Reading

From 1997 to 2000 the National Reading Panel (NRP) carried out a review of research-based knowledge about reading and instruction, especially in the early elementary grades. The research topics relevant to early reading and instruction that the NRP concentrated on were phonemic awareness instruction, phonics instruction, fluency, vocabulary instruction, text comprehension instruction, teacher preparation, and comprehension strategies instruction.

The group that focused on text comprehension instruction located more than 500 studies on the teaching of comprehension instruction. Using scientific criteria such as whether a study of a strategy instruction included a control group, they found that just over 200 of these studies were conducted sufficiently well to be confident that the conclusions based on them are scientifically trustworthy.

Comprehension strategy instruction fosters active reading. The strategies are designed to guide a reader to become more self-aware of one's self-understanding during reading, to become more in control of that understanding, to create images related to contents, to make graphic representations, to write summaries, and to answer or to make up questions. Depending on what type it is, a strategy can be implemented before, during, or after the readingof a text.

Skilled readers may invent strategies that help them understand and remember what they read. Most readers, however, do not spontaneously invent these strategies. Unless they are explicitly taught to apply cognitive procedures they are not likely to learn, develop, or use them. Readers at all levels, in fact, can benefit from explicit comprehension strategy instruction. A teacher begins by demonstrating or modeling a strategy. In some cases, the instruction is reciprocal or transactional, meaning that the teacher first performs the procedures and then the students gradually learn to implement them on their own. The process by which a student adopts the strategy–a process that is called "scaffolding"–is often a gradual one. Readers are first able to experience the construction of meaning by an expert reader, the teacher. As readers learn to take control of their own reading by practicing and acquiring cognitive strategy procedures, they gradually internalize the strategies and achieve independent mastery.

History of Comprehension Strategy Instruction

Interest in reading comprehension strategies began to grow as a part of the new scientific understanding of cognition that emerged in the latter decades of the twentieth century. In 1978 Walter Kintsch and Teun A. van Dijk observed that a reader is an active participant with a text and that a reader "makes sense" of how ideas based on the text relate to one another by interpretive interactions between what the reader gleans from the text and what the reader already knows. They proposed that a reader actively builds meaning as mental representations and stores them as semantic interpretations held in memory during reading. These representations enable the reader to remember and use what had been read and understood.

In a landmark 1979 study Ellen M. Markman wondered whether readers would detect obvious logical contradictions in passages they read. She gave readers a passage about ants that indicated that when ants forage away from their hill they emit an invisible chemical with an odor that they use to find their way home. The passage also indicated, however, that ants have no nose and are unable to smell. Would readers notice that the passage did not make sense? Would they recognize that they did not understand the passage? What would they do? Her disturbing finding was that young and mature readers alike overwhelmingly failed to notice either logical or semantic inconsistencies in the texts. What instruction would help readers to be more conscious of their understanding and to learn strategies that would over-come these comprehension failures?

At about the same time Dolores Durkin observed reading instruction in fourth-grade classrooms over the course of a school year. For many student readers, fourth grade is a transition year from "learning to read" to "reading to learn." In a 1979 article Durkin reported that there was very little comprehension instruction in the classrooms. Teachers assigned questions and told students about content. But in seventy-five hours of reading instruction Durkin observed that year, teachers devoted only twenty minutes, less than 1 percent of the time, to teaching readers how to comprehend and learn new information from reading. Her studies and the others cited above anticipated an intense interest in helping students learn strategies to comprehend and learn from reading.

In the 1970s and early 1980s investigators generally focused on teaching an individual strategy to help readers construct meaning. There were literally hundreds of studies of individual comprehension strategies. One example is Abby Adams and colleagues' 1982 research applying the SQ3R (survey, question, read, recite, and review) technique to fifth-grade classrooms. SQ3R is a text pre-reading graphic organizer instruction developed in 1941 for World War II military personnel undergoing accelerated courses. It is considered a "text previewing" comprehension strategy instruction in that it guides readers to look for the meaning before reading the text. In this instruction, readers learn to use the text's headings, subheads, introductions, and summaries to construct graphic schemata of the text content. As did many of the other comprehension strategy instruction researchers, Adams and her colleagues obtained positive results, finding that students with the pre-reading instruction performed significantly higher on factual short-answer tests than did control group students.

Generally, many types of individual comprehension strategy instructions appeared to be successful in improving readers' ability to construct meaning from text. With the observed success of various individual strategy applications, there were several reviews of this growing body of scientific literature. In 1983 P. David Pearson and Margaret C. Gallagher categorized cognitive strategies by what teachers do to teach the strategies, and Robert J. Tierney and James W. Cunningham's 1984 review subdivided the cognitive strategies into pre-reading, during-reading, and post-reading activities.

With the success of individual strategy instruction in improving reading comprehension measures documented by research, focus shifted to using combinations of strategies to facilitate text comprehension, primarily in experimental situations rather than in natural classrooms. Among these was a very influential 1984 study of "reciprocal teaching" of comprehension by Annemarie S. Palincsar and Ann L. Brown. Reciprocal teaching is a method that in volves the gradual release of responsibility for carrying out a strategy to the readers. It combines teacher modeling and student practice on four cognitive strategies: prediction, clarifying, summarizing, and question generation. Students who received this instruction showed marked improvement on a number of comprehension measures.

Success of teaching multiple strategies led to the study of the effectiveness of preparing teachers to teach comprehension strategies in natural, classroom settings. Two approaches are noteworthy, namely Gerald G. Duffy and Laura R. Roehler's 1987 direct explanation model and Rachel Brown, Michael Pressley, and colleagues' 1996 transactional instruction approach. Direct explanation emphasizes teacher-directed problem solving, whereas transactional instruction, similar to reciprocal teaching, employs teacher-directed actions with interactive exchanges with students in classrooms. Both direct and transactional approaches to training teachers have produced positive results.

Strategies that Work

The NRP identified twelve categories of comprehension instruction that have scientific support for the conclusion that they help readers to construct meaning and thereby improve reading comprehension, including two categories involving the preparation of teachers in cognitive strategy instruction. These strategies stimulate both audio and visual perception, activate memory and semantic processing, enhance perception, engage syntactic knowledge and processing, teach narrative structure, and promote reasoning. The strategies of active listening, comprehension monitoring, and prior knowledge use all serve to promote listening and awareness of one's thinking or "inner speech," a process emphasized by the Russian psychologist Lev Vygotsky in the 1920s. Mental imagery, mnemonic, and graphic organizer instruction, on the other hand, make use of readers' visual imagination and memory. Vocabulary instruction increases word and semantic knowledge and problem solving. Question answering and question generation require the access of what is known or understood and the prediction of future events. Story structure and summarization instruction create awareness of the organization of ideas and what is important. Finally, multiple strategy instruction combines the use of several of these processes together in flexible and appropriate ways. Research conducted in the late 1990s also suggests that teachers can learn to integrate these kinds of strategy instructions in classroom settings and that peers working in cooperative learning situations can effectively tutor each other in comprehension strategies.

Active listening. To instruct active listening, teachers guide readers in learning to listen while others read. The listening reader follows the text as another student reads aloud. The teacher may also pose questions for the readers to answer while they listen. Active-listening training improves listening and reading comprehension. It increases a reader's participation in discussions, engenders more thoughtful responses to questions, increases memory for the text, and focuses the reader's attention and interest on material. For example, in Gloria M. Boodt's 1984 study of training critical-listening strategies with fourth-grade to sixth-grade remedial readers, there was a gradual increase over the eighteen weeks of the study in students' willingness to participate in group discussions and provide more thoughtful responses to direct questions. Overall, four studies of this strategy met NRP scientific criteria. The students in the active listening studies ranged from first grade through sixth grade; they improved in critical listening, critical reading, and general reading comprehension.

Comprehension monitoring. One can learn to listen to one's own reading and to monitor one's own comprehension. Instruction in comprehension monitoring during reading helps readers manage their inner speech as they read. Self-listening and self-monitoring of one's own understanding during reading promote more careful reading and better comprehension.

To teach comprehension monitoring, a teacher, when reading aloud to a class, demonstrates the strategy by interrupting her own reading to "think aloud." She articulates to the class her own awareness of difficulties in understanding words, phrases, clauses, or sentences in a text. When a text poses potential comprehension breakdowns, such as unfamiliar concepts or logical inconsistencies in a passage, the teacher might look back in the text to try to solve a problem, restate the text content in more familiar terms, or look forward in the text to find a solution. After observing a teacher model the comprehension monitoring strategy, readers are encouraged to carry out the same procedures–first with teacher scaffolding and then on their own. Eventually the student readers take responsibility for recognizing comprehension difficulties and for demonstrating ways to overcome them (e.g., by guessing and looking back or reading forward in the text).

The teaching of comprehension monitoring is very effective. The NRP found twenty studies of comprehension monitoring instruction with readers in grades two through six that met scientific criteria. In them, readers who were taught to self-monitor comprehension improved one of the following: their detection of text inconsistencies, their memory for the text, or their performance on standardized reading comprehension tests.

Prior knowledge. Prior knowledge instruction is designed to assist readers in bringing to mind their own knowledge that is relevant to understanding the text. A teacher can activate prior knowledge by asking students to think about topics relevant to the passage, by teaching the requisite relevant knowledge, by using pre-reading activity on related but better-known topics, by having the readers predict what will happen in the text based on personal experience, by having readers make associations during reading, and by previewing the story or text.

In fourteen studies with students spanning grades one through nine reviewed by the NRP, prior knowledge instruction helped readers improve on recall, in question answering, and in content area and standardized reading comprehension performance. For example, in a 1988 study, Teresa A. Roberts found that prior knowledge instruction had a positive effect on both factual and inferential comprehension performance with students in grades five and nine.

Mental imagery. Mental imagery instruction teaches readers to construct images that closely represent the content of what was read and understood. In 1986, after instructing less-skilled fourth-and fifth-grade readers in imagery training, Linda B. Gambrell and Ruby J. Bales had them read stories with inconsistencies like those in the 1979 Markman study mentioned above and instructed them "to make a picture in your mind to help determine if there is anything that is not clear and easy to understand about the story." Control students, that is, those without the imagery training, were simply asked to "do whatever you can do to help determine if anything is not clear and easy to understand about the story." The results were that imagery-trained readers were more likely to detect inconsistencies than the controls. In four studies with students in grades two through eight, the NRP found that mental imagery instruction led to modest increases in memory for the text that was imaged and improved reader detection of text inconsistencies.

Mnemonics. Like mental imagery instruction, mnemonic instruction teaches readers to use an external memory aid, but unlike mental imagery instruction, the mnemonic image can be one that does not necessarily closely represent the text. A teacher demonstrates how to construct a picture, keyword, or concept as a proxy for a person, concept, sentence, or passage–such as using an image of a "tailor" to remember the name "Taylor." These keywords and images aid later recall. In five studies examined by the NRP, mnemonic instruction improved reader memory of the assigned keywords and recall for the passages read. For example, in 1986 Ellen E. Peters and Joel R. Levin gave mnemonic instruction to good and poor readers and then gave them passages about "famous" people. As compared to the control subjects, the mnemonic-trained students were more likely to learn and remember information about new concepts and people who were unfamiliar to them.

Graphic organizers. Graphic organizer instruction shows readers how to construct displays that organize one's ideas based on a reading of the text. Graphic organizers aim at creating awareness of text structures, concepts and relations between concepts, and tools to represent text relationships visually. They also assist readers in writing well-organized summaries. Diagrams, pictorial devices, and story maps can all be used to outline the relationships among text ideas. This instruction is useful for expository texts in content areas such as science or social studies.

In eleven studies reviewed by the NRP that used graphic organizers with readers in grades four through eight, readers generally benefited in remembering what they read, in improved reading comprehension, or in improved achievement in social studies or science courses. For example, in 1991 Bonnie Armbruster and her colleagues compared the effectiveness of a graphic organizer instruction that taught fourth-and fifth-grade social science students to visually represent the important ideas in a social science text. In contrast, the control students' instruction consisted of workbook activity directions recommended in the teacher's edition of the social science textbook. The fifth-, but not necessarily the fourth-grade students, who received the graphic organizer cognitive strategy instruction scored higher on recall and recognition measures than the controls who received the workbook activity instruction.

Vocabulary instruction. There are many studies on teaching vocabulary but few on the relationship between vocabulary instruction and comprehension. In the context of comprehension strategy instruction, vocabulary instruction promotes new word meaning knowledge by teaching readers semantic processing strategies. For example, students learn to generate questions about an unknown word by examining how it relates to the text or noticing how a word changes meaning depending on the context in which it occurs. The teacher may model being a "word detective," looking for contextual clues to find a word's meaning, analyzing words and word parts, and looking at the surrounding text for clues to a word's meaning. For instance, the word comprehension combines com, meaning "together" with pre-hension, meaning "able to grasp in one's hand." From this, an operational definition of comprehension can be constructed (e.g., putting together individual word meanings to grasp an idea).

In three studies of vocabulary instruction in a cognitive strategy context with fourth-grade students reviewed by the NRP, the instruction led to success in learning words, in use of word meanings, and in increased story comprehension. For example, in a 1982 study involving fourth-graders receiving vocabulary instruction, Isabel L. Beck and her colleagues taught the students to perform tasks designed to require semantic processing. These students performed at a significantly higher level than pre-instruction matched controls on learning word meanings, on processing instructed vocabulary more efficiently, and in tasks more reflective of comprehension. In the three NRP-reviewed studies, however, learning to derive word meanings did not always improve standardized comprehension performance.

Question answering. Question answering focuses the reader on content. Why or how questions lead the student to focus on causes and consequences. Question answering guides students and motivates them to look in the text to find answers. Instruction on question answering leads to improvement in memory for what was read, to better answering of questions after reading, or to improvement in finding answers to questions in the text during reading.

In a 1985 study, Taffy E. Raphael and Clydie A. Wonnacott trained fourth-grade and sixth-grade readers to analyze questions, distinguishing those questions that could be answered by information in the passage from questions that required prior knowledge or information not in the text. The results were that students who had received this instruction provided higher quality responses to questions than a control group of students. In seventeen studies examined by the NRP for this strategy, the results were usually specific to experimenter tests of question answering and were greater for lower-grade than for upper-grade readers and greater with average and less-skilled readers than with high-achieving readers.

Question generation. Teachers demonstrate this strategy by generating questions aloud during reading. Readers then practice generating questions and answers as they read the text. Teachers provide feedback on the quality of the questions asked or assist the student in answering the question generated. Teachers teach the students to evaluate whether their questions covered important information, whether questions related to information provided in the text, and whether they themselves could answer the questions.

The scientific evidence that question generation cognitive strategy instruction is effective is very strong. In 1996 Barak Rosenshine and his colleagues conducted meta-analysis of twenty-six question generation studies with students from third grade through college. Like individual experimental studies, a meta-analysis applies scientific criteria to obtain a quantitative assessment of an instruction's effectiveness. A meta-analysis differs from single studies, however, in that it obtains a quantitative impact of a particular strategy by looking at effectiveness across a group of studies. In addition to the Rosenshine meta-analysis, the NRP examined twenty-seven question generation studies with students from grades three through nine. Question generation instruction during reading benefited reading comprehension in terms of improved memory, in accuracy in answering questions, or in better integration and identification of main ideas. The evidence that it improved performance on standardized comprehension tests is mixed.

Story structure. Story structure instruction is designed to help readers understand the who, what, where, when, and why of stories, what happened, and what was done and to infer causal relationships between events. Readers learn to identify the main characters of the story, where and when the story took place, what the main characters did, how the story ended, and how the main characters felt. Readers learn to construct a story map recording the setting, problem, goal, action, and outcome of the story as they unfold over time.

Story structure instruction improves the ability of readers to answer questions, to recall what was read, and to improve standard comprehension test performance. The instruction also benefits recall, question answering, and identifying elements of story structure. For example, in 1983 Jill Fitzgerald and Daisy L. Spiegel found that instruction in narrative structure enhanced story structure knowledge and had a strong positive effect on reading comprehension with average and below-average fourth-grade students who had been identified as lacking a keen sense of narrative structure.

The NRP examined seventeen studies using story structure instruction with readers ranging from third grade through sixth grade. Story structure instruction improved readers' ability to answer short-answer questions and retell the story. In three of the studies, standardized tests were used for assessment. Story structure instruction led to improved reader scores in two of those studies.

Summarization. Teaching readers to summarize makes them more aware of how ideas based on the text are related. Readers learn to identify main ideas, leave out details, generalize, create topic sentences, and remove redundancy. Through example and feedback, a reader can be taught to apply these summarization rules to single-or multiple-paragraph passages by first summarizing individual paragraphs and then constructing a summary or spatial organization of the paragraph summaries.

In eighteen studies on summarization with students from grades three to eight examined by the NRP, readers improved the quality of their summaries of text not only by identifying the main ideas but also by leaving out detail, including ideas related to the main idea, generalizing, and removing redundancy. Further, the instruction of summarization improves memory for what is read, both in terms of free recall and answering questions. For example, in 1984 Thomas W. Bean and Fern L. Steenwyk examined whether training sixth-grade readers in rules for summarization developed in 1983 by Ann L. Brown and Jeannie D. Day would improve comprehension. They found that readers receiving summarization instruction either by rule-governed or intuitive-summarization techniques performed better than controls who were told to find main ideas but who had no explicit instruction. The summarizationtrained students significantly outperformed the control group in the quality of their summaries and on a standardized test.

Multiple-strategy instruction. Readers can learn and flexibly coordinate several comprehension strategies to construct meaning from texts. Palincsar and Ann L. Brown's reciprocal teaching method, described in 1984, instructs readers to use four main strategies during reading: generating questions, summarizing, seeking clarification, and predicting what will occur later in the text. Additional strategies may also be introduced, including question answering, making inferences, drawing conclusions, listening, comprehension monitoring, thinking aloud, and question elaborating. The teacher models strategies and, in some cases, explains them as they are modeled. Then the reader, either alone or as a leader of a group, applies the strategies.

The evidence indicates that demonstration and repeated use of the strategies leads to their learning by readers and improvement in comprehension. In 1994 Rosenshine and Carla Meister conducted a meta-analysis of sixteen reciprocal teaching studies with students in grades one through eight. Most of the readers were above grade three. Weaker and older readers benefited most from reciprocal teaching. In eleven studies of reciprocal teaching in grades one through six reviewed by the NRP but not covered by Rosenshine and Meister, reciprocal teaching produced clear positive improvement on tasks that involve memory, summarizing, and identification of main ideas.

Multiple-strategy programs that do not use reciprocal teaching mainly have the student practice strategies with modeling and/or feedback from the teacher. In explicit, direct approaches, the teacher always explains a strategy before the teacher models it during reading.

Teacher preparation for text comprehension instruction. Teachers have to learn strategy instruction in order to interact with students at the right time and right place during the reading of a text. Teachers also need to know about cognitive processes in reading and how to teach strategies through explanation, demonstration, modeling, or interactive techniques; how to allow readers to learn and use individual strategies; and how to teach a strategy in conjunction with several other strategies.

Four studies conducted in the late 1980s and 1990s indicated that teachers who learn multiple comprehension strategy instruction and use it in their classrooms improve the reading comprehension of their students, especially those who are below average in skill. Improvements occurred in subject matter learning and in performance on standardized reading comprehension tests. In 1996 Rachel Brown, Michael Pressley, and colleagues taught teachers to use transactional strategy instruction in a yearlong program where students made comprehension gains. Transaction instruction involves teacherdirected actions with interactive exchanges with students in classrooms.

Cooperative learning by peers. Readers may learn best when they are in social situations in which they are actively engaged with other learners who are near their same level of understanding. Cooperative learning involves readers reading together with a partner or in small groups. As they read aloud and listen to others, the teacher can guide them to use any of the various strategies for effective reading comprehension. At first the teacher may model reading through her demonstrated use of a strategy. Then the student readers carry out the demonstrated activities with a partner or in small reading groups. Readers take turns reading and listening, asking questions, answering questions, summarizing, recognizing words, predicting, and clarifying. The readers are encouraged to tutor each other on strategies. Group cooperative instruction has been found to promote intellectual discussion, increased student control over their learning, increased social interaction with peers, and savings in teacher time.

For example, in 1998 Janette K. Klingner, Sharon Vaughn, and Jeanne S. Schumm investigated the effectiveness of a cooperative learning approach designed to encourage culturally and linguistically diverse general education fourth-grade students to use strategic reading by employing various summarization and clarification procedures during reading. Students in the cooperative learning classes made greater gains in reading comprehension and equal gains in content knowledge than controls in measures that included a standardized reading test, a social studies unit test, and audiotapes of group work.

In ten studies on cooperative learning of comprehension strategies reviewed by the NRP, students successfully learned the reading strategies. Cooperative learning can also be effective for integrating students with academic and physical disabilities into regular classrooms. The social interaction increases motivation for learning and time spent by the learners on tasks.

Implications for Future Research and Practice

Instruction of cognitive strategies for reading comprehension has been successful across a wide number of studies for readers in grades three to eleven. Despite these successful demonstrations, there are many unanswered questions. Among these are whether certain strategies are more appropriate than others for readers of certain ages or different abilities, whether comprehension strategy instruction would improve performance and achievement in all content areas, whether successful instruction generalizes across different types of texts, and whether comprehension strategies work better if what is being read engages the readers' interests. Researchers also need to find out more about important teacher characteristics that influence successful instruction of reading comprehension, especially in regard to decision-making processes (e.g., knowing when to apply what strategy with which particular student[s]). Finally, there has been little research that directly compares different methods of teaching comprehension. More needs to be known about "best approaches" to comprehension instruction and the circumstances under which they are successful. How does one best develop independent readers who have the abilities to understand what they read on their own?

Conclusion

Cognitive strategy instruction does work to improve readers' comprehension performance. In her 2000 address to teachers, Carol Minnick Santa, president of the International Reading Association, noted that "teaching [comprehension] is a lot harder and more abstract than teaching phonemic awareness or language structures. Moreover, effective comprehension instruction … demands extensive teacher knowledge." In 1993, after a five-year study of teaching teachers to implement comprehension strategy instruction, Gerald G. Duffy, a developer of the direct-instruction approach to cognitive strategy instruction, concluded that teaching students to acquire and use strategies requires a fundamental "change in how teacher educators and staff developers work with teachers and what they count as important about learning to be a teacher" (p. 244). Successful comprehension teachers must be strategic themselves, coordinating individual strategies and altering, adjusting, modifying, testing, and shifting tactics appropriately until readers' comprehension problems are resolved. For readers to become good reading strategists requires teachers who have appreciation for reading strategies.