The Relationship between Reading and
Cognitive Proeessing of Television and Radio
Kathy Pezdek, Ariella Lehrer, and Sara Simon
Clarem,ont Graduate School
and SIMON, SARA, The Relationship between Reading and Cognitive Processing of Television and Radio. GHILD DEVELOPMENT, 1984, 55, 2072-2082, This study compares the relationship between comprehension and memory for text and television, and comprehension and memory for text and radio, 2 stories were edited to create 3 matched versions of each—-(1) a
storybook with pictures, (2) a radio version, and (3) a television version. Third and sixth graders read
1 story and were presented either the television or radio version ofthe other story, counterbalanced
for story and order of presentation, A battery of memory and comprehension tests was administered
following both the reading and the television or radio conditions. The principal result is that across a
range of tasks, performance in the television and reading conditions was not significantly correlated;
however, performance in the radio and reading conditions was positively correlated and generally
significant. The second important finding is that the absolute levels of performance were similar in
the television and reading conditions, and performance was better in both of these conditions than in
the radio condition. Hypotheses are discussed to explain why performance in the television condition was not related to reading ability.
PEZDEK, KATHY; LEHRER, ARIELLA;
This study compares comprehension and
memory for infonnation presented in various
forms of media. More specifically, this study
compares the relationship between comprehension and memory for material read
with comprehension and memory for material
presented on television and on radio. Although there has been a great deal of interest
in recent years in the effects of television and
other forms of media on the viewer (see Murray, 1980, for a review of the television literature containing 2,886 citations, 60% of which
were published after 1975), little of this work
has examined the cognitive processes involved in media processing.
The majority of the previous research on
the relationship between reading and media
processing has compared reading ability with
the amount of time spent viewing television,
(There is little previous research on radio processing.) These correlational studies have reported that the amount of television viewing
is negatively correlated with both reading
ability and more general school achievement
in junior high—age subjects (Homik, 1978;
Morgan, 1980; Ridder, 1963). Similarly, Nelson (1973) reported significant negative correlations between the amount of television
viewed and several indices of early language
development in children as young as 18
months old. It is difficult, however, to make
causal inferences from these results. For example, the contributing effects of IQ and SES
have been largely unexamined. In one study
that did covary out the effects of demographic
background, IQ, and parents' education, no
relationship was found between reading ability and television viewing habits in third to
fifth graders (Zuckermann, Singer, & Singer,
1980).
The present study compares comprehension and memory for material read with comprehension and memory for material presented on television and on radio. In this
study, two folk tales were edited such that
there were three matched versions of each—
(1) a storybook with pictures; (2) an auditory-only, "radio" version; and (3) an audi-
This research was supported by the National Institute of Education (NIE G-81-0049) and the
Haynes Foundation. Portions of these findings were presented at the meeting of the Society for
Research in Ghild Development, Detroit, 1983, and the meeting of the Psychonomic Society, San
Diego, 1983, We are very grateful to Patricia Greenfield for the use of her materials. We acknowledge
the assistance of Mark Runco with the data analysis and Allan Ribbler for data collection. We also
appreciate the cooperation of Jack Hassinger, Robert Vislay, and the students and staff at Kingsley,
Lehigh, Margarita, Monte Vista, and Moreno Elementary Schools in Ontario-Montclair, Galifomia,
Finally, we thank Dale Berger, Ed Teyber, Jessica Beagles-Roos, Kelly Godfrey, Sidney Fox, and
Bob Lunn for tiieir critical reading of this article ^nd Jane Gray for preparing the manuscript. The
address of the first author is: Department of Psychology, Glaremont Graduate School, Glaremont,
Galifomia 91711.
'
[Child Development, 1984,55,2072-2082, © 1984 by the Society for Research in Child Development, Inc,
All dghts reserved, 0009-3920/84/5506-0015$01,00]
Pezdek, Lehrer, and Simon
2 0 73
sistent with those of Beagles-Roos and Gat
(1983).
Second, a number of studies in the reading literature have compared memory for information read by subjects with memory for
information read to subjects (an approximation of a radio condition). Mixed findings have
resulted. Smiley, Oakley, Worthen, Campione, and Brown (1977) tested seventh graders and reported that recall of story propositions as a function of the structural importance
of these units was similar in these two conditions. King (1968) visually or aurally presented stories to college-age subjects. Although the conditions did not differ in terms
of the recall of factual details, subjects recalled more gist statements when they read
than when they listened. On the other hand,
Horowitz and Berkowitz (1967) presented college-age subjects a story in a reading or listening condition. The listeners recalled more
ideas, omitted fewer important ideas, and
more accurately recalled stylistic features of
The present study also avails a compari- the story than the readers. Taken together,
son of the absolute levels of performance in few consistent conclusions follow from the litthe television, radio, and reading conditions. erature comparing absolute levels of perforThis comparison offers an answer to the ques- mance among media conditions. This is
tion, from which medium do children at dif- primarily because of methodological differferent ages comprehend and retain more— ences across studies, and because text, televitelevision, radio, or text? Although a number sion, and radio have not been compared
of researchers have addressed portions of this within one study.
question, these three media conditions have
The primary purpose of the present study
not previously been compared on similar mea- is to compare the relationship between comsures within a single study.
prehension and memory for television and
First, several studies have compared tele- text, and comprehension and memory for
vision and radio processing. Barrow and West- radio and text. The design of this study also
ley (1959), using a recall test with elementary entails a comparison of the absolute levels of
school children, and Wetstone and Friedlan- performance among the television, radio, and
der (1974), using a multiple-choice com- text conditions.
prehension test with children in kindergarten
through third grade, reported better perfor- Method
mance in an auditory-visual than auditorySubjects and design.—The subjects were
only condition. Beagles-Roos and Gat (1983) 96 children recruited from racially mixed
compared television and radio presentation public schools in middle-class neighborhoods
conditions and reported that recall of explicit in Ontario-Montclair, California. There were
story content did not differ between these 48 third graders (M = 8.9 years) and 48 sixth
conditions. Children age 6-10 years were graders (M = 11.8 years). Equal numbers of
tested in this study. However, in the televi- males and females participated in each condision condition, children recalled more story tion. The children were brought to the univerdetails and were more likely to draw infer- sity campus by a parent and were paid $10.00
ences based on story actions. In the radio con- to participate.
dition, children more accurately recognized
The primary independent variables in
expressive language from the story and were
the
experiment
were age and media condimore likely to draw inferences based on world
knowledge not presented in the story. Simi- tion. Half of the subjects participated in a
larly, Meringoff (1980) compared 6-10-year- reading and television condition, and half parold children's memory for stories presented ticipated in a reading and radio condition.
on television with their memory for stories
Materials.—Two stories were utilized in
read to them. The results were generally con- this study to enhance the generalizability of
tory-visual, "television" version. The auditory
track in the television and radio conditions
was the same as the transcript of the book.
Third graders (beginning readers) and sixth
graders (more advanced readers) read one
story and were presented either the radio or
the television version of the other story. Five
tasks were selected that assessed diverse aspects of the complex comprehension and
memory construct. These five tasks were (1)
comprehension questions that involved
probed recall, (2) sentence recognition, (3)
picture recognition, (4) sentence reordering,
and (5) picture reordering. The battery of tests
was adminstered following both the reading
condition and the radio or television condition. This allows within-subject comparisons
of performance on each task, following a reading condition and a television or radio condition. The relationship between comprehension and memory for television and text, and
also comprehension and memory for radio and
text, can thus be compared.
2074
Child Development
the findings. These stories were "Strega
Nona" and "A Story, A Story," both folk tales.
The stories were novel to all subjects. The
original materials were produced by Weston
Woods. These materials were edited such that
there were three matched versions of each
story: (1) a storybook with colored pictures, (2)
a radio version, and (3) a television version in
color. These materials were the same as those
used by Beagles-Roos and Cat (1983) and
originally by Creenfield, Ceber, BeaglesRoos, Farrar, and Cat (Note 1). The audio
tracks in the television and radio conditions
were identical, and these were the same as
the transcript of the book. The same material
was thus included in each condition, with the
addition of still pictures in the storybook version and animation in the television version.
Each subject read one story aloud and was
presented either the radio or television version ofthe other story, with the assignment of
story to condition counterbalanced across
subjects. Children read the stories aloud to
ensure that they read the complete story and
decoded words correctly. The television and
radio productions of "Strega Nona" lasted 7
min, 44 sec; the productions of "A Story, A
Story" lasted 8 min, 29 sec.
In addition to the narrative materials,
each child was presented an expository segment. Because narrative materials are known
to have a unique structure (see Mandler &
Johnson, 1977), the present study included expository material to test if the effects are consistent with different materials. Also, normal
television and radio programming typically
includes both types of segments. The expository segment utilized was an edited version of
"3,2,1 = Contact" (8 min, 38 sec in length), a
production of Children's Television Workshop. In this segment, information about plant
grov^ili, mental hygiene, and the color of skin
was presented. All children were presented
this same expository segment. However, children in the radio condition were presented
only the auditory track from the original segment. Children in the television condition
were presented the intact television version of
the program. The expository segment was selected such that it was entirely comprehensible from the auditory track alone, and all test
questions that followed could be answered
from information presented on the auditory
track alone.
To summarize the above, half of the subjects read one story and were presented the
radio version of the other story and tiie radio
version of the expository segment. Half of the
subjects read one story and were presented
the television version ofthe other story and
the television version of tHe expository segment. The order of presenting the (1) story to
be read, (2) media story, and (3) media expository segment was randomly arranged across
subjects.
Dependent measures.—Following each
story presentation, a battery of five tests was
administered. This included (1) a set of comprehension questions, (2) a sentence recognition test, (3) a sentence reordering test, (4) a
picture recognition test, and (5) a picture reordering test. The comprehension test was always presented first. The order of presenting
the two sentence tests and the two picture
tests was counterbalanced across subjects, but
for each set, the recognition test always preceded the reordering test.
Sixteen comprehension questions were
generated from each story. Four types of questions were arrived at by probing information
that was (1) central or peripheral to the plot of
the story, and (2) explicitly or implicitly presented in the story. Four questions of^ each
type were generated for each story. Three experimenters verified the fit of each question to
one of these four categories of comprehension
questions. The comprehension accuracy was
analyzed for each of the four types of questions separately, as well as across all 16 questions. Each answer was scored on the following three-point scale: two points if the child
answered correctly, one point if a verbal
prompt from the experimenter was necessary
before the child gave the correct answer, and
no points if the child did not give the correct
answer even with the prompt. Prompts were
essentially restatements of the original questions but with an additional piece of information given. If the child did not answer correctly when prompted, the experimenter gave
the correct response before proceeding. The
acceptable "correct answer" for each question
was specified in advance, making the scoring
of answers quite clear-cut.
In the sentence recognition test, subjects
were presented 12 sentences, one at a time,
on cards. Their task was to i-ead each sentence aloud and then indicate whether the
sentence had been in the story. Half of the
sentences were from the story and half were
distractor sentences. The distractor sentences
were constructed to be syntactically similar to
sentences in the story, but based on details
semantically inconsistent with the story. The
recognition data were scored in terms of the
signal detection measure of d'. The values of
d' reflect subjects' recognition sensitivity in
distinguishing original items from distractor
Pezdek, Lehrer, and Simon
items (see Banks, 1970, for an explanation of
signal detection theory). The procedure outlined by Hochhaus (1972) was followed for
calculating d' values.
2075
Procedure.—Each child was tested individually in a comfortably furnished room. The
parent and child were seated together in the
testing room while the study was explained.
The parent left the child alone with the experFollowing the sentence recognition test, imenter once the procedure began. The child
the experimenter selected the six original sen- was instructed that he or she would (1) read a
tences, shuffled them, and instructed subjects story to the experimenter, (2) receive another
to place them in the order they were pre- story (either on the radio or on television), and
sented in the story from beginning to end. (3) receive a lesson (either on the radio or on
The reordering task was scored on the basis of television). The order of these tasks was ranthe number of displacement errors, that is, domly arranged. Immediately following each
how far each item was placed from its correct task, subjects were presented the appropriate
position in the series. Consider the example memory and comprehension tests. The televiwhere the items were ordered 1, 3, 4, 2, 5, 6. sion segments were presented on a 19-inch
Items 1, 5, and 6 are in their correct positions. color television. The "radio" segments were
Items 3 and 4 are each displaced by one space presented on a portable cassette tapefrom their correct positions, and item 2 is dis- recorder.
placed by two spaces. The total displacement
error in tiiis example is thus four.
Results
The picture recognition and reordering
The presentation of the results is orgatests were structured similarly to the sentence
recognition and reordering tests. Pictures nized around three issues: (1) comparisons of
from the storybook were placed individually the absolute levels of performance in the teleon cards. The artist for the storybook pictures vision, radio, and reading conditions; (2) comalso did the visuals for the television version, parisons of the correlations between perforso the basic pictures and style were the same. mance in the reading condition and the
In the radio condition, the picture recognition television and radio conditions; and (3) results
of the multiple regression analyses using
and picture reordering tests served as inferen- questionnaire
data to predict radio and televitial tests rather than memory tests, since these sion comprehension.
subjects had not previously seen the pictures.
Comparisons of performance in the teleFollowing the expository segment, subvision,
radio, and reading conditions.—The
jects were presented a series of nine comprehension questions. Answers to these ques- mean performance on each dependent varitions were scored on the same 0-2 scale able in the television and radio narrative conditions and in the reading condition for third
indicated above.
and sixth graders is presented in Table 1. The
Several additional measures were col- significance level for each comparison is p <
lected to use as predictor variables. First, each .05. First, a 2 (television vs. radio) x 2 (third
child's reading achievement was recorded us- vs. sixth graders) analysis of variance was pering the standardized reading test that was ad- formed on each of three dependent variables
ministered and scored in the schools. This test separately. The picture recognition and picwas the Cooperative Primary Test for third ture displacement measures were not ingraders and the Test of Basic Skills for sixth cluded in these analyses because subjects in
graders. In addition, a questionnaire was com- the radio condition were not presented the
pleted by parents. The questionnaire probed original visual materials. Performance on the
the following information: parents' education, picture recognition and displacement meafamily income, mean amount of time per day sures in the radio condition are presented
that die child watches television (calculated only for consideration as baseline data. On the
from 1 week television viewing diaries), mean comprehension test questions, performance
amount of time per day that the attending par- was better in the television than radio condient watches television (calculated from 1 tion, F(l,92) = 13.33, MSe = 0.15; sixth gradweek television viewing diaries), mean ers performed better than third graders,
amount of time per day that the child reads F(l,92) = 11.12; and these factors did not sig(exclusive of school homework), mean num- nificantly interact. On the sentence recogniber of books per month that the child reads, tion test, scored in terms of the signal detecmean time per day that the attending parent tion measure of d', again, performance was
reads to him/herself, mean number of books better in the television than radio condition,
per month that the attending parent reads to F(l,92) = 20.24, MSe = 1-61; sixth graders
him/herself.
performed better than third graders, F(l,92) =
2076
Child Development
TABLE 1
MEAN PERFORMANCE AND VARIANCE ON EACH DEPENDENT VARIABLE IN THE
READING AND TELEVISION OR RADIO CONDITIONS FOR THIRD AND SIXTH
GRADERS
TELEVISION GROUP
Third graders:
Comprehension^
Sentence d'
Sentence displacement..
Picture d'
Picture displacement....
Sixth graders:
Comprehension**
Sentence d'
Sentence displacement..
Picture d'
Picture displacement....
RADIO CROUP
Television
Reading
Radio
Reading
1.40
(.09)
2.33
(1.33)
2.79
(15.04)
4.38
(1.09)
.92
(1.73)
1.28
(.11)
2.77
(2.59)
2.58
(12.34)
4.06
(1.70)
3.79
(23.48)
1.03
(.18)
1.16
(1.26)
4.79
(23.04)
2.44^
(2.30)
8.00^
(24.44)
1.22
(.17)
2.84
(1.79)
4.17
(21.19)
1.58
(.12)
3.30
(1.88)
2.25
(13.33)
3.89
(1.64)
1.42
(6.78)
1.50
(.07)
3.07
(3.19)
1.96
(8.04)
4.39
(1.69)
2.21
(11.65)
1.38
(.20)
2.45
(1.99)
2.50
(18.17)
2.49*^
(2.00)
5.67^
(28.41)
1.44
(.16)
3.27
(1.82)
1.17
(8.32)
3.69
(2.03)
4.58
(21.04)
3.75
(2.30)
2.33
(9.28)
^ Mean score per question; maximum = 2.0.
^ Because subjects in the radio condition were never presented the pictures, performance on the picture recognition and displacement measures in this condition are considered as baseline data.
13.66; and these factors did not significantly
interact. No effects were significant on the
sentence displacement measure.
The next set of analyses involves withinsubject comparisons of performance in the
television versus reading conditions. Half of
the children are included in this analysis. Separate 2 (television vs. reading) x 2 (third vs.
sixth graders) analyses of variance were performed on each of the five dependent variables. Few effects were significant in these
analyses. Only on the picture displacement
measure was Üiere a media effect; the picture
displacement error was greater in the reading
than television condition, F(l,46) = 8.39, MSe
= 12.20. The only test on which there was a
significant effect of grade was the comprehension test; sixth graders were more accurate
than third graders, F(l,46) = 8.62, MSe =
0.11. Crade and media did not significantly
interact on any measure.
In the next set of analyses, within-subject
comparisons of performance in the radio and
reading conditions were analyzed. The other
half of the children are included in this analy-
sis. Separate 2 (radio vs. reading) X 2 (third
vs. sixth graders) analyses of variance were
performed on the three relevant dependent
variables. On the comprehension test questions, subjects were more accurate in the reading than radio condition, F(l,46) = 4.03, MSe
= 0.09 (this effect is marginally significant, .05
< p < .06); sixth graders were more accurate
than third graders, F(l,46) = 6.94, MSe =
0.27; and these factors did not interact. Similarly, on the sentence recognition test, subjects were more accurate in the reading than
radio condition, F(l,46) = 11.63, MSe = 3.21;
sixth graders were more accurate than third
graders, F(l,46) = 4.64, MSe = 3.81; and
these factors did not interact. On the sentence
displacement measure, the error rate was
greater for third graders than for sixth graders,
F(l,46) = 7.68, MSe = 21.87, and no other
effects were significant.
The above comparisons of the absolute
levels of performance in the television, radio,
and reading conditions allow some general
conclusions. First, between-groups comparisons of performance in the television and
Pezdek, Lehrer, and Simon
2077
radio conditions yield a superiority of televi- in the television and reading conditions was
sion over radio on the comprehension and picture displacement, and the correlation was
sentence recognition tests. It is important to significantly negative for third graders (r =
note that the between-groups comparisons of — .41) and significantly positive for sixth gradperformance in the reading conditions did not ers (r = .58). On the other hand, as can be
significantly differ for any measure. (These seen in Table 2, there was a more consistent
data are presented in the second and fourth and positive relationship between perforcolumns in Table 1.) Thus, differences be- manee in the radio and reading conditions
tween the results in the television and radio across all measures. The correlations between
conditions are not likely a result of differences performance in the radio and reading condibetween the two groups of subjects. Second, tions were positive and significant for third
within-subjects comparisons of performance graders (r = .54) and sixth graders (r = .48) on
in the television and reading conditions re- the comprehension measure and for third
veal similar levels of performance on most graders on the sentence recognition measure
measures. Third, within-subjects comparisons (r = .60).
of performance in the radio and reading conditions reveal a superiority of the reading
The magnitude of the differences in corcondition over the radio condition on the relations between the television and reading
comprehension and sentence recognition condition as compared with the radio and
measures. Thus, across a broad range of tasks, reading condition was next tested for each
subjects generally performed at similar levels measure and each grade separately. Each of
in the television and reading conditions, and the six pairs of corresponding correlations in
better in both of these conditions than in the the top three rows of Table 2 was compared
radio condition.
using z tests for the difference between two
independent correlations. None of these difCorrelations between media processing ferences was significant. However, these comand reading tasks.—The above analyses com- parisons lacked sufficient power due to the
pared the overall levels of performance in the relatively small sample sizes for this particular
television, radio, and reading conditions. The test (i.e., with N = 24 per group, a difference
next question ofinterest is, what is the nature in correlations of .60 is necessary for
of the within-subject relationship between significance). Nonetheless, the conclusion
performance on each task following the read- that performance in the television and reading
ing versus the television and radio condi- conditions is not correlated but performance
tions? No previous study has examined this in the radio and reading conditions is corissue. In order to address this question. Pear- related must be qualified by the finding that
son product-moment correlation coefficients the correlations in these two conditions did
were computed for each dependent measure not differ from each other,
between performance in the reading condition and performance in the television and
As reported above, comprehension in the
radio conditions for third and sixth graders reading condition was significantly correlated
separately. These data are presented in Table with comprehension in the radio condition
2. The only measure on which there was a but not the television condition. (This result
significant correlation between performance can be seen in the top row of Table 2.) To test
TABLE 2
GORRELATIONS BETWEEN PERFORMANCE ON EACH MEASURE IN THE READING
AND TELEVISION OR RADIO GONDITIONS FOR THIRD AND SIXTH GRADERS
TELEVISION VS.
READINC
Third
Graders
Gomprehension
12
Sentenced'
36
Sentence displacement
- .02
Pictured'...
• .15
Picture displacement.,,,.. - .41*
* p < ,05,
RADIO VS,
READING
Sixth
Graders
Third
Graders
Sixth
Graders
.18
.08
.04
.02
.58*
,54*
.60*
.23
.48*
.25
.26
2078
Child Development
the consistency of this finding, additional cor- reading condition (rho = .43), and second,
relation coefficients were computed between sentence displacement by third graders was
perfonnance in the reading condition and per- significantly correlated in the radio and readformance in the television and radio condi- ing condition (rho = .40). Thus, the extreme
tions by third and sixth graders for each of the scores apparently did not alter the overall patfouj types of comprehension questions. For tern of results.
both age groups, the conrelations between
^j^.^¿ ^
.^^^ limitation of the above
comprehension accuracy m the readmg and ^^^^j^^ .^ ^^^ ^ ^ conclusions may be retelevision conditions were significant only for ^^^^^^ ^^ ^^^^^^ ^^ narrative materials only,
penpheral explicit items. However com- ^j^.^ .^ ^ ^ ^ ^ ^ ^ ^ ^^^^^^^ ^ ^ ^ ^ ^ ^ ^^^^^¿
prehension accuracy m the reading and radio ^ ^ j ^ ^ ^ ^ ^ ^^ ^^^^ ^ ^
^^^^^^
^^^^
conditions was sigmficantiy positively cor- Mandler & Johnson, 1977), and because regurelated for both age groups for all four types of j ^ television and radio programming includes
comprehension itenris. Thus, the pattem of re¿^
as weU as nanative segments. The
suits on the comprehension measure is gener- ^i^vision and radio versions of the expository
ally consistent for each of the four specific ^
^ ^^^^ .^^1^¿^¿ .^ ^ . ^ ^ ^ ^ ^ ^ ^ ^J^
types of comprehension questions.
^ ^ generalizability of the results beyond narThe above results suggest that com- native materials. The correlations between
prehension and memory for text are related to comprehension accuracy in the narrative concomprehension and memory for material pre- ¿ition and the expository condition were posisented on radio but not on television. How- ^ve and significant for third graders in the
ever, several additional analyses were con- television condition (r = .47), third graders in
ducted to further test this conclusion. First, it ^^ radio condition (r = .53), sixth graders in
is possible that the correlations between per- the television condition (r = .41), and sixth
formance on each measure in the television graders in the radio condition (r = .53). Thus,
and reading conditions were low because of a ^^ pattem of comprehension results appears
restricted range of scores in the television con- similar for narrative and expository segments,
dition. To test this interpretation, the variance
Fourth, to strengthen the results of the
in the television and reading conditions was present study, an additional measure of chilcompared on each of the five measures. These dren's reading ability was obtained. This meadata are presented in parentheses in the first s^^e was perfonnance on the standardized
two columns of Table 1. Tests of the differ- reading test administered by the public
ence between two related variances revealed schools. These standardized reading test
that only the difference in variances for the scores were con-elated with (1) comprehenthird graders on the picture displacement mea- sion accuracy in the reading condition in the
sure was significant, i(22) = 10.37. On all present study, (2) comprehension accuracy in
other measures, there was no significant dif- the nanative television and radio conditions
ference between the variance in the television and (3) comprehension accuracy in the exposiand reading conditions for either third or sixth tory television and radio conditions. These
graders. Further, the variance in tlie television data are presented in Table 3.
and radio conditions was compared. These
data are presented in parentheses in the
first
First, as can be seen in the first column of
and third columns of Table 1. No significant Table 3, the conelations between the standifferences between the variance in the tele- dardized reading test and the reading comvision and radio conditions resulted for either prehension measure utilized in the present
third or sixth graders. Thus, this alternative study were positive and significant in all conexplanation for the results is unsupported.
ditions except for the third graders in the television condition. Thus, the reading comSecond, the correlations between perfor- prehension test utilized in thé present study
manee in the television and reading condition yields results generally similar to those inand the radio and reading condition were dexed by the standardized reading test Sectested for the effects of extreme scores. For ond, as can be seen in the second column of
this purpose, the correlations reported in Table 3, performance on the standardized readTable 2 were recalculated using Spearman ing test was not significantly correlated with
rank-order correlations. The pattem of results comprehension of narrative segments in the
that followed from this analysis was the same television condition, but was correlated with
as that reported in Table 2 except for two of comprehension of narrative segments in the
the 16 correlations. These exceptions are, first, radio condition by third graders only. Third,
sentence recognition by third graders was the results in the third column of Table 3 sugsignificantly correlated in the television and gest that performance on the standardized
Pezdek, Lehrer, and Simon
2079
TABLE 3
CORRELATIONS BETWEEN STANDARDIZED READING TEST SCORES AND THREE MEASURES IN THE
PRESENT STUDY
Television group:
Third grade
Sixth grade
Radio group:
Third grade
Sixth grade
Standardized Reading
with Reading
Comprehension
Standardized Reading
with Comprehension of
Narrative Segment
Standardized Reading
with Comprehension of
Expository Segment
.29
.41*
.24
.14
.07
.28
.50*
.65*
.52*
.18
.82*
.55*
*o < .05.
reading test was not significantly correlated
with comprehension of the expository segment
in the television condition, but was highly correlated with comprehension of the expository
segment in the radio condition. With the exception of the sixth graders in the radio condition of the narrative segment, these results
confirm the results reported above. That is,
comprehension and memory for text are related
to comprehension and memory for material
presented on radio but not on television.
Results of the multiple regression analyses including questionnaire data.—Two
stepwise multiple regression analyses were
performed to assess which descriptive variables significantly predicted (1) radio comprehension accuracy, and (2) television comprehension accuracy. The predictor variables
were: standardized reading score, amount of
television viewing by the child, amount of
television viewing by the parent, parents'
education, and family income. These measures are described in the Method section. Because all parents did not complete the questionnaire, these analyses include 40 subjects
in the radio condition and 38 subjects in the
television condition.
ability is significantly correlated with memory
and comprehension for material presented on
radio but not on television.
Discussion
This study was designed to address two
issues. First, this study examines the relationship between comprehension and memory for
material read and comprehension and memory for material presented on television and
radio. Second, this study includes a comparison ofthe absolute levels of performance in
the television, radio, and reading conditions
on a range of dependent measures. The results are discussed in terms of these two issues.
The principal finding is that comprehension and memory for material read are related
to comprehension and memory for material
presented on radio but not on televisiori. This
conclusion follows from the result that the correlations between performance in the television and reading conditions were not
significantly different from zero on any measure except picture displacement; on this measure the correlation was negative for third
Radio comprehension was significantly graders and positive for sixth graders. On the
related to standardized reading score (simple other hand, the correlations between perforr = .55; cumulative R^ = .31). No other vari- mance in the radio and reading conditions
able significantly predicted radio comprehen- were more consistent and positive. In this
sion. In tlie second analysis, only the amount study, a battery of five tasks was utilized to
of television viewing by the child was assess diverse aspects of the complex comsignificantly related to television comprehen- prehension and memory construct. The gension accuracy (simple r = .25; cumulative R^ eral consistency ofthe findings across the five
= .12). It is interesting that the significant re- measures suggests that the results are robust
lationship between television comprehension and not restricted to more limited comprehenand amount of television viewing by the child sion and memory processes. These results
suggests that for children, television com- must be qualified, however, by the finding
prehension may improve with practice. Al- that correlations in the television and reading
though little new information resulted from cotidition did not significantly differfromcorthese two analyses, the findings do further relations in the radio and reading condition on
support the previous conclusion that reading any measure.
2080
Child Development
The above pattern of results was corroborated with several additional analyses.
First, the correlations between comprehension in the radio and reading conditions were
positive and significant for each of the four
types of comprehension questions, but the
correlations between performance in the television and reading conditions were not
significant except for peripheral-explicit questions. Second, the findings generalize beyond
narrative materials. Comprehension accuracy
in the narrative and expository segments was
significantly correlated in the radio and television conditions. Third, the correlational results do not appear to be confounded by a restricted variance in the television condition or
by the presence of extreme scores. Finally,
standardized reading test scores were significantly correlated with performance on the
reading comprehension test utilized in this
study, and were significantly correlated with
comprehension accuracy in the radio but not
the television condition. Together, these results validate the reading comprehension measure utilized in this study and further support
the conclusion that reading comprehension is
related to comprehension of radio but not television.
There are at least three possible interpretations of the finding that the relationship between comprehension and memory for text is
related to comprehension and memory for material presented on radio but not television.
First, diese results may reflect that the cognitive processes underlying memory and comprehension for text are similar to the cognitive
processes underlying memory and comprehension for material presented on radio but
not television. Second, the same cognitive
processes may underlie memory and comprehension for all three types of media, but
different individuals may be differentially
facile at processing various types of information. Third, the same cognitive processes may
underlie memory and comprehension for all
three types of media, but these same cognitive
processes result in different types of information being retained from different forms of
media. The results of the present study do not
differentiate among these three interpretations. Future research will be necessary to
more specifically determine tiie extent to
which the effects are the result of differences
in the processes, the processors, and the type
of information that results from the processes.
processing différences between television
and reading reside at the attention and perceptual levels. Although this interpretation is
not specifically tested in tiiis study, it seems
unlikely because oral readers necessarily
have to attend to the text 100% of the reading
time, but they can and do attend to television
far less. Yet despite the fact that attention to
text was greater than attention to television,
memory and comprehension were generally
better for television than for text (see Table 1).
The media differences thus appear to be at a
higher level of processing.
Hypotheses regarding the nature of the
differences between remembering and comprehending television and text can be
generated in light of the findings by Pezdek
and Stevens (1984). That study reported that
the visual material on television is more salient and memorable than the auditory material; however, the audio and video information on television are processed together as
part of a single integrated stimulus and should
not be considered as independently processed.
The results of Lorch, Anderson, and Levin
(1979) and Pezdek and Hartinan (1983) ñirther
support this notion by showing that comprehension of auditory information on television is highly correlated with visual attention
to the television. Thus, the audio and video
channels appear to work together synergistically such that features of the audio channel
serve to direct attention to important aspects
of the video channel, and also, features of the
video channel serve to highlight and direct
attention to important aspects of the audio
channel. These findings suggest that "good
television comprehension" involves the ability to benefit from the mutually supporting information presented in parallel on the video
and audio television channels. Because this
process does not have a counterpart in reading, this factor may account for the memory
and comprehension differences between television and reading.
A second factor that may account for
memory and comprehension differences between television and reading is the extent to
which the top-down processing is driven by
visual versus auditory verbal information. In
other words, although both television processing and reading appear to involve schemadriven processing (see Anderson & Lorch,
1983), perhaps visual information is more effective in guiding schema generation on teleThe result of particular interest is the vision, whereas verbal information is more effinding that performance was uncorrelated in fective in guiding schema generation in
the television and reading conditions. One in- reading. If so, then good comprehenders of
terpretation of this result is that the critical television would be those who primarily
Pezdek, Lehrer, and Simon
utilized visual information to generate
schemata. Because there is not a good
counterpart to this process in reading, a good
comprehender of television would not necessarily be a good comprehender of text.
The second important finding in this
study is that the absolute levels of performance were generally similar in the television
and reading conditions. However, performance was better in both of these conditions
than in the radio condition. This was found
with measures of comprehension accuracy
and sentence recognition. The sentence displacement measure was generally insensitive
to media differences.
The pattern of results was generally similar for third graders (beginning readers) and
sixth graders (more advanced readers). On
most measures, sixth graders performed better
than third graders, but no significant interactions were found. This suggests that although
the sixth graders were processing the information in each condition more effectively, the
two age groups were operating similarly on
the material provided.
The finding that comprehension and retention were better in the television than
radio conditions is consistent with previous
studies in which comparable presentation formats were utilized (see Barrow & Westley,
1959; Wetstone & Friedlander, 1974). The result that television is better retained than
radio is particularly interesting given that the
test items in the present study were based
primarily on auditorily presented information.
This result, together with the results of Pezdek and Stevens (1984), suggests that the accompanying visual information on television
makes the auditory-verbal material more
memorable than it is in the auditory-only condition.
The previous studies that have compared
performance in reading and radio (i.e., listening) conditions have reported either no difference in the recall of details (King, 1968) or
superior recall of listeners over readers
(Horowitz & Berkowitz, 1967). The present
results were that comprehension and sentence recognition were significantly better in
the reading than radio condition. However, an
important methodological difference between
this study and previous studies is the use of
pictures in the books that were read. This difference suggests that for children, pictures aid
comprehension and retention of information
in books. Finally, the result in the present
study that information is retained at comparable levels in the television and reading condi-
2081
tions has not been reported or tested elsewhere.
The present study concludes that across a
range of memory and comprehension measures, performance in the reading and television processing conditions was not significantly correlated. Thus, although reading
and comprehending television appear to be
similar in many ways, there are sufficient differences that reading ability does not discriminate good from poor television comprehension. Research in the direction of articulating
the locus of the cognitive processing differences between memory and comprehension
of television versus test is our high priority for
the friture.
Reference Note
1.
Greenfield, P. M., Geber, B., Beagles-Roos, J.,
Farrar, D., & Gat, I. Television and radio experimentally compared: Effects of the medium
on imagination and transmission of content.
Paper presented at the meeting of the Society
for Research in Child Development, Boston,
April 1981.
References
Anderson, D. R., & Lorch, E. P. Looking at television: Acting or reacting? In J. Bryant & D. R.
Anderson (Eds.), Children's understanding of
television. New York: Academic Press, 1983.
Banks, W. P. Signal detection theory and human
memory. Psychological Bulletin, 1970, 74, 8 1 99.
Barrow, L. C, & Westley, B. H. Comparative teaching effectiveness of radio and television. Audio
Visual Communciation Review, 1959,7,14—23.
Beagles-Roos, J., & Gat, I. The specific impact of
radio and television on children's story comprehension. Journal of Educational Psychology, 1983, 75, 128-137.
Hochhaus, L. A table for the calculation of á' and B.
Psychological Bulletin, 1972, 77, 375-376.
Homik, R. C. Television access and the slowing of
cognitive growth. American Educational Research Journal, 1978, 15, 1—15.
Horowitz, M. W., & Berkowitz, A. Listening and
reading, speaking and vmting: An experimental investigation of differential acquisition and
reproduction of memory. Perception and Motor
Skills, 1967, 24, 207-215.
King, D. J. Retention of connected meaningful material as a function of modes of presentation and
recall. Journal of Experimental Psychology,
1968, 77, 676-683.
Lorch, E. P., Anderson, D. R., & Levin, S. R. The
relationship of visual attention to children's
comprehension of television. Child Development, 1979, 50, 722-727.
2082
Child Development
Mandler, J, M., & Johnson, N, Remembrance of
things parsed: Story structure and recall, Cognitive Psychology, 1977, 9, 111-151.
MeringofF, L. K, Influence of the medium on chil. dren's story apprehension. Journal of Educational Psychology, 1980, 72, 240-244.
Morgan, M. Television viewing and reading: Does
more equal better? Journal of Communication,
1980, 30,159-165,
Murray, J, P, Television and youth. Boys Town,
Neb,: Boys Town Press, 1980,
Nelson, K, Structure and strategy in leaming to talk.
Monographs of the Society for Research in
Child Development, 1973, 38 (1, Serial No,
149),
Pezdek, K,, & Hartman, E, F, Ghildren's television
vievkdng: Attention and comprehension of auditory versus visual infonnation. Child Development, 1983, 54,1015-1023,
Pezdek, K., & Stevens, E, Ghildren's memory for
auditory and visual information on television.
Developmental Psychology, 1984,20,212-218,
Ridder, J, M, Pupil opinion and the relationship of
television to academic achievement,/oumaZ of
Educational Research, 1^3, 57, 204-206,
Smiley, S, S,, Oakley, D, D., Worthen, D,, Gampione, J, G,, & Brown, A. L, Recall of thematically relevant material by adolescent good and
poor readers as a function of written versus oral
presentation. Journal of Educational Psychology, 1977, 69, 381-387,
Wetstone, H, S,, ôc Friedlander, B, Z, The effect of
live, TV, and audio story narration on primary
grade children's listening comprehension.
Journal of Educational Research, 1974,68,3235,
Zuckermann, D, M,, Singer, D. G,, & Singer, J, L,
Television viewing, children's reading and related classroom behavior, Joumoí of Communication, 1980, 30, 166-174,
This document is a scanned copy of a printed document. No warranty is given about the accuracy of the copy.
Users should refer to the original published version of the material.