Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

218

ABSTRACT

Reports from research and the larger educational community demonstrate that too many students have limited ability to

comprehend texts. The research reported here involved a two-year study in which standardized comprehension instruc-

tion for representations of two major approaches was designed and implemented. The effectiveness of the two experi-

mental comprehension instructional approaches (content and strategies) and a control approach were compared. Content

instruction focused student attention on the content of the text through open, meaning-based questions about the text.

In strategies instruction, students were taught specific procedures to guide their access to text during reading of the text.

Lessons for the control approach were developed using questions available in the teacher’s edition of the basal reading

program used in the participating classrooms. Student participants were all fifth graders in a low-performing urban district.

In addition to assessments of comprehension of lesson texts and an analysis of lesson discourse, three assessments were

developed to compare student ability to transfer knowledge gained. The results were consistent from Year 1 to Year 2. No

differences were seen on one measure of lesson-text comprehension, the sentence verification technique. However, for

narrative recall and expository learning probes, content students outperformed strategies students, and occasionally, the

basal control students outperformed strategies students. For one of the transfer assessments, there was a modest effect in

favor of the content students. Transcripts of the lessons were examined, and differences in amount of talk about the text

and length of student response also favored the content approach.

he pursuit of more precise understandings of

comprehension instruction has been an active,

ongoing area of research, at least since Dolores

Durkin’s (1978–1979) well-known criticism that very

little went on in classrooms that could be called com-

prehension instruction. The research reported here ad-

dresses the need for more precise understandings of

present-day comprehension instruction through the

implementation of standardized lessons on common

texts for two approaches to comprehension instruc-

tion, a strategies approach and a content approach, and

a comparison of their effects. The strategies approach

centers on the direct teaching of specific procedures,

such as summarizing, making inferences, and gener-

ating questions, and using them in working with text.

The other approach to comprehension, which we have

labeled a content approach, focuses on keeping students’

attention directed toward the content of what they are

reading and working through the text to build a repre-

sentation of the ideas through discussion.

Given that comprehension is such a complex cog-

nitive endeavor and is affected by, at least, the reader,

the text, and the context, comprehension research has

considered many features as contributing to student

outcomes. Here we will provide a glimpse into some

major areas of comprehension instructional research

as a way of illustrating the ancestry of the approaches

we are investigating. The framework that we will use

to provide this glimpse is the traditional perspective of

before, during, and after reading.

Considerations of what activities should happen be-

fore reading have centered on upgrading background

knowledge as a way to support students as they read.

Studies have examined background knowledge both in

terms of how it functions (Anderson & Pearson, 1984;

Johnson-Laird, 1983; Van Dijk & Kintsch, 1983) and

Rethinking Reading Comprehension

Instruction: A Comparison of

Instruction for Strategies and Content

Approaches

Margaret G. McKeown, Isabel L. Beck, Ronette G.K. Blake

University of Pittsburgh, Pennsylvania, USA

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

219

how it can be upgraded. The majority of studies that

have examined effects of background knowledge has

demonstrated that upgrading can enhance students’

compre hen sion. T hi s re su lt h as b ee n fou nd w it h pr im ar y-

grade children (Beck, Omanson, & McKeown, 1982;

Pearson, Hansen, & Gordon, 1979), intermediate-grade

students (McKeown, Beck, Sinatra, & Loxterman,

1992), middle school students (Graves, Cooke, &

LaBerge, 1983), and high school students (Hood, 1981).

Two other prominent facets that have been pursued in

the before-reading arena are purpose for reading and

instruction of vocabulary to be encountered in an up-

coming selection. Research on providing students, or

asking students to develop, a purpose for reading is

sparse, but the studies that have been done show posi-

tive effects (Tierney & Cunningham, 1984). Despite the

scant amount of research, purpose for reading has be-

come an ingrained practice in conventional reading les-

sons. Teaching vocabulary can enhance comprehension

of text if the kind of instruction provided helps students

build meaningful associations to their knowledge base

and more than a brief definition is provided (Baumann,

Kame’enui, & Ash, 2003).

Before considering during-reading activities, which

are the focus of this paper, we will touch on after-

reading activities. After-reading activities often involve

one or another form of questioning. After Anderson

and Biddle (1975) reported that questions asked after

reading yielded better comprehension than merely re-

viewing the text did, attention turned to investigating

which kinds of questions were most effective. Examples

of questions singled out for being more effective include

those that ask about the most important text informa-

tion (Rickards, 1976), application questions (Rickards

& Hatcher, 1977–1978), and high-level questions (Yost,

Avila, & Vexler, 1977). Beck and McKeown (1981) de-

veloped a procedure to help teachers create questions

based on the most important information and the se-

quence of that information throughout the text.

Another avenue explored in developing effective

after-reading activities is interpretive discussion. Such

discussions typically focus on prompting students to

respond to a “big question” that arises from the text,

with an eye toward fostering critical-reflective think-

ing about text ideas (Wilkinson, Soter, & Murphy, in

press). Approaches to interpretive discussion that are

backed by evidence of success include Junior Great

Books (Great Books Foundation, 1987), Collaborative

Reasoning (Anderson, Chinn, Waggoner, & Nguyen,

1998), Philosophy for Children (Sharp, 1995), and

Grand Conversations (Eeds & Wells, 1989).

During-reading interventions emerged as efforts to

influence readers’ ongoing interactions with text. The

historical roots of during-reading interventions are

manifested in studies of inserted questions. Studies by

Watts and Anderson (1971) and Rothkopf (1966, 1972)

suggested that when students respond to questions

during reading, their understanding of the text is stron-

ger than it is if they simply read the text. Tierney and

Cunningham (1984), in their review of comprehension

instruction, suggested that deeper understanding of

during-reading questions was a worthy avenue to pur-

sue but that it needed to be tied to models of the text, of

the reader, or of mental processes.

Theoretical Foundations

and Current Status of Strategies

and Content Approaches

The development of the two approaches used in the re-

search reported here—strategies and content—came in

response to considering models of mental processing

as suggestions of ways to intervene for comprehension

development. Specifically, as is discussed in the next

section, the strategies approach developed from models

of thinking and learning processes and the content ap-

proach from models of text processing. A crucial impli-

cation of processing models is that learners need to be

mentally active to process text successfully. A common

feature of both the strategies and content approaches is

that they aim to engender active student engagement

with reading.

A major distinction between the two approaches is

that strategy instruction encourages students to think

about their mental processes and, on that basis, to ex-

ecute specific strategies with which to interact with

text. In contrast, content instruction attempts to engage

students in the process of attending to text ideas and

building a mental representation of the ideas, with no

direction to consider specific mental processes.

Strategies

The notion of providing instruction in strategies, rou-

tines for dealing with text, arose from work in devel-

opmental psychology that had established the active,

strategic nature of learning that developed as children

matured. Studies were then conducted that taught strat-

egies for general learning tasks. The strategies used in

such studies included rehearsal, categorization, and

elaboration (Brown, Bransford, Ferrara, & Campione,

1983).

Based on the developmental work, Brown and her

colleagues investigated the extent to which students

used various strategies for studying, such as note-

taking and underlining (Brown, 1981, 1982b; Brown

& Smiley, 1977). From their work, Brown and her col-

leagues surmised that it might be possible to improve

comprehension of young children or less-able learners

Reading Research Quarterly • 44(3)

220

by teaching them effective study strategies (Brown &

Smiley, 1978). The eventual manifestation of this line of

work in relation to reading was reciprocal teaching, an

approach that taught young students to apply strategies

of summarizing, questioning, clarifying, and predicting

(Palincsar & Brown, 1984).

Strategies instruction also finds roots in models of

thinking. Symons, Snyder, Cariglia-Bull, and Pressley

(1989) traced notions of strategy teaching to the theo-

ries of Baron (1985) and Sternberg (1979, 1982), both of

whom emphasized that during the process of thinking

in problem-solving, competent thinkers employ strat-

egies such as identifying their goal, monitoring their

progress, and evaluating evidence. The reasoning that

followed was that providing young students with some

procedures they could employ while reading could fa-

cilitate their comprehension. These roots led Pressley et

al. (1992) to develop transactional strategies instruction,

an approach in which the teacher explains and models

strategies and uses these strategies to guide dialogue

about text.

In addition to reciprocal teaching (Palincsar &

Brown, 1984) and transactional strategies (Pressley et

al., 1992), programs of strategy instruction by Paris,

Cross, and Lipson (1984) and Duffy and Roehler (1989)

have also had a sustained impact on the area of strate-

gies instruction. Paris et al. focused their instructional

approach, informed strategies for learning, on develop-

ing awareness of the goals of reading and the value of

using strategies to pursue those goals. Their approach

is designed to teach students to evaluate, plan, and reg-

ulate as they build awareness of their processing. The

strategies taught in the informed strategies for learning

approach include understanding the purposes of read-

ing, activating background knowledge, allocating at-

tention to main ideas, evaluating critically, monitoring

comprehension, and drawing inferences. The work of

Duffy (1987) emphasized self-regulation and self-mon-

itoring, focusing on using strategies to remove block-

ages to comprehension. They emphasized the role of

direct explanation of strategies and the role of explicit

modeling in the instruction. Beyond these influential

programmatic approaches to strategies instruction, the

strategies literature comprises additional strategies pro-

grams (see, e.g., Anderson & Roit, 1993; Block, 1993;

Klingner, Vaughn, & Schumm, 1998) and numerous

smaller studies on a variety of strategies—implemented

both individually and in combination.

The most prominent review of the strategies lit-

erature was part of the National Reading Panel (NRP)

report (National Institute of Child Health and Human

Development [NICHD], 2000) that concluded that “the

past two decades of research appear to support the

enthusiastic advocacy of instruction of reading strate-

gies” (p. 4-46). The report identifies seven individual

strategies that the panel found to be supported by solid

evidence for improving comprehension: comprehension

monitoring, cooperative learning, graphic and semantic

organizers, question answering, question generation,

story structure, and summarization. The report sum-

marizes the effectiveness of the studies in each area,

providing a picture of overall success with using the

strategies.

Content

Although models of thinking and general learning

underlie strategies instruction, models developed to

explain specifically how a reader processes text (see,

e.g., Graesser, Singer, & Trabasso, 1994; Kintsch, 1974;

Trabasso, Secco, & van den Broek, 1984; van den Broek,

Young, Tzeng, & Linderholm, 1998) are the roots of a

content approach to comprehension. Text-processing

models take the perspective that the mental process-

es in reading focus on the development of coherence

based on organizing the meaningful elements of the

text. From a text-processing perspective, a reader moves

through text identifying each new piece of text informa-

tion and deciding how it relates to information already

given and to background knowledge (see Kintsch & van

Dijk, 1978). The focus is on what readers do with text

information to represent it and integrate it into a coher-

ent whole.

A text-processing perspective on comprehension

suggests that a content orientation may be a productive

direction for instructional intervention. That is, com-

prehension enhancement might derive from a focus on

continually striving for meaning as reading of the text

moves along rather than considering when and how to

call up specific routines to deal with new information. A

number of researchers have speculated that such an ap-

proach might be an alternative to direct strategies teach-

ing (Baker, 2002; Carver, 1987; Dole, Duffy, Roehler,

& Pearson, 1991; Gersten, Fuchs, Williams, & Baker,

2001; Kucan & Beck, 1997; Pearson & Fielding, 1991).

The body of research on content-focused approaches

is smaller than that of strategies approaches, and ques-

tioning the author (QtA; Beck & McKeown, 2006; Beck,

McKeown, Sandora, Kucan, & Worthy, 1996) may be

the approach that is most explicitly oriented toward a

text-processing view. The text-processing approach that

QtA connects with most directly is Kintsch’s (1998)

construction–integration model, in which there are two

phases: the construction phase, in which readers acti-

vate textual information, and the integration phase, in

which the activated ideas are integrated.

Other approaches that center on meaningful talk

about a text include in their orientation a sociocogni-

tive perspective, where the group is seen as forming an

interpretive community that jointly constructs mean-

ing. Such approaches generally fall under the label

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

221

collaborative discussion and typically initiate discussion

by focusing on a theme from the text or an issue-related

question, such as a question about a character’s motives.

These approaches include instructional conversations

(Saunders & Goldenberg, 1999), collaborative reason-

ing (Chinn, Anderson, & Waggoner, 2001), dialogic

instruction (Nystrand, 1997), and Junior Great Books

(Dennis & Moldof, 1983).

Summarizing approaches that focus on meaning-

ful talk about text, Applebee, Langer, Nystrand, and

Gamoran (2003) noted that although the approaches

have different vocabulary and routines, the form and

focus of the interventions significantly overlap, and

“results converge to suggest that comprehension of dif-

ficult text can be significantly enhanced by replacing

traditional I-R-E [Initiation-Response-Evaluation] pat-

terns of instruction with discussion-based activities”

(Applebee et al., 2003, p. 693). Other researchers have

found that discussion around text can promote problem-

solving, comprehension, and learning (Anderson et

al., 1998; Nystrand, 1997; Wegerif, Mercer, & Dawes,

1999). Discussion that leads to such outcomes fea-

tures open questions, student control of interpretive

authority, more student than teacher talk, and teacher

responses that are based on students’ responses (see,

e.g., Beck et al., 1996; Chinn, O’Donnell, & Jinks,

2000; Nystrand, Wu, Gamoran, Zeiser, & Long, 2003).

Thus, there appears to be a convergence of results that

discussion-based practices are effective for comprehen-

sion improvement, similar to the convergence of results

on the effectiveness of strategies instruction.

What We Still Don’t Know About Both

Approaches

An issue with both approaches is that there is still not

clear guidance on how to proceed instructionally. In

the case of strategies instruction, it seems that to make

instructional decisions we need to know which strate-

gies to use, how they should be taught, and how they

should be used in the course of reading. The available

research leaves all these factors in doubt. In terms of

which strategies are key, a large number of candidates

have been identified, with the National Reading Council

report (Snow, Burns, & Griffin, 1998) and the NRP re-

port (NICHD, 2000) listing overlapping but different

sets of strategies that claim effectiveness. The more re-

cent work on strategies has advocated the instruction

of multiple strategies and the flexible coordination of

them. But which set of strategies should go into the mix

is not clear.

How strategies should be taught is not easily de-

rived from the research. One problem here is that what

goes on under a strategy label is not consistent from

study to study. In some cases, the same strategy label is

given to very different sets of activities. Our analysis of

the 18 studies highlighted in the NRP report (NICHD,

2000) as having shown positive results from instruc-

tion in summarizing showed a variety of tasks and

activities. For instance, in one summarization study,

students were taught steps for creating a summary, in-

cluding the following: (1) select main information, (2)

delete trivial information, and (3) relate to supporting

information (Rinehart, Stahl, & Erickson, 1986). In an-

other summarization study, students were asked direct

questions about the literal content of the text for the

purpose of leading students to draw an inference about

a character’s actions (Carnine, Kame’enui, & Woolfson,

1982).

A similarly confounded picture emerges from ex-

amination of the studies labeled comprehension moni-

toring. A study in that category by Schmitt (1988)

instructed students in activating prior knowledge, set-

ting purposes, generating and answering prequestions,

forming hypotheses, verifying or rejecting hypotheses,

evaluating predictions, and summarizing. In contrast,

a study by Miller (1985) in the same category included

teaching students to ask themselves questions as they

read, such as “Is there anything wrong with the story?”

and to underline problems they found. Thus, not only

do activities under comprehension monitoring vary

widely but studies also include activities that are the

domain of another strategy, such as summarizing and

asking questions.

The foregoing discussion casts no aspersions on

the quality of the activities used within the studies but

rather points out that the sum total of studies leaves

us without a consistent picture of which strategies are

effective and what is effective about them. Typically, re-

search in content approaches to comprehension provide

only general directives on how students were brought

into discussion or how teachers learned to question and

to respond to students’ contributions in ways that were

productive toward building a coherent representation. It

may be that practice focusing on what is important and

making connections initiates readers’ mental engage-

ment with such strategies as summarization and infer-

ence, but they are not dealt with explicitly. Graesser

(2007) suggested that in the construction–integration

model, “strategies exist but they do not drive the com-

prehension engine. Instead, the front seat of compre-

hension lies in the bottom-up activation” (p. 11) of text

ideas and the integration of those ideas.

Outcomes measured in content approaches include

quality of discussion and comprehension of a story that

students had read and discussed and only rarely in-

vestigate subsequent achievement. For one content ap-

proach, QtA, although the focus of research has been

on changes in classroom discourse, three studies have

also included investigations of achievement (Beck et

Reading Research Quarterly • 44(3)

222

al., 1996; Beck & McKeown, 1998; Sandora, Beck, &

McKeown, 1999). Sandora et al. compared the effects

on students’ comprehension of QtA and Junior Great

Books and found that students in the QtA approach had

greater recall and higher scores on answers to interpre-

tive questions than did those in the Great Books group.

The Beck et al. and Beck and McKeown studies included

an individually administered comprehension task on a

novel text passage that measured growth in comprehen-

sion monitoring and comprehension of the text. Both

the Beck et al. and Beck and McKeown studies showed

advantages for QtA students; in Beck et al., students im-

proved in monitoring, and in Beck and McKeown, both

monitoring and comprehension increased. However, the

results are limited. In the Beck et al. study, there was no

control group—students were given a pretest and post-

test. In Beck and McKeown, fifth and sixth graders were

compared with business-as-usual control students from

the same school. QtA fifth graders showed improvement

in monitoring, and sixth graders showed improvement

in both monitoring and comprehension. But because

of the transitory nature of the school population, there

were only 11 students in each approach in fifth grade

and 13 in each approach in sixth grade.

Thus, the findings about a content-oriented ap-

proach are limited by the number of studies, the small

sample size of those studies, and the sparse evidence

of improvement on tasks beyond the classroom dis-

cussions. This last characteristic is the opposite of a

drawback of the strategies findings. That is, for QtA

and discussion-based approaches, the results center on

interactions during discussion, whereas in strategies re-

search, there are few examples of what the classroom

interactions were like and thus little insight about what

led to the results.

Rationale for the Study

A limitation of the evidence for both strategies and

content approaches is that, with the exception of the

Sandora et al. (1999) study, the approaches have been

compared with only traditional instruction or unknown

instruction—whatever happened to be going on in

those other classrooms at the time. Little information

is available on what the instruction or texts were like

in the comparison groups. Additionally, strategies and

content approaches have never been compared with

each other.

Most importantly, studies of both types of ap-

proaches suffer from lack of standardization of what

teachers tell students, what students do, and how the

interactions proceed. In comparison, in research on

lower-level reading processes, the evidence for the role

of phonics was not simply that “phonics instruction is

good” but that, to be good, instruction needed to be

systematic and sequential and include activities such as

blending (NICHD, 2000). Of course, teacher variation

plays a large role even with prepared materials, but in

the case of strategies and content approaches, the teach-

er is the only instructional agent. Thus, the research on

strategies and content approaches provides little guid-

ance on what in the instruction was responsible for the

outcomes. It could be the case that simply more time

and attention to text is the key that leads to improve-

ment (Carver, 1987), and the instructional activities and

prompts do not matter. We doubt that is the case; it is

more likely that some activities are more effective than

are others.

Overview of the Study

The research reported here involved a two-year study

in which we developed, implemented, and compared

standardized instruction for representations of two

major approaches targeted to enhance comprehension.

Student comprehension of common texts was examined

under the following approaches: strategies instruction,

content instruction, and basal instruction. Lessons were

designed around texts being used in the classrooms,

and thus students in all approaches dealt with the same

texts. These texts were the five core selections that

made up one theme within the basal reader in use in the

school. Each selection was the core text in a weekly les-

son plan that also included reading of another shorter

text and activities in fluency, writing, vocabulary, word

study, self-selected reading, and other typical language

arts practices.

The part of the five-day lesson plan labeled guided

comprehension consisted of reading the core text with

program-provided, scripted teacher questions. In this

study, we used the guided comprehension slice of the

weekly lesson plan to implement the lessons that we de-

veloped. Each lesson occupied a total of 45 to 75 minutes

a week within one of the daily 90-minute reading blocks.

During the other reading periods over the week, the

teachers engaged in whatever other activities they chose

from the basal reader, in accordance with their customary

classroom practice. Thus, this study was situated within

authentic contexts of classroom reading instruction.

In the first year of the study, this comprised lessons

for five narrative selections. For the second year of the

study, the same lessons for the five narrative selections

were implemented with a new cohort of students and

three expository text lessons were added. We hypothe-

sized that this longer period of instruction might make it

feasible to evaluate the effects of the instruction on com-

prehension of texts beyond those used in the classroom.

Thus a task for assessing transfer was developed.

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

223

The goal of this work is a greater understanding of

the extent to which instruction focused on content or

on strategies enhances students’ comprehension of spe-

cific texts and their independent comprehension abil-

ity. Findings from the study have the potential to yield

greater understanding because the use of standardized

lessons on common texts that represent both approach-

es allows a rather precise picture of what students do

during instruction and are able to do following instruc-

tion. Transcripts of videotaped lessons within the two

approaches afford opportunities to identify the kinds

of interactions that seem to be the most productive for

enhancing students’ abilities to deal with text ideas.

Investigation of comprehension of texts beyond those

used in instruction provides insight about the potential

long-term effects of the approaches.

Year 1

Method

The purpose of the study was to compare the effective-

ness of two experimental comprehension instructional

approaches and a control approach. One experimental

approach, the content approach, centered on readers’ de-

velopment of a coherent representation of the text. The

other experimental approach, the strategies approach,

involved direct teaching of explicit comprehension strat-

egies. The control approach, the basal approach, was

based on the instruction presented in the 2003 basal

reader that was used in the school district.

Materials

Lessons for each of the three approaches were devel-

oped based on 5 fifth-grade narrative text selections

from the basal reading program in use. The titles of the

five stories and a brief description of each story are pro-

vided in Appendix A.

Lesson Design

The format in each approach was teacher-directed,

whole-class instruction in which portions of the text

were read aloud, mostly by students, and were inter-

rupted with questions that elicited student discussion.

This is in keeping with customary guided comprehen-

sion provided in basal programs to the extent that ques-

tions are scripted for the teacher to deliver, and their

placement is designated within the reading selection.

The number of stops—points at which reading was

stopped and discussion initiated—was similar for each

of the two experimental approaches. For the basal les-

sons, we used questions supplied in the teacher’s edi-

tion and their associated stopping points. Thus all three

approaches included questions interspersed within the

reading. The design of the questions for initiating and

guiding discussion in each approach is discussed in the

following sections. An example of a script for each ap-

proach is presented in Appendix B.

Content Approach

Content instruction focused student attention on the

content of the text through general, meaning-based

questions about the text. Reading was stopped and dis-

cussion initiated at purposely selected points when, for

example, a key character was introduced, some impor-

tant event had occurred, or where we judged some con-

fusion might arise for readers. This method is based on

the QtA approach (Beck & McKeown, 2006; Beck et al.,

1996). At each stopping point, the teacher provided an

initiating question to start the discussion. These open-

ended questions were designed to make public the im-

portant idea in a text segment (e.g., “What’s going on

here?” “How does all this connect with what we read

earlier?”). Most stops also included a follow-up question

that the teacher could use to focus on the key ideas from

that portion of text. These follow-ups (e.g., “Can any-

one add to that?” “What does that mean?”) were used as

needed to elicit additional information.

Strategies Approach

In strategies instruction, students were taught to use

specific procedures to guide their access to text during

reading. An important notion behind teaching strate-

gies is that strategies instruction is productive because

it can be applied to any piece of text. At stopping points,

the teacher used a strategy to prompt discussion and

reminded students how to apply the strategy.

Strategies Selection

We chose the strategies to use for the present study af-

ter considering those highlighted in two recent reports

on reading, the National Reading Council report (Snow

et al., 1998) and the NRP report (NICHD, 2000). The

National Reading Council report focuses on the fol-

lowing strategies: summarizing, predicting, drawing

inferences, and monitoring for coherence and misun-

derstandings. The NRP report lists seven individual

strategies for which there was evidence of improved

reading comprehension after instruction: comprehen-

sion monitoring, summarization, question generation,

question answering, cooperative learning, story struc-

ture, and graphic and semantic organizers.

In selecting among recommended strategies, we

considered which procedures might most naturally be

called on as a reader works through a text to understand

the content and also which strategies were most appro-

priate for a discussion-based lesson. Our thinking was

that readers may summarize important information as

Reading Research Quarterly • 44(3)

224

they move through text, develop a sense of what may

be coming next, draw inferences to create connec-

tions, and form questions to check that they are on

track. Additionally, effective readers monitor their un-

derstanding. We therefore selected summarizing, pre-

dicting, drawing inferences, question generation, and

comprehension monitoring for strategy-based lessons.

The strategy to be used at each stopping point was

selected depending on which strategy fit best at that

point in the text (e.g., if the story reached a turning

point, we would ask the students to predict, or if the

story presented confusing information, we asked the

students to check their understanding).

Expert Review

Because our expertise is in content instruction, to de-

velop the strategy lessons, an intensive process of les-

son review and revision was undertaken. We designed

a three-stage review process that elicited feedback from

strategies experts on how well our strategy lessons rep-

resented strategies instruction as understood within the

reading research community and from teachers on is-

sues of classroom implementation.

The first review cycle involved three strategies ex-

perts from the national research community. This cycle

was designed to elicit extensive open-ended comments

on strategy selection and early drafts of the lessons.

Based on the comments, lessons were revised. The sec-

ond cycle involved 10 reading specialists/coaches. This

stage was designed to elicit reading practitioners’ evalu-

ation of ease of lesson use. The third cycle involved six

additional nationally recognized strategies experts. This

stage was designed to elicit more structured responses

from a larger group of strategies researchers on our most

complete version of the lessons.

Thus, in the first stage of review, the experts pro-

vided comments on our initial strategy lessons in terms

of usability and fidelity to strategies instruction. In the

second and third stages, reviewers were asked to rate

strategies instruction in the lessons on the following six

dimensions, as well as to provide written comments:

1. Coherent, student-friendly language

2. Adequate representation of the strategy

3. Teachers’ ability to instruct the strategy

4. Students’ ability to follow teacher directives

5. Adequate practice using the strategy

6. Appropriate student scaffolding toward

independent strategy use

Reviewers were asked to rate, on a 4-point scale,

how well each of the five implemented strategies (e.g.,

summarize) was operationalized across each of the six

dimensions, providing 30 ratings from each reviewer.

The scale was as follows: 4 = almost always, 3 = usu-

ally, 2 = sometimes, 1 = does not occur. The average

rating from the reading practitioners across the six di-

mensions for the five strategies (second stage of review)

was 3.73. The six nationally recognized strategies ex-

perts (third stage of review) were more critical (as was

expected). Their ratings averaged 3.26. See Table 1 for

means and standard deviations of the ratings from read-

ing practitioners and national experts for each of the

six dimensions. The experts provided specific feedback

about the lesson, noting that if their suggestions were

incorporated their ratings would be higher. Lessons

were revised accordingly.

Basal Approach

For the lessons in the basal approach, we developed

scripts based on the questions in the teacher’s edition

of the basal reading program used in the participating

classrooms. Because so much material is presented in

the teacher’s edition, we included only questions that

were presented for use during reading of the selection

Dimension

Reading practitioners

M (SD)

Strategy experts

M (SD)

Coherent language 3.65 (0.41) 3.80 (0.40)

Adequate strategy representation 3.85 (0.35) 3.37 (0.69)

Teachers’ ability to instruct 3.92 (0.21) 3.13 (0.79)

Students’ ability to follow directives 3.60 (0.41) 3.17 (0.67)

Adequate strategy practice 3.73 (0.30) 3.25 (0.81)

Appropriate scaffolding 3.63 (0.52) 2.83 (0.58)

Table 1. Practitioner and Expert Ratings of Strategy Lessons by Dimension on a 4-Point Scale

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

225

and limited the questions to those directed to compre-

hension. For example, we omitted questions about word

recognition, grammar, and so on, but asked the ques-

tions about text indicated at the stopping points in the

teacher’s edition. Thus the approach we created was not

a business-as-usual control and thus was more accu-

rately called basal comprehension because we may have

provided a stronger focus on comprehension during the

reading of the basal’s major weekly text.

Participants

Students

Student participants were all fifth graders from six intact

classrooms who took reading in their regular classroom

in one school in a small, urban district in southwestern

Pennsylvania. The district in which the study took place

had been under court-ordered desegregation from 1981

until 2000. In 2004–2005, the school was identified as

“in need of improvement” by the Pennsylvania System

of School Assessment, with 48% of the fifth graders

scoring at basic or below basic in reading according to

the 2005 assessment. A majority of the students were

African American (58%) and about half of the students

(49%) qualified for free or reduced-price lunch.

To the best of our knowledge, the students’ reading ex-

periences had been through commercial basal programs.

We base this on knowledge of the curriculum in the dis-

trict, mobility patterns in the district, and the curriculum

in adjoining districts. First, there was mobility among the

student population within the district, but the same basal

program was used throughout the district. Second, there

was some mobility with two adjoining districts, a large ur-

ban district and a small, high-poverty urban district. Both

of these districts also had basal programs in place.

Assignment to Approach

Three sets of test results were used to determine com-

parability of the six classrooms and to assign two class-

rooms to each of the three instructional approaches.

The first were the Terra Nova reading comprehension

(McGraw-Hill, 2000) scores for 132 students who had

taken the test in the spring of their fourth-grade year

and were expected to enter fifth grade in the fall. After

we removed from the study pool students who trans-

ferred out of the school before fifth grade began and

special education students who did not take reading

with their regular class and added 13 students who

transferred into the school to begin fifth grade, we had

a final sample size of 119 students, with Terra Nova

reading comprehension scores for 105 of these stu-

dents. We then administered two Woodcock-Johnson

(Woodcock, Mather, & Schrank, 2004) decoding tests

(Word Identification and Word Attack) to these 119

students.

To evaluate whether there were significant differ-

ences among the classes in reading ability, one-way

analyses of variance (ANOVAs) were done for the two

Woodcock-Johnson tests and the Terra Nova test. There

was not a significant difference among the six classes for

the Word Identification test, F(5, 111) = 1.39, p = .234,

or for the Word Attack test, F(5, 111) = 1.64, p = .156.

For the Terra Nova scores, the ANOVA indicated that

classes were not equivalent, F(5, 100) = 2.64, p < .03.

Post-hoc tests revealed that one classroom’s scores dif-

fered from two other classrooms at a p = .08 level of

significance. See Table 2 for the pretest scores for all

individual measures.

To assign the six classes (and the corresponding

teachers) to the three instructional approaches and

to make them as comparable as possible, we ranked

Approach Class #

Word Identification test

M (SD)

Word Attack test

M (SD)

Reading comprehension

M (SD)

Standard score

M (SD)

Content 1 4.95 (1.33) 5.99 (3.40) 8.84 (8.13)

2 5.08 (1.67) 5.54 (3.87) 5.40 (3.61)

Overall 5.02 (1.51) 5.75 (3.63) 6.73 (5.93) 0.25 (2.70)

Strategies 3 6.13 (2.71) 8.04 (5.25) 9.40 (5.70)

4 4.72 (1.53) 5.41 (3.95) 6.05 (3.01)

Overall 5.38 (2.25) 6.65 (4.73) 7.68 (4.76) 0.39 (2.64)

Basal comprehension 5 5.60 (1.93) 6.96 (4.79) 6.85 (3.15)

6 5.04 (1.91) 4.58 (2.94) 5.07 (2.28)

Overall 5.33 (1.91) 5.81 (4.12) 5.96 (2.85) 0.10 (2.57)

Table 2. Mean (SD) Scores for Three Baseline Assessments and Overall Standard Scores of Reading Ability for Each

Approach for Year 1

Reading Research Quarterly • 44(3)

226

the classrooms from highest (Classroom 1) to lowest

(Classroom 6) according to the mean classroom scores

on the Terra Nova and Woodcock-Johnson tests. We

also observed the classrooms to get a sense of class-

room dynamic (e.g., student behavior, teacher discipline

style), and these factors were considered when yoking

the classrooms. In particular, when the classes were

close on the other criteria, we took into account class-

room dynamic, especially student behavior. Given that

three classrooms exhibited a relatively high number of

student disruptions, we included one of these class-

rooms in each approach.

To test for comparability across the three pairs of

classrooms for the two Woodcock-Johnson subtests and

the Terra Nova assessment, we combined the three scores

into one standard score by converting each individual

score from each of the three tests into a z score and then

summing the three z scores. After a standard score was

computed for each student, an ANOVA was conducted.

Standard scores by approach are shown in Table 2 on

page 224. No significant differences among the three ap-

proaches were found, F(2, 116) = 0.620, p = .540.

Teachers

All six fifth-grade classroom teachers and three sup-

port teachers were invited to participate in the study

and all agreed to do so. The nine teachers each received

a $300 stipend for their participation, which included

a half-day initial training, three after-school meetings,

preparation time for conducting the lessons, e-mail cor-

respondence with research staff, and exit interviews.

Each of the classroom teachers was assigned to ap-

proach by virtue of the class he or she taught, per the

process described above. Each pair of classes (and the

corresponding teachers) was randomly assigned to one

of the three approaches. The three supporting teachers

(two Title I teachers and one special education inclusion

teacher) who would conduct make-up lessons were also

assigned to approaches. One teacher, the special educa-

tion inclusion teacher, had already been assigned to one

of the classrooms by the school district so she had to

be kept with that classroom and the other in the pair.

The other two support teachers were not previously as-

signed, so they were randomly assigned to the remain-

ing two approaches.

Eight teachers were women (one was a man); all were

European American. The number of years of teaching

experience ranged from 2 years to 23 years. Specifically,

the two teachers in the content approach had 2 years’

and 18 years’ experience; the two teachers in the strate-

gies approach had 15 years’ and 23 years’ experience,

and the two teachers in the basal-comprehension ap-

proach had 5 years’ and 21 years’ experience. All teach-

ers had spent their entire teaching careers in the district

in which the study occurred.

As mentioned, a basal reading program was in use

in the district and basal readers had been the core of

the reading curriculum for decades. However, to get a

sense of how the teachers implemented the guided com-

prehension portion of the weekly reading lesson, two

members of the research staff observed a guided read-

ing lesson for each teacher. The observations focused

on the kinds of questions teachers asked, the extent to

which they engaged students in discussion, and how

they dealt with student responses. The observations in-

dicated the following:

• Five of the six teachers used predominately ques-

tions from the basal. The sixth teacher did not

use the basal questions but asked questions that

elicited student experiences and inserted a good

many of his own experiences that were tangen-

tially related to the text (this teacher was assigned

to the content approach).

• Four of the five teachers that used basal questions

conducted the lesson mainly through an Initiate-

Respond-Evaluate pattern. That is, they posed the

basal question; accepted a brief, mostly literal, re-

sponse from a student; and went on.

• The fifth basal questioning teacher frequently

followed up student responses by elaborating on

them or probing students for more information

(this teacher was assigned to the basal approach).

The points above indicated that in four of the six

classrooms, there was essentially no discussion. In the

classroom in which the teacher focused on students’ ex-

periences, the discourse was a collection of student ex-

periences but was not a discussion, in that there was no

building of relationships among student comments. In

the remaining classroom, the teacher’s follow-up of stu-

dent responses provided some threads of connection,

and thus the discourse had elements of a discussion, in

contrast to the Initiate-Respond-Evaluate pattern.

Measures

We used a variety of measures to assess the outcomes

of the three approaches. There were two categories of

assessments: lesson-text comprehension and beyond-

lesson-text-assessments. The lesson-text comprehen-

sion assessments included the sentence verification

technique (SVT; Royer, Hastings, & Hook, 1979) for

each text of the five lessons implemented in this study

and included story recall for two of those lesson texts.

We selected the two lesson-comprehension measures to

represent different levels of comprehension. The SVT

requires recognition of text content, and recall requires

recalling content from memory and constructing lan-

guage to express it. We also analyzed transcripts of

classroom discussions from two of the five lessons to

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

227

consider differences in discourse during the lessons

among the approaches. The beyond-lesson assessments

included a comprehension-monitoring task and a task

that assessed knowledge of strategies.

Lesson-Text Comprehension

SVT

The SVT was developed and used extensively by Royer

and his colleagues (1979). Because of its extensive track

record and because it is a measure that can be devel-

oped for specific texts, it fit our purposes well. In this

task, students examine sentences and determine wheth-

er each is true about the text (Royer, Carlo, Dufresne, &

Mestre, 1996).

Task Development. In the SVT version we de-

veloped, the items were either paraphrases (PPs) or

inferences of ideas from the text. This format was de-

veloped because theory suggests that performance on

paraphrased items represents a more text-bound level

of comprehension, whereas performance on inference

items represents comprehension of gist. For each item, a

true and false (PP), local inference (LI), and global infer-

ence (GI) of an idea from the text were designed. An LI

required students to draw an inference about a localized

portion of the text, whereas a GI required students to

make a connection across a large portion of the text.

Because the stories were of different lengths, the

number of items across the tests varied, ranging from

36 items to 60 items. To exemplify the types of items

on our SVTs, we created items for a well-known story,

The Three Little Pigs, and those items are presented in

Appendix C.

Pilot Work. We conducted two phases of pilot work

on the SVTs. The first involved students from a different

school in the same district in which the present study

was implemented. We administered the assessment to

a group of students who read one of the stories with-

in a strategies lesson and a group who did so within a

content lesson. There were 14 students in each group.

Additionally, we piloted the SVT items for all five texts

with students from a similar demographic who were at-

tending a summer school program. We divided the 40

students into two groups and administered the strat-

egies lessons to one group and the content lessons to

the other. With information from the pilot results, we

revised the measures, removing floor and ceiling items

and adding new items as necessary.

Administration. For three of the lessons, the SVT

was administered immediately following the reading

lesson. For the two stories for which oral recalls were

taken, the SVT was administered after oral recalls were

collected. The SVTs were administered to an entire class

by a research team member who read each item aloud.

Scoring. Each SVT was scored for total items cor-

rect, GIs correct, LIs correct, and PPs correct. Overall,

SVT comprehension scores of the texts were calculated

by averaging the scores from the five stories, and this

was done for each scoring dimension. This yielded four

scores, one for each item type (GI, LI, PP) and a total

score. Internal consistency of the total scores for the five

narrative SVTs, as assessed by Cronbach’s Alpha, was

equal to .901.

Oral Recall

Oral recalls were collected from all students for two

of the five stories used in the lessons, Off and Running

(Soto, 2005) and The Fun They Had (Asimov, 2005).

Administration. After each class had completed the

lesson for the two recall-targeted stories, research team

members took recalls from each individual student.

The recall protocol involved four prompts: one initiat-

ing prompt and three additional prompts to encourage

students to recall as much as possible. Recalls were re-

corded and subsequently transcribed.

Scoring. Recalls were scored for length and qual-

ity. Scoring for both dimensions involved dividing

the story texts and student recalls into content units.

Content units are units of text about one clause in

length (Omanson, 1982), a unit length we have used in

previous work (Beck et al., 1982). The number of con-

tent units in each student recall that represented a text

content unit was used as the length measure, and the

level of importance of each of these units was the basis

for the quality measure.

Importance was determined by categorizing each

unit as major, support, or detail. A major unit was de-

fined as containing ideas that are very important to the

story and a key component of a good summary. A sup-

porting unit, although not necessarily key, holds main

ideas together and adds elaborative information to the

major units. A summary could do without these ideas,

but it would be less complete. A detail unit was defined

as less-essential information. These ideas do not hold

the story together, and a summary would be complete

without them.

Off and Running was divided into 282 content units,

with 17 major ideas, 36 supporting ideas, and the rest

considered detail units. The Fun They Had was divided

into 182 content units, with 12 major ideas, 21 sup-

porting ideas, and the rest detail units. The units were

defined and labeled (as major, support, or detail) by one

researcher, after which all researchers discussed the cat-

egorization and reached consensus.

Prior to scoring, the recalls were blinded as to in-

structional approach. After blinding, each student recall

Reading Research Quarterly • 44(3)

228

was divided into content units, and those that men-

tioned a text content unit were counted and used as the

length score.

For quality scoring, points were assigned to each

unit: major units were assigned a value of three points,

supporting units two points, and details one point.

Recalls were scored by one team member and 20%

were rescored by a second team member to determine

inter-rater reliability. Differences for the 20% were then

resolved through consensus. For Off and Running, 88%

agreement was reached, and for The Fun They Had, 86%

agreement was reached.

As was done for the SVTs, an overall comprehension

score was calculated for each student for the recalls.

This score was calculated by averaging the recall scores

from the two stories.

Pilot Work. To get a sense of how student recalls

would align with the text content units, we collected

oral recalls for Off and Running and The Fun They Had

from the same group of summer school students with

whom we had piloted the SVTs. The information gath-

ered contributed to the development of rules for scoring

students’ wording that did not match text wording but

represented text ideas.

Lesson Discourse

As a way to understand the role of discussion during

the lessons in promoting comprehension, we exam-

ined transcripts of the lessons on the texts for which

students had provided recalls. We viewed promoting

comprehension as the extent to which the discussion

focused on building meaning from the text and used

two metrics to capture this. One was the proportion of

the discussion that was text based, which was instanti-

ated as talk that matched any of the content units, and

the other was the length of student responses, which

was instantiated as number of words per student re-

sponse. Proportion of text-based talk seemed a direct

measure of focus on the text. Regarding length of stu-

dent response, we have noted in previous research that

as discussion becomes more meaning focused, students’

responses become longer (Beck et al., 1996; McKeown,

Beck, & Sandora, 1996).

Beyond-Lesson-Text Assessments

Comprehension Monitoring

To explore whether students’ awareness of their com-

prehension (i.e., comprehension monitoring) was differ-

ent across the three instructional approaches, we used

a task based on one we have used previously in which

students read a short text containing some inconsisten-

cies, were stopped at several intervals, and asked wheth-

er there was anything that did not make sense (Beck

et al., 1996). The rationale for this comprehension-

monitoring task is that the ability to make decisions

about whether comprehension is proceeding success-

fully would seem to be a precursor to being able to take

steps to address problems. We assessed whether there

was any change in students’ comprehension monitoring

through a pre- and posttest on nonlesson stories that

were counterbalanced.

Task Development. Two 473-word stories with

a Flesch-Kincaid 3.3 grade-level readability (Flesch,

n.d.) were developed. Each had five paragraphs, with

three of the paragraphs including one inconsistency,

one paragraph including two inconsistencies, and one

paragraph with no inconsistencies. The stories were

composed in a familiar, fable-like style using student-

friendly language.

Administration. The comprehension-monitoring

task was individually administered, audiotaped, and

subsequently transcribed. The researcher read the text

aloud as the student followed along, a format that was

used to prevent decoding from confounding compre-

hension. The researcher stopped after each paragraph

and asked whether anything did not make sense. If the

students did not detect an error in the paragraph, the

researcher prompted them further by rereading two ad-

jacent sentences, one of which contained an error for

the four paragraphs that had errors, and asking the stu-

dent to explain those sentences.

Scoring. The comprehension-monitoring measure

was scored on a 2-point scale for each paragraph. In the

error paragraphs, students were given two points if they

detected the target error without additional prompt-

ing and one point if they detected the error only after

the researcher reread the sentences in which the error

existed. For the error-free paragraphs, two points were

awarded if students recognized there were no errors

and also were able to explain two important target sen-

tences. A second rater scored 20% of the responses and

an inter-rater reliability of 95% was obtained.

Strategies Task

The strategies approach provided intensive instruction

and practice during which each strategy was explicitly

named, described, and discussed. Given this instruc-

tion, we might expect the students in the strategies ap-

proach to show greater knowledge of strategy terms and

ability to apply them to text. Yet, strategy terms are also

introduced in basal reading lessons, including the basal

series used in our study, and from our interactions with

school personnel, we learned that the fifth-grade stu-

dents had been exposed to this instruction. Thus we

wondered whether knowledge of strategies would in-

deed differ across approaches.

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

229

To compare strategies knowledge across approaches,

we developed an additional task that was administered

as a posttest. The task was designed to measure each

student’s knowledge of strategies used in this study. The

task was based on one developed by Kozminsky and

Kozminsky (2001).

Task Development. Six stories were developed, av-

eraging 349 words, with an average 3.6 Flesch-Kincaid

grade level (Flesch, n.d.). For each story, a correspond-

ing four-question multiple-choice test was developed.

The questions covered four of the five strategies taught

in the strategies approach: asking a question about im-

portant information, predicting, inferring, and summa-

rizing. We did not include comprehension monitoring

because it did not lend itself to multiple-choice questions

(and it was assessed with a separate comprehension-

monitoring test, as described above).

The test questions were identical for all stories and

presented in the same order:

1. Which question might a teacher ask to know if a

student had understood the important ideas from

the story?

2. What would be a good prediction about what

might happen next in the story?

3. Reading between the lines, what could you infer

from this story?

4. Which of these is a good summary of the story?

Administration. All six strategy-task stories were ad-

ministered, two per classroom, in the first year of the

study. Four different stories were tested per approach,

so that every story was administered in two classrooms,

each in a different approach of the study. The task was

administered to the whole class. To prevent decoding

issues from confounding students’ strategy knowledge,

the stories, the four questions, and four choices were

read aloud.

Scoring. Scoring was straightforward. Students were

given one point for each correct answer, with a possible

total of eight points across two tests.

Procedures

Teacher Training

Prior to implementation of the study approaches, the

research team provided training to teachers. Training

began with a short overview of the study’s purpose.

Then teachers for each approach met separately for a

half day to receive intensive training in their assigned

approach—content, strategies, or basal comprehension.

Teachers received a notebook containing all lesson ma-

terials for their approach. Finally, the scripted lessons

were reviewed and discussed. Teachers were asked not

to share their lessons and materials with teachers in the

other approaches, and we found no indication that they

had done so.

One week after training, each participant group ob-

served a demonstration lesson of their approach by one

researcher in one of their fifth-grade classrooms. The

teachers had a copy of the scripted lesson and an ob-

servation sheet that prompted them to attend to certain

features of the instruction. A question-and-answer ses-

sion followed.

Student Preparation

Students were prepared for the two experimental ap-

proaches through practice lessons developed by the

research team and delivered by the teachers. Student

preparation for the strategies approach consisted of

three 45-minute lessons over three days. Each of the

five strategies targeted for the study was introduced and

practiced first with several sentences and then with a

short piece of text. Two strategies were introduced and

practiced in the first lesson and two more in the second

lesson. For the final lesson, the last strategy was intro-

duced and all five strategies were practiced. Students

were given a reminder sheet that briefly described each

strategy. Students were told to use these sheets during

the preparation lessons as well as in their classrooms

when the five story lessons were implemented.

Student preparation for the content approach was

done in one session in which the teacher told students

that during some upcoming reading lessons they would

be going through a story together, reading a portion of

the text, and stopping to talk about what they had read.

They then practiced doing this on a short text. There was

no preparation for the basal-comprehension approach

because the lessons were based on the comprehension

questions available in the guided comprehension portion

in the teacher’s edition of the basal reader being used.

Implementation of Lessons

and Assessments

The three instructional approaches were implemented in

the six fifth-grade classrooms within the selected school,

with two classes per approach. The instruction began in

the middle of January, and over the next five weeks, the

teacher in each classroom taught a story lesson at the

beginning of each week. Every lesson was observed by

a researcher. During observations, the researcher com-

pared the script of the lesson to the implementation in

the classroom. Feedback was only called for after the ini-

tial lesson, in which one teacher in the content approach

and one teacher in the strategies approach made minor

deviations from the script. During all subsequent obser-

vations, we judged the lessons to be implemented with a

high degree of fidelity for all teachers.

Reading Research Quarterly • 44(3)

230

To check fidelity, the transcripts of two subsequent

lessons, the second and the fifth, were analyzed for the

percentage of scripted initial questions asked, the extent

to which the questions elicited an appropriate response,

and whether useful follow-up questions were asked.

The analyses indicated that teachers implemented the

lessons with a high degree of fidelity.

For both stories across all three approaches, the ini-

tial scripted question was asked 100% of the time and

elicited an appropriate response from a student 89% of

the time in the content approach and 100% of the time

in the strategies and basal-comprehension approaches.

Teachers posed a follow-up question when needed,

which were scored appropriate 94% of the time for the

content approach and 100% of the time for the strate-

gies and basal-comprehension approaches. The results

were virtually the same for the fifth story. These very

high scores are not surprising given the scripted nature

of the lessons.

Two of the stories were completed within one class-

room period of 45 minutes; the other three stories were

divided into two parts that were delivered on consecu-

tive days. All lessons were audiotaped for later tran-

scription. Two of the lessons were also videotaped. At

the completion of three of the lessons (lessons 1, 3, and

4) all students were administered the SVT by a research

team member. After the other two lessons (lessons 2

and 5), students were first individually administered

oral recalls, which were audiotaped for subsequent

transcription. When the oral recalls were completed for

a class, the students were administered the SVT.

After all story lessons and their corresponding as-

sessments were completed, the strategies test was admin-

istered to each classroom by a research team member.

Finally, students were individually administered the

posttest of the comprehension-monitoring task.

Results

Questions of interest in the first year included two ques-

tions addressed to lesson comprehension:

1. Were there differences in comprehension among

approaches for lesson texts?

2. Did patterns in the lesson discourse help to ex-

plain any differences found?

Two other questions addressed more general outcomes:

3. Were there differences in students’ ability to mon-

itor comprehension on nonlesson texts among

approaches?

4. Were there differences in students’ knowledge of

comprehension strategies among approaches?

Using the student as the unit of analysis, data were

analyzed with one-way ANOVAs (analyses of covari-

ance were also run using the calculated standard score

of reading ability as a covariate; all findings were con-

sistent with the ANOVA results), with Tukey’s honestly

significant difference (HSD) tests run for post-hoc dif-

ferences (we examined our data for classroom effects

with a nested analysis, and there were no significant

effects; in considering this, we decided to maintain our

one-way ANOVA because our power would be drasti-

cally reduced due to the small number of classrooms in

our study if we analyzed our data using a nested analy-

sis). The exception was the comprehension-monitoring

pre- and post measure, which was analyzed with a re-

peated measures ANOVA with approach (content, strat-

egies, basal comprehension) as the between-subjects

variable and test time (pre or post) as the within-sub-

jects variable. Effect sizes are reported using a Cohen’s d

standardized effect size, which is calculated as the dif-

ference in group means divided by the pooled standard

deviation.

Lesson-Text Comprehension Measures

Two comprehension assessments were used for lesson

texts: SVT and story recall. The SVT was administered

after each of the five lessons. Story recalls were collected

after two of the lessons. Scores for these assessments

were calculated by averaging scores across stories.

SVT

The SVT data for each of the five stories were analyzed

for differences in four areas: total score, GIs correct, LIs

correct, and PPs correct. Analyses showed that there

were no significant differences across approaches for any

of the four areas. The results are presented in Table 3.

As can be seen from the table, all three groups

did moderately well on the SVT tests, with a mean

high of 81.20% correct and a low of 76.94% correct.

Additionally, 98% of the scores were above 50%, and

about half of the scores were above 80%.

Recall

The oral recalls collected for two of the five stories were

analyzed on two dimensions: length of recall and qual-

ity of recall. The results of these analyses are presented

in Table 4.

As indicated in the table, the approaches had sig-

nificantly different recall lengths and quality scores.

Post-hoc tests (Tukey’s HSD) on these results indicated

that the recalls of the content group (M = 16.20) were

significantly longer than were those of the strategies

group (M = 10.33; p = .001). The length of the basal-

comprehension students’ recalls (M = 14.13) was also

significantly higher than was that of the strategies

group (p = .05). The length of the basal-comprehension

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

231

students’ recalls was not significantly different from that

of the content students’ recalls (p = .373; d = .270).

Analysis of data for quality of the oral recalls, pre-

sented in Table 4, revealed that the scores of the ap-

proaches also differed significantly on this dimension.

Post-hoc analyses indicated that content students

(M = 33.40) produced higher quality recalls than did

the strategies students (M = 22.97; p = .001). The quality

score of the basal-comprehension students (M = 29.20)

was not significantly different from the average scores of

the content group (p = .292; d = .310) or the strategies

group (p = .09; d = .627).

Lesson Discourse

The analysis of lesson transcripts from the two lessons

from which we had recalls involved examining two

patterns. The first was the proportion of talk that direct-

ly reflected text ideas. The two lessons were combined,

and the number of text-based words was tallied at each

stop. The second was the amount of talk in each student

turn during the discussion. Differences were analyzed

separately for student and teacher talk.

As seen in Table 5, the ANOVA for text-related stu-

dent talk revealed a significant difference. Post-hoc tests

(Tukey’s HSD) indicated that content lessons had a sig-

nificantly higher proportion of text-based student com-

ments (94%) than did strategies lessons (75%; p = .005)

and that the basal-comprehension lessons also had a

significantly higher proportion of text-based comments

(86%) than did the strategies lessons (p = .002). The

content and basal-comprehension lessons were not sig-

nificantly different from each other (p = .209; d = .463).

Content

(n = 44)

Strategies

(n = 37)

Basal comprehension

(n = 38)

Score type M (SD) M (SD) M (SD) F p d

Total % 78.30 (9.75) 78.42 (9.07) 79.41 (9.05) 0.168 .846 .107

Global inferences % 76.94 (9.63) 77.03 (8.11) 78.13 (8.67) 0.281 .804 .139

Local inferences % 78.52 (9.69) 78.83 (10.02) 78.61 (10.20) 0.010 .990 .026

Paraphrases % 79.17 (11.42) 79.22 (10.40) 81.20 (9.44) 0.474 .624 .181

Table 3. Mean (SD) Scores and Analysis of Variance Results for Percent of Sentence Verification Technique Items Correct

for Narrative Texts by Approach for Year 1

d = Cohen’s d standardized effect size, calculated as the difference in group means divided by the pooled standard deviation.

Content

(n = 44)

Strategies

(n = 37)

Basal comprehension

(n = 38)

Score type M (SD) M (SD) M (SD) F p

Length 16.20 (8.96) 10.33 (4.55) 14.13 (6.14) 7. 29 .001

Quality 33.40 (16.14) 22.97 (9.61) 29.20 (10.27) 6.91 .001

Table 4. Means (SD) and Analysis of Variance Results of Length (by Number of Idea Units) and Quality of Recall Scores

for Narrative Texts by Approach for Year 1

Content

(n = 32)

Strategies

(n = 30)

Basal comprehension

(n = 35)

Participant M (SD) M (SD) M (SD) F p

Student % 94.08 (12.08) 74.88 (23.24) 85.86 (21.99) 7. 36 .001

Teacher % 50.36 (25.84) 26.79 (16.74) 47.18 (28.76) 8.35 .000

Table 5. Mean (SD) Percent and Analysis of Variance Results of Words About Text per Stop by Students and Teachers

for Year 1

Note. n = total number of stops across both stories.

Reading Research Quarterly • 44(3)

232

Table 5 also presents the amount of teacher text-

based talk, which showed similar results. In the content

lessons, teachers used a significantly higher propor-

tion (50%) of words about the text than the strategies

teachers did (27%; p = .000). The proportion of teacher

talk about the text in the basal-comprehension lessons

(47%) and the strategies lessons was also significantly

different (p = .000; d = .866). The content and basal-

comprehension lessons were not significantly different

from each other (p = .858; d = .116).

Examination of the second pattern, words per stu-

dent response, is presented in Table 6. The average num-

ber of words that a student spoke during each response

in the discussion differed by approach. Students aver-

aged many more words per response on average in the

content lessons (M = 24.38) than they did in the strate-

gies lessons (M = 11.74) and the basal-comprehension

lessons (M = 13.29; p = .000). The strategies lessons and

basal-comprehension lessons were not significantly dif-

ferent (p = .802; d = .180)

Outcomes Beyond Lesson Texts

In addition to assessments of comprehension of the les-

son texts and an analysis of lesson discourse, two as-

sessments that were independent of the lesson texts

were developed to compare student ability to transfer

knowledge gained in the study lessons. Specifically,

we collected data on student ability to monitor reading

comprehension and student knowledge of and ability

to apply the comprehension strategies stressed in the

strategies approach.

Comprehension Monitoring

Table 7 presents data on the comprehension monitoring

pre- and posttest, which had a total possible score of 12

points. These data were analyzed with a two-way, re-

peated measures ANOVA. The interaction between time

and approach was not significant, F(2, 115) = 0.451,

p = .638, which indicates that the mean change from

pre- to posttest was not significantly different across ap-

proaches. However, there was a significant main effect

of time, F(1, 115) = 9.40, p = .003, meaning that, regard-

less of approach, there was a significant score increase

from pretest (M = 5.01) to posttest (M = 5.87). The main

effect for approach was not significant, F(2, 115) = 0.169,

p = .845, but was of little interest because the change

across approaches was the focus of this analysis.

Strategies Task (Posttest Only)

The purpose of this task was to compare students’ strat-

egies knowledge across the three approaches. This as-

sessment had a total possible score of eight points. The

ANOVA revealed that the scores on these tests were not

significantly different for the three approaches following

the study implementation. These results are presented

in Table 8.

Discussion

The results from the first year showed no differences

on one measure of lesson-text comprehension, the SVT,

but did show differences on the other lesson-text mea-

sure, recall. On this measure, both content students and

basal-comprehension students outperformed strategies

students in terms of length, and content students out-

performed strategies students in terms of overall qual-

ity. These results are not necessarily in conflict, as recall

is a productive measure that may capture a higher level

of comprehension than would a multiple-choice test,

such as the SVT. Even though the differences among

approaches were limited to the recall measure, we

view these results as meaningful because the contrast

Approach n M (SD) F p

Content 32 24.38 (11.99) 15.523 .00

Strategies 30 11.74 (9.18)

Basal

comprehension

35 13.29 (8.14)

Table 6. Means (SD) and Analysis of Variance Results of

Number of Words per Student Response in Narrative

Lessons for Year 1

Note. n = total number of stops across both stories.

Pretest Posttest

Approach n M (SD) M (SD)

Content 43 5.14 (2.61) 5.72 (2.72)

Strategies 37 4.97 (2.57) 6.20 (2.46)

Basal

comprehension

38 4.90 (2.76) 5.71 (2.58)

Marginal 118 5.01 (2.63) 5.87 (2.59)

Table 7. Mean (SD) Comprehension-Monitoring Pre- and

Posttest Scores by Approach for Year 1

Condition n M (SD) F p

Content 40 4.53 (1.32) 1.86 .16

Strategies 39 4.90 (1.62)

Basal

comprehension

37 4.24 (1.50)

Table 8. Mean (SD) Strategies-Knowledge Test Scores

by Approach for Year 1

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

233

between the lessons was restricted, providing a very

stringent test. All three approaches had the same les-

son format of interspersed questions and discussion

during reading, and the lessons were scripted and fol-

lowed with fidelity. Thus, the differences in the lessons

were limited to the type of questions asked at the speci-

fied stops. Viewing the recall results as meaningful also

makes sense in light of findings from the analysis of

lesson discourse, which showed that discussion in the

content approach was more focused on the text and stu-

dents gave longer responses.

The lack of difference on the SVTs may indicate that

all approaches achieved adequate comprehension of the

texts. As noted in the Results section, the means on

the SVTs were fairly high. Moreover, in looking at the

spread of the scores across the stories, 98% of the stu-

dent scores were above 50% and about half of the scores

were above 80%. One interpretation of high scores is

that the test was easy for the students. As for the assess-

ments that tapped abilities beyond the specific lessons,

we found no differences among approaches for either the

comprehension-monitoring task or the strategies task.

For comprehension monitoring, we did find gains from

pretest to posttest across approaches. This gain suggests

that all three instructional approaches had positive ef-

fects on students. However, the fact that this change

occurred for all students regardless of approach is in

contrast to previous findings with a similar task where

we found differences for students in QtA classrooms but

no differences in control classrooms in which instruc-

tion was business as usual (Beck & McKeown, 1998).

As for the strategies task, the lack of difference

seems to suggest that the strategies approach did not

provide students an advantage for applying the spe-

cific strategies independently. It seems that students in

all approaches had equal familiarity with the strategy

terms and how to apply them.

A final question is why results for the basal-

comprehension approach did not differ significantly

from the content approach or the strategies approach.

In trying to create an appropriate comparison approach,

we may have inadvertently made teachers’ use of the

basal program more effective. Our development of these

lessons stems from our observation of each of the six

teachers’ presentations of a story lesson before the study

began. We noted that none of them followed the kind

of lesson we planned to provide—that is, reading inter-

spersed with questions about the story and discussion.

Our thinking was that such a format needed to be held

constant in all approaches. Further, we provided teach-

ers in the basal-comprehension approach with a script

that contained only the during-reading questions from

the basal guided comprehension section. Asking teach-

ers to follow our script also alleviated decisions about

other kinds of activities that surround the story in the

basal that could compete with story reading and discus-

sion. Thus, our format may have provided a streamlined

and focused approach to the story lesson.

Year 2

Method

Year 2 of the study was designed to replicate our meth-

odology from Year 1, with the addition of random as-

signment of students to classrooms. We also investigated

the effects of the three instructional approaches with

expository text and explored transfer to texts that stu-

dents read without instructional support. Additionally,

we included a passage-comprehension baseline test, a

pretest in addition to a posttest assessment of strategies

use, and teacher exit interviews.

Participants

Year 2 implementation took place in the same school

with the same six fifth-grade teachers as in Year 1. In

Year 1 of the study, classrooms were intact at the start

of the study implementation and, therefore, random as-

signment of students was not possible. For Year 2, how-

ever, we received permission from the school district

to randomly assign students to classrooms, with the

stipulation that we maintain balance by race and spe-

cial education status (including gifted students) among

classrooms. First, we randomly assigned the minority

students in the district (European American and Asian

American students) to the classrooms. The remaining

students (African American) were then randomly as-

signed to the six classrooms. Classroom assignments

were then reviewed to check that special education stu-

dents were represented equally in the six classrooms.

Given that the six fifth-grade teachers kept their Year

1 instructional-approach assignments for Year 2, stu-

dent assignment to classroom was, in essence, assign-

ment to approach. Total sample size was equal to 116

participants.

To test for equivalence in reading ability across ap-

proaches, we used the same Woodcock-Johnson tests of

decoding (Word Attack and Word Identification) used

in Year 1 of the study. In addition, we assessed compre-

hension ability using the Woodcock-Johnson Passage

Comprehension test. Standard scores were calculated,

and a one-way ANOVA was not significant, indicating

that reading ability and comprehension were not sig-

nificantly different across the six classrooms or three

approaches.

To evaluate whether there were significant differ-

ences among approaches in reading ability, one-way

ANOVAs were done for the three Woodcock-Johnson

tests. There was not a significant difference among

Reading Research Quarterly • 44(3)

234

the three approaches for Word Identification, F(2,

113) = 0.17, p = .844; Word Attack, F(2, 113) = 0.292,

p = .748; or Passage Comprehension, F(2, 113) = 0.41,

p = .665. See Table 9 for the pretest scores for all indi-

vidual measures and the standard score of overall read-

ing ability.

Measures

Expository Lessons

To investigate the effects of the instructional approaches

on the learning of content material, three lessons us-

ing expository texts were developed from the book Bees

Dance and Whales Sing by Margery Facklam (1992). The

topic, animal communication, did not overlap with the

topics students were studying in their science or social

studies classes. These additional lessons were imple-

mented in the classrooms after the completion of the

five narrative lessons.

Additional Measures

Expository Data

For all three expository texts, students were assessed

with the SVT. Internal consistency of the total scores for

the three expository SVTs, as assessed by Cronbach’s

Alpha, was equal to .57. Oral recalls were also collected

for two of the three expository texts. To assess learning

of specific scientific concepts in the texts after students

completed their recall, three knowledge-probe questions

were presented (e.g., “What are infrasonic sounds?”).

To develop a scoring scheme, we listed the major

text concepts that addressed each question, assigning

each one point. Then for each major text concept, sup-

porting ideas were listed, and each was assigned one

half point. The total possible score was 17.75, which was

far from what we anticipated students would obtain. But

we needed to include every possible detail that students

might mention. Answers to probed questions were

scored by one team member, and 20% were rescored by

a second team member to determine inter-rater reliabil-

ity. Differences for the 20% were resolved through con-

sensus. For the chapter “Barks, Chirps, and Melodies,”

85% agreement was reached, and for “Messages by the

Mile,” 85% agreement was reached.

Transfer Task

One new goal for Year 2 was to assess transfer effects.

Successful transfer involves being able to recognize ele-

ments in a new situation that are similar to those from

the learning situation, something that young and less

able learners find difficult to do (Brown, 1982a; Brown

& Campione, 1981; Rumelhart & Norman, 1981). Thus,

for young learners, transfer of learning may be unlikely

unless there is explicit instruction aimed at transfer

(Brown et al., 1983). To facilitate students’ abilities to

transfer the effects of the instructional approaches, we

designed a sequence of five lessons about such topics

as lighthouses, the Chicago fire, and the “War of the

Worlds” radio broadcast. The first four lessons were de-

signed to scaffold students to transfer what they had

learned from teacher-led text discussions to building

comprehension on their own. The sequence of lessons

allowed for gradual release of responsibility to students

by fading teacher-led discussion prompts and increas-

ing prompts for silent thinking. The final lesson was

used to assess transfer through individual oral recall

after the final text was read silently. The number and

type of prompt for each lesson in the sequence is shown

in Table 10.

The prompts followed the format of the instruction-

al approach being implemented in the classroom:

• In the content approach, the teacher reminded

students that they were trying to understand what

they had just read and prompted them to describe

what they were thinking about.

• In the strategies approach, students were asked

to choose a strategy that would fit best at each

stopping point, followed by discussion of how the

strategy had been applied.

• In the basal-comprehension approach, the teacher

prompted students to think about a strategy that

could be used at each stop or a question that could

Approach

Word Identification test

M (SD)

Word Attack test

M (SD)

Passage Comprehension test

M (SD)

Standard score

M (SD)

Content 5.0 (1.4) 6.0 (2.8) 4.2 (2.0) –0.05 (2.26)

Strategies 5.2 (2.3) 5.7 (3.6) 4.4 (2.0) 0.06 (2.96)

Basal comprehension 5.1 (1.7) 6.3 (3.7) 4.0 (1.3) –0.01 (2.47)

Table 9. Mean (SD) Scores for Three Baseline Assessments and Overall Standard Scores for Each Approach

for Year 2 Students

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

235

be asked. This was followed by a discussion of

how to apply the strategy or answer the question.

Prompts for silent thinking were similar, but inter-

actions did not include oral responses or discussion.

Students were given time to consider how they would

respond and then reading resumed.

At the end of each lesson in the sequence, students

were reminded that the kind of thinking they had done

during the lessons should be done when reading on

their own. For all of the lessons, students were allowed

to refer to a reminder sheet designed by the researchers

to prompt students to use the instructional approach

used in their classroom. The transfer lessons were pre-

sented by the classroom teachers toward the end of the

study and were spread over a two-week period.

Strategies Task

For Year 2, revisions to the strategies task were made

to items that showed floor and ceiling effects, and the

four texts with the strongest sets of items were identi-

fied. We judged strong items to be those that had either

performed well in Year 1 or had been easy to revise, that

is, we had been able to hypothesize why a particular

item had been problematic and could readily overcome

the problem.

The other change from Year 1 was that we admin-

istered the strategies task in a pre- and posttest design.

Each classroom administered two of the stories for the

pretest and the other two as the posttest, with the sto-

ries counterbalanced within approach.

Exit Interviews

At the conclusion of the study, teachers were inter-

viewed to determine their overall satisfaction with their

prescribed instructional approach. To ensure honest

responses, we hired an individual experienced in con-

ducting interviews who was not associated with the

study. All teachers were told that their names would not

be connected to their comments. The interviewer used

a protocol of eight questions that encouraged teachers

to talk about their overall experience while implement-

ing the study. The interviews were audiotaped and then

transcribed, and each teacher’s comments were blinded

to ensure anonymity.

Results

As in Year 1, we collected data in two areas: lesson com-

prehension and comprehension beyond the lessons.

Questions of interest in the second year included three

questions addressed to lesson comprehension:

1. Were differences in comprehension of narratives

found in Year 1 among approaches replicated?

2. Did patterns in the lesson discourse help to ex-

plain any differences found?

3. Were there differences in comprehension among

approaches on the expository texts and in the

learning of expository material?

Three further questions related to outcomes beyond the

lesson texts—Were there differences among approaches

on any of the following:

4. The comprehension-monitoring task?

5. The transfer task?

6. The strategy task?

All data, with the exception of the comprehension-

monitoring task and the strategies task, were analyzed

with one-way ANOVAs, with Tukey’s HSD tests com-

puted for post-hoc differences. The comprehension-

monitoring data and strategies-task data were analyzed

with two-way repeated measures ANOVAs, with ap-

proach as the between-subjects variable and test time

(pre or post) as the within-subjects variable.

Lesson-Text Comprehension Measures

As in Year 1, two comprehension assessments were used

for lesson texts: SVT and story recall. The SVT was ad-

ministered after each of the five narrative lessons and

three expository lessons. Story recalls were collected for

two of the narrative lessons and two of the expository

lessons.

SVT Comprehension Assessment

SVT scores were averaged across tests for the five narra-

tives and separately for the three expository texts. The

scores were analyzed with one-way ANOVAs for four

different measures: percent correct for the total score,

GIs, LIs, and PPs. The average scores were not sig-

nificantly different among approaches for any of these

measures. The results of these analyses are presented in

Table 11 for the narrative lessons and Table 12 for the

expository lessons.

Lesson

# of oral-discussion

prompts

# of silent-thinking

prompts

1 5 0

2 3 2

3 2 3

4 0 5

5 0 5

Table 10. Type of Prompt in Transfer Lessons

Note. Lesson 5 was assessed for transfer.

Reading Research Quarterly • 44(3)

236

Narrative Recall

The results of the analyses from the oral recalls for the

narrative texts are presented in Table 13. The results from

the length analysis (by number of content units recalled)

for Year 2 showed overall significance for the narrative

texts. Post hoc tests indicate that, as in Year 1, the mean

length of the content recalls (M = 17.32) was significantly

higher than was the mean length of the strategies recalls

(M = 11.77; p = .005). The mean length of basal-com-

prehension students’ recalls (M = 15.63) was not signifi-

cantly different from the content recalls (p = .59; d = .20)

or the strategies recalls (p = .08; d = .66).

Analysis of the data for quality revealed a similar

pattern as that seen for length. The mean quality score

for the students in the content approach (M = 33.76) was

significantly higher than was the mean quality score

for the students in the strategies approach (M = 24.79;

p = .01). The basal-comprehension students’ mean score

(M = 31.15) was not significantly different from the con-

tent (p = .66; d = .19) or strategies (p = .10; d = .59) mean

scores.

Expository Recall

We collected oral recalls for two expository texts that,

just as for narratives, were scored in terms of length and

quality. These data are presented in Table 14. Although

the mean length and quality scores for the content

group were slightly larger, differences were not signifi-

cant. However, there were differences in the scores for

the knowledge probes, which were direct questions

that queried specific scientific content. These data are

also presented in Table 14. As noted earlier, the total

possible score was 17.75 points, which accounts for ev-

ery possible detail about the probed content. The low

scores need to be considered in light of the inflated

maximum.

The average score on the probes of the content group

(M = 4.27) was significantly higher than was the average

score of the strategies group (M = 3.15; p = .02). The aver-

age score of the basal-comprehension group (M = 4.30)

was also significantly higher than was the average of

the strategies group (p = .02). The average scores of the

content group and the basal-comprehension group were

not significantly different (p = .99; d = .02).

Lesson Discourse

As in Year 1, the transcripts of the lessons assessed with

oral recalls were analyzed. The two expository lessons

for which recalls were collected were added to the anal-

ysis, for a total of four lessons. This analysis involved

Content

(n = 41)

Strategies

(n = 38)

Basal comprehension

(n = 37)

Score type M (SD) M (SD) M (SD) F p d

Total % 77.17 (8.72) 76.02 (11.34) 76.66 (7.37) 0.15 .86 .103

Global inference % 76.72 (9.21) 75.21 (11.18) 76.53 (7.65) 0.29 .75 .143

Local inference % 76.22 (9.86) 76.71 (11.03) 75.12 (7.78) 0.27 .77 .138

Paraphrase % 77.57 (10.15) 75.92 (13.45) 78.09 (8.60) 0.41 .67 .170

Table 11. Means (SD) and Analysis of Variance Results for Percent of Sentence Verification Technique Items Correct

for Narrative Texts by Approach for Year 2

d = Cohen’s d standardized effect size, calculated as the difference in group means divided by the pooled standard deviation.

Content

(n = 41)

Strategies

(n = 38)

Basal comprehension

(n = 37)

Score type M (SD) M (SD) M (SD) F p d

Total % 77.38 (8.83) 75.26 (9.12) 75.08 (7.67) 0.88 .42 .250

Global inference % 75.92 (10.29) 73.83 (10.54) 74.11 (10.40) 0.48 .62 .184

Local inference % 73.07 (10.88) 72.63 (10.35) 71.93 (8.71) 0.13 .88 .096

Paraphrase % 83.46 (10.01) 80.38 (10.44) 80.26 (10.98) 1.19 .31 .290

Table 12. Means (SD) and Analysis of Variance Results for Percent of Sentence Verification Technique Items Correct

for Expository Texts by Approach for Year 2

d = Cohen’s d standardized effect size, calculated as the difference in group means divided by the pooled standard deviation.

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

237

examining the same two measures examined in Year

1 (student and teacher words about text and length of

student response). Data for words about the text are pre-

sented in Table 15.

The pattern of results for Year 2 was similar to

the pattern seen in Year 1. As shown in Table 15, the

ANOVA for student talk revealed a significant differ-

ence. Post-hoc tests (Tukey’s HSD) indicated that con-

tent lessons had a significantly higher mean percent of

text-based student comments (97%) than did strategies

lessons (66%; p < .0005) and basal-comprehension les-

sons (84%; p = .002). The basal-comprehension lessons

also had a significantly higher mean percent of text-

based student comments than did the strategies lessons

(p < .0005).

Table 15 also presents the amount of teacher text-

based talk. The content lessons had a significantly high-

er mean percent of teacher talk about the text (42%)

than did the strategies lessons (30%; p = .04) and the

basal-comprehension lessons (27%; p = .01). The basal-

comprehension lessons and the strategies lessons were

not significantly different in terms of the mean percent

of teacher talk about the text (p = .815; d = .183).

The data from analysis of length of student response

are presented in Table 16. The average number of words

per student response in the discussion differed by ap-

proach. Students produced significantly more words per

response on average in the content lessons (M = 26.55)

than they did in the strategies lessons (M = 9.50) or

basal-comprehension lessons (M = 13.84; p < .0005).

Content

(n = 41)

Strategies

(n = 38)

Basal comprehension

(n = 37)

Score type M (SD) M (SD) M (SD) F p

Length 17.32 (10.12) 11.77 (5.71) 15.63 (5.98) 5.41 .01

Quality 33.76 (16.85) 24.79 (11.35) 31.15 (10.06) 4.75 .01

Table 13. Means (SD) and Analysis of Variance Results for Length (by Number of Idea Units) and Quality of Recall Scores

for Narrative Texts by Approach for Year 2

Content

(n = 41)

Strategies

(n = 38)

Basal comprehension

(n = 37)

Score type M (SD) M (SD) M (SD) F p

Recall length

(number of units)

15.65 (8.60) 12.19 (5.80) 14.22 (5.33) 2.56 .08

Quality 29.19 (14.16) 23.98 (10.34) 27.31 (8.95) 2.08 .13

Probes 4.27 (1.86) 3.15 (1.98) 4.30 (1.53) 5.028 .008

Table 14. Means (SD) and Analysis of Variance Results for Length and Quality of Recall Scores and Probes for Expository

Texts by Approach for Year 2

Content

(n = 56)

Strategies

(n = 52)

Basal comprehension

(n = 54)

Approach M (SD) M (SD) M (SD) F p

Student % 97.08 (5.59) 66.05 (26.09) 83.95 (22.83) 32.37 .00

Teacher % 41.83 (32.74) 29.73 (21.25) 26.70 (10.14) 5.43 .005

Table 15. Means (SD) and Analysis of Variance Results for Percent of Words About Text per Stop by Student and Teacher

in Four Lessons

Note. n = total number of stops across both stories.

Reading Research Quarterly • 44(3)

238

The strategies and basal-comprehension lesson means

did not differ significantly (p = .141; d = .59).

Having obtained similar results in Years 1 and 2 for

the oral-recall measures and for the analyses of lesson

discourse, we conducted a further analysis to try to dis-

cern the roots of those differences within the discourse

itself. This analysis was done using transcripts from one

classroom in each of the experimental approaches of the

discussion of Isaac Asimov’s (2005) The Fun They Had.

Below, we considered the initial 163-word text segment,

which features two children, Margie and Tommy, in the

year 2157 who discover an old printed book and are

stunned by it because “the words stand still,” in contrast

to the books they read on their television screens. The

text follows:

Margie even wrote about it that night in her diary. On the

page headed May 17, 2157, she wrote, “Today Tommy found

a real book!” It was a very old book. Margie’s grandfather

once said that when he was a little boy his grandfather told

him that there was a time when all stories were printed on

paper. They turned the pages, which were yellow and crin-

kly, and it was awfully funny to read words that stood still

instead of moving the way they were supposed to—on a

screen, you know. And then, when they turned back to the

page before, it had the same words on it that it had had

when they read it the first time. “Gee,” said Tommy, “what a

waste. When you’re through with the book, you just throw

it away, I guess. Our television screen must have a million

books on it, and it’s good for plenty more. I wouldn’t throw

it away.” (“The Fun They Had,” copyright  1957 by Isaac

Asimov, from Isaac Asimov: The Complete Stories Of Vol. I

by Isaac Asimov. Used by permission of Doubleday, a divi-

sion of Random House, Inc.)

In the strategies approach, the discussion for this seg-

ment began with the teacher directing an activity for

monitoring comprehension:

Teacher: Oh, good. Stop there, Honey. This is a

good place to stop and check our under-

standing. What might be confusing here?

Holly.

In response, a student identified a line of text that of-

fered a confusion. When the teacher followed up by

asking what the student could do to help herself under-

stand the confusion, the student replied with a strategic

procedure rather than a way to address that particular

confusion:

Holly: On the page headed May 17, 2157.

Teacher: Oh, 2157. That might be confusing to a

person reading this and what could you

do to help yourself understand 2157?

Student: Ask a question.

The teacher called on another student, who offered a

way to address confusion that is relevant to this par-

ticular confusion, which is about dates, but he framed

it hypothetically:

Kyle: Maybe you could like to tell if it’s a date or

what. Just like if it’s a date, you could um,

see how many years from now it is.

In acknowledging Kyle’s comment, the teacher also took

the opportunity to point out that he had made an in-

ference. After Kyle confirmed doing so, the teacher re-

turned to the original question of possible confusions

in the text:

Teacher: In the future, good. You made an infer-

ence. Did the author once say this is tak-

ing place in the future? Did the author

state that?

Students: No.

Teacher: But yet you knew it, didn’t you, Kyle?

Kyle: Yeah.

Teacher: Based on what you read. And then you

even did some work in your mind, trying

to figure out how far into the future it was

so that was good. Is anyone in the room

confused about anything we read here?

What did you think about that business

with the books having print?

Another student then offered a different potentially

confusing text concept. Again, when the teacher asked

how the student cleared up the confusion, the student

replied procedurally:

Student: I was confused about moving on the tele-

vision screen...

Student: Yeah, that one.

Student: ...the part when it said they turned the

pages, which were yellow and crinkly, and

it wasn’t, it was awfully funny to read the

words that stood still instead of moving

Approach n M (SD) F p

Content 56 26.55 (17.23) 31.134 .00

Strategies 52 9.50 (8.29)

Basal

comprehension

54 13.84 (6.20)

Table 16. Means (SD) and Analysis of Variance Results

for the Number of Words per Student Response in Four

Lessons

Note. n = total number of stops across both stories.

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

239

the way that they were supposed to on a

screen.

Teacher: That was confusing to you, huh? What did

you do to clear up your confusion?

Student: Ask a question, read on, reread.

The teacher asked the student which he chose. In reply-

ing, the student offered a most authentic moment:

Teacher: You could do...which one would you pick?

Student: Ask a question.

Teacher: Yeah. What would you be asking yourself?

Student: What the heck are they talking about?

The teacher responded to the question and then asked

students to make an inference:

Teacher: [Laughs] That’s funny. That’s good. Well, I

think the, the irony of that part is that we

read books where the print is on the pages

and it stands still, don’t we?

Students: Yes.

Teacher: Isn’t that what we are doing now?

Students: Yes.

Teacher: So make an inference about...what these

kids must be thinking in the future?

Students offered responses and then reading resumed:

Student: How weird they were.

Student: How dizzy.

Although the students in this discussion selected im-

portant concepts from the text, for the most part the

concepts are not used for building meaning but rath-

er are treated as instances of how a strategy could be

applied.

In the content approach, the discussion for this seg-

ment began with the teacher posing a question and a

student summarizing the text segment, including key

ideas about how Tommy and Margie’s concept of a book

from their 2157 vantage point differs from the old book

they found.

Teacher: So what’s, what’s this all about? What’s

going on here? What’s going on? Tajae,

what’s going on?

Tajae: Tommy found a book and they’re looking

in it and they’re saying the pages are crin-

kly and stuff and they’re thinking that if

you read the book, you can go back in and

it will be totally different about it but it’s

all still the same and they say that after

you read it one time, you might as well

throw it away cause you’ll, cause if you

read it and you know what it’s about, if

[inaudible] TV one cause if you turn on

the TV and then you watch something,

the next day it won’t be the same page.

The teacher briefly acknowledged the response and

called on another student whose hand was raised who

added other relevant ideas, including the future setting

of the story:

Teacher: Yeah, all that’s very true. Catherine.

Catherine: It’s in the future. It says that it’s 2157, so

the book is really old because now, like

in the future, it’s saying that they don’t

read, um, books. They don’t read books.

They read like on television screens and

they’re shocked because the book is re-

ally old and stuff.

The teacher then responded in a way that integrated the

two students’ comments, followed by a third student

who offered elaboration:

Teacher: Yeah, and you, you added to what Tajae

already said. Yeah, this is set in the future.

And the book and the screen are two to-

tally different things.

Student: Um, the words don’t, um, move around

when they’re like...the way, um, they think

they’re, they were supposed to.

In this discussion, students presented ideas from the

text very much in their own language. Even though that

language is a bit rough around the edges, each subse-

quent student contribution seemed to take account of

and build on the prior ones, such that movement to-

ward coherence is evident.

Outcomes Beyond Lesson Texts

As in Year 1, outcomes beyond the lesson texts were

also assessed. In Year 2, this included comprehension

monitoring, a strategies task, and a transfer task. The

strategies task was administered as a pretest/posttest as-

sessment rather than only the posttest used in Year 1.

Comprehension Monitoring

Table 17 presents data for the comprehension monitor-

ing pre- and posttest. As in Year 1, the total possible

score for this assessment was 12 points. These data were

analyzed with a two-way, repeated, measures ANOVA.

The analysis of the Year 2 data revealed a similar

pattern of differences as was seen in Year 1. The inter-

action between time and approach was not significant,

F(2, 113) = 0.830, p = .439, which indicates that the

Reading Research Quarterly • 44(3)

240

mean change from pretest to posttest was not signifi-

cantly different across approaches. However, there was

a significant main effect of time, F(1, 113) = 13.57,

p < .0005, meaning that, regardless of approach, there

was a significant increase from pretest (M = 4.67) to

posttest (M = 5.63). The main effect for approach was

not significant, F(2, 113) = 1.215, p = .301, but was of

little interest as the change across approaches was the

focus of this analysis.

Strategies Task

The strategies data were analyzed with a two-way,

repeated-measures ANOVA. Data for these results are

presented in Table 18. The total possible score for this

assessment was 8 points.

The interaction between test time and approach was

not significant, F(2, 113) = 1.21, p = .302, which indi-

cates that the mean change from pretest to posttest did

not differ by approach. The main effect of time was not

significant, F(1, 113) = 0.340, p = .561, which indicates

that there was not an overall change in mean scores

from pretest to posttest. The main effect for approach

was also not significant, F(1, 113) = 0.540, p = .584.

Transfer Task

The transfer task consisted of individual student oral

recalls collected after students silently read the final text

in a sequence of five passages in which discussion over

the first four texts was faded and then eliminated. The

fifth selection was the transfer text.

Data from the transfer task are presented in Table 19.

The analysis of the length scores revealed a significant

difference in the scores among the approaches. Further

post-hoc analyses were conducted and although the ap-

proaches were not found to be significantly different at

the p = .05 level, there appeared to be a trend of higher

scores in favor of the content group over the strategies

group (p = .054; d = .490). For the quality scores, the

difference among approaches was not significant at the

p = .05 level (p = .055; d = .460), but as seen for length

scores, the score differences again suggested the trend

of higher scores in favor of the content group over the

strategies group.

Exit Interviews

To ascertain teachers’ attitudes toward the lessons, we

hired an experienced interviewer to conduct exit in-

terviews. The interview was rather loosely structured

around three major areas, teachers’ comfort with the

approach—or how natural it felt—its effect on stu-

dents, and thoughts about the approach in comparison

to other approaches they used or were familiar with.

There were two or three follow-up questions for each

area, with a total of eight questions. Because the con-

versation that each teacher had with the interviewer

was open and free-ranging, all six teachers did not nec-

essarily address each question. But they did all have

responses about each of the three major areas. We

summarize those responses below for the two teachers

in each approach: strategies (S), content (C), and basal

comprehension (B).

“How natural did the approach feel?”

S1 The approach felt very natural to be honest.

S2 It was pretty natural to me because I think we touch those

strategies in every lesson we do.

C1 Not natural at first, but the more I did them I felt a lot

more comfortable. I always wanted to put my two cents

in.

C2 It’s not natural to not go deeper. It’s hard to just let them

think on their own and not pull the information from

them.

B1 I really felt sort of restricted. I wasn’t able to give the kids

my background knowledge into the subject that we were

reading or my experiences.

B2 Totally unnatural because I had to stick to that script.

Because I’m constantly probing for more information and

more meaning.

As can be noted above, the teachers’ reactions ranged

widely among approaches but was rather consistent

Pretest Posttest

Condition n M (SD) M (SD)

Content 39 4.77 (2.73) 6.07 (2.96)

Strategies 38 4.81 (2.19) 5.88 (2.56)

Basal

comprehension

37 4.42 (2.71) 4.92 (2.91)

Marginal 114 4.67 (2.54) 5.63 (2.84)

Table 17. Mean (SD) Comprehension-Monitoring Pre-

and Posttest Scores by Approach

Pretest Posttest

Approach n M (SD) M (SD)

Content 41 4.24 (1.69) 4.76 (1.92)

Strategies 38 4.87 (1.76) 4.76 (1.95)

Basal

comprehension

37 4.57 (1.71) 4.49 (1.66)

Marginal 116 4.55 (1.72) 4.67 (1.84)

Table 18. Mean (SD) Strategies-Knowledge Pre- and

Posttest Scores by Approach

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

241

within approach. That is, the strategies teachers were

quite satisfied, the content teachers were somewhat sat-

isfied, and the basal-comprehension teachers were not

satisfied as far as the approach feeling natural to them.

It is interesting to note an issue that arose in each ap-

proach regarding why the approach was not completely

natural to them. The teachers referred to their usual ap-

proach in a way that suggests that they often provide a

great deal of information to the students (e.g., “my two

cents”; “give the kids my background knowledge...or

my experiences”) or pose questions in a way that leads

students to the responses (“pull the information from

them”).

“What was the effect on students?”

S1 I think it helped reach out to some kids that were on

the fence, and I think it definitely helped to push them

over. There was still that percentage of kids that seemed

clueless.

S2 I think it helped them a lot. Sometimes maybe the kids

were a little confused, but it helped them with asking

themselves the questions.

C1 Some responded quickly and they were on top of it.

Others you could tell the thought process was taking

place, so they didn’t respond as well. But all in all, I think

they responded pretty well.

C2 It wasn’t just listing facts. They really had to think about

what the author was trying to say. Some of them couldn’t

quite grasp what was being asked of them. They had a

difficult time responding. The students that did respond

did well.

B1 I believe it does help them to grasp the meaning,

however, I believe that you also need to ask other

questions.

B2 The discourse was labored. You’d think they were bored.

I [usually] have those kids with an open notebook, and I

stop and pose a question, and they got to respond, write

something.

As can be noted above, five of the six teachers indi-

cated that they saw some benefit in the approach for at

least some of their students, even though it might have

been difficult for students to respond at first.

“How did the approach compare with others you’ve used or

know about?”

S1 Prior to the strategy study, if we were reading a novel,

I would focus on the story elements. But this was more

meaty. This got them to follow a recipe that will never fail

them.

S2 I thought it was very effective in getting them to engage

in the story and then again to know what to do. I like this.

I liked it a lot.

C1 [In my previous instruction] the children got to view the

story three times. Once on audio, once I read it to them,

and the third time we had the students read the story. I

don’t think that they need all that repetition. They need to

realize that whenever you do the story, whether it’s one

time or three times, that you need to pay attention at all

times.

C2 The basal has a wider range or variety of questioning.

So the student that likes to stick straight to the point, that

doesn’t understand context deeply, there’s stuff for them.

And then there’s also other types of questioning for the

other students.

B1 My approach was close to the basal approach that I used,

except that I found the approach to be limiting somewhat

to the types of questions.

B2 The length of time that’s spent on the story [was

different]. With the basal series, they have the tapes, and

I will often have them listen to the tape before we even

read or partner read.

Two teachers referred to their habitual approach as

“doing the story three times,” including listening to it

on tape initially. Two teachers referred to the need for

“other types of questions,” and the way that one of the

content teachers explained it, it seems that she means

that some students need more literal questions to be

able to respond. The responses seem to indicate that

both strategies teachers and at least one content teacher

would be happy to continue to use aspects of the ap-

proach in their future teaching. Perhaps surprising is

that the teachers using a modified version of their own

basal were the least satisfied.

Discussion

We now summarize the results of Year 2 (a discussion

of all aspects of the study follows). The second year of

Content

(n = 40)

Strategies

(n = 37)

Basal comprehension

(n = 36)

Score type M (SD) M (SD) M (SD) F p

Length 16.35 (8.40) 12.41 (7.31) 12.75 (6.16) 3.39 .037

Quality 31.72 (14.51) 25.08 (13.85) 25.75 (10.71) 2.98 .055

Table 19. Means (SD) and Analysis of Variance Results for Transfer-Task Length (by Number of Idea Units) and Quality

Scores by Approach for Year 2

Reading Research Quarterly • 44(3)

242

the study showed a similar pattern as that found in the

first year, despite having random assignment of stu-

dents in the second year. That is, the content approach

showed an advantage over the strategies approach in

recall length and quality for narrative texts, but no dif-

ferences were found for scores on the SVT. Also similar

were the results of the lesson-discourse analysis, with

content discussion being more focused on text than

was strategies discussion and content students provid-

ing longer responses than did both strategies and basal-

comprehension students. The lesson-discourse analysis

included two expository texts that were added for the

second year. Similarities were also found in beyond-

lesson-texts assessments. That is, again, no differences

were found among approaches for the comprehension

monitoring or the strategies task.

The results for the expository texts reflected the pat-

tern for the narrative texts in recall as well as lesson

discourse. Scores for both length and quality of recall

were higher for content than they were for strategies

and about equal to that of the basal-comprehension

group, although these differences were not significant.

Analysis of the knowledge-probe data for the exposi-

tory texts did show significant differences, however, for

the content and basal-comprehension groups over the

strategies group.

A beyond-lesson measure that was added for the

second year of the study, recall of a transfer text, add-

ed to the picture of differences among approaches.

Differences in quality scores and length of recall favored

the content group, with mean scores of the strategies

and basal-comprehension groups being nearly identical.

Although these differences did not reach the p = .05

level of significance, they do seem to confirm an impor-

tant trend throughout the data. That is, any differences

in recall scores favored the content group.

General Discussion

The goal of the present study was to compare the results

of two comprehension instructional approaches (strate-

gies and content) and a control approach (basal compre-

hension) under circumstances that would allow fairly

precise description of what was going on in classrooms

within each approach. To do so, we implemented rig-

orously designed representations of the two approach-

es. The rigorous design included using common texts

within each approach, scripting what the teacher said

and asked students to do, providing teacher training

and feedback, and conducting fidelity checks for every

lesson in each approach. Thus, we attempted to ensure

as much as possible that only the issues of interest were

varied, specifically, how teachers prompted students

to think about text and how teachers responded to stu-

dents’ comments during reading lessons.

Results were remarkably consistent across the two

years, an indication that the lesson design held in-

structional approaches constant over two cohorts of

fifth graders, despite lack of random assignment in the

first year and whether it was the initial or second time

through the lessons for the teachers. At a general level,

our results lead us to conclude that all the instructional

approaches provided for adequate comprehension, and

a small but consistent pattern of differences occurred

that favored the content approach.

In the discussion that follows, we consider what un-

derlies the similarities and differences found, focusing

on the two experimental approaches, content and strat-

egies. We then discuss how the basal-comprehension

approach functioned in this study and finally turn to

limitations of the study and implications for research

and practice.

SVT

Strategies and content did not differ on the SVTs, which

are essentially recognition tests, for both years, and the

findings, as noted earlier, suggest that students in all

approaches were moderately successful (Tables 3 and

11). Our speculation for why there were no differences

is that the format of interspersed oral reading and dis-

cussion that was held constant across approaches was a

powerful influence. Certainly, oral reading of the texts

helped to address decoding difficulties that might have

hindered some students’ understanding had the texts

been read independently. A more interesting theoretical

reason is that the interspersed reading-and-discussion

format might have scaffolded components of text pro-

cessing for students by giving students the opportunity

to focus on important information and make connec-

tions to prior text and relevant outside information.

Thus, for the level of comprehension assessed through

a recognition test, such as the SVT, the substance of

the questions that prompt discussion (e.g., a content

prompt, such as “What’s going on here?,” or a strategies

prompt, such as “What inference can you make here?”)

may be less important than the interspersed reading-

and-discussion format.

Oral Recall

The substance of questions asked during lessons appar-

ently did make a difference for the level of comprehen-

sion required by a recall task. Differences in recall for

narratives favored the content group across both years

in terms of length and quality of recall (Tables 4 and

13). Consider that a recall task requires a constructed

response, in contrast to a recognition task. In construct-

ing a response, a reader has to bring forth information

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

243

from memory, decide which information to use, and put

it into language. A recognition task, on the other hand,

calls for matching information provided with informa-

tion stored in memory. Thus, recall represents a higher

level of comprehension.

An answer as to why the content approach showed

advantage on recall may lie in the nature of the dis-

cussions prompted by the substance of the questions.

Discussions between approaches differed in terms of

what students and teachers talked about and how much

students said (Tables 5, 6, 15, and 16). As our findings

demonstrate, in the content approach, a greater propor-

tion of talk focused on text ideas, and students’ respons-

es were longer. To a large extent, this result suggests that

students do what is asked of them. If questions directly

prompt students to talk about text content, they talk

about text content, and thus remember more text ideas

than they do if questions prompt them to access text

content through strategies. Indeed, strategies prompts

may split focus between talking about strategies and

talking about content.

Lesson Discourse

How the differences found in oral recall played out in

discussion can be gleaned from transcripts of the lesson

discussions in the strategies and content classrooms.

The discourse excerpts that we examined suggested

that the prompts to apply strategies may not lead stu-

dents to consider text content directly but indirectly,

taking them first through the strategy routine. For in-

stance, when asked to summarize, students’ focus may

go to the who, what, when, where structure in contrast

to “what’s going on” in this section. Or when asked to

think about a question to ask, students may consider

what makes a good question rather than what is impor-

tant content to be questioned.

The teacher questions asked in the content lessons,

on the other hand, seemed to encourage students to

express and integrate what they had understood from

text, as evidenced in their recalls of the text. Why is

this good? The notion is that the discussion provides a

kind of external model of comprehension, that is, going

through text, selecting what is important, and connect-

ing those ideas to build an understanding.

Expository Text

We had three dependent measures for expository text

that we implemented only in the second year: SVTs, re-

call, and learning probes. As was the case with narra-

tive texts, there were no differences on SVT measures

(Table 12). In contrast to narrative recalls, there were no

differences among the approaches for expository-text

recall (Table 14). The latter may relate to the more dif-

ficult content in expository text in terms of knowledge

about the world. Anticipating such potential difficulty,

we included knowledge prompts, which asked students

to talk about the key concepts in the texts (e.g., “Tell me

about infrasonic sounds”). The idea was that it may be

too difficult to go into memory and retrieve new con-

cepts, but when probes are provided, students may be

able to bring forth relevant information. Thus, while the

potential explanation for no differences on the SVTs is

that students in each approach did moderately well, the

lack of differences in recall may indicate that all stu-

dents had more difficulty because the expository-text

content was unfamiliar. Yet, the prompts seemed to pro-

vide enough support for the content students to access

what they had read relative to the strategies students.

Comprehension-Monitoring Task

The comprehension-monitoring task, in which stu-

dents identified anomalous content, was intended to be

a measure of transfer that might have developed from

experiencing the comprehension approach used in their

classroom. We used this particular task for two reasons:

First, identifying obstacles to comprehension is key to

monitoring one’s comprehension, and second, we had

used a similar task in previous research and had found

differences between instructional groups. Yet the results

of the comprehension-monitoring task in the current

study show no differences among approaches in either

year.

In considering what to make of the comprehension-

monitoring results, two features that also appeared

in both years are of note. First, there was significant

increase on the measure across time regardless of ap-

proach, and second, the scores were very low overall.

The increase in scores suggests that the task did capture

something about comprehension ability and supports

the idea that all approaches had positive effects. Yet the

scores themselves suggest that the task was quite dif-

ficult for students. Our conclusion then is that the task

we used may be of value, but redesign is called for be-

fore it could be a useful measure in further studies.

Strategies Task

Another attempt to assess learning beyond the lesson

text was the strategies task. In the strategies task, we

attempted to measure students’ application of strate-

gies on a paper-and-pencil task. Students read short

texts and answered questions including “Which is the

best summary?” and “Which is an inference you could

draw?” Results show that there were no differences

across approaches (Tables 8 and 18). This result might

be interpreted to mean that the instruction in the strat-

egies approach did not provide students an advantage

in explicit knowledge of strategies compared with what

they had been exposed to in their regular classroom

Reading Research Quarterly • 44(3)

244

instruction. We cannot draw this conclusion, however,

because the scores were very low overall, and in the

second year, when there was a pre- and posttest, no dif-

ferences were found over time regardless of approach.

We would have expected the strategies approach to

show an advantage. (As can be seen from the example

strategy-lesson excerpt, students did seem to know the

strategies.) The lack of an advantage might be attributed

to the way strategies were measured, requiring some-

thing different from the way students were taught about

strategies. The assessment task required students to read

a short text and then be reflective about “which of four

possibilities...is a good summary...[or] is an inference

you could draw,” in contrast to developing a summary or

making an inference through questions and interactions

with the text, teacher, and peers. The task requirements

may have been different enough from what students

were doing in the lesson that students were not able to

transfer their knowledge of strategies to a more analytic

task. Although Kozminsky and Kozminsky (2001) used

a very similar task and found significant differences

among groups of ninth graders of different educational

levels (e.g., academic, vocational, learning disabilities),

those subjects had not engaged in an intervention, and

as such, the task may not be sensitive to instruction.

Transfer Task

Another transfer task included in Year 2 scaffolded

students toward using the comprehension approach

they had experienced in class with a text read with-

out such support. In considering the possible outcomes

of a transfer measure, it is of note that an underlying

motivation for teaching strategies is that strategies are

seen as generative and that learning strategies should

provide students with tools they can use on their own

and apply to any text. We would contend that a content

approach is generative also, as the discussion simulates

the comprehension process, and experience with such

discussions may promote the development of a kind of

mental template of that model that students incorporate

into their own repertoires.

The results for the transfer task showed no signifi-

cant advantage by approach for allowing students to

transfer what they had learned (Table 19). However,

the trend of higher mean scores for the content group

was once again in evidence for both length and quality

of recall and was associated with moderate effect sizes

(d = .490 and d = .460, respectively).

Role of Basal-Comprehension Approach

Why did the basal-comprehension approach deliver re-

sults nearly equal to those of the content approach and in

some cases better than those of the strategies approach?

Our expectation was that the basal-comprehension

approach would produce discussions that ranged fur-

ther from text content than did the experimental ap-

proaches. We based that position on our familiarity

with conventional basal lessons, which tend to target

questions to a variety of issues, including direct inqui-

ries about text content, establishing purpose for read-

ing, character feelings and motivations, connections to

personal experience, applying strategies to text ideas,

and questions beyond guiding comprehension, such as

word recognition, grammar, and progress-monitoring

assessments. Moreover, a host of other activities sur-

rounds the text, such as connections to other content

areas, activities for students at different skill levels, flu-

ency activities, and suggestions for literary analysis, as

well as postreading questions and suggested assessment

strategies. The scripted version of the basal-comprehen-

sion lessons we developed included only questions in-

tended to be asked during reading and only those that

pertained to comprehension. As such, the questions did

not trigger as diverse a discussion as might have oc-

curred had all questions and activities been used.

Thus, the basal-comprehension approach in this

study, which we had originally viewed as conventional

instruction, was quite constrained compared with a

business-as-usual basal control. Constraining the basal-

comprehension approach was done deliberately so we

could understand what was going on in this control ap-

proach. However, we may have inadvertently created

a somewhat advantaged control because the focus on

comprehension might have been stronger than is typical

of basal instruction.

Beyond the scripted and prescribed questions, the

other feature that likely increased the basal’s effective-

ness was the interspersed reading-and-discussion for-

mat, which we have already acknowledged as a feature

that supported each group. Even though it appears that

the intent of basal publishers is to offer questions to be

used during reading, our observations at the school

used in this study and beyond reveal that teachers rare-

ly conduct the lessons in that way. Rather, they assign

the text to be read silently and then ask questions.

Limitations and Implications

What can we say about our results on a general level

that is meaningful to issues in the field, both in terms

of research and classroom practice? First, the studies

reported here are not meant as definitive answers about

strategies and content instruction. The instantiation

of strategies is only one example of what strategies in-

struction might look like. There are other choices that

could be made in the strategies selected, the way they

were introduced, and the way in which they were ap-

plied in the lessons. There are, as well, other instantia-

tions of content instruction that could be employed. For

example, one might use interpretive questions, such as

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

245

those in Junior Great Books (Dennis & Moldof, 1983),

or global questions meant to draw a variety of student

judgments, such as those in collaborative reasoning

(Anderson et al., 1998).

A question that remains open is whether student

ability or cognitive factors might affect the results.

Future researchers could provide such insight by devel-

oping studies that are designed to consider examination

of treatment by ability interactions.

Another limitation to the work reported here is our

selection of dependent measures. What is measured

naturally governs the effects that are shown, and in any

one effort, there are limits to the number and kind of

effects that are feasible to measure. Specific to the cur-

rent studies, the measurement of transfer was limited.

Clearly it would be desirable to measure further trans-

fer, for example, what students could do if given a text

to read completely on their own. Also desirable would

be to measure such transfer following some lapse of

time after instruction.

Despite the limitations, we see the work as having

implications that are meaningful to research and prac-

tice. We acknowledge that the differences we found that

favor the content approach were small, yet they showed

a consistent pattern. Second, given the very stringent

design approaches in this study, it is not surprising that

the differences we found were small. All three approach-

es used scripted lessons of high-quality instruction; all

teachers were trained, observed, and given feedback;

and each approach was based on interspersed reading

and discussion, which, as noted, provides a theoreti-

cally strong foundation for comprehension. Indeed, that

format may well have been the most positively influen-

tial feature in the instructional design.

An issue for practice is that interspersed reading

and discussion may be a departure from how reading

lessons are typically taught in intermediate grades. It

follows then that our recommendation for practice is

that at least some of the time teachers work through

a text with students, alternating reading with discus-

sion. The value of this practice includes enabling stu-

dents with poor decoding skills to have access to the

text, providing an external model of the comprehension

process, allowing students to have access to their peers’

thinking, assisting students in building meaning that

may be provided by collaboration, and having an expert

other—the teacher—support students’ meaning build-

ing. The expert other also benefits from this format, for

instance, in being able to observe confusion as it occurs

and consider what the source of confusion might be.

The teacher also gains understanding of the complexity

of the comprehension process and students’ individual

differences in engaging in that process. This provides

the teacher a basis for deciding how to intervene to sup-

port students.

Within the reading-and-discussion format, the ma-

jor question that prompted this study arises: Which

better supports comprehension, a strategy focus or a

content focus? Even if results had indicated that strat-

egies and content produced no differences in student

outcomes, there is reason to question the productivity

of strategies instruction if an equally effective approach

allows students to consider text meaning directly rather

than indirectly through strategies. Our results indi-

cate that going directly for meaning is feasible and at

least as effective as pursuing meaning by going through

strategies.

Part of the rationale underlying both content and

strategies approaches is that they encourage active pro-

cessing. But what is active processing? Must it include

conscious and deliberate attention to the process or

just consciousness that a process exists—that active ef-

fort needs to be made toward understanding? Sinatra,

Brown, and Reynolds (2002) highlighted the importance

of an active stance in contrast to explicit knowledge of

strategy application. They explained their choice by

saying that allocation of deliberate attention while read-

ing may not involve deliberate awareness—competent

readers don’t consciously tell themselves “pay attention

to what’s important.” Further, Sinatra et al. considered

that continued attention to deliberate use of strategies

may undermine comprehension. This is because com-

prehension takes significant mental resources, which

are limited. So if some resources are devoted to calling

up strategies, adequate resources may not be directed

toward the actions needed for comprehension because

students are being asked to do something in addition to

making sense of the text.

Our findings suggest that getting students to ac-

tively build meaning while reading does not necessitate

knowledge of and focus on specific strategies, but, rath-

er it may require attention to text content in ways that

promote attending to important ideas and establishing

connections between them. We acknowledge that the

consensus in the field is that strategies instruction is

useful (Baker, 2002; Block, Gambrell, & Pressley, 2002;

Gersten et al., 2001; NICHD, 2000). But an overall posi-

tive picture does not provide insight into what it is about

strategy instruction that may enhance student reading.

Such a general claim is inadequate for guiding teachers

as to effective ways of using strategies in the classroom.

Our review of the literature demonstrates that how

strategies should be taught is not easily derived from

the research. Among the problems is that strategy labels

do not represent a consistent set of activities, and there

is little description of how students interact with strate-

gies in the course of reading.

It can be argued that the strategies instruction we

implemented was not optimal. Given our design and

iterative review process, however, we are satisfied that

Reading Research Quarterly • 44(3)

246

our lessons were a reasonable representation of strate-

gies instruction. A claim that there are better represen-

tations of strategies instruction carries with it a need

to not only provide such representations but also show

what it is about them that may lead to better outcomes.

Our specifications of instructional procedures and re-

cord of every lesson enable detailed understanding of

what transpired during the lessons and allow reason-

able hypotheses to be made about the learning conse-

quences of the lessons. Additionally, our lesson design

provides a touchstone that can be scrutinized, com-

pared, and evaluated.

So what is the appropriate role of comprehension

strategies in the classroom? Surely, no one would want

to omit from the curriculum the ability to develop sum-

maries or be aware of the need to draw inferences. Our

recommendation is that such strategies be taught but

that they not lead the process of building understanding

of a text. Rather, we suggest concepts such as summa-

ry, inference, and prediction be introduced to students

with examples using short texts. Then those terms can

be used during discussion to talk about such text inter-

actions as they occur. For example, if a student sum-

marizes a section of text during discussion, the teacher

could comment that the student provided a good sum-

mary, or a teacher might point out that a student had

drawn an inference.

In the present study, lesson transcripts showed that

strategies questions did prompt students to bring key

ideas into the discussion but that students spent as

much time focusing on the strategic actions themselves

as on the content of what they were reading and seemed

less likely to connect the ideas. The importance of mak-

ing connections among ideas is paramount. Focusing on

strategies during reading may leave students less aware

of the overall process of interacting with text, especially

in terms of the need to connect ideas they encounter

and integrate those ideas into a coherent whole. Content

instruction, because it focuses directly on important

ideas and making connections, may be more likely to

support students in building a coherent representation.

Notes

We gratefully acknowledge the Institute for Education Sciences of

the U.S. Department of Education for its support of the research

described in this article. The opinions expressed do not necessarily

reflect those of the institute, and no official endorsement should be

inferred. The authors wish to express thanks to the teachers, stu-

dents, and administrative team whose efforts and cooperation made

this study possible. The authors also thank Timothy Shanahan, who

was instrumental in the conception and design of the study; Carol

Baker, whose reviews of data analyses were indispensable; and Maria

Barron for all of her work in data collection and materials design.

References

Anderson, R.C., & Biddle, W.B. (1975). On asking people ques-

tions about what they are reading. In G.H. Bower (Ed.), The

psychology of learning and motivation (Vol. 9, pp. 90–132). New

York: Academic.

Anderson, R.C., Chinn, C., Waggoner, M., & Nguyen, K. (1998).

Intellectually stimulating story discussions. In J. Osborn & F.

Lehr (Eds.), Literacy for all: Issues in teaching and learning (pp. 170–

186). New York: Guilford.

Anderson, R.C., & Pearson, P.D. (1984). A schema-theoretic view

of basic processes in reading comprehension. In P.D. Pearson, R.

Barr, M.L. Kamil, & P. Mosenthal (Eds.), Handbook of reading re-

search (pp. 255–291). New York: Longman.

Anderson, V., & Roit, M. (1993). Planning and implementing collab-

orative strategy instruction for delayed readers in grades 6–10. The

Elementary School Journal, 94(2), 121–137. doi:10.1086/461755

Applebee, A.N., Langer, J.A., Nystrand, M., & Gamoran, A. (2003).

Discussion-based approaches to developing understanding:

Classroom instruction and student performance in middle and

high school English. American Educational Research Journal, 40(3),

685–730. doi:10.3102/00028312040003685

Asimov, I. (2005). The fun they had. In I.L. Beck, R.C. Farr, &

D.S. Strickland (Eds.), Trophies: Distant voyages (pp. 584–592).

Chicago: Harcourt.

Baker, L. (2002). Metacognition in comprehension instruction.

In C.C. Block & M. Pressley (Eds.), Comprehension instruction:

Research-based best practices (pp. 77–95). New York: Guilford.

Baron, J. (1985). Rationality and intelligence. Cambridge, England:

Cambridge University Press.

Baumann, J.F., Kame’enui, E.J., & Ash, G.E. (2003). Research on

vocabulary instruction: Voltaire redux. In J. Flood, D. Lapp,

J.R. Squire, & J.M. Jensen (Eds.), Handbook of research on teach-

ing the English language arts (2nd ed., pp. 752–785). Mahwah, NJ:

Erlbaum.

Beck, I.L., & McKeown, M.G. (1981). Developing questions that

promote comprehension: The story map. Language Arts, 58(8),

913 –918.

Beck, I.L., & McKeown, M.G. (1998). Transforming knowledge into

tangible resources to support pedagogical change: Final report to

the Spencer Foundation.

Beck, I.L., & McKeown, M.G. (2006). Improving comprehension with

Questioning the Author: A fresh and expanded view of a powerful ap-

proach. New York: Scholastic.

Beck, I.L., McKeown, M.G., Sandora, C., Kucan, L., & Worthy, J.

(1996). Questioning the Author: A yearlong classroom implemen-

tation to engage students with text. The Elementary School Journal,

96(4), 385–414. doi:10.1086/461835

Beck, I.L., Omanson, R.C., & McKeown, M.G. (1982). An instruc-

tional redesign of reading lessons: Effects on comprehension.

Reading Research Quarterly, 17(4), 462–481. doi:10.2307/747566

Block, C.C. (1993). Strategy instruction in a literature-based read-

ing program. The Elementary School Journal, 94(2), 139 –151.

doi:10.1086/461756

Block, C.C., Gambrell, L.B., & Pressley, M. (Eds.). (2002). Improving

comprehension instruction: Rethinking research, theory, and classroom

practice. San Francisco: Jossey-Bass; Newark, DE: International

Reading Association.

Brown, A.L. (1981). Metacognition and reading and writing: The

development and facilitation of selective attention strategies for

learning from texts. In M.L. Kamil (Ed.), Directions in reading:

Research and instruction (pp. 21–43). Washington, DC: National

Reading Conference.

Brown, A.L. (1982a). Learning and development: The problems of

compatibility, access and induction. Human Development, 25(2),

89–115.

Brown, A.L. (1982b). Learning how to learn from reading. In J.A.

Langer & M.T. Smith-Burke (Eds.), Reader meets author/bridg-

ing the gap (pp. 26–54). Newark, DE: International Reading

Association.

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

247

Brown, A.L., Bransford, J.D., Ferrara, R.A., & Campione, J.C. (1983).

Learning, remembering, and understanding. In P.H. Mussen

(Series Ed.), J.H. Flavell & E.M. Markman (Vol. Eds.), Handbook

of child psychology: Vol. 3. Cognitive Development (4th ed., pp. 77–

166). New York: Wiley.

Brown, A.L., & Campione, J.C. (1981). Inducing flexible thinking:

The problem of access. In M. Friedman, J.P. Das, & N. O’Connor

(Eds.), Intelligence and learning (pp. 515–529). New York: Plenum.

Brown, A.L., & Smiley, S.S. (1977). Rating the importance of struc-

tural units of prose passages: A problem of metacognitive devel-

opment. Child Development, 48(1), 1–8. doi:10.2307/1128873

Brown, A.L., & Smiley, S.S. (1978). The development of strate-

gies for studying texts. Child Development, 49(4), 1076 –1088.

doi:10.2307/1128747

Carnine, D.W., Kame’enui, E.J., & Woolfson, N. (1982). Training

of textual dimensions related to text-based inferences (Brief

Research Report). Journal of Reading Behavior, 14(3), 335–340.

Carver, R.P. (1987). Should reading comprehension skills be taught?

In J.E. Readence & R.S. Baldwin (Eds.), Research in literacy: Merging

perspectives. 36th yearbook of the National Reading Conference (pp.

115–126). Rochester, NY: National Reading Conference.

Chinn, C.A., Anderson, R.C., & Waggoner, M.A. (2001). Patterns of

discourse in two kinds of literature discussion. Reading Research

Quarterly, 36(4), 378–411. doi:10.1598/RRQ.36.4.3

Chinn, C.A., O’Donnell, A.M., & Jinks, T.S. (2000). The structure

of discourse in collaborative learning. Journal of Experimental

Education, 69(1), 77–97.

Dennis, R.P. & Moldof, E.P. (1983). A handbook on interpretive reading

and discussion. Chicago: Great Books Foundation.

Dole, J.A., Duffy, G.G., Roehler, L.R., & Pearson, P.D. (1991). Moving

from the old to the new: Research on reading comprehension in-

struction. Review of Educational Research, 61(2), 239–264.

Duffy, G.G., & Roehler, L.R. (1989). Why strategy instruction is so

difficult and what we need to do about it. In C.B. McCormick,

G.E. Miller, & M. Pressley (Eds.), Cognitive strategy research: From

basic research to educational applications (pp. 133–157). New York:

Springer-Verlag.

Duffy, G.G., Roehler, L.R., Sivan, E., Rackliffe, G., Book, C., Meloth,

M.S., et al. (1987). Effects of explaining the reasoning associated

with using reading strategies. Reading Research Quarterly, 22(3),

347–368. doi:10.2307/747973

Durkin, D. (1978–1979). What classroom observations reveal about

reading comprehension instruction. Reading Research Quarterly,

14(4), 481–533. doi:10.1598/RRQ.14.4.2

Eeds, M., & Wells, D. (1989). Grand conversations: An exploration

of meaning construction in literature study groups. Research in the

Teaching of English, 23(1), 4–29.

Facklam, M. (1992). Bees dance and whales sing. San Francisco: Sierra

Club Books for Children.

Flesch, R. (n.d.). How to write plain English. Retrieved March 17,

2008, from www.mang.canterbury.ac.nz/writing_guide/writing/

flesch.shtml

Gersten, R., Fuchs, L.S., Williams, J.P., & Baker, S. (2001). Teaching

reading comprehension strategies to students with learn-

ing disabilities. Review of Educational Research, 71(2), 279–320.

doi:10.3102/00346543071002279

Graesser, A.C. (2007). An introduction to strategic reading compre-

hension. In D.S. McNamara (Ed.), Reading comprehension strate-

gies: Theories, interventions, and technologies (pp. 3–26). New York:

Erlbaum.

Graesser, A.C., Singer, M., & Trabasso, T. (1994). Constructing

inferences during narrative text comprehension. Psychological

Review, 101(3), 371–395. doi:10.1037/0033-295X.101.3.371

Graves, M.F., Cooke, C.L., & LaBerge, M.J. (1983). Effects of pre-

viewing difficult short stories on low ability junior high school

students’ comprehension, recall, and attitudes. Reading Research

Quarterly, 18(3), 262–276. doi:10.2307/747388

Great Books Foundation. (1987). An introduction to shared inquiry.

Chicago: Author.

Hood, M. (1981). The effect of previewing on the recall of high school

students. Unpublished master’s thesis, University of Minnesota,

Minneapolis.

Johnson-Laird, P.N. (1983). Mental models. Cambridge, MA: Harvard

University Press.

Kintsch, W. (1974). The representation of meaning in memory. Hillsdale,

NJ: Erlbaum.

Kintsch, W. (1998). Comprehension: A paradigm for cognition. New

York: Cambridge University Press.

Kintsch, W., & van Dijk, T.A. (1978). Toward a model of text com-

prehension and production. Psychological Review, 85(5), 363–394.

doi:10.1037/0033-295X.85.5.363

Klingner, J.K., Vaughn, S., & Schumm, J.S. (1998). Collaborative

strategic reading during social studies in heterogeneous fourth-

grade classrooms. The Elementary School Journal, 99(1), 3–22.

doi:10.1086/461914

Kozminsky, E., & Kozminsky, L. (2001). How do general knowl-

edge and reading strategies ability relate to reading compre-

hension of high school students at different educational levels?

Journal of Research in Reading, 24(2), 187–204. doi:10.1111/1467

-9 817.00141

Kucan, L., & Beck, I.L. (1997). Thinking aloud and reading com-

prehension research: Inquiry, instruction, and social interaction.

Review of Educational Research, 67(3), 271–299.

McGraw-Hill. (2000). Terra Nova test of basic skills (2nd ed.).

Monterey, CA: Author.

McKeown, M.G., Beck, I.L., & Sandora, C.A. (1996). Questioning

the Author: An approach to developing meaningful classroom

discourse. In M.F. Graves, P. van den Broek, & B.M. Taylor (Eds.),

The first R: Every child’s right to read (pp. 97–119). New York:

Teachers College Press.

McKeown, M.G., Beck, I.L., Sinatra, G.M., & Loxterman, J.A.

(1992). The contribution of prior knowledge and coherent text

to comprehension. Reading Research Quarterly, 27(1), 79 –93.

doi:10.2307/747834

Miller, G.E. (1985). The effects of general and specific self-

instruction training on children’s comprehension monitoring

performances during reading. Reading Research Quarterly, 20(5),

616–628. doi:10.2307/747947

National Institute of Child Health and Human Development. (2000).

Report of the National Reading Panel. Teaching children to read: An

evidence-based assessment of the scientific research literature on read-

ing and its implications for reading instruction (NIH Publication No.

00-4769). Washington, DC: U.S. Government Printing Office.

Nystrand, M. (with Gamoran, A., Kachur, R., & Prendergast, C.).

(1997). Opening dialogue: Understanding the dynamics of language

and learning in the English classroom. New York: Teachers College

Press.

Nystrand, M., Wu, L.L., Gamoran, A., Zeiser, S., & Long, D.A.

(2003). Questions in time: Investigating the structure and dy-

namics of unfolding classroom discourse. Discourse Processes,

35(2), 135–198. doi:10.1207/S15326950DP3502_3

Omanson, R.C. (1982). An analysis of narratives: Identifying cen-

tral, supportive, and distracting content. Discourse Processes,

5(3–4), 195–224.

Palincsar, A.S., & Brown, A.L. (1984). Reciprocal teaching of

comprehension-fostering and comprehension-monitoring activi-

ties. Cognition and Instruction, 1(2), 117–175.

Paris, S.G., Cross, D.R., & Lipson, M.Y. (1984). Informed strategies

for learning: A program to improve children’s reading aware-

ness and comprehension. Journal of Educational Psychology, 76(6),

1239–1252. doi:10.1037/0022-0663.76.6.1239

Reading Research Quarterly • 44(3)

248

Pearson, P.D., & Fielding, L. (1991). Comprehension instruction.

In R. Barr, M.L. Kamil, P.B. Mosenthal, & P.D. Pearson (Eds.),

Handbook of reading research (Vol. 2, pp. 815–860). White Plains,

NY: Longman.

Pearson, P.D., Hansen, J., & Gordon, C. (1979). The effect of back-

ground knowledge on young children’s comprehension of ex-

plicit and implicit information. Journal of Reading Behavior, 11(3),

201–209.

Pressley, M., El-Dinary, P.B., Gaskins, I., Schuder, T., Bergman, J.L.,

Almasi, J., et al. (1992). Beyond direct explanation: Transactional

instruction of reading comprehension strategies. The Elementary

School Journal, 92(5), 513–555. doi:10.1086/461705

Rickards, J.P. (1976). Stimulating high-level comprehension by in-

terspersing questions in text passages. Educational Technology,

16(11), 13–17.

Rickards, J.P., & Hatcher, C.W. (1977–1978). Interspersed meaning-

ful learning questions as semantic cues for poor comprehenders.

Reading Research Quarterly, 13(4), 538–553. doi:10.2307/747511

Rinehart, S.D., Stahl, S.A., & Erickson, L.G. (1986). Some effects

of summarization training on reading and studying. Reading

Research Quarterly, 21(4), 422–438. doi:10.2307/747614

Rothkopf, E.Z. (1966). Learning from written instructive materials:

An exploration of the control of inspection behavior by test-like

events. American Educational Research Journal, 3(4), 241–249.

Rothkopf, E.Z. (1972). Structural text features and the control of

processes in learning from written materials. In R.O. Freedle &

J.B. Carroll (Eds.), Language comprehension and the acquisition of

knowledge (pp. 315–335). Washington, DC: Winston.

Royer, J.M., Carlo, M.S., Dufresne, R., & Mestre, J. (1996). The as-

sessment of levels of domain expertise while reading. Cognition

and Instruction, 14(3), 373–408. doi:10.1207/s1532690xci1403_4

Royer, J.M., Hastings, C.N., & Hook, C. (1979). A sentence verifica-

tion technique for measuring reading comprehension. Journal of

Reading Behavior, 11(4), 355–363.

Rumelhart, D.E., & Norman, D.A. (1981). Analogical processes in

learning. In J.R. Anderson (Ed.), Cognitive skills and their acquisi-

tion (pp. 335–360). Hillsdale, NJ: Erlbaum.

Sandora, C., Beck, I.L., & McKeown, M.G. (1999). A comparison

of two discussion strategies on students’ comprehension and

interpretation of complex literature. Reading Psychology, 20(3),

177–212. doi:10.1080/027027199278394

Saunders, W.M., & Goldenberg, C. (1999). Effects of instruc-

tional conversations and literature logs on limited- and fluent-

English-proficient students’ story comprehension and thematic

understanding. The Elementary School Journal, 99(4), 277–301.

doi:10.1086/461927

Schmitt, M.C. (1988). The effects of an elaborated directed read-

ing activity on the metacomprehension skills of third graders. In

J.E. Readence & R.S. Baldwin (Eds.), Dialogues in literacy research:

Thirty-seventh yearbook of the National Reading Conference (pp. 167–

181). Chicago: National Reading Conference.

Sharp, A.M. (1995). Philosophy for children and the development

of ethical values. Early Child Development and Care, 107(1), 45–55.

doi:10.1080/0300443951070106

Sinatra, G.M., Brown, K.J., & Reynolds, R. (2002). Implications

of cognitive resource allocation for comprehension strategies

instruction. In C.C. Block & M. Pressley (Eds.), Comprehension

instruction: Research-based best practices (pp. 62–76). New York:

Guilford.

Snow, C.E., Burns, M.S., & Griffin P. (Eds.). (1998). Preventing

reading difficulties in young children. Washington, DC: National

Academy Press.

Soto, G. (2005). Off and running. In I.L. Beck, R.C. Farr, & D.S.

Strickland (Eds.), Trophies: Distant voyages (pp. 492–506).

Chicago: Harcourt.

Sternberg, R.J. (1979). The nature of mental abilities. American

Psychologist, 34(3), 214–230. doi:10.1037/0003-066X.34.3.214

Sternberg, R.J. (1982). A componential approach to intellectual de-

velopment. In R.J. Sternberg (Ed.), Advances in the psychology of

human intelligence (Vol. 1, pp. 413–463). Hillsdale, NJ: Erlbaum.

Symons, S., Snyder, B.L., Cariglia-Bull, T., & Pressley, M. (1989).

Why be optimistic about cognitive strategy instruction? In C.B.

McCormick, G.E. Miller, & M. Pressley (Eds.), Cognitive strategy

research: From basic research to educational applications (pp. 3–32).

New York: Springer-Verlag.

Tierney, R.J., & Cunningham, J.W. (1984). Research on teaching

reading comprehension. In P.D. Pearson, R. Barr, M.L. Kamil, &

P.B. Mosenthal (Eds.), Handbook of reading research (pp. 609–655).

White Plains, NY: Longman.

Trabasso, T., Secco, T., & van den Broek, P.W. (1984). Casual cohe-

sion and story coherence. In H. Mandl, N.L. Stein, & T. Trabasso

(Eds.), Learning and comprehension of text (pp. 83–111). Hillsdale,

NJ: Erlbaum.

van den Broek, P., Young, M., Tzeng, Y., & Linderholm, T. (1998).

The landscape model of reading: Inferences and the on-line con-

struction of a memory representation. In H. van Oostendorp &

S.R. Goldman (Eds.), The construction of mental representations dur-

ing reading (pp. 71–98). Mahwah, NJ: Erlbaum.

Van Dijk, T.A., & Kintsch, W. (1983). Strategies of discourse compre-

hension. New York: Academic.

Watts, G.H., & Anderson, R.C. (1971). Effects of three types of in-

serted questions on learning from prose. Journal of Educational

Psychology, 62(5), 387–394. doi:10.1037/h0031633

Wegerif, R., Mercer, N., & Dawes, L. (1999). From social interac-

tion to individual reasoning: An empirical investigation of a pos-

sible socio-cultural model of cognitive development. Learning and

Instruction, 9(6), 493–516. doi:10.1016/S0959-4752(99)00013-4

Wilkinson, I.A.G., Soter, A.O., & Murphy, P.K. (in press). Developing

a model of quality talk about literary text. In M.G. McKeown & L.

Kucan (Eds.), Bringing reading researchers to life: Essays in honor of

Isabel L. Beck. New York: Guilford.

Woodcock, R.W., Mather, N., & Schrank, F.A. (2004). Woodcock-

Johnson III Diagnostic Reading Battery. Itasca, IL: Riverside.

Yost, M., Avila, L., & Vexler, E.B. (1977). Effect on learning of

postinstructional responses to questions of differing degrees of

complexity. Journal of Educational Psychology, 69(4), 399–408.

doi:10.1037/0022-0663.69.4.399

Submitted April 8, 2008

Final revision received February 20, 2009

Accepted March 3, 2009

Margaret G. McKeown is a clinical professor of education

and a senior scientist at the Learning Research and

Development Center, University of Pittsburgh, PA, USA;

e-mail mckeown@pitt.edu.

Isabel L. Beck is professor emerita in the School of

Education at the University of Pittsburgh, PA, USA; e-mail

ibeck@pitt.edu.

Ronette G.K. Blake is a research specialist at the Learning

Research and Development Center at the University of

Pittsburgh, PA, USA; e-mail rgk6@pitt.edu.

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

249

Narrative Texts From Basal Series

Frindle by Andrew Clements

This story is classified as realistic fiction. It is one of

the simpler stories in this theme.

Synopsis: The witty class clown invents a new word

for pen and comes up with a plan to get all of his class-

mates to use the word.

Off and Running by Gary Soto

This story is classified as realistic fiction. A number

of Spanish words have been incorporated into the text

of this story.

Synopsis: A girl is running for class president against

the class clown. In her struggle to win, she seeks advice

from a female relative who was formerly mayor of a city

in Mexico.

Little by Little by Jean Little

This story is classified as an autobiography. This sto-

ry is set apart from the others as it invokes background

knowledge about World War I.

Synopsis: Jean Little, the partially blind author, tells

a story about being teased at school for having poor eye-

sight and struggling with math.

Dear Mr. Henshaw by Beverly Cleary

This story is classified as realistic fiction. Instead of

being written in standard narrative form like the other

stories in the theme, this story is written as a series of

journal entries.

Synopsis: A boy who aspires to be an author copes

with anxieties at school and at home by writing in his

journal. He designs an elaborate alarm for his lunch-

box that wins him notoriety at school, and after writing

about the alarm for a school writing contest, he gets a

chance to meet a published author.

The Fun They Had by Isaac Asimov

This story is classified as science fiction. This is the

only science fiction story in this theme, and it requires

more abstract thought than do the other stories.

Synopsis: A boy living in the future finds a book

about school in our time. After sharing the book with

his friend, Margie, she decides that she would rather go

to school with other children and with a human teacher

like in our present time rather than alone and with a

robotic teacher as in her future time.

Expository Texts

Body Language by Margery Facklam

This text describes how a variety of animals use

body language to communicate with each other. It also

describes how we can read this body language and use

it for our own advantage. The concepts addressed in this

text are concrete and familiar to students. Therefore,

this text appears to be the easiest of the three exposi-

tory texts.

Barks, Chirps, and Melodies by Margery Facklam

This text describes how Vervet monkeys make dif-

ferent warning calls for different types of dangers. It

also addresses the different calls that birds use to com-

municate with each other.

Messages by the Mile by Margery Facklam

This text describes how whales and elephants com-

municate through infrasonic messages, sounds that

humans cannot hear but can feel. This text also talks

about scientific research on these sounds. Due to the

abstract nature, this text appears to be the most difficult

of the three expository texts.

Appendix A

Summaries of Narrative and Expository Texts

Used in Lessons

Reading Research Quarterly • 44(3)

250

Off and Running:

Strategies

Prompt 8: Question Generation

1. Let’s think back about the important parts of what

we just read, and ask a question that will make

us think more about those parts. Ask a question

like a teacher would ask on a test. Remember we

should be able to answer these questions from the

text we just read.

• One student responds with a question. Accept all

reasonable responses.

2. Is this question getting at the important ideas,

and can we answer it from what we just read?

3. Now let’s think some more about what’s impor-

tant in what we just read and ask another ques-

tion. Remember, we have to be able to answer the

question from what we just read.

• One student responds with a question. Accept all

reasonable responses.

4. Is this question getting at the important ideas,

and can we answer it from what we just read?

5. So why do we ask ourselves questions?

[student response]

Yes! So, did our questions make our brains think

more about the important parts of what we’ve

been reading?

6. If students have trouble generating questions

How can we can help ourselves come up with a

question?

[If needed, hold up cheat sheet/card as a clue]

That’s right, think of our question words. Look on

your cheat sheet and say the words with me: who,

what, when, where, why, or how. So, who can use

one of those words to ask a question here?

Call on one student to respond; proceed as in #4.

Possible questions:

What is Miata’s promise if she wins the election?

Content

Prompt 8

To establish that Miata plans to improve school ap-

pearance if she becomes president and to establish that

Miata feels that her ideas are silly after hearing Dona

Carmen:

What’s this all about?

If students don’t mention point above:

Reread:

“After listening to the old woman’s story, Miata was

afraid that she had nothing really to offer.”

What does this tell us?

Basal

Prompt 10: Make Judgments

Do you agree with Miata’s statement that the things

she wants to do are just little things? Why or why

not?

Possible response: No, cleaning up graffiti, repairing bro-

ken equipment, caring for lawns, and planting flowers

are big jobs for a fifth grader.

The Fun They Had:

Strategies

Prompt 2: Infer

1. Let’s stop here and make an inference about the

teacher.

• Students respond with an inference.

2. If students responded with an appropriate inference:

What did you do to make an inference?

• If students can’t describe the process

To infer, we put together an idea we just read with

something we read earlier, or know already, to

come up with an idea that’s not written in the text

Appendix B

Example Prompts From Lesson Scripts

(From Off and Running and The Fun They

Had) for Each Approach

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

251

but the author wanted us to think. Sometimes in-

ferring is called reading between the lines.

3. With this in mind, did we make a good inference?

(If needed, use the definition to help students evaluate

the inference, and if it’s not appropriate, to revise it.)

4. If students cannot come up with an appropriate

inference

We need to put together what we just read with

something that we read earlier to make an in-

ference. So we just read that the teacher is me-

chanical—which means it’s a kind of machine.

We read that the teacher can be taken apart and

repaired, and it has a screen where you can see

lessons.

• So if we read between the lines, what can we

infer from those two things?

Students respond with an inference.

• If students still can’t come up with a good inference

Since the teacher is mechanical, can be repaired,

and has a screen where you can see lessons,

we can infer that the teacher is some kind of

computer.

Content

Prompt 2

To clarify that the teacher is not just a machine but some

kind of computer:

So what’s the author telling us now about the

teacher?

If students repeat phrases from the text:

That’s what the author says, but what does he mean

by that?

Basal

Prompt 3: Main Idea

Why did Margie hate putting her homework and

tests into the slot most of all?

Possible response: Because it was impersonal, pro-

grammed, and uninteresting; she had to use a punch code,

which meant her grade was calculated immediately.

Note. The script examples are for the three separate instructional

approaches and thus vary in style and format.

Reading Research Quarterly • 44(3)

252

True False

Paraphrase The wolf blows down the first little pig’s

house and then eats him.

The wolf blows down the first little pig’s

house and puts him in a cage.

Local inference The wolf could not blow down the brick

house because the bricks were too heavy.

The wolf could not blow down the brick

house because he was too hungry.

Global inference The third little pig avoids being eaten by the

wolf because he is smarter than the wolf.

The third little pig avoids being eaten by the

wolf because he is stronger than the wolf.

Appendix C

Example of Sentence Verification Technique

Item Design Based on The Three Little Pigs

Rethinking Reading Comprehension Instruction: A Comparison of Instruction for Strategies and Content Approaches

253

Go-Kart

When I was 11 years old, I got in the biggest trouble of

my life. It was all because my father bought me a go-

kart. The go-kart was fire engine red and had a chain-

saw motor that was a screaming terror. My family lived

in a neighborhood with winding dirt roads. Before long

I was blasting through turns, kicking up gravel in the

go-kart.

The roads weren’t the only things that were dirt. So

were the driveways. But one morning, an asphalt truck

pulled up to our house. Soon our dusty, rutted drive

became a perfect, smooth, black ribbon.

A few days later my mom and dad had to go out.

Before they left, my dad reminded me of my promise.

I promised I would never use the go-kart if he wasn’t

around. If I did, no more go-kart.

After my parents left, one of my friends came over.

In no time we were rolling the kart out on the drive.

I figured a little ride wouldn’t hurt. Besides, my dad

would never know.

The gas tank on the go-kart was empty. We kept ex-

tra gas in a giant, 10-gallon gas can. We dragged the full

can across the driveway and lifted it up, but it was too

heavy for us. We ended up pouring one gallon of gas into

the go-kart and about nine gallons onto the driveway.

The next thing I knew, the go-kart was sitting in a

puddle of gooey black muck. And that muck just melted

away as we watched. We just stared at the crater like it

was a science experiment gone very wrong.

I knew I was going to be in big trouble. And as it

turned out later, I was right. But for right now, not only

had I broken my promise about not using the go-kart,

but I’d also messed up our brand-new driveway. I felt so

bad. I just rolled the go-kart back to the garage.

While I waited for my parents to come home, I felt

worse and worse. Finally, they pulled into the drive and

parked right over the hole. They hadn’t noticed it. Was

I lucky!

I remembered hearing that cars can leak gas. Maybe

my dad would believe that was what happened to our

drive. If he believed that, I was saved!

1. What might a teacher ask on a test to see if you

understood important ideas from the story?

a. Why was the asphalt driveway important?

*b. Why will the boy’s parents be upset with

him?

c. What did the kid do with the go-kart after

filling it with gas?

d. Why did the boys play with the go-karts

when parents weren’t home?

2. What would be a good prediction about what

might happen next in the story?

a. The kid’s parents will believe their car caused

the driveway hole, and the kid will only be in

trouble for not telling them the gas can was

empty.

b. The kid will tell Mom that his friend caused

all the trouble.

*c. When Dad moves the car into the garage, he’ll

start yelling about the driveway.

d. The kid will tell Dad everything, but he won’t

get in trouble because he told the truth.

3. Reading between the lines, what could you infer

from the story?

*a. When gas gets on fresh asphalt, the asphalt

melts away.

b. Some cars have gas tanks that leak.

c. A lot of kids in the neighborhood had go-

karts.

d. You have to take special care of asphalt drive-

ways so they stay in good shape.

4. Which of these is a good summary of the story?

a. A kid broke his promise by filling up the gas

tank on his go-kart, and he made a mess of

his new driveway.

b. A kid wanted to use his go-kart, but every-

thing went wrong because he was out of gas.

c. A kid’s friend talked him into breaking the

rules with his go-kart, and because of that

they messed up the driveway.

*d. A kid broke his promise and tried to use his

go-kart without his dad, and he ruined their

new driveway doing it.

Note. * indicates correct choice.

Appendix D

Example of Strategies-Task Text

and Questions