The Reading Strategies of Adult Basic Education Students

Pat Campbell and Grace Malicky

University of Alberta

Abstract: This large-scale SSHRC funded study examined the word identification and comprehension strategies of 344 adult basic education students enrolled in 58 full-time and part-time programs situated in urban and rural communities across Canada.

Background to the Study

The reading behaviours, skills, and processes of adult literacy learners is a relatively new field of academic study, despite the ever-growing awareness of and support for adult literacy. A recent study examining the characteristics of research on reading between 1969 to 1998 found that only 3 percent of the research involved adults as research participants (Guzetti, Anders & Neuman, 1999). Over the last 16 years, a small body of research has been conducted on the reading strategies and processes of the adult population. However, some of this research is problematic because it focuses on phonological and orthographic skills, and has employed assessment procedures that do not reflect authentic reading (Hiebert, Valencia & Afflerbach, 1994). The major purpose of our study was to examine the word identification and comprehension strategies used by 344 adult basic education students.

Methodology

The Participants

The participants in this study consisted of 344 adult basic students enrolled in 34 community colleges and school board programs and 24 community-based programs. These programs were situated in large communities (population greater than 65,000) and small communities (population less than 65,000) in Alberta, Ontario, the Northwest Territories, and Nova Scotia.

The participants were Canadian-born individuals whose mother tongue was English. There were 219 females and 125 males, and the average age of the participants was 33 years. The majority of the adults (71 percent) were full-time students attending adult basic education programs offered by colleges and school boards, while 7 percent were part-time students. The remaining adults (22 percent) were attending volunteer tutoring or community-based programs on a part-time basis. Adults ranged in reading levels from those at a beginning level to those who were able to read high school material with adequate comprehension.

Data Collection

We randomly assigned a set of two narrative and/or expository passages to each of the 344 students. We used passages from the Canadian Adult Reading Assessment (Campbell & Brokop, 2000), an informal reading inventory containing passages grouped into 9 levels, with increased difficulty based on factors such as length, number of questions, and readability (See Table 1). Modified versions of the Dale-Chall (Chall & Dale, 1995) and Fry (1977) formulas were applied to establish the readability levels of each passage.

Table 1: Nature of Passages at each Level

Level 1 Level 2 Level 3 Level4 Level 5 Level 6 Level 7 Level 8 Level 9
Word Length

34-86

97-143 146-170 208-251 260-289 297-347 389-451 461-498 450-535
# of Questions

5-6

7 8 8-10 9-10 10 10 10 10
Readability

Grade 1

Grades 1-2 Grades 2-3 Grades 3-4 Grades 4-5 Grades 5-6 Grades 7-8 Grades 9-10 Grades 11-12

A team of 19 adult basic education instructors administered a set of 2 passages to each of the 344 students. Each student was asked to read the passage silently, and then orally. The instructor recorded the student's oral miscues as he/she read aloud. After the passage was read, the instructor recorded the student's retelling and asked him/her a set of comprehension questions. In order to be included in the study, the student needed to score between 65 and 90 percent on the comprehension questions.

In order to examine word identification strategies, we used miscue analysis. The students' oral miscues were analyzed using categories adapted from Goodman, Watson, and Burke (1996) and from the Canadian Adult Reading Assessment (2000) to determine whether readers used print and/or meaning to decode unfamiliar words. Each miscue was classified as print-based, meaning-based, integrative, or non-integrative. After each miscue had been classified, we calculated mean scores and standard deviations for each of the four types of miscues for students at each of the 9 reading levels.

In order to examine comprehension strategies, we collected 2 types of information: unaided retellings and answers to comprehension questions. After scoring each comprehension question, we compared the student's performance on inference vs. factual questions. Then, we calculated mean scores and standard deviations for each type of question for students at each of the nine reading levels.

Results and Discussion

Oral Reading Miscues

There were few differences when we compared the adults' uncorrected miscues across the 9 reading levels (See Table 2). The results of MANOVA confirmed that these differences were either not significant or significant only at the .05 level (F=1.58, p=.021) for one of the randomly generated samples. From the descriptive statistics, it is clear that for adult literacy students at each of the 9 reading levels, the majority of uncorrected miscues were meaning-based or integrative while a small percentage of uncorrected miscues were print-based and non-integrative. Overall, this suggests that the beginning readers (Levels 1 to 3) were using reading strategies similar to those of the intermediate (Levels 4 to 6) and advanced readers (Levels 7 to 9).

Table 2: Means and Standard Deviations for Uncorrected Miscues

TYPE OF MISCUE

LEVEL MEAN S.D. N
Print-based

1

2

3

4

5

6

7

8

9

Total

.15

.15

.10

.21

.13

.15

.13

.23

.14

.15

.17

.15

.15

.21

.12

.15

.12

.17

.16

.16

40

40

40

38

34

37

39

36

40

344

Meaning-based

1

2

3

4

5

6

7

8

9

Total

.39

.39

.45

.39

.38

.43

.43

.36

.36

.40

.34

.27

.33

.25

.24

.22

.20

.19

.21

.26

40

40

40

38

34

37

39

36

40

344

Integrative

1

2

3

4

5

6

7

8

9

Total

.33

.38

.39

.35

.42

.35

.37

.33

.45

.38

.26

.24

.29

.23

.20

.17

.16

.19

.22

.22

40

40

40

38

34

37

39

36

40

344

Non-integrative

1

2

3

4

5

6

7

8

9

Total

.11

.08

.04

.05

.08

.07

.06

.08

.05

.07

.18

.13

.08

.08

.08

.06

.06

.08

.07

.10

40

40

40

38

34

37

39

36

40

344

When results were combined across the 9 levels, the majority of miscues were meaning-based (40 percent) or integrative (38 percent) while a smaller percentage of miscues were print-based (15 percent), and even fewer were non-integrative (7 percent). The discrepancy between proportions of meaning-based and print-based miscues suggests a greater reliance on knowledge than on print by adult literacy students at all levels of reading development. A relatively high percentage of miscues fell into the integrative category, which indicates that even beginning readers were able to use their knowledge and print cues together at least some of the time. In summary, the nature of the adults' uncorrected miscues across the 9 levels was very similar when mean scores were examined. However, the high standard deviation scores presented in Table 2 suggest considerable individual difference among adults within each reading level. It appears that differences in reading strategies used by adult within levels are greater than differences across levels.

Question Data

The results of the MANOVA indicated that there were significant differences across reading levels in the proportion of correct responses to factual and inference questions (F=3.8 to 4.79, p=.000). Post hoc tests revealed that the most consistent difference across the 3 randomly generated samples was on inference questions where adults at Level 1 were less successful in providing correct answers than were adults at other reading levels. Consistency in performance across other reading levels on both factual and inference questions is evident in the descriptive statistics presented in Table 3.

Table 3: Means and Standard Deviations for Inference and Factual Questions

Level

Mean S.D. N
Factual

1

2

3

4

5

6

7

8

9

Total

.86

.81

.78

.84

.77

.78

.77

.79

.81

.80

.09

.12

.11

.13

.12

.08

.12

.09

.11

.12

40

40

40

38

34

37

39

36

40

344

Inference

1

2

3

4

5

6

7

8

9

Total

.63

.76

.78

.74

.79

.75

.74

.78

.82

.75

.18

.12

.11

.13

.09

.10

.16

.11

.10

.13

40

40

40

38

34

37

39

36

40

344

The results on answers to questions confirm that adults at all levels of reading development were able to use both text information and background knowledge to construct meaning from what they read. However, as indicated on the post hoc tests, adults at Level 1 made less effective use of knowledge-based than text-based information. They were able to answer correctly 86 percent of factual questions correctly, as compared to 63 percent of inferential questions. This may be more a reflection of how they were taught in literacy classes than of their ability to make inferences, since they did not differ in their ability to make inferences from adults at higher reading levels on the retelling task. Perhaps Level 1 students had developed expectations about how to answer questions from their experiences in beginning literacy classes, but they hadn't developed expectations about how to retell a passage. Traditionally, educators have assumed that making inferences is a higher-level task than is comprehending details (e.g., Bloom, 1956) and that beginning readers will have difficulty with inferential tasks. Hence, educators may not ask many inference questions or teach beginning readers how to answer this type of question. As a result, adults at the beginning stages of literacy development may have developed the expectation that questions asked following the reading of passages will require them to recall details from the text, rather than to use their knowledge along with text to draw inferences.

Implications

Overall, the results on the miscue and comprehension tasks indicate that adults at all stages of literacy development are able to make effective use of their knowledge as they read. This suggests that programs emphasizing and reinforcing use of knowledge to identify words and construct meaning are appropriate for adult literacy students from the beginning to the advanced stages of literacy proficiency. The results of this study do not support literacy programs that stress strategies for processing print information prior to focussing on meaning. The advantage of literacy programs that emphasize use of knowledge-based information is that instruction begins with what adults bring to programs--what they already know--rather than with what they did not know. This does not mean that strategies for processing print should be neglected. However, the goal of an effective literacy program is to help readers to integrate reading strategies rather than to rely excessively on any one strategy.

The results of this study support integrated programs for all adult literacy students. However, this does not mean that there is no need for individualization of instruction for adult literacy students or that "one size fits all." The large standard deviations, particularly evident in the miscue data, imply that there are significant differences in reading strategies among adults within every level of reading proficiency. Hence, while integrated programs will be appropriate for all adults, the relative emphasis on print-based, text-based, and knowledge-based strategies will vary depending on the needs of specific individuals. For those who rely too heavily on print-based cues, there needs to be a heavier focus on meaning. For those who rely so heavily on their knowledge that the meanings they construct bear little resemblance to that intended by the author, there needs to be a heavier focus on strategies for processing print and text information. This will require that literacy instructors teach diagnostically and that they have an effective diagnostic tool to assess the needs of their students. Informal reading inventories are an appropriate diagnostic tool to help instructors observe how adults read in the actual reading situation. The Canadian Adult Reading Assessment (Campbell & Brokop, 2000) has been designed to serve this function in the Canadian context.

Finally, the fact that there were few differences in reading strategies used by adults at different levels of reading proficiency implies the need for a spiral rather than sequential curriculum. Print-based, text-based, and knowledge-based strategies can be taught to adults at every level of literacy development. It is not necessary to begin with print, and then move to a focus on text and knowledge-based information. It is the level and complexity of material that needs to change as adults develop increasing reading proficiency, rather than the focus of strategy instruction.

References

Bloom, B. (1956). Taxonomy of educational objectives: Cognitive domain. New York: McKay.

Campbell, P. & Brokop, F. (2000). Canadian adult reading assessment: Instructor's Edmonton, Alberta: Grass Roots Press.

Chall, J.S. & Dale, E. (1995). Manual for the new Dale-Chall readability formula.Massachusetts: Brookline Books.

Fry, E. (1977). Fry's readability graph: Clarifications, validity, and extension to level 17. Journal of Reading, 21, 242-252.

Goodman, K.S. (1970). Behind the eye: What happens in reading. In K.S. Goodman & O.

Niles (Eds.), Reading: Process and program (pp. 3-38). Urbana, IL: National Council of Teachers of English.

Goodman, Y.M., Watson, J., & Burke, C.L. (1996). Reading miscue analysis: Alternative procedures. New York: Richard C. Owen Publishers.

Guzzetti, B. Anders, P.L, & Neuman, S. (1999). Thirty years of JRB/JLR: A retrospective of reading/literacy research. Journal of Literacy Research, 31(1), 67-92.

Hiebert, E.H., Valencia, S.W. & Afflerbach, P.P. (1994). Definitions and perspectives. In

S.W. Valencia, E.H. Hiebert & P.P. Afflerbach (Eds.), Authentic reading assessment: Practices and possibilities (pp. 6-21). Newark, Delaware: International ReadingAssociation.


Return to our Conference Information Centre

Return to Proceedings Index

Click here to return to the CASAE-ACÉÉA Home Page

Revised May 2001 by the ACÉÉA/CASAE Internet Working Group.