a
formula advocated by Myers and Well
(1991,
p. 258). All
analyses were repeated deleting this participant; there were
no differences in significance of effects.
Looking. The mean percentage of coder uncertainty for
looking was 0.57%, with a range from 0% to 15.17%. Ah
Age (3) X Gender (2) X Exposure Number (5) mixed-
design ANOVA revealed no significant effects of uncer-
tainty, indicating that it was approximately evenly distrib-
uted across conditions. Periods of uncertainty were not
included in the looking data. If, for example, a child had a
cumulative total of
1
min of uncertainty during a given type
of content, the denominator for calculating percentage
looking was reduced by that amount.
Percentage looking was measured as the number of
frames spent looking at the screen divided by the total
number of frames in the program times 100. An Age (3) X
Gender (2) X Content Type (2) X Repetition Number (5)
mixed-design MANOVA trend analysis yielded a significant
Content Type X Repetitions (linear trend) interaction, F(l,
30) = 11.77,/? <
.01,
which is illustrated in Figure 1. There
was a significant
Age
X Repetition (linear trend) interaction,
F(2,
30) =
6.03,
p <
.01,
which was qualified by an Age X
Gender X Repetition (linear trend) interaction, F(2, 30) =
4.33,
p < .05. That is, the linear trend over repetitions
differed by age and gender.
The interaction of repetition with age and gender was
analyzed by performing Gender X Exposure trend analyses
for each age group, collapsing across content type. There
were no significant effects for the 3- or 4-year-olds; the
linear function of repetition was L
—
72.16% +
1.55/?
for
the 3-year-olds, and L = 84.59% - 2.06/? for the 4-year-
olds,
where L is percentage looking and R is repetition
number. The analysis for the 5-year-olds revealed the source
of the omnibus interactions involving age and gender. There
was a significant main effect of repetitions, F(l, 10) =
20.44,
p < .001, and a marginally significant interaction
634CRAWLEY, ANDERSON, WILDER, WILLIAMS, AND SANTOMERO
100 1
Figure
1. Mean percentage looking as a function of repetition
number
and
content type.
with gender, F(l, 10) =
4.93,
p =
.051.
The 5-year-old boys'
linear function of repetition was L =
88.91%
- 4.88/?,
whereas the 5-year-old girls' function was L = 75.89 -
1.67/?.
Looking declined with repetition among 5-year-olds,
but more so for boys (a loss of nearly five points per
repetition) than girls (a loss of less than two). It should be
noted, however, that boys initially looked more at the
program than did girls.
Interactions directed at the television. The interactions
were coded as verbal or nonverbal in nature. Verbal interac-
tions were further coded as answers, imitations, comments
about the program, and other. Nonverbal interactions were
coded as answers (usually pointing), imitations, and other.
As with looking, there were occasional points at which a
coder was uncertain about the behavior or utterance. When
the child was not visible, nonverbal interactions were not
coded. Verbal interactions were coded in these cases because
the child could still be heard. Nonverbal uncertainties were
coded if the child made a nonverbal motion that was
ambiguous. There was a mean of
0.15
episodes of nonverbal
uncertainties, with a range from 0 to 2. Verbal uncertainties
were coded if the child made a verbalization and the coder
could not understand what was said. There was a mean of
1.22 episodes of verbal uncertainties, with a range from 0 to
10.
Verbal and nonverbal uncertainties are ignored for the
rest of
the
analyses.
The analysis for total interactions (see Table 1) revealed a
significant linear trend of repetitions, F(l, 30) = 4.46, p <
.05,
which varied as a function of content type, F(l, 30) =
8.60, p <
.01,
as illustrated in Figure
2.
Separate analyses for
each content type showed a significant linear trend for the
educational content, F(l, 30) = 5.96, p < .05, / = .38 +
.17/?,
where / is interactions per minute. The trend for
entertainment content was not significant, / = .47 + .08/?.
We further analyzed the interactions according to whether
they were verbal or nonverbal.
For verbal interactions per minute, there was a significant
linear trend for repetitions, F(l, 30) = 4.52, p < .05, which
varied as a function of content type, F(l, 30) = 7.75, p <
.01.
Separate analyses for each content type revealed a
significant linear trend for the educational units, F(l, 30) =
5.85, p < .05, / = .24 + .14/?, but not for the entertainment
units,/= .31 + .07/?.
Verbal interactions were coded as answers, imitations, and
talking about the program. Because some cells contained
very few responses, the data for each type of verbal
interaction were analyzed collapsed across gender of partici-
pant. The analyses were thus Age (3) X Content Type (2) X
Repetitions (5) MANOVAs with trend analysis. Analyses
indicated that talking about the program did not increase
with repetition whereas imitations did, F(l, 33) =
6.81,
p <
.05,
/= -.02 + .06/?. Of more educational importance with
respect to the educational goals of the program were verbal
answers. There was a significant interaction (linear trend) of
repetitions with content type, F(l, 33) = 5.75, p < .05.
Verbal answers increased with repetition, F(l, 33) = 7.95,
p <
.01,
but more so for educational, F(l, 33) = 7.16, p <
.05,
/ = .13 + .07/?, than for entertainment content, F(l,
33) =
5.98,
p <
.05,
/ = .02 + .014/?.
Table 1
Total Interactions per Minute as a Function of Age, Content Type, and Repetition Number
Age
3
4
5Total
3
4
5Total
1
0.63 (1.00)
0.51 (1.02)
0.40 (0.80)
0.51 (0.92)
0.96(1.19)
0.57 (0.51)
0.33 (0.55)
0.61 (0.81)
Repetition number
2 3
Educational content
0.60 (0.92) 0.79 (0.83)
0.77(1.33)
1.18(1.93)
0.70(0.95)
1.13(1.33)
0.69(1.07)
1.04(1.43)
Entertainment content
0.65 (0.66) 0.58 (0.52)
0.73 (0.83) 0.88 (0.84)
0.29 (0.42) 0.59 (0.66)
0.56 (0.67) 0.69 (0.69)
4
0.78 (0.86)
1.61 (2.11)
0.71 (0.83)
1.05 (1.46)
0.70 (0.50)
1.14(1.09)
0.50 (0.60)
0.79(0.81)
5
0.79 (0.91)
1.44(2.47)
1.29(2.09)
1.19(1.94)
0.74 (0.61)
0.99 (1.36)
0.91 (1.35)
0.89 (1.15)
Note. The values represent the mean interactions per minute with the standard deviations in
parentheses. There were significant linear trends for both educational and entertainment content but
no effects of age.
REPEATED VIEWING OF BLUE'S CLUES 635
1.5 -,
-O-
Educational
-•- Entertainment
2 3 4 5
Exposure Number
Figure
2.
Mean interactions
per minute as a
function of repetition
number and content type.
Verbalizations about the television program varied only
by type of content, F(l, 33) =
10.33,
p < .01. The means
and standard deviations were .11 (.27) for the educational
content and .22 (.31) for the entertainment content.
For total nonverbal interactions per minute, a linear trend
of repetitions varied significantly by content type, F(l,
30) = 4.87, p < .05, which was further complicated by an
interaction of linear trend of repetitions with gender and
content type, F(l, 30) = 6.57,p < .05. This interaction was
due to a linear increase in girls' nonverbal interactions with
educational content, / = .10 + .04/?, whereas there was no
such increase for entertainment content, / = .15 - .003/?.
Boys showed no significant increase in nonverbal interac-
tions for either type of content, / = .16 + .02/?, for both
educational and entertainment content.
Nonverbal interactions were coded as answers, imitations,
and other (such as dancing to the music). Because there was
substantially less nonverbal interaction than verbal interac-
tion and because the overall analysis of nonverbal interac-
tions indicated that the interactions varied by gender of
viewer, we concluded that there were too few data for
meaningful analysis.
Comprehension
The comprehension test yielded five separate scores.
Summing the scores on the five types of items produced a
total score. The highest possible total score was 49 (22
points were possible for educational, 10 for entertainment,
11 for far transfer, and 3 points each for strategy use in the
two matching games). Free recall to the question "What did
you just see?" was also coded for how many units of
information were provided by the child. This was scored
only for
the
Blue's
Clues
viewers and was analyzed separately.
Mean percentages of the total possible scores (with
standard deviations) for each of the five question types may
be seen in Table 2. An Age (3) X Gender (2) X Conditions
(3) X Question Type (5) MANOVA, with question type as a
repeated measure, was used to analyze percentage of correct
answers. There was a main effect of age, F(2, 90) = 21.56,
p < .001, due to better comprehension with age. A main
effect of condition, F(2, 90) = 16.16, p < .001, was due to
better performance following five repetitions of Blue's Clues
than after one repetition, which in turn was better than
performance by the group that saw a different program.
There was a main effect of question type, F(4, 360) =
168.53,
p < .001, with far transfer strategy use being the
most difficult and far transfer items being easiest (see Table
2).
There was also an Age X Item Type interaction, F(8,
360) = 4.14, p < .001, and a Conditions X Item Type
interaction, F(8, 360) = 3.48, p < .01. To determine the
source of these interactions, we ran separate between-
subject ANOVAs, Age (3) X Gender (2) X Conditions (3),
on each comprehension category.
For educational items, there were main effects of
age,
F(2,
90) = 26.10,/? <
.001,
and condition, F(2,90) = 27.66,/? <
.001,
and for entertainment items, main effects of age, F(2,
90) =
22.73,
p <
.001,
and condition, F(2,90) = 25.53,/? <
.001.
These effects were due to better performance with age,
as well as better performance following five repetitions than
one repetition, and both were superior to the group that saw a
different program.
The analysis for far transfer (same tasks but different
problems than in the episode) yielded only a main effect of
age,
F(2, 90) = 25.96, p < .001, due to better performance
with age. There were main effects of condition for strategy
use (due to greater use of the demonstrated strategy by the
repeated Blue's Clues viewers) in both the original matching
game, F(2, 90) = 4.11, p < .05, and the far transfer
matching game, F(2, 90) = 4.11, p < .05. There was low
performance in some of the cells used for the analysis of
strategy use, leading to the problem of artificially con-
strained variability. Therefore, the Kruskal-Wallis one-way
ANOVA, a nonparametric test, was also used to test the
effects of condition on strategy use. Again, there was a
significant effect of condition on both the original matching
game, x2(2, N= 108) = 9.7, p <
.01,
and for the far transfer
matching game, x2(2, N = 108) = 10.2,/? < .01.
An Age (3) X Gender (2) X Condition (2) ANOVA on
free recall by the two Blue's Clues groups revealed only a
significant age effect, F(2, 60) =
11.71,
p < .001. The
means and standard deviations of the number of recalled
units were .59 (.73), 1.85 (2.13), and 3.17 (2.33) for the 3-,
4-,
and 5-year-olds.
Discussion
Although preschooler demand for repetition of books and
videos is widely known among parents and other caretakers
of young children (Mares, 1998), the phenomenon has been
the subject of remarkably little formal research. We exam-
ined the effects on 3- to 5-year-old viewers of repeating an
educational television program, Blue's Clues, for
5
consecu-
tive days.
Blue's Clues is designed to provide cognitive challenges
and teach cognitive skills to preschool viewers while
636CRAWLEY, ANDERSON, WILDER, WILLIAMS, AND SANTOMERO
Table 2
Percentage Comprehension
Age&
question type
by Question
C
Type,Age,
lx
and Condition
Group
5xTotal
3-year-olds
Far transfer matching
Original matching
Entertainment
Educational
Far transfer
Total comprehension
4-year-olds
Far transfer matching
Original matching
Entertainment
Educational
Far transfer
Total comprehension
5-year-olds
Far transfer matching
Original matching
Entertainment
Educational
Far transfer
Total comprehension
All participants
Far transfer matching
Original matching
Entertainment
Educational
Far transfer
Total comprehension
2.78 (9.62)
11.11 (29.59)
10.83 (12.40)
31.82(15.14)
46.97 (30.73)
27.89 (12.98)
11.11(21.71)
0.00 (0.00)
31.67 (14.67)
46.21 (9.85)
71.97 (23.72)
44.05 (8.98)
8.33 (20.72)
5.56(12.97)
28.33 (12.67)
48.86 (12.58)
81.06(16.65)
46.77 (9.31)
7.41 (18.01)
5.56(18.69)
23.61 (15.88)
42.30 (14.49)
66.67 (27.83)
39.57 (13.30)
0.00 (0.00)
15.15 (22.92)
20.91 (16.40)
41.74(17.58)
36.36 (22.64)
32.10(13.78)
9.52 (20.37)
11.90(30.96)
45.00(22.10)
63.96 (18.48)
77.27 (18.78)
56.56 (15.67)
0.00 (0.00)
9.09(30.15)
55.45 (18.09)
75.21 (10.63)
84.30 (17.76)
64.56 (8.07)
3.70(13.28)
12.04 (27.78)
40.83 (23.47)
60.61 (20.76)
66.92 (28.26)
51.53 (18.61)
15.15(22.92)
15.15(31.14)
37.27 (21.02)
57.44 (13.68)
56.20 (24.33)
47.87 (13.08)
25.64 (38.86)
35.90 (44.02)
54.62(16.13)
70.98 (15.80)
76.92(18.78)
64.05(11.41)
22.22 (35.77)
30.56 (38.82)
67.50 (15.45)
79.55 (8.99)
81.82 (13.98)
71.09(5.49)
21.30(33.00)
27.78 (38.63)
53.61 (21.00)
69.70 (15.67)
72.22 (21.73)
61.45 (13.98)
5.88 (15.29)
13.73 (27.36)
22.65 (19.74)
43.32 (18.50)
46.52 (26.74)
35.71 (15.54)
15.38 (28.46)
16.24 (34.09)
44.10(19.96)
60.84(18.17)
75.52 (20.02)
55.21 (14.67)
10.48 (25.27)
15.23 (30.62)
50.29 (22.56)
67.66 (17.42)
82.34 (15.74)
60.70 (12.98)
10.80 (24.04)
15.12(30.70)
39.35 (23.65)
57.53 (20.52)
68.60 (25.99)
50.85 (17.77)
Note. The values represent the mean percentage correct with the standard deviations in parentheses.
C = Busy
World
of Richard Scarry comparison group; lx = Blue's Clues one-exposure group; 5x =
Blue's Clues repeated-exposure group.
entertaining them. More than any prior television program,
Blue's Clues encourages audience participation. Based on
ideas that learning would improve and the audience would
feel empowered by meeting the program's cognitive chal-
lenges, the Nickelodeon network adopted a novel broadcast
strategy of showing an identical episode on 5 consecutive
days each week. This strategy seems well founded. We
found that, with repetition, visual attention remained con-
stant, with the exception of some decline among 5-year-old
boys.
Verbal and nonverbal interactions, especially answers
and imitations, greatly increased with episode repetition.
Comprehension improved, and children increased their
application of a demonstrated problem-solving strategy to
problems both shown and not shown by the program.
With respect to theories of television viewing, the results
are least consistent with Singer's (1980) orienting response
theory that predicted attention should decline with repetition
at all ages. The Anderson and Lorch (1983) and Huston and
Wright
(1983,
1989) theories, which emphasize the impor-
tance of the viewers' engagement with the content, provided
predictions that are consistent with observed results, al-
though not in detail. The results are also broadly consistent
with Salomon's (1983) theory of invested mental effort.
Because Blue's Clues overtly invites viewer participation
and active problem solving, we hypothesized that invested
mental effort would increase as the children became familiar
with the program's formats and demands. Consequences of
this increased mental effort would be increased overt viewer
participation, maintained visual attention, and improved
comprehension, all of which were found.
Practically speaking, repetition has not hurt and may have
helped audience ratings for Blue's Clues. Since the begin-
ning of the second season (October 1997), Blue's Clues has
attracted more preschool viewers than any other preschool
show on television (Nielsen Media Research, based on K2-5
weekly cume AA [000]). In both educational and audience
terms,
therefore, the repetition strategy appears to be a
success.
There are several important empirical questions that are
left unanswered by the present research. Because Blue's
Clues is unusual both in its cognitive challenge and in its
ability to elicit audience participation, we do not know what
the effects of episode repetition might be for more traditional
television programs. Based on Salomon's (1983) theory of
invested mental effort, furthermore, traditional educational
television programs that do not invite active participation
may not be perceived by young viewers as requiring
sustained mental effort. These programs may not bear
repeating to the same extent that Blue's Clues does.
We also do not know if the beneficial effects of repetition
are limited to the ages examined in the present study.
Repeated viewing of videos at home declines among school-
REPEATED VIEWING OF BLUE'S CLUES 637
age children (Mares, 1998), and repetition of an educational
television program for older children may not have the
desired effects.
Theories of child development and cognitive theories of
learning have remarkably little to say about the effects of
repetition other than that comprehension and memory should
benefit. We suggest that any such theory must incorporate
notions of optimal comprehensibility, active participation,
and a need for cognitive mastery. Regardless of the form
such a theory may take, the present research indicates that
repetition can be a useful tool to increase enjoyment and
learning from media for young children.
References
Anderson, D. R., & Levin, S. R. (1976). Young children's attention
to "Sesame Street." Child
Development,
47,
806-811.
Anderson, D. R., & Lorch, E. P. (1983). Looking at television:
Action or reaction? In J. Bryant & D. R. Anderson (Eds.),
Children's understanding of television: Research on attention
and comprehension (pp.
1-33).
New York: Academic Press.
Beaver, J. M. (1982). Say it! Over and over. Language Arts, 59,
143-148.
Field, D. E., & Anderson, D. R. (1985). Instruction and modality
effects on children's television attention and comprehension.
Journal of Educational
Psychology,
77, 91-100.
Huston, A. C, & Wright, J. C. (1983). Children's processing of
television: The informative functions of formal features. In J.
Bryant & D. R. Anderson (Eds.), Children's understanding of
television: Research on attention and comprehension (pp. 35-
68).
New York: Academic Press.
Huston, A. C, & Wright, J. C. (1989). The forms of television and
the child viewer. In G. Comstock (Ed.), Public communication
and behavior
(pp.
103-158). Orlando, FL: Academic Press.
LaPierre, T. (1996). A trip back in time (M. Da Silva, Director). In
M. Charest (producer), The busy world of Richard
Scarry.
New
York: Nickelodeon.
Mares, M. L. (1997, May). Re-examining children's comprehension
of television: The potential effects of the VCR. Paper presented at
the annual conference of the International Communication
Association, Montreal, Quebec, Canada.
Mares, M. L. (1998). Children's use of VCRs. Annals of the
American Academy of Political and Social Science, 557, 120-
131.
Martinez, M. (1983). Exploring young children's comprehension
through storytime talk. Language
Arts,
60, 202-209.
Martinez, M., & Roser, N. (1985). Read it again: The value of
repeated readings during storytime. Reading
Teacher,
38, 780-
786.
Morrow, L. M. (1988). Young children's responses to one-to-one
story readings in school settings. Reading Research Quarterly
23,
89-107.
Myers, J. L., & Well, A. D. (1991). Research design and statistical
analysis. New York: HarperCollins.
Peracchio, L. A. (1992). How do young children learn to be
consumers? A script processing approach. Journal of Consumer
Research, 18, 425^140.
Salomon, G. (1983). Television watching and mental effort: A
social psychological view. In J. Bryant & D. R. Anderson (Eds.),
Children's understanding of television: Research on attention
and comprehension (pp. 181-198). New York: Academic Press.
Salomon, G., & Leigh, T. (1984). Predispositions about learning
from print and television. Journal of Communication, 34,
119-135.
Samuels, S. J. (1979). The method of repeated readings. Reading
Teacher,
32, 403-408.
Santomero, A. C. (1996). Blue's snacktime (T. Kessler, Director).
In T. Kessler (Producer), Blue's Clues. New York: Nickelodeon.
Sell, M. A., Ray, G. E., & Lovelace, L. (1995). Preschool children's
comprehension of a Sesame Street video tape: The effects of
repeated viewing and previewing instructions. Educational
Technology
Research and
Development,
43(3), 49-60.
Singer, J. L. (1980). The power and limitations of television: A
cognitive-affective analysis. In P. H. Tannenbaum (Ed.), The
entertainment functions of television (pp. 31-65). Hillsdale, NJ:
Erlbaum.
Snow, C. E. (1983). Literacy and language: Relationships during
the preschool years. Harvard Educational
Review,
53, 165-189.
Snow, C. E., & Goldfield, B. A. (1983). Turn the page please:
Situation-specific language acquisition. Journal of Child Lan-
guage, 10, 551-569.
Sulzby, E. (1985). Children's emergent reading of favorite story-
books: A developmental study. Reading Research
Quarterly,
20,
458^*81.
van der Molen, J. H., & van der Voort, T. H. (1997). Children's
recall of television and print news: A media comparison study.
Journal of Educational
Psychology,
89,
82-91.
Wright, J. C, Kotler, J. A., Hughes, K. P., & Donley, S. M. (1997,
April).
Brief educational segments on PBS' Ready-to-Learn
series: Effects of program relevance and repetition. Poster
session presented at the biennial meeting of the Society for
Research in Child Development, Washington, DC.
Received July 23, 1998
Revision received January
21,
1999
Accepted January
21,
1999
