Chapter 8

Reading Comprehension

Definition

Reading comprehension can be described as understanding a text that is read, or the process of constructing meaning from a text (NRP, 2000b, p. 4-5). Why do those who have studied reading comprehension describe it as a process of "construction" or "making meaning?" First of all, reading comprehension involves all of the elements of the reading process, described in earlier sections of this review, acting together. As comprehension takes place, words are decoded and associated with their meanings in the reader's memory, and phrases and sentences are processed rapidly or fluently enough so that the meanings derived from one word, phrase, or sentence are not lost before the next is processed.

Second, the writer who composed the text being read put together a whole, hopefully coherent, network of thoughts for the reader. This more or less coherent whole is recreated as needed, piece-by-piece, in the reader's memory without the benefit of live conversation, relying only on what is derived from the text and the reader's own prior knowledge or past experiences, also stored in memory. This complex network of ideas that represents a text in memory is constantly modified as the reading progresses. Problems in creating this representation, or understanding a specific text, may be encountered as the text is processed. Recognizing these problems and reasoning through and resolving them is a part of the comprehension process. Comprehension is an active process and the reader must interact and be engaged with the text for it to work well.

Comprehension is a strategic process and these strategies can be taught. Strategies are

procedures that guide students as they attempt to read and write. For example, a reader may be taught to generate questions about the text as it is read. These questions are of the why, what, how, when, or where variety; and by generating and trying to answer them, the reader processes the text more actively (NRP, p. 4-40).

Typically, instruction of cognitive strategies employed during reading consists of:

1. The development of an awareness and understanding of the readers' own cognitive processes that are amenable to instruction and learning
2. A teacher guiding the reader or modeling for the reader the actions that the reader can take to enhance the comprehension processes used during reading
3. The reader practicing those strategies with the teacher assisting until the reader achieves a gradual internaliza-- tion and independent mastery of those processesÉ (p. 4-40)

Rationale

Why should reading comprehension be taught? Comprehension is the purpose of reading. In order to get information from a text, remember it later, and use it effectively, whether for work or for pleasure, reading comprehension is essential. Many readers are not aware of comprehension strategies and are not likely to develop them on their own (NRP, p. 4-40).

Assessment

Students read extended texts when their reading comprehension is assessed. These may consist of just a few sentences for beginning readers to long passages for more advanced readers. Students who have constructed good representations of a text they have read will be able to recall and make inferences from specific ideas in the text. The most common form of assessment is question-asking, although many other tasks may be used. Multiple choice questions, short answer questions, cloze tests, and summarizing are examples of tasks used to assess comprehension.

Other measures may be used to determine whether students have learned specific comprehension strategies such as question generation, error detection (detection of comprehension errors while reading), and other forms of comprehension monitoring.

The NRP distinguishes between two broad types of assessment. Just as a teacher may use teacher-designed tasks or published tests to assess student reading comprehension, researchers conducting experiments may also use both researcher-designed and standardized tests. For both teachers and researchers, the assessments that they design may be more closely related than standardized tests to what they are teaching or investigating.

...experimenter tasks reflect near transfer [to practiced tasks or those used during instruction] and standardized tests reflect far transfer (generalization of what has been learned) (p. 4-42).

Reading Comprehension Assessment:
Principles and Trends

Question

Based upon assessment results, what are ABE learners' strengths and needs in reading comprehension? Is one form of assessment best?

There are several studies that evaluate ABE students' reading comprehension ability. One of these also provides at least some information about ABE students with a reading disability and ABE students whose first language is not English. In addition to the above, comprehension is the only major topic area where research was found that begins to consider questions related to the quality of the tests used for assessment.

Principle 11:

Adults who qualify for ABE have poor functional literacy comprehension achievement.

Although they may be able to perform simple comprehension tasks such as recalling ideas from simple stories and locating a single piece of information in a simple text, they are often unable to combine (integrate and synthesize) information from longer or more complex texts. (Gold, 1983; Kirsch, Jungeblut, Jenkins, & Kolstad, 1993)

By definition, ABE students are those adults reading at Levels 1 and 2 (out of a total of five Levels) on the NALS measures of prose, document, and quantitative functional literacy (Adult Education and Family Literacy Act, Title II of the Workforce Investment Act, PL105-220, 1998; Kirsch et al., 1993). The NALS results are based on a sample of about 26,000 adults, aged 16 and older. These include a representative sample of the whole adult population in the United States (13,600), roughly 1,000 from each of 11 states, and 1,100 inmates in correctional institutions. NALS scores are arranged on a 500 point scale, with Level 1 scores ranging from 0 - 225 and Level 2 scores from 226 - 275. The average score for those taking the NALS is roughly 270, or the high end of Level 2 (averages of 272, 267, and 271 on the Prose, Document, and Quantitative scales respectively). The average score for those completing high school is 270 on the Prose scale. ABE students, then, are those whose functional reading comprehension ability is roughly the same as or lower than those completing 12 years of school (those receiving a high school diploma).

Those scoring at Level 1 on the NALS range from those who are unable to demonstrate an understanding of simple texts to those who are able to perform simple tasks such as locating a specific piece of information in a text. Seventy-eight percent of adults with 0-8 years of education score at Level 1.

Those scoring at Level 2 are able to locate information in a text, make simple inferences, and integrate pieces of information in simpler texts. They are generally not able to combine (integrate or synthesize) information from longer or more complex texts.

Additional research suggests that intermediate adult readers (those scoring at GE 4 -- 6 on a standardized reading comprehension test) are able to recall information from simple stories (Gold, 1983).

Trend 12:

When different assessment instruments are used to measure gain in reading comprehension achievement, or when the same instrument is used at several points over the course of instruction, results related to reading comprehension achievement may be extremely variable. Some comprehension measures may be more reliable than others. (Perin & Greenberg, 1993; Venezky, Bristow, & Sabatini, 1994)

Research: One experimental study, in which ABE students' reading comprehension was measured with two different tests at four points in time (once before, twice during, and once after instruction) found significantly larger gains for one group at time two on one measure, but not on another (Perin & Greenberg, 1993). In addition, one group's gain from time one to time four was significant on only one of the two measures. Growth on both measures was extremely variable over time. In a descriptive study, involving one group of ABE learners who were administered three tests of reading comprehension at three points in time, learners showed gain on all three measures from time one to time two, but on only one from time two to time three (Venezky, Bristow, & Sabatini, 1994). The three measures used were the TABE Reading Comprehension and the TALS Document and Quantitative tests. An analysis of the TABE and TALS found the TALS to be a more reliable measure of reading comprehension.

Trend 13:

Experimental results from one study in which ABE learners and skilled college readers were interviewed about their knowledge of reading comprehension suggest that ABE readers' metacomprehension ability is more like that of beginning readers described in the literature (Gambrel & Heathington, 1981). Interview results suggest that both groups are aware of the influence of motivation, interest, and prior knowledge on comprehension. ABE readers, however, are less aware of text structure, or how paragraphs and stories are organized, and strategies that can be used to resolve comprehension failure. Like younger beginning readers, they are more likely to associate reading with decoding as opposed to comprehending text.

ESOL

Questions

Is one form of reading comprehension assessment best for ABE students with lower levels of English language ability, particularly students in ESOL programs? Based upon assessment results, what are the strengths and needs in reading comprehension for ABE students with lower levels of English language ability?

Principle 12:

ESL adults, on average, tend to have lower functional literacy comprehension achievement in English; the percentage of ESL adults among the ABE target population is greater than the percentage among the general adult population. (Kirsch, Jungeblut, Jenkins, & Kolstad, 1993)

Adults born outside this country, and likely to have learned a language other than English as their first language (ESL adults) score lower on the NALS than those born in the United States (Kirsch et al., 1993).

Learning Disability

Questions

Is one form of reading comprehension assessment best for ABE students with a learning disability? Based upon assessment results, what are the strengths and needs in reading comprehension for ABE learners with a learning disability?

Principle 13:

Adults with a learning disability tend, on average, to have lower functional literacy comprehension achievement and are over-represented within the ABE target population. (Kirsch, Jungeblut, Jenkins, & Kolstad, 1993)

A little over 80% of all adults reporting that they have a learning disability score at Level 1 or 2 on the NALS assessment of functional literacy comprehension (Kirsch et al., 1993).

Reading Comprehension Instruction:
Principles and Trends

There is more research on reading comprehension instruction than there is on any of the other major topics. The only subtopics for which there is no reading comprehension research are those related to learner characteristics, where only motivation studies were found.

Overall

Question

Does participation in adult basic education increase ABE students' reading comprehension achievement?

Principle 14:

Participation in an adult literacy program may lead to an increase in reading comprehension achievement. (Alessi, Siegel, Silver, & Barnes,1982; Boudett & Friedlander 1997; Conti 1985; Brooks, Davies, Ducke, Hutchison, Kendal, & Wilkin, 2001; Curtis & Chmelka, 1994; Darkenwald & Valentine, 1985; Fitzgerald & Young, 1997; Friedlander & Martinson, 1996; Gerber & Finn, 1998; Gorman & Moss, 1981; Hayes, 1989; Gretes & Green, 1994; Lazar, Bean, & Van Horn, 1998; McDonald, 1997; Mikulecky & Lloyd, 1997; Nickse, 1988; Perin & Greenberg, 1993; Philliber, Spillman, & King, 1996; Rich & Shepherd, 1993; Sheehan-Holt & Smith, 2000; Smith, 1996; Sticht, Armstrong, Hickey, & Caylor, 1987, and Sticht, 1989, 1997; Venezky, Bristow, & Sabatini, 1994)

Results from experimental research that evaluates ABE programs' overall effectiveness in increasing reading comprehension achievement are mixed, with positive results slightly outnumbering negative results. One study reporting negative results evaluated five ABE programs in a large metropolitan area and found overall increases in reading comprehension in only one of the programs (Boudett & Friedlander, 1997, a re-analysis of data from Friedlander & Martinson, 1996). Another study, using a time series design (adults were tested at four times during the course of instruction), found no evidence for increased reading comprehension ability on one measure, a reading test. Based on structured teacher observations, however, it found some, albeit slight, evidence for increased comprehension achievement (Perin & Greenberg, 1993).

Five experimental studies reported positive results. One evaluation of thirty-two programs in ten cities found overall gains in reading comprehension achievement (the measure used was "total reading" from a standardized test, which included tests of vocabulary and comprehension; Philliber, Spillman, & King, 1996). It should be noted that the design for this study included a post-hoc analysis with no control for initial group differences (and might have used grade equivalent gain scores as the unit of analysis). One of four classroom intervention studies found support for a particular teaching style (Conti, 1985), although the measure used tested math ability as well as reading comprehension. The three other classroom interventions found strong support for teaching specific comprehension strategies: summarizing and self-questioning (Rich & Shepherd, 1993), locating information and paraphrasing (Alessi, Siegel, Silver, & Barnes, 1982), and scanning, inferencing, organizing, summarizing, question answering, and vocabulary development (Gretes & Green, 1994).

These mixed, slightly positive results from the experimental research suggest that it is not an easy task to construct an adult literacy program that will lead to gains in reading comprehension for participating adults.

Results from non-experimental research tend to support those experimental results discussed above that find positive effects for reading comprehension instruction. Gains in reading comprehension following a wide variety of instructional interventions are reported in some studies, but control groups and/or significance testing were not used (Curtis & Chmelka, 1994; Brooks, Davies, Ducke, Hutchison, Kendal, & Wilkin, 2001; Darkenwald & Valentine, 1985; Gerber & Finn, 1998; Gorman & Moss, 1981; Lazar, Bean, & Van Horn, 1998; McDonald, 1997; Mikulecky & Lloyd, 1997; Nickse, 1988). Of these, the strongest positive result comes from a large-scale study of ABE programs in Britain (Brooks et al.). Using a test of functional literacy comprehension with 1224 adults in 71 basic skills programs, the study found significant increases in reading comprehension after an average of approximately thirty hours of instruction.

One influential study that has not appeared in a peer-reviewed journal, but that is cited regularly, suggests that adult literacy programs can increase reading comprehension achievement (Sticht et al., 1987, and Sticht, 1989, 1997). Finally, one correlational study finds an association between program participation and increases in reading comprehension achievement (Fitzgerald & Young, 1997) while another reports an association between practicing reading in natural settings and increases in reading achievement (Smith, 1996).

On the other hand, another correlational study finds no relationship between reported program participation and reading comprehension achievement (Sheehan-Holt & Smith, 2000) and the evaluation of one ABE literacy program yields results similar to those of Perin & Greenberg (1993), where increases in reading comprehension achievement were obtained on one measure of reading comprehension, but not on another (Venezky, Bristow, & Sabatini, 1994).

A relatively large number of the research studies cited above have used reading comprehension to measure the overall effects of participation in ABE programs. These include, in addition to the studies with experimental results mentioned above: Brooks, Davies, Ducke, Hutchison, Kendal, & Wilkin, 2001; Fitzgerald & Young, 1997; Lazar, Bean, & Van Horn, 1998; Mikulecky & Lloyd, 1997; Hayes, 1989; Nickse, 1988; Darkenwald & Valentine, 1985; and Venezky, Bristow, & Sabatini, 1994. Although some of this research has focused on specific techniques that can be used to improve reading comprehension, most has not. Instructional approaches used in these programs are often not described in detail. Some of the research aggregates data across multiple programs that may use multiple approaches to reading comprehension instruction. Consequently, some of the research studies suggesting that improvement in comprehension is possible are unable to provide specific suggestions for actually improving the reading comprehension of individual adults in ABE programs. One final issue related to these studies is that a few use a standardized test's "total reading" score to measure growth, which may include measures of vocabulary and word recognition as well as reading comprehension. Although they may provide some of the best evidence available regarding reading comprehension achievement, these studies, for all of the reasons mentioned above, need to be interpreted cautiously.

Trend 14:

Change in reading comprehension achievement resulting from instruction may be extremely variable and any change observed may depend on the test used to measure achievement. (Perin, & Greenberg, 1993; Venezky, Bristow, & Sabatini, 1994)

Only one of the experimental studies discussed above involved multiple measurements over time (Perin & Greenberg, 1993). It found that adults' gains in reading achievement over time are variable, and that they may depend upon the type of comprehension test administered. This is supported by results reported in one non-experimental study (Venezky, Bristow, & Sabatini, 1994).

Goals and Setting

Questions

Does participation in a program specifically aligned with one of the three major ABE goals or settings lead to a greater increase in reading comprehension achievement than participation in another type of program? Does setting affect the degree to which comprehension instruction increases achievement in other aspects of reading?

The three major ABE goals or settings are general functional literacy, family literacy, and workplace literacy. Several research studies have compared the reading comprehension achievement of adults in one type of setting to the reading comprehension achievement of adults in another. The trend resulting from these studies will be discussed first. Whether reading comprehension can be improved within any one of these types of programs, without regard to any other type of program, is a different question. This question will be addressed in the sections that follow (under the headings workplace literacy, family literacy, and general functional literacy).

Trend 15:

In some situations, participation in a workplace literacy or family literacy program may lead to greater increases in reading achievement than participation in other types of programs. (Philliber, Spillman, & King, 1996; Sheehan-Holt & Smith, 2000; Sticht, Armstrong, Hickey, & Caylor, 1987, and Sticht, 1989, 1997)

Experimental results from one study suggest that adult learners in family literacy programs at thirty-two locations in ten cities increased their reading comprehension achievement more than those in non-family literacy programs in another city's program that used the same measure of comprehension (the measure used was "total reading" from a standardized test, which included tests of vocabulary and comprehension; Philliber, Spillman, & King, 1996). It should be noted that the design for this study included a post-hoc analysis with no control for initial group differences (and might have used grade equivalent gain scores as the unit of analysis). One strand of influential research, not published in a peer-reviewed journal but included here because it is widely cited in the ABE literature, finds that some work-related literacy programs that use job-specific content during instruction lead to (a) the same increase in general literacy as those using non-content based instruction, and (b) a much higher increase in the comprehension of work-related material (Sticht et al., 1987, and Sticht, 1989, 1997).

In contrast to the above, one non-experimental study (Sheehan-Holt & Smith, 2000, an ex post facto regression analysis using the NALS data) compares adults who report participating in job-related programs with those reporting participation in community-based tutoring programs and finds no relationship between reading comprehension achievement and type of program.

Workplace Literacy

Question

Within a workplace literacy program, is it possible to increase reading comprehension?

Trend 16:

It may be possible to increase reading comprehension in a workplace literacy program. (Lazar, Bean, & Van Horn, 1998; McDonald, 1997; Mikulecky & Lloyd, 1997; Perin & Greenberg, 1993; Sticht, Armstrong, Hickey, & Caylor, 1987, and Sticht, 1989, 1997)

When adults completing a work-related program of instruction in an experimental study are compared with those who do not attend as long, their overall reading comprehension is higher on one of two measures of reading comprehension (Perin & Greenberg, 1993). Using a more traditional experimental design, another study identified several specific characteristics of workplace literacy programs that lead to increased reading comprehension achievement (Mikulecky & Lloyd, 1997).

Two non-experimental studies using reading comprehension pretests and posttests, along with a strong non-peer-reviewed strand of research, also describe methods that lead to increases in reading comprehension (Lazar, Bean, & Van Horn, 1998; McDonald, 1997; Sticht et al., 1987, and Sticht, 1989, 1997). All together, the five studies suggest that it is possible to increase ABE adults' reading comprehension in a wide variety of work-related literacy programs, including hospital settings, a vocational classroom, military job-training programs, manufacturing plants, an insurance company, and a women's prison.

Family Literacy

Question

Within a family literacy program, is it possible to increase reading comprehension?

Trend 17:

It may be possible to increase reading comprehension in a family literacy program. (Hayes, 1989; Nickse, 1988; Philliber, Spillman, & King, 1996)

Two non-experimental and one experimental study found that literacy training in family literacy settings can lead to increases in reading comprehension ability. Each of three studies reports gains of about one Grade Equivalent (GE) for anywhere from approximately 40 to 130 hours of instruction (Hayes, 1989; Nickse, 1988; Philliber, Spillman, & King, 1996, using a combined comprehension and vocabulary test score result, a post-hoc analysis with no control for initial group differences, and, perhaps, grade equivalent gain scores as the unit of analysis).

General Functional Literacy

Question

Within a functional literacy program, is it possible to increase reading comprehension?

Trend 18:

It may be possible to increase reading comprehension in a general functional literacy program. (Brooks, Davies, Ducke, Hutchinson, Kendall, & Wilkin, 2001; Darkenwald & Valentine, 1985; Fitzgerald & Young, 1997; Greenberg, 1998; Gretes & Green, 1994; Purcell-Gates, 1993; Scully, 1991; Venezky, Bristow, & Sabatini, 1994)

One experimental and virtually all of the non-experimental studies conducted within functional literacy programs report positive results on measures of reading comprehension.

An experimental study of a computer-based program for intermediate ABE readers (reading at GE 4 -- 9 based on TABE total reading scores) finds that reading comprehension can be increased with instruction in several specific comprehension strategies using functional reading material, such as texts that provide health, consumer, and voting information for adults (Gretes & Green, 1994).

Four non-experimental studies reporting results from single groups receiving pretests and posttests find that on four out of five measures (one study used two measures), adult ABE learners' reading comprehension increased after they participated in functional literacy programs.

An analysis of 71 basic skills programs in Britain suggests that ABE learners, especially those attending for more than 50 hours, benefit slightly from instruction in general functional literacy programs (Brooks et al., 2001). Another study involved students in ABE and ESL programs in 20 states (Fitzgerald & Young, 1997). It found an increase of roughly 1 GE on the TABE and CASAS for those involved in pre-secondary and secondary ABE classes.

In a survey conducted by telephone, a large percentage of a random sample of adults who had participated in one state's ABE programs reported that their reading was better after having participated (Darkenwald & Valentine, 1985). A fourth non-experimental study looked at one program in more detail and found that, overall, students' reading comprehension increased on a measure of functional reading comprehension (the TALS Document test) but not on a measure focusing more on basic reading skills (the TABE) (Venezky et al., 1994).

In addition to studies of large ABE programs, three case studies report that one-to-one tutoring by experienced teachers leads to increases in adults' ability to read and understand functional texts based on a variety of measures, including teacher observation (Greenberg, 1998; Purcell-Gates, 1993; Scully, 1991).

Instructional Methods and Material

Teaching Strategies

Question

What specific teaching strategies or techniques for reading instruction can be used to increase reading comprehension achievement?

Eight experimental studies dealing directly with reading comprehension teaching strategies address this question. Because there is very little overlap between the issues considered in these eight studies, five separate trends and two principles are drawn from the studies. The two principles are supported by five experimental results and each trend is supported by at least one experimental result (one study has two experimental results). The results are divided roughly into three categories: (1) direct reading comprehension instruction, (2) multiple-components instruction, and (3) enabling instruction or setting. Most research related to ABE reading comprehension teaching strategies fits into the enabling category. Rather than addressing reading comprehension directly, other instructional variables are manipulated in order to address it indirectly.

Direct Reading Comprehension Instruction

Principle 15:

Providing explicit instruction in reading comprehension strategies may lead to increased reading comprehension achievement. (Alessi, Siegel, Silver, & Barnes,1982; Mikulecky & Lloyd, 1997; Rich & Shepherd, 1993)

Results from a study of instructional strategies used in workplace literacy settings with adult learners who are not beginning readers (roughly GE 6 or higher) suggest that providing explicit, as opposed to incidental, instruction in reading comprehension strategies leads to increased reading comprehension achievement (Mikulecky, 1997). Comprehension strategies taught include skimming, reading more carefully in order to monitor comprehension, using headings, and focusing on topics. Although this approach was not used in isolation, it was evaluated in relative isolation statistically to determine its effect on reading comprehension, controlling for the influence of other factors.

In another experimental study, instruction in the use of two specific reading comprehension strategies was effective in improving intermediate adult readers' comprehension (adults with GE scores from high 4 to 7 on a standardized test of reading comprehension) (Rich & Shepherd, 1993). Students in one group were taught how to ask themselves questions about a text as they read (who, what, when, where, how, and why questions). Students in another group were taught how to verbally summarize a text as they read. A third group was taught how to use both strategies as they read. All students worked in small groups, receiving guided instruction from a teacher who gradually had students take on the teacher's role -- leading group questioning or summarizing practice. Students were told how to use the strategies, the rationale for using them, and how to check on or monitor their understanding as they used them. In addition to small group work, students spent a significant portion of each 45-minute lesson, 15 minutes, working on their own applying a strategy as they took practice comprehension tests. Those in the group that learned how to use both strategies outperformed control groups that read the same texts or took the practice tests but did not receive strategy instruction.

The combined approach was found to be effective with two measures of comprehension: free recall (the number of ideas from the text that were mentioned as the passage was recalled) and questions (multiple choice and short answer). The summarizing and questioning strategies were effective by themselves on the questioning test. Contrary to what the researchers had predicted, the combined summarizing and questioning strategy was not more effective than the use of either strategy by itself. Self-questioning seemed to be an especially effective strategy. The total number of sessions, six in all, was small. More time to learn the combined strategy may have been needed.

In a third experimental study, an early but fairly sophisticated form of computer-based instruction was found to improve adults' performance on two comprehension tasks: locating and paraphrasing information in texts (Alessi, Siegel, Silver, & Barnes, 1982). Using this program, intermediate adult readers (at GE 4-6 based on a standardized test of reading comprehension) completed forty self-paced reading comprehension lessons in twenty hours over a two-month period. Follow-up testing one month later showed that initial gains had been maintained. During instruction, comprehension tasks and concepts were presented clearly and overtly and students read and re-read passages and answered questions. They learned how to locate information in order to answer literal comprehension questions (who, what, when, where, why, and how questions that could be answered by looking back at the text) and how to recognize paraphrases of sentences in the passages that were read. The computer kept track of progress, and simpler tasks were mastered before more difficult ones were presented. Passages and questions were sometimes presented more than once, with the computer making small changes in the text to draw student attention to important information and text features. Specific feedback was given and computer graphics such as boxes, arrows, and underlining were used to direct student attention and model correct responses. Although students improved their ability to locate and paraphrase information, their new skills did not transfer to a comprehension task that was not taught, recognizing main ideas in passages.

Multiple-components Instruction

Principle 16:

Combining comprehension instruction with instruction in various other components of reading may lead to increased reading comprehension achievement. (Curtis & Longo, 1997; Gretes & Green, 1994; Gold & Horn, 1982 and Gold & Johnson, 1982)

One multiple-components strategy (reported in Gold and Horn, 1982, and Gold and Johnson, 1982) includes many elements, and any one or combination of elements could be responsible for the increase in reading comprehension found for beginning readers (those scoring below GE 5 on measures of reading comprehension). These elements include: (a) one-to-one instruction by (b) trained tutors that includes (c) Directed Listening (the instructor reads a text aloud and students answer questions and discuss the passage), followed by (d) the generation of student texts based on the previous discussions, using a Language Experience Approach (students dictate to the instructor, who produces the texts), followed by (e) basic reading skills instruction using the student-generated texts, including whole-word phonics, a multi-sensory approach to word recognition (the VAKT method), and reading comprehension instruction. The authors stress the use of listening comprehension instruction followed by basic skills instruction using texts in the content area of interest to adults.

A newer, computer-based program for teaching reading comprehension uses a multiple-components approach that includes comprehension strategy instruction along with vocabulary instruction and word recognition or fluency practice (Gretes & Green, 1994). Intermediate ABE readers using the program, those who scored between GE 4 and 9 on a standardized test of reading comprehension, increased their reading achievement (based on TABE total reading scores, a combined measure of vocabulary and comprehension). This self-paced program presents instruction and practice in several specific comprehension strategies: scanning a text for information, making inferences from information in a text, organizing information, summarizing information, and answering questions about a text. Before beginning the comprehension exercises, however, learners can choose to listen to the text while reading along and also study key vocabulary in the passage. A natural sounding (digitized) voice reads text presented on the computer screen, including the passages that will be used for comprehension instruction, individual words, vocabulary definitions, and instructions for exercises. In this experimental study, the effects of comprehension strategy instruction were not separated from the effects of the vocabulary instruction and what may amount to word recognition and fluency practice using the spoken text.

Results from a non-experimental study suggest that another multiple-components strategy may lead to increased reading comprehension achievement in a special program for out-of-school adolescents. In this program, developmental strengths and weaknesses in each component of reading are diagnosed in order to focus instruction appropriately (Curtis & Longo, 1997).

Enabling Instruction or Setting

Trend 19:

The degree to which a literacy program is learner-centered may differentially affect students at different ability levels. (Conti, 1985; Fitzgerald & Young, 1997)

In one experimental study, twenty-nine ABE, ESL, and GED teachers completing a survey were found to use a more teacher-centered approach as opposed to a more "traditional" ABE learner-centered approach (Conti, 1985). Results from the study suggest that the degree to which these teachers' classes are learner-centered (or collaborative) affects students differently, depending on their level of reading and math ability (the measure used tested math ability as well as reading ability). Weaker teacher-centered approaches, those that incorporate more learner-centered activities, appear to lead to increased comprehension and math ability among adults who score below about GE 9, and among ESL adults. Stronger teacher-centered approaches appear to be more effective with adult students in GED classrooms (those working on high school level reading tasks).

One non-experimental study may support this result. In an exploratory correlational study based on students in ABE programs in 20 states, a relationship was found between the use of a highly individualized (as opposed to pre-structured and fixed) curriculum and increases in adults' reading comprehension (Fitzgerald & Young, 1997). This may have been mediated by the use of experienced, full-time instructors who typically use a more individualized approach. Individualized approaches might be equated, to some degree at least, with learner-centered approaches.

Trend 20:

With adult readers at the intermediate level (reading around GE 6), a meaning-based diagnostic-prescriptive approach to teaching may lead to increased reading comprehension achievement. (Cheek & Lindsey, 1994)

In an experimental study of two contrasting teaching styles, one style was found to be more effective for teaching inferential reading comprehension skills, but not for teaching literal reading comprehension skills. Three measures of reading comprehension were used: the Literal, Inferential, and Total Comprehension subtests of the Stanford Diagnostic Reading Test. Literal comprehension tests measure a learner's ability to recall specific ideas or pieces of information from a text that has been read. Inferential tests measure a learner's ability to draw valid inferences from the ideas or information presented in the text.

The style found to be more effective for teaching inferential reading comprehension used a meaning-based, diagnostic-prescriptive approach. The less effective style used a programmed learning approach. The diagnostic-prescriptive approach had several important characteristics: formal and informal assessment to identify learner strengths, needs, and interests in reading; use of these assessment results to develop individualized teaching strategies, methods, and materials for word analysis and reading comprehension instruction; and, language-experience and literature-based instruction emphasizing regular student-teacher interaction, real-life reading material, and reading as a meaning-making activity. The programmed learning approach, on the other hand, emphasized placement of students at their current reading levels in computer-based or print-based programs where they could work independently, at their own pace, and in a step-by-step manner toward a specific word analysis or reading comprehension learning objective. There was no significant difference between groups on the literal comprehension measure.

The seventy-one students who participated in this study were each randomly assigned to one of the two approaches. Random assignment, the fact that the groups' initial reading levels were not significantly different, and the fact that both classes were team-taught by the same pair of teachers, rule out many extraneous factors that could have accounted for the results. Therefore, fairly specific questions for future research studies can be asked: Although the researchers attempted to vary only their teaching style or approach, how different were the specific reading comprehension teaching strategies in each group (especially given that one emphasized "meaning-centered instruction" while the other focused more on "word identification")? What were the separate effects of the diagnostic-prescriptive and the meaning-based aspects of the more successful approach to teaching inferential reading comprehension?

Trend 21:

Requiring adults to attend a literacy program in order to receive welfare benefits may not increase reading comprehension achievement. (Boudett & Friedlander, 1997 and Friedlander & Martinson, 1996)

This result comes from a study of five large programs in a metropolitan area (Boudett & Friedlander, 1997, and Friedlander & Martinson, 1996). In this study, the reading comprehension gains of adults required to attend ABE programs in order to receive welfare benefits were compared to the gains of those who received benefits without this requirement (attendance in a program was optional). Those required to take ABE (including GED or ESL) classes did attend these classes more often. Roughly one-third did, a much higher percentage than those in the other group. However, those required to attend did not have a greater increase in reading comprehension achievement than those who were not required to attend, even when considering only those who actually participated in an adult education program.

Trend 22:

In programs where a teacher has assistance in the classroom, students may make greater gains in reading comprehension achievement. (Brooks, Davies, Ducke, Hutchison, Kendal, & Wilkin, 2001)

An experimental study of 71 programs in Britain found that, when the main teacher in a classroom has assistance from either volunteers or paid assistants, reading comprehension achievement is significantly greater than in classrooms where no assistance is available (Brooks, Davies, Ducke, Hutchison, Kendal, & Wilkin, 2001).

Trend 23:

Dealing briefly but explicitly with issues related to reading self-efficacy and motivation among adult learners in a literacy class may lead to increased reading comprehension achievement. (Mikulecky & Lloyd, 1997)

Results from a study of instructional strategies used in workplace literacy settings with adult learners who are not beginning readers (roughly GE 6 or higher) suggests that dealing briefly but explicitly, as opposed to incidentally, with issues related to reading self-efficacy and motivation leads to increased reading comprehension achievement.

Instructional Materials

Question

Does the use of specific instructional material lead to increased reading comprehension achievement?

Trend 24:

Integrating adult-oriented, contextually relevant material into literacy programs may lead to increased reading achievement. (Aderman, 1987; Curtis & Chmelka, 1994; McDonald, 1997; Mikulecky & Lloyd, 1997; Sticht, Armstrong, Hickey, & Caylor, 1987, and Sticht, 1989, 1997)

In workplace literacy programs with high rates of retention and students reading at about GE 6 or higher, the use of workplace-oriented material at least 20 - 30% of the time during classroom instruction appears to increase reading comprehension of work-related material (Mikulecky & Lloyd, 1997). This is supported by the non-experimental results from another study (McDonald, 1997) in which the integration of job-oriented tasks and material with language instruction in an ESL vocational class leads to a greater increase in general reading comprehension (measured by the ABLE) than non-integrated approaches. In addition, one frequently cited study (not published in a peer-reviewed journal) also finds that work-related literacy programs using job-specific content during instruction appear to increase job-specific reading comprehension achievement. It may also increase non-job-specific reading comprehension achievement (as measured by tests that do not use job-specific comprehension passages, for example) (Sticht et al., 1987, and Sticht, 1989, 1997).

One more non-experimental study supports the use of material that is relevant to adults. In this study, modifying a popular adult literacy phonics and sight word instructional program (Laubach) so that it includes practice with more challenging, adult-oriented words incorporating the letter-sound correspondences being taught appears to lead to a faster rate of growth in reading comprehension (Curtis & Chmelka, 1994). Although the use of adult-oriented content may be more motivating and relevant for adults, its effectiveness in this program is attributed to its complexity. In programs teaching word recognition, adults may rely on their sight word knowledge to read simpler words instead of applying word analysis strategies, such as sounding out. Using more complex words during instruction, words that are not a part of an adult's sight word knowledge, means that these words will need to be sounded out, which requires the application, and therefore practice, of letter-sound knowledge.

Finally, results from a descriptive study suggest that re-writing information in a nutrition pamphlet using ABE students' language results in increased recall for beginning level readers compared with a standard version of the pamphlet (Aderman, 1987). This was true for the four students in this study even though the readability levels of the passages were the same.

Intensity and Duration of Instruction

Question

Does more intense literacy instruction, or instruction that is of longer duration, increase reading comprehension achievement?

Ideally, to judge to what degree the intensity or duration of instruction leads to gains in reading comprehension, learners would be assigned to groups that varied in either the intensity of instruction or the length of time they received instruction. Practically, this is a difficult criterion to meet, and the studies reported here have used intact or existing groups, comparing, for example, those who have participated for a certain length of time in an ABE program with those who have not. Selection bias, therefore, cannot be ruled out. Those who stay for a short period of time in a program may share an important characteristic, such as initial level of reading ability, that those who stay for a longer period of time do not share.

Trend 25:

Spending at least 70% of classroom time practicing reading and writing, including the occasional but direct or deliberate discussion of reading strategies, may increase learners' meta-comprehension abilities. (Boudett & Friedlander, 1997 and Friedlander, & Martinson, 1996; Mikulecky & Lloyd, 1997)

Only one experimental study has looked at the effects of intensity of instruction on reading comprehension (Mikulecky & Lloyd, 1997). Results from the study suggest that, up to a point, more intense instruction leads to increases in reading metacomprehension ability. Spending 70% or more of classroom time on literacy practice (reading and writing) leads to an increase in reading metacomprehension abilities, or the ability to deliberately use strategies such as skimming, reading more carefully or monitoring, using headings, and focusing on topics. Classes in the programs studied typically lasted for about one and one-half hours and total class time per week ranged from two to five hours.

Results from a non-experimental study suggest that increasing the intensity of instruction by simply increasing the total number of classroom hours per week does not have a positive effect on reading comprehension. Class time of nine or more hours per week was associated with declines in reading achievement (reported in Boudett & Friedlander, 1997 and Friedlander & Martinson, 1996).

Trend 26:

Reading comprehension achievement may increase as a learner stays longer in a literacy program, although progress may be extremely variable over time. (Brooks, Davies, Ducke, Hutchison, Kendal, & Wilkin, 2001; Boudett & Friedlander 1997 and Friedlander & Martinson, 1996; Fitzgerald & Young, 1997; Gretes & Green, 1994; Perin & Greenberg, 1993; Philliber, Spillman, & King, 1996)

Studies comparing groups of learners receiving a different number of hours of instruction in reading have mixed results regarding reading comprehension achievement. Overall, the results slightly favor programs that last longer over those that are shorter (one experimental and two non-experimental results were positive while one experimental result and one non-experimental were not).

The results from one experimental study (Philliber, Spillman, & King, 1996) suggest that staying in a literacy program longer increases reading achievement. The Philliber et al. finding, based on a combined comprehension and vocabulary GE gain score, found that 51 or more hours of attendance were needed to produce significantly greater gains. The larger, non-experimental study by Brooks et al., based on a national evaluation of literacy programs in Britain, also reports that those who attended for more than 50 hours benefited most (these were also the students who attended regularly). In the other experimental study, adults who completed more than a single lesson in a computer-based reading comprehension program had greater gains on a measure of comprehension achievement, although there was no difference between those completing 2 to 3 lessons and those completing 4 to 6 (Gretes & Green, 1994).

Results from another experimental study (Perin & Greenberg, 1993) are mixed. Length of stay affected achievement on one measure of reading comprehension (teacher ratings) but not on another (reading tests). Results from non-experimental studies are also mixed. Two correlational studies, both involving large numbers of ABE learners, produced seemingly contradictory results. One found a positive relationship between the number of hours in an ABE program and reading comprehension achievement (Fitzgerald & Young, 1997). The other study found that number of hours was positively associated with gains only for ESL students, and that this effect was very small (Boudett & Friedlander, 1997, and Friedlander & Martinson, 1996).

Teacher Preparation

Question

Does teacher preparation lead to increases in learner reading comprehension achievement?

Trend 27:

Staff with more experience or training may have a better chance at improving reading comprehension achievement. (Brooks, Davies, Ducke, Hutchison, Kendal, & Wilkin, 2001; Fitzgerald & Young 1997)

Although several studies mention teacher preparation, only two studies were found that addressed it directly, one non-experimental and one experimental. An exploratory path analysis of ABE program data from twenty states found that the degree of staff teaching experience is positively associated with reading comprehension achievement (Fitzgerald & Young 1997). A national study of basic skills programs in Britain (Brooks, Davies, Ducke, Hutchison, Kendal, & Wilkin, 2001) found that for those programs in which all teachers are qualified (with certification or a Bachelor's degree in education), students make significantly greater gains in reading comprehension.

Learner Characteristics

Of the four learner characteristic factors, a trend was found only for motivation.

Reading Level

Question

Are certain forms of reading comprehension instruction more effective for students reading at a particular reading level (for beginning versus more advanced readers, for example)?

No trends related to reading level were identified. Although research studies describe approaches to reading comprehension instruction that are used with ABE students at different levels of ability, very few compare the use of one approach across ability levels.

Overall, even the weak evidence that exists is mixed. Two correlational studies arrive at contradictory conclusions. One finds that a learner's initial level of reading ability has the most influence on reading achievement (Fitzgerald & Young, 1997), while the other finds no strong relationship between ABE program participants' reading achievement and their initial reading level (Boudett & Friedlander, 1997, and Friedlander & Martinson, 1996). Another study finds that teaching style is an important factor in GED versus pre-GED learners' achievement, but uses a measure that includes math as well as reading (Conti 1985). One experimental study finds no difference in achievement based on reading level for those completing a computer-based reading comprehension program (Gretes & Green, 1994).

ESOL

Question

Are certain forms of reading comprehension instruction more effective for ABE students with different levels of language ability, particularly ESL ABE students?

Only one comprehension study was found that dealt with ABE students in ESOL classes, and so no trends were identified. McDonald's non-experimental results suggest that the use of job-oriented tasks and content with ESL students in a vocational class is effective in increasing their general functional reading ability (measured with the ABLE) (McDonald, 1997). It is not clear how students in the comparison groups varied in language ability.

Motivation

Question

Does instruction that improves ABE students' motivation or self-esteem also lead to increased reading comprehension ability?

Trend 28:

The direct or deliberate discussion of learners' literacy beliefs and plans in order to deal with issues of reading self-efficacy and motivation may increase reading comprehension achievement. (Boudett & Friedlander, 1997 and Friedlander & Martinson, 1996; Mikulecky & Lloyd, 1997)

Most ABE programs assume that adult learners' motivation and feelings of self-efficacy are important issues. However, very few studies investigate whether or not attempting to improve motivation and feelings of self-efficacy will have a positive affect on adult learners' reading ability. Mikulecky and Lloyd (1997) describe workplace literacy programs in which learners' feelings of literacy self-efficacy, or how good they perceive their reading and writing to be, and how important reading and education are in their future plans, is evaluated and then dealt with briefly, but directly, during classroom literacy instruction. Learners in their study who participated in these classroom discussions had much greater gains in their ability to comprehend workplace related texts than those who did not.

Another study (reported in Boudett & Friedlander, 1997, and Friedlander & Martinson, 1996) attempted to increase ABE students' motivation by requiring attendance in an ABE program to receive welfare benefits. However, even though those required to attend did attend more often, they did not have a greater increase in reading comprehension achievement than those who did not.

Ideas for Reading Comprehension Instruction from K-12 Research

ABE reading comprehension assessment research indicates that adults who qualify for ABE services have poor functional literacy comprehension and that their knowledge about reading comprehension (their meta-comprehension) may be similar to that of children who are beginning readers. This suggests that ABE reading comprehension instruction should provide ways to help adults understand what they already do as they read for meaning, and what alternative strategies are available.

The strongest result from research related to reading comprehension instruction is that participation in ABE programs may lead to increases in ABE learners' reading comprehension achievement. Results related to other topics and subtopics are not as strong because fewer studies have addressed them. However, several trends do emerge from the research.

Principles and trends in the ABE reading comprehension research suggest that it is important for adults to spend time actually practicing reading and discussing reading strategies. They also stress that the learning situation or context is important for ABE reading comprehension instruction, and that it is important for adults to deal briefly but directly with the issues of motivation and self-efficacy as they work on their reading.

While the NRP results do not directly support these trends, they do not contradict them either and might be seen as complementary. Although the NRP analysis of text comprehension research does not address topics and subtopics related to goals and setting, instruction material, intensity and duration of instruction, ESOL classrooms, learning disability, and motivation, it does present a considerable body of research related to specific approaches to reading comprehension instruction that may provide some guidance for ABE educators. Unlike the ABE research, with its mixed results related to the effects of specific strategies for reading comprehension instruction in ABE programs, research at the K-12 level finds very strong evidence to support teaching several specific reading comprehension strategies.

Instructional Methods and Material

In addition to describing specific strategies that may be taught, the NRP review of research at the K-12 level also describes research related to teacher preparation and how it affects growth in student reading comprehension. The NRP review does not present research directly related to the ABE subtopics having to do with materials and the intensity or duration of instruction. Results from the NRP review of other aspects of instruction (alphabetics, fluency, and vocabulary) also suggest ways in which reading comprehension may be improved. In what follows, the first idea presented deals with these results.

Teaching Strategies

Idea 18:

To improve ABE learners' reading comprehension, use a multiple-components approach to instruction in which all aspects of the reading process are addressed, as needed, including phonemic awareness, word analysis, and vocabulary as well as reading comprehension.

K-12 Research. The NRP review found that reading comprehension achievement can be improved indirectly by teaching skills that enable comprehension. Teaching phonemic awareness to beginning readers leads to improved reading comprehension (pp. 2-4, 2-5). Small-group instruction is especially effective (pp. 2-4, 2-5, 2-20). PA training is also effective in improving comprehension for disabled readers at higher reading levels (through at least GE 6) (p. 2-94). Systematic, as opposed to non-systematic, phonics instruction improves reading comprehension for beginning readers (p.2-94), and for older readers with a reading disability (p. 2-116). Teaching fluency using repeated and guided oral reading leads to increases in reading comprehension (p. 3-3). Finally, pre-teaching important vocabulary words before reading can also improve reading comprehension (p. 4-4).

Idea 19:

To improve ABE learners' comprehension of texts used during instruction (those ABE learners reading above Grade Equivalent 3), teach them a strategy that can be used during the reading process and that enables them to become actively engaged in understanding a text. Eight effective strategies have been identified: comprehension monitoring, cooperative learning, graphic organizers, story structure, question answering, question generation, summarization, and multiple strategies (using a combination of strategies when appropriate).

K-12 Research. The NRP identifies 16 categories of instruction and finds that eight appear to have "a firm scientific basis for concluding that they improve comprehension of normal readers" (p. 4-42). These eight include seven specific strategies and a multiple-strategy approach. All of these strategies appear to improve students' comprehension of texts they read in the classroom while practicing comprehension, as measured on tests of recall, question answering, question generation, and summarizing (p. 4-6). It should be noted that the NRP research reviewed involved students in grades 3 and up, not those in the lower grades.

The eight kinds of instruction that appear to be effective and most promising for classroom instruction are (p. 4-6):

1. Comprehension monitoring in which the reader learns how to be aware or conscious of his or her under-
2. Cooperative learning in which readers work together to learn strategies in the context of reading.
3. Graphic and semantic organizers that allow the reader to represent graphically (write or draw) the meanings and relationships of the ideas that underlie the words in the text.
4. Story structure from which the reader learns to ask and answer who, what, where, when, and why questions about the plot and, in some cases, maps out the time line, characters, and events in stories.
5. Question answering in which the reader answers questions posed by the teacher and is given feedback on the correctness of their answers.
6. Question generation in which the reader asks himself or herself what, when, where, why, what will happen, how, and who questions.
7. Summarization in which the reader attempts to identify and write the main or most important ideas that integrate or unite the other ideas or meanings of the text into a coherent whole.
8. Multiple-strategy teaching in which the reader uses several of the procedures in interaction with the teacher over the text. Multiple-strategy teaching is effective when the procedures are used flexibly and appropriately by the reader or the teacher in naturalistic contexts.

Idea 20:

To improve ABE learners' general reading comprehension achievement (those ABE learners reading above Grade Equivalent 3), teach them to use a repertoire of several strategies that they can use consciously and flexibly as needed while reading and that enable them to become actively engaged in understanding a text. Combinations of the following strategies are suggested by the research: comprehension monitoring, cooperative learning, graphic organizers, story structure, question answering, question generation, and summarization.

K-12 Research. Based on studies of students in grades 3 and up, several specific strategies (comprehension monitoring, cooperative learning, graphic organizers, story structure, question answering, question generation, and summarization) appear to improve students' comprehension of texts they read in the classroom while practicing comprehension, as measured by non-standardized, researcher-made tests of recall, question answering, question generation, and summarizing (p. 4-6). Teaching the use of more than one strategy for reading comprehension can lead to increases on standardized measures of reading comprehension, or general reading comprehension achievement that is not tied to the specific texts used during instruction (pp. 4-6, 4-47). This suggests that teaching a multiple-strategy approach to comprehension generalizes to reading outside the classroom for children. This is an especially important goal for adult literacy programs.

Teacher Preparation

One note of caution raised in the NRP report (p. 4-49) that may be relevant for ABE settings, where teachers may be volunteers, paraprofessionals, or otherwise lack training, is:

In spite of heavy emphasis on modeling and metacognitive instruction, even very good teachers may have trouble implementing, and may even omit, crucial aspects of strategic reasoning. The research suggests that, when partially implemented, students of strategy teachers will still improve. But it is not easy for teachers or readers to develop readers' conceptions about what it means to be strategic. It takes time and ongoing monitoring of success to evolve readers into becoming good strategy users.

Comment 7:

To improve ABE learners' general reading comprehension achievement, train their teachers to teach the awareness and use of multiple strategies for reading and understanding a text.

K-12 Research. Unlike the NRP review of other aspects of reading instruction (alpabetics, fluency, and vocabulary), the review of reading comprehension instruction did locate and present several studies having to do with the preparation of teachers for reading comprehension instruction. Although these four studies represent a relatively small body of research when compared with other areas reviewed by the NRP, several trends were observed. The most important trend is that teachers can be taught to teach reading comprehension to students and, when they are, their students become aware of comprehension strategies, use the strategies, and improve their reading (p. 4-8).

Comment 8:

To improve ABE teachers' knowledge of reading comprehension instruction, use both preservice and inservice training, and to improve their students' reading comprehension achievement directly, use inservice training.

K-12 Research. Correlational and experimental studies were also reviewed to examine the effects of pre-service and in-service teacher education. Preservice education typically occurs before teachers are certified, while inservice education typically involves professional development opportunities that occur after a teacher has begun teaching (p. 5-4). Trends from this research suggest that: (1) teachers learn the reading instruction strategies and techniques that they are taught during preservice education (p. 5-1), and (2) inservice education appears to lead to improved teacher knowledge and improved reading achievement for the teachers' students.

Learner Characteristics

Functional Reading Level

The NRP review presents just a few results related to learner characteristics, all having to do with students' reading levels. The studies reviewed included children in, roughly, grades three through six (pp. 4-42 -- 4-46).

The question generation and multiple strategies instruction studies included those at higher grade levels (up to GE 9 and 11, respectively).

Idea 21:

For intermediate adult readers (Grade Equivalents 3-6), improve comprehension of narrative texts by teaching story structure, or the typical content and organization of stories.

K-6 Research. The NRP review found that poor readers (in grades 3 -- 6) benefited more from instruction in story structure than good readers (p. 4-46).

Idea 22:

To improve the general reading comprehension achievement of adult intermediate and advanced readers, teach the flexible use of multiple reading comprehension strategies.

K-6 Research. The NRP review found that good readers and students in the 7th grade or higher benefited most from multiple strategy instruction (p. 4-46).