Talk in the Classroom.
Linguistics, Education
- 2001
Traditionally, language has been divided into four main skills – listening, speaking, reading and writing. Until the mid-19 century, schools tended to focus on the written language. Foreign languages…
Reading Psychology, 26:55–80, 2005
Copyright C
2005 Taylor & Francis Inc.
0270-2711/05 $12.00 + .00
DOI: 10.1080/02702710590910584
THE RELATIONSHIP BETWEEN LISTENING AND READING
COMPREHENSION OF DIFFERENT TYPES OF TEXT AT
INCREASING GRADE LEVELS
IRENE-ANNA N. DIAKIDOY
Department of Psychology, University of Cyprus, Nicosia, Cyprus
POLYXENI STYLIANOU
Agrokipia Elementary School and University of Cyprus, Nicosia, Cyprus
CHRISTINA KAREFILLIDOU and PANAYIOTA PAPAGEORGIOU
Department of Psychology, University of Cyprus, Nicosia, Cyprus
This study examined the hypotheses that [a] the relationship between listening
and reading comprehension becomes stronger after decoding mastery; [b] the dif-
ference between listening and reading decreases with increasing grade level; and
[c] similar patterns of relationship and difference are obtained with narrative
and expository texts. The sample included 612 students in Grades 2, 4, 6, and
8. Students read and listened to two narratives and two expository texts and
completed corresponding comprehension tests that were in the form of sentence
verification tasks. The findings confirmed the first two hypotheses but not the
third one. In the case of expository text, the relationship between listening and
reading comprehension was weaker than the corresponding one with narrative
text, and performance levels were comparable across all elementary grades. More-
over, reading comprehension levels were higher than listening comprehension levels
in Grade 8, regardless of text type. The implications of these findings with respect
to the dominant unitary process model and the assessment and instruction of oral
and written language comprehension are discussed.
By the time children enter school, they have acquired to various
degrees the skills that enable them to use and understand oral lan-
guage in a variety of contexts [Snow, 1983; Wells, 1986]. Despite
the acknowledged differences between oral and written language
This research was supported by an Elva Knight Research Grant from the International
Reading Association to the first author.
We thank George Spanoudes for his help in data analyses, and the students for their
enthusiastic participation and their teachers and principals who made their participation
possible.
Address correspondence to Irene-Anna N. Diakidoy, Department of Psychology, Uni-
versity of Cyprus, P.O. Box 20537, Nicosia CY-1678, Cyprus. E-mail: eddiak@ucy.ac.cy
55
56 I.-A. N. Diakidoy et al.
[Perfetti, 1987; Tannen, 1985], such skills have been commonly
associated with later reading achievement. This assumption un-
derlies much of the emergent literacy research [Eller, Pappas, &
Brown, 1988; Feitelson, Goldstein, Iraqi, & Share, 1993; Senechal,
LeFevre, Thomas, & Daley, 1998] and the assessment of reading, as
indicated by the inclusion of listening comprehension measures in
various standardized tests and reading inventories [Burns & Roe,
1993; Johns, 1994; Kertoy & Goetz, 1995]. Listening comprehen-
sion figures prominently as a component in theoretical models of
reading [e.g., Hoover & Gough, 1990; Joshi & Aaron, 2000] and
has been considered to provide the base for the acquisition of
reading comprehension skills [Sticht & James, 1984]. The extent,
however, to which this is true across the reading tasks that differ-
ent text types represent has not received any attention. Therefore,
the present study examined the relationship between listening and
reading comprehension at increasing grade levels, and the extent
to which it is influenced by the type of text.
Differences between oral and written language concern both
the course of their development as well as the nature of the lin-
guistic stimulus they present. The acquisition of oral language pre-
cedes the acquisition of written language, and it is a natural process
that takes place gradually in highly contextualized situations. On
the other hand, the acquisition of written language depends on
formal, systematic instruction and occurs in more decontextual-
ized situations [Snow, 1983]. Although spoken language is more
likely to contain fragments instead of complete sentences, listeners
can take advantage of extra-linguistic clues, such as gestures and
prosodic information, such as stress and intonation [Akinnaso,
1985; Sinatra, 1990]. Moreover, listeners and speakers share the
same context, and, therefore, spoken language is characterized
by greater involvement and interaction [Akinnaso, 1985; Biber,
1982]. Consideration of such differences has led to the advance-
ment of the dual process view which postulates that, although read-
ing and listening comprehension share some common elements,
they essentially represent different processes [Samuels, 1987; see
also Sinatra, 1990].
On the other hand, Sticht and his colleagues have advanced
the position that listening comprehension and reading compre-
hension are highly interrelated [Sticht, Beck, Hauke, Kleinman,
& James, 1974]. Their extensive review of the available studies led
The Relationship Between Listening and Reading 57
them to conclude that both listening and reading comprehen-
sion depend on the same general comprehension process. There-
fore, after decoding skills have been mastered, performance on
listening and reading comprehension tasks should be compara-
ble. Moreover, listening comprehension level represents a poten-
tial for reading comprehension. Specifically, they have argued that
listening comprehension skill facilitates the acquisition and pre-
dicts the level of skill that will be achieved in reading [Sticht et al.,
1974; Sticht & James, 1984]. Their position essentially represents a
contrasting unitary process view, according to which the same com-
prehension process underlies both listening and reading [Sinatra,
1990].
Support for a unitary process view has come from studies that
have shown input modality [listening vs. reading tasks] to have no
effect on the recall of meaningful and syntactically well-structured
word strings [Guthrie & Tyler, 1976] and on the recall and sum-
marization of stories [Kintsch & Kozminsky, 1977; Smiley, Oakley,
Worthen, Campione, & Brown, 1977]. Moreover, Sinatra [1990]
found a facilitative effect in comparison times when an identical
auditory stimulus preceded the presentation of two visual stim-
uli. Since that effect was manifested for meaningful sentences,
syntactic word strings, and random word strings but not for ran-
dom nonword strings, Sinatra [1990] concluded that listening and
reading processing converge at the lexical level as well. Additional
support for the unitary process view has also come from studies
that have shown directly listening comprehension to predict read-
ing comprehension skill [Curtis, 1980], and instructional inter-
ventions targeting listening comprehension to generally have pos-
itive effects on reading comprehension performance [Pearson &
Fielding, 1982; Sticht et al., 1974].
Although the relationship between listening and reading com-
prehension is assumed to be unidirectional early on, it is also ex-
pected that differences between listening and reading compre-
hension diminish over time and as a result of print exposure
[Perfetti, 1987; Sticht & James, 1984]. On the basis of their re-
view, Sticht et al. [1974] concluded that up to Grade 7, listen-
ing comprehension ability is higher than reading comprehension
ability, whereas in adulthood the direction of the difference is re-
versed. Curtis [1980] found that the relationship between reading
skill and listening comprehension changed as a function of age,
58 I.-A. N. Diakidoy et al.
with correlations at the Grade 2 level being significantly differ-
ent from the correlations at the Grade 5 level. More interestingly,
Hendrick and Cunningham [1995] found that wide reading was
associated with greater listening comprehension skill at Grade 4.
Although their study was not specifically designed to test direction-
ality, their results support the possibility of a reciprocal relationship
[Hendrick & Cunningham, 1995]. The present study examined
the relationship and the differences between listening and read-
ing comprehension levels at Grades 2, 4, 6, and 8. On the basis
of previous research and claims, we expected the relationship to
become stronger and the differences to decrease with increasing
grade level.
The studies that examined the effects of presentation mode
on the recall and summarizing of text [i.e., Kintsch & Kozminsky,
1977; Smiley et al., 1977], as well as those that examined the effects
of listening to text being read on the acquisition of literacy and
learning in general [Eller et al., 1988; Feitelson et al., 1993], have
all used narrative texts. It can be argued, however, that printed nar-
rative text is closer to oral discourse with narrative characteristics,
such as contextual situatedness and temporal/causal sequences
[Biber, 1982; Graesser, Singer, & Trabasso, 1994], and which oc-
curs frequently in everyday conversation, such as when one nar-
rates what happened at work or over a weekend. Moreover, young
children can be typically expected to have had more exposure
to oral narrative text structures in highly interactive contexts that
guide and facilitate comprehension, such as story reading at home
[Neuman, 1996]. Finally, the information in narrative text is more
likely to be familiar and its organization more likely to be pre-
dictable [Graesser, Swamer, Baggett, & Sell, 1996]. Therefore, it
is not clear whether the strong relationship between listening and
reading comprehension that has been observed so far is also due
to mediating factors, such as the similarity of the communicative
task[s] that oral and printed narration represent, their high fre-
quency of occurrence in both modalities, and/or the familiarity
and concreteness of their content.
Kintsch and Kozminski [1977] and Sinatra [1990] have sug-
gested the possibility that listening and reading processes may
diverge when it comes to processing lengthy and difficult text.
Expository text is more likely to contain unfamiliar information
and to have an underlying abstract and logical structure [Singer,
The Relationship Between Listening and Reading 59
Harkness, & Stewart, 1997]. Moreover, expository text structures
are diverse [Englert & Hiebert, 1984], and, therefore, readers and
listeners cannot rely on the activation of a single schema the way
they can rely on the activation of a story schema to guide their com-
prehension of narrative text. Although students are increasingly
expected to read and learn from expository text with increasing
grade level, they have little opportunity to encounter and inter-
act with this text type in the early grades [Duke, 2000]. Finally,
oral expository discourse, such as lectures and public speeches, is
much more infrequent than oral narrative discourse. It can be ar-
gued, then, that these factors—unfamiliar content and structure
and low frequency of occurrence—increase the relative difficulty
of expository text in terms of processing demands and in com-
parison to narrative text. Therefore, in the present study both ex-
pository and narrative texts were presented orally and in printed
form.
In summary, the purpose of this study was to evaluate [a] the
relationship between listening and reading comprehension at in-
creasing grade levels, and [b] the effects of text type on listening
and reading comprehension levels. With respect to this second
goal, it was predicted that both listening and reading comprehen-
sion performance will be higher for narrative than expository text.
However, it was also expected that with increasing grade level the
differences between listening and reading comprehension of nar-
rative and expository text will decrease as a result of increased
exposure to oral and written expository structures. Finally, of inter-
est was also the extent to which the relationship between listening
and reading comprehension evolves in a similar manner across the
different text types, as would be predicted by the unitary-process
view.
Method
Participants
The sample included 612 students enrolled in three elementary
schools and two middle schools located in different medium-
sized towns in the island nation of Cyprus. Overall, there were
135 second graders [65 males and 70 females], 151 fourth graders
[79 males and 72 females], 151 sixth graders [70 males and
60 I.-A. N. Diakidoy et al.
81 females], and 177 eighth graders [95 males and 82 females].
All students were fluent speakers of Greek. In order to obtain
an indication of Receptive Language Ability across grade levels,
the method developed by Levin et al. [1997] was adapted and
employed in this study. Specifically, the names of the students in
each participating classroom were printed on separate index cards.
Then each classroom teacher was asked to evaluate each student’s
written and oral language comprehension ability by sorting the
corresponding card into one of five categories: very high, high,
average, low, very low [Levin et al., 1997]. In cases where teach-
ers classified the majority of their students into one or two cat-
egories only, they were asked to discuss at length any language
comprehension differences between the students within a cate-
gory and to reclassify accordingly. Teacher ratings ranged from
1to5with higher scores assigned to higher ability categories.
This measure indicated that the students, as a sample, represented
a wide range of Receptive Language Ability levels [M=3.15,
SD =1.23]. In addition, their school achievement levels ranged
from low [3.5] to high [10.0] as indicated by previous year’s Grade
Point Averages [M=7.87, SD =1.50]. The correlation coeffi-
cient between Teacher ratings and Grade Point Average was 0.77
[p=.00].
Materials
TEXTS
A preliminary list of forty short stories, articles, and excerpts
from books, magazines, and encyclopaedias was compiled. Seven
independent teachers, whose teaching experience ranged from
seven to 22 years, read and rated each text for topic familiarity,
number of potentially unfamiliar words, and overall text difficulty
for each grade level. Texts that were judged by all teachers to have
an unfamiliar topic but relatively few unfamiliar words, and to be
moderately difficult but appropriate for a given grade level were
included in the final list [see Appendix]. From this list, two nar-
ratives [271 to 376 words long] and two expositories [262 to 356
words long] were selected for each grade. The expository texts ei-
ther described and elaborated on the characteristics of animals,
physical locations, archaeological findings, astronomical phenom-
ena, and groups of people [Appendix, Texts 1, 5, 8, 11, 13, and 16]
The Relationship Between Listening and Reading 61
or presented facts and details related to socio-cultural topics and
issues [Appendix, Texts 6 and 14]. Therefore, the predominant
expository structures were description and enumeration [Englert
& Hiebert, 1984]. The selected texts were parsed into idea units
[van den Broek, 1989] by two independent graduate assistants. In-
terrater agreement ranged from 79% to 93%, and all differences
were resolved in conference. The selected narratives and exposi-
tories were comparable with respect to the number of idea units
per sentence [M=1.63, SD =.30, and M=1.77, SD =.36 re-
spectively, t[14] =−0.88, p=.39]. Each selected text was printed
in a separate booklet and tape recorded.
COMPREHENSION TESTS
All comprehension tests were in the form of a sentence verifi-
cation task. For each text, a list of literal and inferential statements
was constructed. Literal statements represented single idea units
extracted verbatim from the text [Pearson & Johnson, 1978]. In-
ferential statements represented implicit referential, causal, super-
ordinate goal, generalization, and comparison/contrast relation-
ships between idea units [Graesser et al., 1994]. Two independent
raters evaluated each statement in each list as either literal or infer-
ential. Statements for which there was 100% agreement comprised
the preliminary comprehension tests whose length ranged from 17
to 22 statements.
Fifty-four undergraduate students enrolled in a cognitive psy-
chology course listened to each tape-recorded text and completed
the corresponding preliminary comprehension test. Items that
were not answered correctly by two or more students [18% of the
items, on the average, across tests] were excluded. From the re-
maining items, 14 were selected for inclusion in a final comprehen-
sion test for each text. Specifically, each final comprehension test
included seven literal and seven inferential statements, each fol-
lowed by a Yes/No option. In each test there was an equal number
of Yes and No correct responses. Each correct response received
a score of 1 and each incorrect response a score of 0. The scores
received by each student for each statement were summed and con-
verted to proportions, yielding six measures: Listening Compre-
hension, Reading Comprehension, Narrative Listening Compre-
hension, Expository Listening Comprehension, Narrative Reading
Comprehension, and Expository Reading Comprehension. The
62 I.-A. N. Diakidoy et al.
order of statement presentation in each test was randomized, and
two comprehension test versions with different orders were created
for each text.
Procedure
For the purposes of the study, the students remained in their in-
tact classrooms, and data were collected during the first half of
the school year. The study was completed in two 40 min. sessions
scheduled from two to seven days apart. In each session, students
read and listened to one narrative and one expository text. As a
result, half of all the students in each grade level read a particular
text while the rest heard that same text. Presentation Mode [oral
vs. written], Text Type [narrative vs. expository], and order of pre-
sentation were counterbalanced between classrooms and sessions.
Students listened to or read each text only once. Subsequently, the
written text booklets were collected, and the students completed
the corresponging comprehension test. They were instructed to
read each statement carefully, to decide whether it was true on the
basis of the text they had just read or listened, and to circle Yes or
No accordingly. Comprehension test version was counterbalanced
between students.
Results
Relationship Between Listening and Reading Comprehension
Preliminary analyses indicated that all dependent variables,
Listening Comprehension, Reading Comprehension, Narrative
Listening Comprehension, Narrative Reading Comprehension,
Expository Listening Comprehension, and Expository Reading
Comprehension had normal distributions [skewness .05]. Moreover,
one-way Analysis of Variance indicated that School had no signif-
icant effect, and, therefore, this variable was excluded from sub-
sequent analyses. Listening and Reading Comprehension Scores
were significantly correlated with each other at all grade levels [ p