* Reading is the most frequently mentioned academic subject in
which students with learning disabilities experience failure (Lindsey
& Kerlin, 1979). Although their problems in reading are generally
attributed to poor decoding skills (McCormick & Samuels, 1979;
Perfetti & Hogaboam, 1975), many of these students have difficulty
in finding main ideas and important supporting details (Graves, 1986;
McGee, 1982; Worden & Nakamura, 1983). Several researchers have
reported that poor readers have comprehension deficits textual materials
consisting only of words that could be decoded accurately (Guthrie,
1973; Smiley, Oakley, Worthen, Campione, & Brown, 1977).
LEARNING STRATEGIES AND COMPREHENSION
Failure to employ appropriate learning strategies is often a
critical component of learning disabilities (Alley & Deshler, 1979).
Students with learning disabilities have often been characterized as
"inactive learners" who fail to select, implement, and monitor
effective learning strategies spontaneously (Torgesen, 1982). As groups,
poor readers display less efficient textscanning strategies (DiVesta,
Hayward, & Orlando, 1979; Garner & Reis, 1981), less efficient
comprehension-monitoring strategies (Bos & Filip, 1984), and less
sensitivity to text structure (Smith & Friend, 1986). Moreover, many
such students fail to conceptualize reading as a search for meaning and,
thus, approach the task passively (Bransford, Stein, & Vye, 1982;
Paris & Meyers, 1981).
The generalized deficits in reading comprehension of many students
with learning disabilities suggest the importance of systematic
instruction in learning strategies. Three strategies have empirical
support for use with poor readers: self-questioning (Clark, Deshler,
Schumaker, Alley, & Warner, 1984; Wong & Jones, 1982; Wong,
Wong, Perry, & Sawatsky, 1986), paraphrasing (Hansen, 1978; Jenkins,
Heliotis, Haynes, & Beck, 1986; Schumaker, Denton, & Deshler,
1984), and visual imagery (Clark et al., 1984).
Summarization is a fourth learning strategy that can help students
used deletion and superordination to construct and retain a succinct
summary of important propositions from text. Kintsch and van Dijk (1978)
hypothesized that during the process of comprehension, readers form a
macrostructure, or gist, from the microstructure (i.e., the semantic
content of sentences in the text). Although some information may be
recalled explicitly, it is the text's macrostructure that a reader
primarily remembers and uses as a cue to recall other information from
the text. Kintsch and van Dijk have specified three macrorules for
condensing information: (a) deletion--any proposition that denotes an
accidental property of a discourse referent may be deleted; (b)
generalization--an immediate superconcept may be substituted for a
sequence of micropropositions; and (c) construction--a global
proposition that denotes normal conditions, components, or consequences
may be substituted for a sequence of propositions that makes them
explicit. Application of these macrorules allows the reader to reduce
the number of textual propositions and to extract the macrostructure.
Brown and Day (1983) proposed five basic rules of summarization.
The first two rules require the deletion of unnecessary material. The
third rule, superordination, requires the substitution of a
superordinate term for a list of items or actions. The fourth and fifth
rules deal with topic sentences for each paragraph: selection or
invention of a topic sentence.
Students without disabilities have successfully been taught to
summarize expository reading passages (Armbruster, Anderson, &
Ostertag, 1987; Bean & Steenwyk, 1984; Hare & Borchardt, 1984;
Palincsar, 1982; Rinehart, Stahl, & Erickson, 1986; Taylor, 1982;
Taylor & Beach, 1984). Moreover, summarization training was found to
improve reading comprehension. The purpose of this study was to assess
the effects of instruction in a summarization strategy on the
comprehension of expository material by students with learning
disabilities. The maintenance of the strategy over time and its transfer
to a new situation were also investigated.
The study was conducted in two phases. In the first phase, the
effects of summarization strategy on reading comprehension were assessed
by contrasting three groups of students: one group of students with
average reading ability and two groups of students with learning
disabilities who had been assigned randomly to one of two conditions
(those trained in summarization and those not trained). In the second
phase, maintenance and generalization of the summarization strategy were
assessed for students with learning disabilities who had been trained to
use the strategy. Two dependent measures were used: (a) comprehension
scores on criterion tests prepared by the authors and (b) comprehension
scores from a commercially prepared, norm-referenced reading test
(MacGinitie, 1978). Subjects
Four learning disabilities resource teachers in rural central
Pennsylvania nominated 63 sixth-through ninth-graders who had been
classified as learning disabled and were adequate decoders but poor
To ensure that these were students who had difficulties in reading
comprehension, they were required to meet for additional criteria.
First, in order to test decoding skills, students had to read a 200-word
expository passage written at the fourth-grade level with 90% accuracy.
Second, students' standardized scores on the reading comprehension
subtest of the Gates-MacGinitie Reading Test had to be at least 2 years
below grade level but not lower than Grade 4. (This criterion was set as
the minimum acceptable comprehension skill for instructional purposes
and was consistent with the minimum standards established for training
in learning strategies at the Institute of Research in Learning
Disabilities, University of Kansas.) Third, students' performances
on reproduction of main ideas, a summarization measure, were required to
be below 40%. Fourth, students' performance on two criterion tests
of comprehension was required to be below 40%. Each criterion test
consisted of 10 multiple-choice questions on an expository passage. Two
different expository passages were used to control for subjects'
familiarity with the content material.
Participants were 30 students with learning disabilities in Grades
6 through 9 from three different schools in two school districts, who
met the criteria previously specified. These students were stratified by
reading levels on the Gates-MacGinitie Reading Test and then randomly
assigned to experimental and control groups. Table I contains
descriptive information for the students in these two groups. The groups
appeared equivalent on chronological age, grade placement, Wechsler
Intelligence Scale for Children-Revised intelligence quotients, and
Gates-MacGinitie reading performance.
To provide some normative data on the comprehension tests used in
this study, 15 average readers formed a normal comparison group. This
group was drawn from Grades 6 through 9 in proportion to the study
participants in these grades. These students were not enrolled in any
remedial or accelerated reading program. According to their teacher,
they were reading at, or slightly above, their grade level. Materials
Instructional Materials. A separate set of 10 short paragraphs was
developed to teach students each of the five summarization rules. Each
set of paragraphs focused exclusively on a different rule. In addition
to the paragraphs used to teach the specific summarization skills, six
expository passages were used to train the students to apply the
summarization rules in concert. Passages from Timed Readings Series
(Spargo, Williston, & Browning, 1980) were modified to ensure that
the various summarization rules could be applied. Therefore, lists of
items/events were added to allow superordination; topic sentences were
manipulated to allow their selection or invention; and to allow
deletion, some sentences were paraphrased and added, as were trivial
details. The readability of the passages ranged from 4.0 to 4.6,
according to the Fry (1977) formula; passages ranged in length from 400
to 470 words.
Testing Materials. Six expository passages similar to those
employed for training were prepared for use as pretest, posttest, and
delayed posttest. For each passage, 10 multiple-choice comprehension
questions were constructed. These questions were of two kinds: five
condensation questions (assessing comprehension of main ideas, cause and
effect relationships, concepts, and inferences) and five factual
questions (to assess explicitly stated facts). Further, alternate forms
of the comprehension subtest of the Gates-MacGinitie Reading Test were
administered to all students with learning disabilities before training
and to the experimental group after the training. Testing Procedures
The order of presentation of four passages was balanced across
students and across pretest-posttest conditions so that a passage would
be used as a pretest for some students and posttest for other students.
Pretest and posttest data were collected on two passages per condition
from all students; only one passage was tested per day. Thus, pretesting
required 2 days, as did posttesting. Each student was given a tape
recorder ad a packet containing the expository passage, pencils, red
pen, and scratch paper. On the pretest, students were told what
summarization is and were instructed to read the passage, to underline
the important sentence in pencil, and to cross out the unimportant
sentences with the red pen. On the posttest, the students were asked
only to form a summary. On both pretest and posttest, students were told
to use their marked passage or rough notes to construct and tape-record
their summaries. In addition, students were required to answer the 10
multiple-choice questions per passage. Finally, comprehension scores
from the Gates-MacGinitie Reading Test were available for students with
learning disabilities from their earlier screening.
Using the same procedures as used for the posttest, students in the
experimental group were again tested to assess maintenance of the
summarization strategy. At this time, the students in the experimental
group were also tested for generalization. To assess generalization of
the strategy, the resource room teachers were requested to administer an
alternate form of the comprehension subtest of the Gates-MacGinitie
Reading Test to students in the experimental group. The order of
administration of the delayed posttest and test for generalization was
randomized across students. The delayed posttest was conducted between
25 to 36 days (median = 29) after training. Instructional Procedures
To attribute changes in comprehension to instruction in
summarization, we found it necessary to ensure that the students learned
the strategy. Consequently, a mastery-learning paradigm (Bloom, 1976)
was used to guarantee that each student in the experimental group had
acquired the five summarization rules developed by Brown and Day (1983):
(a) superordination, (b) deletion of redundant information, (c)
selection, (d) invention, and (e) deletion of unimportant information.
These students received training and guided practice in summarizing the
contents of expository passages orally because the poor written-language
skills of many of these students (Cicci, 1983; MacArthur & Graham,
1987) could interfere with summarization.
Thirty-five to 40-minute training sessions were conducted with
small groups of 3 to 4 students each in the learning disabilities
resource room. the total amount of time required for training (not
including pretesting and posttesting time) ranged from 6.5 to 11 hours.
To decrease the threat to the integrity of the independent variable,
scripted formats were used.
Training was based on the principles of explicit or direct
instruction (Pearson, 1984; Rosenshine, 1986; Rosenshine & Stevens,
1984): explicit explanation of the rules, modeling the strategy, guided
practice in controlled materials, monitoring with corrective feedback,
and independent practice. In the first session, the senior author
provided the students with a rationale for learning the summarization
strategy, presented examples of situations in which it could be used,
and informed them about anticipated results. Then she described and
modeled the first rule (superordination). Students were provided with a
set of 10 paragraphs developed exclusively for that rule, and they
practiced the rule until they reached criterion performance.
Students were trained similarly in subsequent sessions to criterion
on each rule of summarization. In each session, the rules learned in the
previous sessions were reviewed prior to teaching a new rule. For each
rule, a different performance criterion was established for mastery. For
superordination and selection of topic sentences, the criterion was set
at 100% accuracy on two consecutive paragraphs because these tasks were
relatively simple. For deletion of redundancies, invention of topic
sentences, and deletion of unimportant information, the criterion was
set at 80% accuracy on two consecutive training pararagphs. After each
rule was mastered in isolation, students received instruction in the
combined use of the five rules. Initially during these instructional
sessions, subjects were not required to form a summary. In the last six
sessions, students were asked to construct oral summaries of the
expository passages using all five rules. They received feedback on
their summarization performance for each session. Students practiced
forming summaries until an evaluation of their summaries indicated that
they had reached mastery on all five rules.
The instruction was also designed to increase the student's
role gradually over the course of the training. In the beginning, the
researcher had the responsibility for direct instruction. However, as
the students learned the summarization rules, they were given
increasingly greater responsibility. By the end of the training,
students had assumed responsibility not only for practicing the rules
but also for checking that each rule had been applied.
Effect of Training on Comprehension
Pretest and posttest raw scores on the multiple-choice
comprehension tests and Normal Curve Equivalents from the
Gates-MacGinitie comprehension subtests are displayed in Table 2.
Performances of the three groups on multiple-choice condensation and
factual questions were analyzed using a 3 (Groups) x 2 (kinds of
questions) x 2 (Test time) analysis of variance (ANOVA) with repeated
measures. The summary of this ANOVA is presented in Table 3. Because the
three-way interaction was significant, F (2, 42) = 15.13, p <.001,
Pretest to posttest x Group interactions were examined separately for
each dependent comprehension measure. Interactions for condensation and
factual questions appear in Figures 1 and 2. [Tabular Data 2 and 3
Follow-up tests (Tukey's Wholly Significant Differences, WSD)
indicated that for both condensation and factual questions, the
performances of the control and the normal comparison groups remained
essentially the same (p <.05) from pretest to immediate posttest
conditions. There were no differences (p <.05) between the control
and experimental groups' performances on condensation and factual
questions on the pretest, whereas both groups of students with learning
disabilities had significantly lower means than the normal comparison
group. However, on the immediate posttest, the experimental group's
performance was significantly greater than that of the control and
normal comparison group's on condensation questions. The
experimental group's performance was also significantly greater
than of the control group but equal to the normal comparison
group's performance on factual questions. Maintenance of
The experimental group's comprehension performance 4 weeks
after termination of training was compared with their performance on the
immediate posttest (Table 2). On condensation questions, the small
difference observed (delayed posttest -- immediate posttest = 0.60) was
not significant, t (28) = 1.46, p > .68. Similarly, for factual
questions, the small difference observed (0.14) was not significant, t
(28) = 0.41, p > .68. Thus it can be inferred that the students
maintained their use of summarization skills, and this resulted in
maintenance of high performance on both kinds of comprehension
questions. Generalization of Summarization Skills
An alternate form of the Gates-MacGinitie Reading Test was
administered to the experimental group to check for generalization of
summarization skills to reading passages different from the controlled
materials used during the training. If generalization occurred, then
subjects' comprehension performance on Form 2 of the
Gates-MacGinitie Reading Test would be significantly greater than their
performance on Form 1 administered during the pretest phase. A t-test
for dependent samples indicated that the observed difference (9.73) on
the alternate forms of the Gates-MacGinitie was significant, t (28) =
5.22, p <.001. From these results, generalization of summarization
skills to the new reading material can be inferred.
These results indicate that students with learning disabilities can
be trained to use summarization rules, that the acquired skills are
maintained, and that spontaneous use of these rules is generalized.
These findings are consistent with previous research conducted with
student without disabilities. Several researchers have reported that
summarization training improved free recall (Taylor & Beach, 1984;
Wong et al., 1986) and cued recall (Armbruster, Anderson, &
Ostertag, 1987; Bean & Steenwyk, 1984; Palincsar, 1982; Rinehart et
al., 1986; Taylor, 1982; Taylor & Beach, 1984). Further, researchers
found that inducing students to write a sentence summarizing the meaning
of each paragraph in a text also facilitated recall compared to a
control reading condition (Bretzing & Kulhavy, 1979; Linden &
Though ample research supports summarization as a
comprehension-fostering strategy, relatively few researchers have
studied the differential effects of summarization training on different
types of questions. In the present study, summarization training was
effective for both condensation and factual questions. The performance
of the students with learning disabilities was comparable to that of
average readers on factual questions, but exceeded the performance of
average readers on condensation questions.
These findings closely parallel those of Palincsar (1982), who
found that explicit instruction in summarizing, questioning, clarifying,
and predicting enhanced comprehension scores for text-explicit,
text-implicit, and script-implicit questions. Moreover, she found the
greatest gain occurred for text-implicit questions. Because her
text-implicit questions are comparable with our condensation questions,
summarization's greater impact on text-implicit questions is also
As a final note, the resource room teachers reported that students
were using the instructed summarization rules spontaneously in different
content areas. Given these student gains, the extensive amount of time
required to train students with learning disabilities in a learning
strategy appears justified.
Alley, G., & Deshler, D. (1979). Teaching the learning disabled
adolescent: Strategies and methods. Denver, CO: Love. Armbruster, B. B.,
Anderson, T. H., & Ostertag, J. (1987). Does text
structure/summarization instruction facilitate learning from expository
text? Reading Research Quarterly, 22, 331-336. Bean, T. W., &
Steenwyk, F. L. (1984). The effect of three forms of summarization
instruction on sixth graders' summary writing and comprehension.
Journal of Reading of Behavior, 16, 297-307. Bloom, B. S. (1976). Human
characteristics and school learning. New York: McGraw-Hill. Bos, C. S.
& Filip, D. (1984). Comprehension monitoring in learning disabled
and average students. Journal of Learning Disabilities, 17, 229-233.
Bransford, J. D., Stein, B. S., & Vye, N. J. (1982). Helping
students learn how to learn from written texts. In M. Singer (Ed),
Competent reader disabled reader: Research and application (pp.
141-150). Hillsdale, NJ: Lawrence Erlbaum. Bretzing, B. B. &
Kulhavy, R. W. (1979). Notetaking and depth of processing. Contemporary
Educational Psychology, 4, 145-153. Brown, A. L., & Day, J. (1983).
Macrorules for summarizing texts: The development of expertise. Journal
of Verbal Learning and Verbal Behavior, 22, 1-14. Cicci, R. (1983).
Disorders of written language. In H. R. Myklebust (Ed.), Progress in
learning disabilities, (pp. 207-232). New York: Grune & Stratton.
Clark, F. L., Deshler, D. D., Schumaker, J. B., Alley, G. R., &
Warner, M. M. (1984). Visual imagery and self-questioning: Strategies to
improve comprehension of written material. Journal of Learning
Disabilities, 17, 145-149. DiVesta, F. J., Hayward, K. G., &
Orlando, V. P. (1979). Developmental trends in monitoring texts for
comprehension. Child Development, 50 97-105 Fry, E. (1977). Fry's
readability graph: Clarification, validity, and extension to level 17.
Journal of Reading, 21, 242-252. Garner, R., & Reis, R. (1981).
Monitoring and resolving comprehension obstacles: An investigation of
spontaneous text lookbacks among upper grade good and poor
comprehenders. Reading Research Quarterly, 16, 569-582. Graves, A. W.
(1986). Effects of direct instruction and metacomprehension training on
finding main ideas. Learning Disabilities Research, I(2), 92-100.
Guthrie, J. T. (1973). Reading comprehension and syntactic responses in
good and poor readers. Journal of Educational Psychology, 65, 294-299.
Hansen, C. L. (1978). Story retelling used with average and LD readers
as a measure of reading comprehension. Learning Disability Quarterly, 1,
62-69. Hare, V. C., & Borchardt, K. M. (1984). Direct instruction of
summarization skills. Reading Research Quarterly, 20, 62-78. Jenkins, J.
R., Heliotis, J., Haynes, M., & Beck, K. (1986). Does passive
learning account for disabled readers' comprehension deficits in
ordinary reading situations? Learning Disability Quarterly, 9, 69-76.
Kinstch, W., and van Dijk, T. A. (1978). Toward a model of text
comprehensionn and production. Psychological Review, 85, 363-394.
Linden, M., Wittrock, M. C., (1981). The teaching of reading
comprehension according to the model of generative learning. Reading
Research Quarterly, 17, 44-57. Lindsey, J. D., and Kerlin, M. (1979).
Learning disabilities and reading disorders: A brief review of the
secondary level literature. Journal of Learning Disabilities, 12,
408-415. MacArthur, C., & Graham, S. (1987). Learning disabled
students composing under three methods of text production: Handwriting,
word processing, and dictation. The Journal of Special Education, 21,
22-42. MacGinitie, W. H. (1978). Gates-MacGinitie Reading Tests (2nd
ed.). Boston: Houghton Mifflin. McCormick, C., & Samuels, S. J.,
(1979). Word recognition by second graders: The unit of perception and
interrelationships among accuracy, latency, and comprehension. The
Journal of Reading Behavior, 11, 107-118. McGee, L. M. (1982). Awareness
of text structure: Effects on children's recall of expository text.
Reading Research Quarterly, 17, 581-590. Palincsar, A. S. (1982).
Improving the reading comprehension of junior-high students through the
reciprocal teaching of comprehension monitoring strategies. Unpublished
doctoral dissertation, University of Illinois at Urbana-Champaign.
Paris, S. G., & Myers, M. (1981). Comprehension monitoring, memory,
and study strategies of good and poor readers. Journal of Reading
Behavior, 13, 5-22. Pearson, P. D. (1984). Direct explicit teaching of
reading comprehension. In G. G. Duffy, L. R. Roehler, & J. Mason
(Eds.), Comprehension instruction: Perspectives and suggestions (pp.
222-233). New York: Longman. Perfetti, C. A., and Hogaboam. T. (1975).
Relationship between single word decoding and reading comprehension
skills, Journal of Educational Psychology, 67, 461-469. Rinehart, S. D.,
Stahl, S. A., and Erickson, L. G. (1986). Some effects of summarization
training on reading and studying. Reading Research Quarterly, 21,
422-438. Rosenshine, B. (1986). Synthesis of research on explicit
teaching. Educational Leadership, 43, 60-69. Rosenshine, B., &
Stevens, R. (1984). Classroom instruction in reading. In P. D. Pearson
(Ed.), Handbook of reading research (pp. 645-798). New York: Longman.
Schumaker, J., B., Denton, P., & Deshler, D. D. (1984). The
paraphrasing strategy. Lawrence: The University of Kansas Press. Smiley,
S. S., Oakley, D. D., Worthen, D., Campione, J. C., & Brown, A. L.
(1977). Recall of thematically relevant material by adolescent good and
poor readers as a function of written vs. oral presentation. Journal of
Educational Psychology, 69, 381-387. Smith, P. L., & Friend, M.
(1986). Training LD adolescents in strategy for using text structure to
aid recall of instructional prose. Learning Disabilities Research, 2,
38-44. Spargo, E., Williston, G. R. & Browning, L. (1980). Timed
readings: Book one. Providence, RI: Jamestown Publishers. Taylor, B. M.
(1982). Text structure and children's comprehension and memory for
expository material. Journal of Educational Psychology, 74(3) 323-340.
Taylor B. M., & Beach, R. W. (1984). The effects of text structure
instruction on middle grade students' comprehension and production
of expository text. Reading Research Quarterly, 19, 135-146. Torgesen,
J. K., (1982). The learning disabled child as an inactive learner.
Topics in Learning and Learning Disabilities, 2, 45-52. Wong, B. Y. L.,
& Jones, W. (1982) Increasing metacomprehension in learning disabled
and normally achieving students through self-questioning training.
Learning Disability Quarterly, 5, 228-240. Wong, B. Y. L., Wong, R.,
Perry, N., & Sawatsky, D. (1986). The efficacy of a self-questioning
summarization strategy for use by underachievers and learning disabled
adolescents in social studies. Learning Disability Focus, 2, 20-35.
Worden, P. E., & Nakamura, G. V. (1983). Story comprehension and
recall in learning disabled vs. normal college students. Journal
Educational Psychology, 74, 633-641. MEENAKSHIGAJRIA (CEC NY Federation)
is an Assistant Professor of Special Education in the Division of
Teacher Education at St. Thomas Aquinas College, Sparkill, New York.
JOHN SALVIA (CEC #405) is a Professor of Special Education and the Head
of the Department of Educational and School Psychology and Special
Education at Pennsylvania State University, University Park.
Means and Standard Deviations for Chronological Age, Grade Placement, Intelligen
ce Quotients, and
Reading Performance for Students With Learning Disabilities
Experimental Group Control Group
(n = 15) (n = 15)
Information Mean (SD) Mean (SD)
CA 14.33 (1.11) 14.70 (1.58)
Grade 7.53 (1.93) 7.20 (1.04)
FSIQ 94.13 (8.95) 95.53 (11.13)
VIQ 88.67 (10.74) 90.27 (11.87)
PIQ 100.73 (10.23) 98.07 (10.95)
NCE 28.00 (4.18) 30.02 (3.90)
ORA (%) 92.53 (1.89) 92.80 (1.83)
Note: CA = Chronological Age; FSIQ = Full Scale IQ;
VIQ = Verbal Scale IQ; PIQ = Performance Scale IQ;
NCE = Normal Curve Equipment Score on Gates-MacGinitie
Comprehension Subtest; ORA = Oral Reading