Fitzgerald, N.B. & Young, M.B. (1997). The influence of persistence on literacy learning in adult education. Adult Education Quarterly, 47(2), 78-91.

Experimental study of

The effect of persistence, or the number of hours of instruction on literacy outcomes in three different adult education settings: ESL, ABE, and adult secondary education. As each type of setting was studied separately, only the design issues and findings relevant to ESL will be reported here.

Learner Participants:

The learners in this study were participants in 44 adult education centers in the United States. As the study made use of existing test scores for participants in adult education literacy classes, subjects were only included in the study if the centers could provide valid, matched pre-test/post-test scores for that individual. In all, data were collected for 614 students. Of these, 349 were ESL learners. Approximately half of the ESL participants were male and half female. Seventy percent were native speakers of Spanish; the remaining thirty percent represent a wide variety of ethnolinguistic backgrounds.

Study Design

The authors collected pre- and post-test scores on the Comprehensive Adult Student Assessment System (CASAS) literacy test for the subjects. They also  collected information on the background of the students and the structure of the adult education program from the from the adult education centers included in the study. Multiple regression analysis was performed on the data. The study looked at possible effects based on the students’ background (ethnicity, years of previous education, and initial English literacy level), on the program structure (whether the program was open or highly structured, the cost of the program per student), and on student persistence (hours per week of instruction, hours of lab instruction, and total hours of instruction).

Findings

For English language learners, persistence had a positive effect on literacy acquisition. On average, the ESL students experienced significant gains in reading comprehension by the end of the instruction. While their average pre-test scores indicated that most had only the literacy skills suitable for entry-level jobs, their average post-test scores indicated that most had attained the literacy level necessary for participation in job training programs that involve comprehension of basic English texts.

Several of the variables studied affected the students’ gains in reading comprehension. The most important variable was the students’ initial level of English literacy. Students with higher initial literacy were much more likely to significantly increase their literacy level during the program (pre-test beta=.66). Students with more years of previous education in their first language benefited more from the instruction than students with less educational background (total beta=.14)). Also instructional intensity (measured in hours per week of instruction) did not contribute directly to improved reading comprehension. Actually, more intense reading instruction had a negative effect on reading comprehension: English reading achievement tended to decline after about nine hours of instruction per week.The authors point out, however, that more intensive instruction was also associated with greater use of the lab and with larger classes. Smaller classes more focused on classroom instruction seem to be more effective than larger classes with a split class/lab approach. Increasing the total hours of instruction, or overall persistence, did positively impact reading comprehension scores (total beta=.20). Students who participated in more total hours of instruction – spread out over a longer period of time -- were more likely to increase their reading comprehension than those who did not.

Several of the other variables addressed also seemed in impact reading comprehension gains.

Comments on the Study

The study makes good use of large amounts of quantitative data to understand the effects of different student background and program variables. The study integrated the investigation of student variables and program variables well. However, large quantitative research design prevented researchers from closely examining different aspects of the actual instruction in these programs. While this indicates that increasing instructional hours in literacy can help ESL learners to improve their reading comprehension, it does not indicate how the instructional time could best be used.  Furthermore, a problem exists with CASAS’s generalizability: as many people have pointed out, pre-post gains seem heavily related to whether the CASAS curriculum was used and not be related to gains on other literacy tests, even similar functional tests like the Test of Awareness of Language Segments (TALS).

Implications for Practice

Literacy instruction does affect adult ESL reading comprehension. Increasing the hours of instruction that a student experiences can be related to reading comprehension. The impact of student persistence underscores the negative effect of absenteeism—by being absent, students decrease the hours of instruction they received and decrease their opportunities for improving their reading comprehension. Although it may seem obvious, these data seem to indicate that teachers should make students aware of the negative consequences of absenteeism. The data from this study suggest that smaller classes, with more highly trained teachers, structured curricula, and support services are more successful in improve the literacy level of the English language learners.

Key Words

Instructional context

Reading skills transfer

Areas for Further Research

A further study could more closely examine the effects of persistence in different types of classrooms, in order to better understand the interaction of persistence and pedagogy.

back to table of contents

Learner Participants

Over 1500 adult ESL students enrolled in high-beginning through high-intermediate classes in community education programs in a major metropolitan area were tested. Subjects were included in the study only if they had taken all three of the tests. In the end, 300 sets of matched test scores were included in the analysis.

Study Design

Reading comprehension sections from three major tests commonly used in community adult literacy programs were administered to the participants. The students took an ESL specific test (the CELSA—Comprehensive English Language Skills Assessment) and two adult basic education tests (the ABLE—Adult Basic Learning Examination—and the TABE—Test of Adult Basic Education). Basic demographic information was collected from the students as well.

Findings

There was a correlation between the scores on the CELSA and the TABE. There was a weaker correlation between the scores on the CELS A and on the ABLE. However, the correlation between the CELSA and the ABE tests was not strong enough to predict scores on one test based on the scores on the other. Higher levels of language proficiency (based on class placement) were associated with stronger correlations between the scores on the two tests.

Comments on the Study

The study includes a careful explanation of the research method and the statistics. Test statistics are responsible and clearly explained.

Implications for Practice

While there is some correlation between the scores of this adult education test and the ESL test, especially for higher level students, this correlation is not strong enough to justify substituting the ABE test for ESL learners. ESL tests should be used rather than adult basic education tests for monitoring student progress in literacy learning. Given the higher correlation between the scores for the higher level learning, it is possible that learners could benefit from earlier mainstreaming into adult basic skills classes. 

Key Words

Reading tests

Community English

Areas for Further Research

Researchers could experiment with the effects of earlier mainstreaming of adult ESL students into other adult education training courses.

back to table of contents

Learner participants

The learners were adult ESL students enrolled in adult basic education classes. All the participants were Asian immigrants. The students sought to improve their English proficiency in order to continue their education at the community college or skills training program level in order to increase their job opportunities in the United States.

Teacher participants

The researcher was the teacher of the class, and had been teaching at the center for several years. He compared his past experiences with student difficulties with reading to the experiences of this class.

Study Design

The researcher implemented new reading techniques in the same curriculum he had taught in past years and reported on the success of the techniques based on comparisons of student reading behaviors from year to year. The techniques he employed were 1) previewing for schema building and 2) teaching students to read for information. Previewing was accomplished by introducing students to the themes of a text and asking them to free write on some of the issues in that text, as well as helping students make predictions about the text. To teach students to focus on the informational content, the teacher helped students to exploit headings, tables, and charts in order to understand the main points of the reading and to find information more efficiently. 

Findings

The researcher found that the students were less likely to be frustrated by the reading and less dependent on bilingual dictionaries if the texts were appropriately previewed before the reading began. He also found that reading for information content helped the students to process larger amounts of text more rapidly and with greater comprehension.

Comments on the Study

The study contains good descriptive detail of the desired reading behaviors and the ways in which the relatively simple innovations encouraged better reading strategies. The claims would be stronger if comprehension measures had been included – and if the researcher had measured the students’ actual rate of reading.

Implications for Practice

Relatively simple reading instructional techniques like previewing texts are often not included in curricula because of a lack of instructional time. This report suggests that it is more effective to use the time in previewing to help students comprehend the main points of the text than to spend time later helping them deal with the text word by word.

Key Words

Previewing

Schema

Predicting

Top-down processing

Areas for Further Research

This work could be replicated with comparison groups and comprehension measures.

back to table of contents

Goodman, K. (1988). The reading process. In P.L. Carrell, J. Devine, & D.E. Eskey (Eds.) Interactive approaches to second language reading, (pp. 11-21). Cambridge: Cambridge University Press.

Theoretical article discussing

The process of reading. Goodman sought to establish a psycholinguistic model of the actual process of connecting meaning to printed language.

Theory

Goodman begins with the premise that reading is more than simply decoding and that better understanding of the actual reading process will lead to better literacy pedagogy. Goodman defines reading as an active process in which the reader uses linguistic cues embedded in the text to reconstruct a message that the writer has encoded in written language.

Study Design

Goodman’s reading model is based on oral miscue analysis. In oral miscue analysis, the reader is asked to read a text aloud. Differences between the written text and the oral reading are considered miscues. It is assumed that miscues are not random mistakes, but rather the result of the reading process unsuccessfully used. Thus, miscues can give researchers insights into the reading process itself.

Findings

Goodman’s model describes the reading process as a cyclical model, in which the mind processes the optical structure (the written symbols), the syntactic structure (relationships among words), and the meaning structure (the message the reader reconstructs). Processing the different structures to arrive at meaning is the goal of each reading cycle. Proficient readers move seamlessly from one cycle to another as they progress through a text. As readers move through cycles of reading, they employ certain processes. These are: recognition-initiation, which occurs when the brain recognizes written symbols as text and begins the reading process; prediction, which is the process of actively predicting what will follow in the text; confirmation, in which the brain seeks to verify previous predictions based on input; correction, which occurs when predictions are not verified and must be modified; and termination, which occurs when the reading is finished or stopped. While initiation and termination normally only occur once in a reading situation, prediction, confirmation, and correction occur cyclically throughout the reading. The success of the reading process is based on confirming or correcting predictions; in other words, successful reading ends in the reconstruction of meaning.

When meaning is not derived from a text, the reading has been unsuccessful. Goodman characterized unsuccessful reading as “short circuits”. When reading is too difficult, for example, readers may short circuit meaning, seeking to understand only the syntactic structures. This is seen in the ESL classroom when readers are able to find the subject and object of a verb, but do not seem to process the meaning of the text.

Comments on the Study

Goodman’s original model of reading was an early psycholinguistic model of the reading process. Goodman asserts that previous models of reading had considered only linguistic cues, leaving out the cognitive processes involved in reading. This revised model makes use of the top-down predicting processes on which the first model was based, but also concedes a place for the bottom-up processing of written symbols. However, the model is still very focused on meaning-based, rather than print-based, aspects of reading. For example, of the five processes Goodman describes as occurring during reading, only one—initiation—is based on the recognition of linguistic cues. The others are more dependent on the meaning that the reader brings to the reading experience.

Any model of reading based on miscue analysis is more likely to explain unsuccessful reading than the processes of successful reading. Furthermore, Goodman’s assertion that he can correctly identify the reason for any given miscue and fit it into his categories for miscue analysis is dubious, at best.

Implications for Practice

Goodman’s model suggests that successful reading depends on good predicting skills. Readers who predict well will need fewer cues to comprehend text than poor predictors. Good predicting skills make the reading process easier and faster. By building good predicting skills, ESL learners may be able to compensate for lower language proficiency. ESL literacy teachers should help their students make predicting meaning part of the reading process. However, since these prediction skills are operative only if the reader can both decode and know the meanings of 90-95% of the words, prediction activities would have the most value if used where the readers know both the topic and the language of the readings.

Key Words

Miscue Analysis
Top-Down Processing
Predicting
Schema

Areas for Further Research

Additional research could focus on the effectiveness of teaching predicting and confirming skills in improving students’ reading comprehension and reading efficiency.

back to table of contents

Gough, P.B. (1985). One second of reading. In H. Singer & R.B. Ruddell (Eds.) Theoretical models and processes of reading (pp. 661-86). Newark, DE: International Reading Association.

Theoretical article discussing

The nature of reading from a bottom-up, information-processing standpoint. Gough attempts to describe what happens in approximately 700 milliseconds, from the time that the eye fixates on the page until a sentence is processed.

Theory

In Gough’s explanation of reading, it begins with eye fixation on text. At this point the visual information is formed into an icon, or a section of the printed page that is processed. This space is assumed to encompass twenty letters in the space of a two-inch by one-inch oval on the page. Once the icon is formed, the process of letter identification begins. Letter identification is extremely rapid—Gough estimates from letter acknowledgement studies that letters are identified at a rate of ten to twenty milliseconds per letter. Combined with eye fixation times, this yields an average reading time of three hundred words per minute. These speeds cause Gough to conclude that we do read letters in sequential order.

Next, the letters are mapped to the phonology system, either as full words or—as Gough suggests—as systematic phonemes (morphophonemic units). The phonemic mapping allows the reader to match the letters to a word from the spoken lexicon. Invoking the phonology system allows for the unification of the lexicon; it assumes that readers have one lexicon for both spoken and written language, and that this lexicon is accessed through the phonology system for both. Once the words have been retrieved, they are transferred to primary, or working memory, where they are analyzed as a sentence or clause. After sentences are understood, they are transferred from working memory to free space for new incoming words. Gough theorizes that the sentences might be held in secondary (or short-term) memory, but in light of the paucity of evidence on this point, labels the transfer-point for processed sentences PWSGWTAU (the Place Where Sentences Go When They Are Understood). These sentences are still available to memory, so that they can be combined with the rest of the text to understand the text as a whole.

Findings

The model gives insights into the necessary equipment for reading. The beginning reader must have a visual (or in the case of blind readers, tactile) system capable of fixation on the textual input. Also, the model assumes that the reader must come to the task of reading with a lexicon, a comprehension device, and a phonological system. While these may not be fully developed, they must exist in at least rudimentary forms for reading to occur.

With these resources in place, the beginning reader must master both letter recognition and decoding (or the conversion of recognized letters into systematic phonemes). Gough suggests that this process is mastered through cryptanalysis, in which the beginning reader is presented with equivalent data in spoken and written form in a sufficient amount over time until connections are made. For children reading in their first language, this may be accomplished when they follow the text of a book while adults read out loud to them. At this point, the child can presumably make form-meaning connections, using the phonological system to call words from the lexicon and processes them as sentences.

Comments on the Study

Gough’s work is one of the most complete and best substantiated of the various bottom-up models of the reading process. Readers must accomplish the tasks he suggests in order to process text. The problem with the model comes from its complete reliance on sequential processing. In light of studies of letter and word string processing, it is clear that information from texts is not uniformly processed in a sequential manner. As Rumelhart points out, sequential systems like this, while they elucidate the necessary component processes of reading, do not allow for non-sequential influence of components on each other. They thereby fail to capture the rich interactions between knowledge and skill systems that characterize proficient reading.

Implications for Practice

This model is useful for teachers, in that it explains the components of reading that can form the basis for strategies readers must develop to be successful. Whether this processing occurs in the sequential, non-interactive order that Gough explicates or not, readers must develop all of these skills in order to read successfully. Second language teachers of non-literate students need to help their students learn the basic processes of letter recognition and decoding; even literate students need practice mapping form to meaning in a second language. While bottom-up processing not constitute the whole of second language reading, it is part of the processes and should not be neglected in instruction.

Key Words

Bottom-up processing

Models of reading

Areas for Further Research

The process and speed of letter recognition, the process of phonemic mapping, and the process of lexical retrieval in ESL reading could be studied. It would be interesting to contrast these processing strategies for learners with different levels of first language literacy.

back to table of contents

The research took place at a community association building, located in the heart of the Cambodian community. The learners were mostly Cambodian females in their twenties and thirties. All lived within a few blocks from the school and many were mothers of children who attended school nearby. Most of the students came from farming families in rural areas of western Cambodia. Few had more than three years of schooling in Cambodia. While most were able to read definitions from a Khmer-English dictionary, there was wide variation in their literacy beyond that. It is unclear how many students were included in the study because the author does not give this information. Rather, he says that 10 –15 students were in the Khmer teachers room daily, while 5 were in the author’s.  It is unclear whether they were always the same 10 – 15 and the same 5 students in class, so the total number of students in this study cannot really be determined.  

Teacher Participants

For the most part, the study draws on observations of two teachers in the center. One is the researcher as well, who did participant observations of his classes. He is an American with very basic knowledge of Khmer. In his intermediate level class, he saw his role as that of a helper and a resource. Similarly, he saw the role of older Cambodian children who accompanied their parents to class as a resource to help their parents learn.

The other teacher was a Cambodian who had learned English over several years in the United States as a Buddhist monk. He closely followed a typical style of Cambodian teaching in which the teachers are very strict. His students did not turn to their children for help with English. The teacher used Khmer extensively to support the English language materials used in class.

Study Design

The data for this study were collected as part of a five-year ethnographic study of a Cambodian community in the United States. The researcher observed and interviewed the participants and their families at home, and, as a teacher in a community center ESL program, was a participant observer of the literacy learning process in two different classes over the course of a year. He also interviewed other teachers and staff from the school, and collected “artifacts of literacy” including student writings.

Findings

Based on the data drawn from field notes and interviews, the researcher created a typology of interactive structures among the Cambodian students. The structures deal with the behavior of the adult students when working with each other, with the children, and with their teachers.

The adults interacted with each other as “knowledge holders.” In this role, they assisted one another with answers in class, on the homework, and on tests. They also interacted as “facilitators,” translating information relating to classroom management such as instructions. To do so, they drew on their knowledge of both oral and written English, as well as oral and written Khmer.

Children frequently accompanied their parents to class. The children often had higher English literacy and better English oral proficiency than their parents, but were not literate in Khmer. They also interacted with the adults as knowledge holders and facilitators, but also functioned as “evaluators,” correcting their parents’ mistakes in class both in written and oral English. While adults often used Khmer writing to exchange knowledge, children only used oral Khmer to give information to their parents.

In addition to fulfilling similar roles as the Cambodian children, the native Cambodian teacher also acted as an authority figure. Having extensive previous education in Khmer and Vietnamese, he was seen as both a Cambodian and English authority. His teaching style included a great deal of copying and oral reading on the part of the learners, which he prompted, corrected, and praised. Like the American teacher, he presented and explained the language, but he did so making use of both Cambodian and English. Because of his ability to fulfill both teaching and translating roles, the interactions in his class were mostly between himself and the students, not between students or between students and their children. Because of this, his classroom more closely reflected the cultural norm of Cambodian classroom, whereas the monolingual teacher’s classroom more closely resembled the students’ home environment.

Comments on the Study

This study reports on extensive data collection with rich, detailed reporting, painting a very clear picture of the classes and the interactions that take place in them. Supplementing the observations with interviews with the students, teachers, and family members clarifies the nature of the interactions that take place. The additional observations at home add more complexity and completeness to the descriptions.  For teachers not familiar with the culture, the information on school attitudes of Cambodians could be useful.

Although this is an excellent descriptive study, it does not document the effectiveness of the different sorts of classroom interactions mentioned in terms of language learning. There is no assessment data of any kind for this.  

Implications for Practice

The classroom is one of the interactional settings for members of a minority community. Teachers should strive to understand the interactions of their students in this and in other settings to understand how the classroom builds on and contributes to the literacy experiences of the students.

Key Words

Instructional context
Family literacy
Community English

Areas for Further Research

The effect of these different interactions and of the different types of classrooms could be studied to determine whether one or the other is more beneficial for literacy acquisition. 

back to table of contents

Sixty-three learners from an intensive English center in the United States volunteered to participate in the research. Learners spoke Spanish, Japanese, or Arabic as their native languages (Tunisian Arabic speakers were considered separately from other Arabic speakers because of their extensive educational background in French). The learners had a variety of English proficiency levels, according to the center’s proficiency exam. Most of the Spanish speakers were clustered in the lower levels of proficiency, while most of the Japanese and Arabic speakers were clustered into the higher levels of proficiency.

Study Design

The students read two stories about animals. The stories had parallel structures and syntax. Each story had two nonsense words embedded in it. One of the words was only used once and could be guessed using the immediate context, through a local context clue (Laufer, 1997). A local clue is a context clue in the same sentences as the unknown word. For example, if a student didn't know the word "partridge", and read the following sentence" The partridge spread her wings, ruffled her feathers, and chirped loudly." the words "wings" "feathers" and "chirped" would all be local context cues. The reader might well be able to infer that the animal being described is a bird. However, in most texts not designed specifically for ESL speakers, sentences do not contain much redundant information in the same sentence. In these cases, there tend to be global cues, which are context cues derived from understanding the meaning of the passage as a whole. An example of this is the passage "It was early morning when the partridge lifted her head. The first rays of the sun fell through the branches of the tree and lit up the leaves surrounding the nest. Through the morning mist, she could see the others taking flight." The reader cannot tell from the text what kind of an animal a "partridge" is and will have to infer the meaning of the word not from the immediate context of the sentence but from the meaning of the passage as whole. In this study, many of the clues were global: All the word was used repeatedly (in each story, it was the type of animal); its meaning had to be guessed using the global context of the story. The students were instructed to read the passages and then retell the stories to the researcher. After reading and retelling both stories, the students were shown each passage again and asked to point out words that made the story difficult to understand. As the students pointed out words, they were asked to guess their meaning and explain the reason for their guesses.

Findings

Across all groups, students were significantly more likely to guess the meaning of the nonsense word correctly when the contextual clues were local rather than global. Their ability to guess the meaning of words in local contexts, however, was undermined in the first passage because many of the students did not know the meaning of the word in the sentence that was used as a contextual clue. This is one of the biggest limitations of the use of contextual clues—several times the researcher found that students were not able to make use of clues because they did not understand the clues themselves. Considering the text as a whole, and not just the nonsense words, students were unlikely to guess word meanings in global contexts; they were better able to guess word meanings when the unknown word appeared along with comprehensible contextual clues.

When contextual clues were unavailable or incomprehensible, students commonly resorted to word analysis to guess meanings. While this was successful with some words, it was problematic for many others. For example, Spanish and Tunisian learners often analyzed the word campfire according to Spanish and French cognates which have military connotations. They then interpreted a passage about a scavenging animal as a story about a soldier running away from battle. This illustrates the negative impact guessing word meanings can have on reading comprehension. Japanese and other Arabic students also made unsuccessful guesses when they sought to analyze words. For example, several Japanese students tried to related habitat to habit, guessing its meaning to relate to habitual actions. Similarly, students from all language backgrounds tried to break the word offspring into its component parts, arriving at guesses like “the end of spring.”

A third problem with guessing came from mismatching background knowledge to the written stimulus; the word they processed from memory was spelled/pronounced differently than the word in the text. For example, several students unsuccessfully guessed the meaning of tapped because they related it to top. Strikingly, many mismatches were syntactically impossible in the context, which could indicate that instructing students to guess at word meaning diverts attention from syntactic processing.

Lastly, processing time was an issue. Students needed time to decide if a word was new or unfamiliar, indicating that vocabulary recognition was not automatic for them in English at this point. Once they determined that they needed to guess word meaning, they needed more time to evaluate and guess. This could indicate that guessing words in context is not a time-saving strategy that allows students to process more input faster, as other reading theorists have asserted.

Comments on the Study

The researcher carefully explains the problems that exist when readers rely on guessing word meanings as the major strategy for reading in a second language. The results are carefully explained and well exemplified.

Because student proficiency levels and language background were not separated as variables in the sampling, it is not possible to draw conclusions on the effect of language proficiency or language background on successful guessing of word meanings.

Implications for Practice

Teachers should be aware that guessing from context may not always be productive and reading comprehension may be impaired students are urged to rely too heavily on guessing meaning from context. Vocabulary development should not be neglected by reading teachers. Use of monolingual and bilingual dictionaries should not be ruled out.

Key Words

Vocabulary
Top-down processing

Areas for Further Research

Researchers could investigate the longitudinal impact of guessing for meaning in context by tracing the development of learners who do and do not use dictionaries in their reading.

back to table of contents

Hilferty, A.G. (1996). Coding decoding: Predicting the reading comprehension of Latino adults learning English. Unpublished doctoral dissertation, Harvard University, Cambridge, MA.

Descriptive study of

The relationship of L2 decoding skills to other reading and language sub-skills in the reading performance of 42 adult Latino learners of elementary English as a second language reading. The study was performed in two phases. Phase I was a correlational study where data from the study was analyzed quantitatively to identify patterns of relationships among selected reading sub-skills and personal variables. Phase II was an observational study of the ESL instruction of adult Latino ESL readers, and explored the utility of the correlational data in the classroom context. 

Learner Participants

Forty-two Spanish-speaking male and female adults attending or recently completing ESL classes for adults in the Cambridge/Somerville area were selected for the study. Participants ranged in age from 16 – 53, with a mean level of 35.7. Level of education ranged from four to 20 years with a mean level of 13.7 years. Length of stay in the United States ranged from three to 216 months with a mean length of stay of 73.31 (about 6 years). Length of attendance in ESL classes ranged from a mean length of one to 60 months. At least 24 of the subjects were employed in full or part-time jobs. Thirty-eight of the learners completed the study.

Study Design

Instruments were chosen to measure language and literacy skills in Spanish and English, including tests of reading comprehension, listening as a measure of oral proficiency, phonemic awareness, decoding through letter/word recognition skills and pseudo-word recognition skills, and a letter-word recognition time measure. Reading comprehension in English was the dependent variable. Decoding skills in English, as measured by letter/word and pseudo-word recognition were the question variables.

The following tests were used:

1. Reading: Bateria Woodcock Psico-Educatovio en Espa.Hol (BWPEE) and the Woodcock-Johnson Psycho- Educational Battery (WJPEB). The BWPEE had been adapted from the WJPEB.

2. Oral proficiency: Because measure of oral proficiency were not yet available for use with Woodcock –Johnson batteries, oral proficiency was measure using Spanish and English listening subtests from the Language Assessment Battery (LAB) of the New York City Public Schools. In the LAB, the subjects select the correct response to an oral stimulus by reading the possible answers silently and marking them in a test booklet.

3. English phonemic awareness: Rosner Test of Auditory Awareness Skills (TAAS). Spanish phonemic awareness: an adaptation of the TAAS developed by staff of the Cambridge Community Language Center. In the TAAS, the subject responds orally and the researcher records the answer.

All tests had corresponding grade levels attached to them. Participants were also interviewed in Spanish to find out information about educational levels in Spanish, time in the United States, and whether or not they were working.

The data were analyzed in two ways: 1. Subjects’ patterns of performance were viewed in a correlation matrix displaying relationships among all the variables in the study. 2. Relationships between the measures and the outcome variable were examined in multiple regression analysis.

Findings from the correlations study

Subject’s mean grade equivalences in English listening were substantially higher and closer to the counterpart Spanish grade equivalents than were their mean grade equivalences in English word attack, reading comprehension, and phonemic awareness.

The four measures that correlated most highly with English reading comprehension were English letter/word identification, English listening, English pseudo-word identification, and English phonemic awareness. In an analysis of the Spanish and English language and reading sub-skills, English decoding accounted for 15% of the subjects’ reading comprehension.

In short, the major reading variables are related to each other across languages: Spanish and English reading comprehension were correlated, and Spanish and English letter/word identification were related to each other.

However the variables related to oral and phonetic aspects of language were not related to each other: Spanish and English listening, Spanish and English phonemic awareness and Spanish and English word attack are not related. Some of the variables that were unrelated to each other across the languages were related to other sub-skills across the language. For example, although Spanish phonemic awareness was not related to phonemic awareness in English, it was related to English reading comprehension, suggesting a Spanish to English effect. Furthermore, oral proficiency and literacy skills in both languages interact with reading sub-skills.

Four predictors that had significant correlation with English reading comprehension were mother’s education, Spanish listening, Spanish reading, and Spanish phonemic awareness.

English reading comprehension had a weak relationship with subject’s education and no relationship with the following: father’s education, age of arrival, the Spanish word-naming variables, Spanish letter/word identification, and Spanish pseudo-word identification.

The Observational Study Design

The implications of the results of the correlational study crossed two areas of application: instruction and research. The two questions to be answered by an observational study were:

1. If ESL decoding is a significant predictor of reading comprehension for L1-litearcy Latino adults reading ESL at approximately a grade 3 level, what are the implications for class groupings and for instruction?

2. What is the utility of research in the basic skills for L2 instruction?

Five participants, all women in their 40s, were selected for the case study research for the observational study. The participants came from three Latin American countries. Two had been in the U.S. for ten years, and had delayed beginning their ESL studies for over eight years after arrival. The other three had lived in the U.S. for less than two years, but had begun studying English within a year of arrival. All had least high school education in their countries some had vocational and one had a college degree. They all had scored fairly high in Spanish oral proficiency and fairly low in English oral proficiency. All had scored low on English reading comprehension.

The observational case studies took place over a period of four months from December 1995 – March 1996. The purpose was to visit the classes of these five learners to see whether creating a double context of statistical and observational data would lead to new information about adult acquisition of second language reading and new ideas for further studies.

Descriptive information was gathered from observations of classroom activities with literacy components by a team of two observers: one English dominant and one Spanish dominant.

The instruments used for observation were the following:

1. A behavioral observations guide

2. An instructional observation guide

3. Two questionnaires (in Spanish) for the students

4. A guideline for teacher interview

5.  A guide for coding observations.

Questions examined in the classroom included the following:

1. How do learners respond to actual ESL instruction?

2. Are their levels of L2 decoding skill apparent in their classroom performance?

3. Does their instruction take decoding levels into consideration?

5.Is it feasible to accommodate decoding in ESL instruction in the adult ESL class?

Findings form the Observational Study

Only one of the classes (which included two of the subjects) included a high number of activities relevant to decoding. Two of the classes had a few activities linked to decoding and one had almost none. Although all classes observed contained some elements of instruction related to decoding, decoding activities were not usually delivered through direct instruction, and were seldom part of teachers’ long-term plans for instruction, as determined through the class observations and the teacher interviews. Furthermore, none of the learners interviewed specified decoding as a goal. Theys were much more focused on listening and speaking and were happy with the instruction they were receiving.

All of the women observed had scored low in English reading comprehension, however those who had delayed their study after arrival in the United States for several years had especially low measures of English letter-word identification and phonemic awareness. 

The following questions were answered:

1. Would some instruction contain elements of the instruction considered optimal for normal child L1 readers?

Some, yes. But in most of the classes, these elements of instruction were incidental to other goals of the teachers and were not part of the longer-term, conscious goals on the part of the teacher.

 2. Would the information from this correlational study be usable in the planning and conducting of the instruction observed? In others words, would it be feasible to teach English decoding in the ESL classroom?

Yes. There were many opportunities to use decoding in the ESL classrooms. There was a lot of environmental print as well as instructional print. Class activities such as one using the song “I just called to say I love you” (Stevie Wonder) provided lots of opportunities for decoding and word recognition that were not exploited.

Implications for Practice

1. Adult Latino learners at the beginning level of reading in English would benefit from instruction and practice in English decoding as well as in English listening. This is consistent with the theory – prevalent from research with L1 reading in children – that suggest that mastery of lower-level reading skills is necessary for the development of higher-level skills in reading.

2. The transition from decoding Spanish to decoding English is likely to be more difficult than the reverse. The evidence of a possible transfer from L1 to L2 decoding suggests that for such students, it may make more sense to master the more transparent Spanish decoding system before tackling the English system. However, there appears to be no full, automatic transfer of Spanish decoding to English decoding skills. It appears that students independently acquire second language word recognitions skills.

3. It is quite possible that the relationship between reading comprehension and decoding is reciprocal. Therefore, classroom activities should include instruction in both skills.

4. Even for subjects who have fairly high levels of education in their first language, working with printed language in English is hard. Oral skills do not automatically transfer to print.  Deliberate instruction is needed in reading.

5. Teacher education should include instruction on the importance of decoding to beginning-level readers of English and in ways to foster this.

 6. Delaying the study of a new language for several years may inhibit the development of word recognition skills.

Comments on the Study

This study seems to be well planned and well executed. The combination of the correlation study and the observations is worthwhile. Of particular interest was the importance decoding appeared to play in reading comprehension and the fact that it did not appear to automatically transfer from Spanish to English. Also of interest was the fact that decoding was not being directly taught in the classrooms.

However, the small sample size (38) and the very small number of case studies  caution against too much generalization from the findings. Also problematic was the use of some of the tests: the English listening test had been normed on children only. Furthermore, it was not a true listening test in that it was a multiple-choice pencil and paper test that required reading ability. Since there were no beginning level literacy learners this was probably not an issue for this study, but would be if students with fewer L1 and L2 literacy skills had been tested.

Key Words

Decoding
L1 literacy
Reading skills transfer
Reading tests
Second language development
Second language proficiency
Teacher Education
Written-oral/reading relationships

Areas for Further Research

More research should be undertaken on the role of decoding in adult ESL acquisition – when and how is it most useful. This study could be replicated with a larger group of learners. It could also be done with learners whose L1 does not use the Roman alphabet. Further research could focus on comparing adult patterns of learning and the patterns of the child L1 learner. Research should be done on whether there are some adults for whom initial instruction in L1 decoding should precede L2. Research is needed on the effects of different kinds of class composition on language learning – whether placement by oral proficiency only meets the needs of learners and whether compartmentalizing English is a good idea, or if integration of skills is not preferable. Research is needed on the effects of delaying instruction for years after arrival in the United States: A study could be done testing the reading skills of adult learners at similar levels of total English proficiency who vary by length of time spent in the United States before beginning instruction.

back to table of contents 

Hsueh-chao, M.H. & Nation, P. (2000). Unknown vocabulary density and reading comprehension. Reading in a Foreign Language, 13(1), 403-430.

Experimental study of

The relationship between vocabulary knowledge and text comprehension for ESL readers. The researchers attempted to determine what percentage of words in the text learners needed to understand in order to engage in reading with comprehension without looking up unknown words.

Learner participants

The learners in the study were adult ESL learners in a pre-university English course in an English-speaking country. Most were preparing to study at an English speaking university. They were chosen based on their scores on the Vocabulary Levels Test (Nation,1990). They needed to score more than 14 of 18 at the 2000 word level as well as more than 10 out of 18 at the 3000 level.

Study Design

Four versions of a text were prepared for the study. The text was 673 words long and was written as a first person narrative. In three versions of the text, low frequency words in the text were replaced by nonsense words (nonsense words were used in order to guarantee that no learners knew the words). One version (95% coverage) had 5% of the words replaced by nonsense words. The 90% coverage version had 10% of the words replaced and the 80% version had 20% of the words replaced. Words were chosen based on Francis and Kucera’s (1982) frequency list. They were replaced in reverse order of their frequency, with the least frequent words replaced first. Words appearing multiple times were always replaced by the same nonsense words. After the replacements were made, the texts were simplified to guarantee that only the nonsense words would be unknown. All words that had not been replaced by nonsense words had to be within the most frequent 2000 words of English (well within the vocabulary level of the subjects).

The students were randomly assigned to read one of the four passages. Following their reading, they each completed three tasks. The first task was a multiple choice test containing fourteen items dealing with explicit and implicit information from the text. The questions dealt with ideas determined by a native speaker to be in the top 50% of the idea of the story based on importance of the idea to the text. One point was given for each correct answer.

The second task was a cued-written recall. Students were given questions as prompts. Their answers would supply information that served to reconstruct the story. The aim of the written recall was to get learners to produce as much as they could about the story. In all, 27 cues (all in the form of questions) were given. The recall was scored by weighting each idea according to its importance in the story. If a student recalled one of the top 25% most important ideas, four points were awarded. For recall of the next 25% most important ideas, three points were awarded, etc.

Finally, learners were asked to predict what their scores on the tests would be. The purpose of this was to determine how the different amounts of unknown vocabulary of the text affected student perceptions of comprehension.

Findings

On the multiple-choice measure, the mean score for each group clearly decreases as the percentage of nonsense words increases. No student who read the 80% coverage text, for example, scored above 10/14, while this is the lowest score for students who read the 100% coverage text. The results for the recall were very similar—as the percentage of nonsense words increased, student ability to recall decreased. Statistical measures indicated that 48% of the variation in student scores could be explained by the difference in the amount of nonsense words, indicating that the percentage of unknown words in a text is an important factor in comprehension. The authors looked for evidence of a vocabulary threshold. A sudden jump in comprehension at a particular level of vocabulary understand would indicate that a threshold had been crossed. The study does not find evidence for a threshold, in that there is no sudden leap in comprehension at any one point, but rather a steady relationship between the percentage of words known and comprehension. Across groups, student predictions of their scores were fairly accurate, indicating that students are aware of miscomprehension caused by unfamiliar vocabulary.

Comments on the Study

The study is very well controlled, and the measures are well designed. The recall measure seems to be somewhat redundant to the multiple choice measure, since the students were asked so many questions that they were more likely to simply answer the questions than attempt any independent recall. The recall could be better considered as a short-answer test than a true recall. This similarity likely accounts for the high correlation of multiple-choice scores and recall scores. The scoring of recalls is somewhat counter-intuitive. The top 25% most important ideas, which were most likely also the most salient ideas, were awarded the most points. It would seem more logical to award more points for comprehending the less salient ideas. No rationale for this decision is given. In addition, the results, while suggestive, are not necessarily the same that would be found if a students encountered similar percentages of unknown actual words, where they could have applied intraword analysis or other word-level strategies to guess meanings.

Implications for Practice

Vocabulary knowledge, while not the only component of reading, seems to be one of the most important knowledge areas for second language readers. Based on the regression analysis performed by the researchers, students reading narrative texts need to know at least 98% of the vocabulary they encounter in the texts in order to read with comprehension without needing to use a dictionary. Other reading skills like guessing meanings in context will not be able to compensate for vocabulary knowledge lower than this level. Given this high level of comprehension needed before reading in context strategies are effective, pre-teaching of unknown vocabulary seems to be indicated for most learners with most texts.

Key Words

Top down processing
Vocabulary
Threshold

Areas for Further Research

This is a very well designed research study. It could be replicated among a different population, or it could be replicated using texts on different topics with different groups to determine how vocabulary knowledge and schematic knowledge interface in reading comprehension.

back to table of contents

Hudson, T. (1982). The effects of induced schemata on the “short circuit” in L2 reading: Non-decoding factors in L2 reading performance. Language Learning, 32, 3-31.

Experimental study of

The effects of inducing schema activation on students’ reading comprehension. Much research has shown that second language readers, in applying background knowledge to a text, will ignore or alter contextual clues to bend the text to the schema they wish to apply. The author wanted to determine whether inducing the correct schema either before or during the reading process could increase learners’ comprehension of a text.

Learner Participants

The participants were ninety-three students enrolled in an intensive English language institute. The students came from various linguistic backgrounds. They were divided into beginning, intermediate, and advanced levels based on the reading class in which they were currently enrolled.

Study Design

Three reading passages dealing with scientific topics were prepared for each proficiency level. For each reading passage, three different types of schema-inducing activities were designed. The first, pre-reading, involved discussion of pictures related to the topic of the reading. The second, vocabulary, involved students receiving and discussing a list of relevant vocabulary before reading. For each of these treatments, the schema-inducing activity preceded reading the article, which was followed by a multiple choice comprehension test. In the third treatment, the students read the article and answered the multiple-choice questions. They were then given the article again and retook the original test. The purpose of this treatment was to see if post-reading activities (answering comprehension questions after reading the text) could induce schema.

Each proficiency group was given all three types of reading treatments for the articles selected for their proficiency level. Within each proficiency level, the students were divided into three groups, so each group performed the activities in a different order and with different articles. The treatment took place during normally scheduled reading labs. Looks like a good design.

Findings

Judging by the scores on the post-test when compared to those of the pre-test, for the beginning students, the pre-reading discussion treatment was significantly more effective than the pre-reading vocabulary treatment or the post-reading treatment. For the intermediate students, both the pre-reading treatment and the vocabulary treatment were more effective than the post-reading treatment. For the advanced learners, however, the post-reading treatment was the most effective, although there were no significant differences among scores received by those the treatments. The researcher interprets these findings as evidence that schema-induction can have an effect on comprehension at different proficiency levels. He also concludes that the reading strategies students use differ by level. The advanced students, for example, seem to apply schema more flexibly, as they are able to modify their schema after reading, while beginning and intermediate students are not. Thus, schema-induction should be attuned to the proficiency level and strategy use of the students.

Comments on the Study

The study is very well designed and the findings are carefully interpreted. The researcher also wisely cautions readers to consider the possible effects of conducting this research in a laboratory, rather than a classroom setting and of using multiple-choice tests as a comprehension measure. A student can possibly have high scores on a multiple-choice test without having successfully reconciled the schema to the reading. Multiple-choice tests reveal information about general comprehension and about test-taking skills, and may not be fine-tuned enough to detect the effects of poorly reconciled schemata. It is likely that the effects for advanced learners were small for all treatments in part because they may have already gotten 90% of the texts, so there was not much room for improvement.

Implications for Practice

Teachers can help students build appropriate schema to comprehend new texts. This could increase student confidence and comfort in reading at any level, and increase the amount of comprehensible input through reading the students receive. However, the level of the student should be taken into consideration when choosing schema-building activities. Beginning and intermediate students may benefit more from activities that encourage them to build schema before reading. Advanced level students appear to be able to benefit from schema-building activities whenever they are presented, whether before or after the reading.

Key Words

Schema
Top-down processing

Areas for Further Research

If students do change their reading strategies as they become more proficient, longitudinal studies could possibly give us information about when and how strategies are changed.

back to table of contents

Huntley, H.S. (1992). The new illiteracy: A study of the pedagogic principles of teaching English as a second language to non-literate adults. Unpublished manuscript. (ERIC Document Reproduction Service No. ED 356685)

Theoretical discussion of

Theory

Much research has indicated that non-literate adult ESL learners have special literacy needs that are often not addressed in traditional ESL classrooms. Huntley begins the study with definitions of literacy in terms of functional ability and with a consideration of the dual burden that non-literate ESL learners face of learning both oral and written language at the same time. After considering proposals that written language should not be taught until after learners have achieved a highly functional command of oral language, she concludes that the most effective manner of instruction for non-literate ESL adults is one that integrates oral and written language learning. She then discusses different types of ESL non-literate learners and addresses pedagogical approaches to meet their needs.

Findings

There are four types of ESL learners who are in need of ESL literacy instruction: pre-literate, non-literate, semi-literate, and non-roman alphabetic literate. Pre-literate ESL learners come from cultures in which literacy is rare or non-existent. These learners have had no or extremely limited exposure to written language. Pre-literate learners generally have a slow-paced learning rate, and require repetition and re-teaching. They need to begin literacy instruction with practice in discriminating different letters and with comparing sight words before they will be ready for literacy instruction. The teacher needs to make it clear to this group of learners that the written symbols are a method of communicating, and that reading will eventually be meaningful.

Non-literate ESL adults come from cultures where literacy is fairly common, but have never learned to read. While they do not read, they have had the opportunity to notice printed language, and are more aware of the role of reading in society. These learners likely feel stigmatized because they cannot read, and many will be unwilling to disclose their illiteracy. Teachers should be very sensitive to these learners’ affective needs.

Semi-literate ESL adults have not progressed beyond elementary level literacy in their native languages. They are likely as well to have acquired basic English literacy skills, but not to have acquired functional literacy to meet their personal or employment needs. Teachers can help these learners build on their literacy skills through extensive reading at an appropriate difficulty.

Non-roman alphabetic ESL adult learners can be highly literate in their first language, but are unused to the Roman alphabet. Their first language literacy is still an important advantage; these learners can still transfer cognitive skills of reading, but need practice with forming letters and numbers and with sound-symbol correspondences.

Comments on the Study

The study gives a good overview of some of the important issues teachers face in adult ESL literacy situations. Reviewing student needs in terms of different types of literacy exposure and literacy needs is a helpful exercise for determining what and how to teach reading to different students. The categories Huntley has suggested should be useful for teachers.

Implications for Practice

Teachers should be aware of the different types of non-literate students in their classes and needs of those students. Teachers and tutors can customize the literacy instruction for students with different literacy needs.

Key Words

Pre-literate learners
Instructional context

Areas for Further Research

Researchers could investigate different pre-literacy techniques to determine the effectiveness for pre-literate students. Very little research has been done on how to help students who read in non-Roman alphabets or those who read ideograms learn to read with the Roman alphabet.

back to table of contents

Joe, A. (1998). What effects do text-based tasks promoting generation have on incidental vocabulary acquisition? Applied Linguistics, 19 (3), 357-377.

Experimental study of

The relationship between reading and recall tasks and incidental vocabulary acquisition.The researcher wanted to see if reader engagement in reading and recall would help the learners acquire vocabulary without specifically focusing on vocabulary learning. The researcher also wanted to determine whether instructing learners on reading strategies to help them integrate new and known information would increase the effectiveness of reading tasks for incidental vocabulary learning.

Learner participants

There were a total of 48 adult English learners who were studying in two different non-credit courses in Wellington, New Zealand. Forty-one were studying in a 3-month intensive English proficiency course; the remaining 7 were participating in an ESL program in a technical institute. Fifteen were males, 36 were females and their ages ranged from 19-46 years. Approximately three-quarters of the subjects were Asian; the remaining fourth were mostly Samoan. They were in low intermediate to advanced-level English classes. At the beginning of the study, their scores on a generalized test of the first three thousand most common words in English, their scores ranged from 23 to 95 out of a possible 102. Only four students scored over the minimum level (86) that indicated mastery.

Study Design

The students all took a pre-test at the beginning of the study. The pre-test consisted of an interview in which the students reported their knowledge of words associated with the topic of “pain.” In the interview, the students’ knowledge of several words from the first and second thousand general word lists was estimated. The students also completed a reading and recall task, and the researcher investigated the subjects’ language proficiency during the interview.

After the pre-tests, the learners in the study were randomly assigned to one of three treatment groups (There were 16 students per group. The groups had an even distribution of students who had scored in the high and low vocabulary range on the pre-test). One group received explicit instruction on integrating new and known information, and then read a text on pain control and retold it without referring to the original text.

The second group did not receive strategy instruction. They read the same text as the first group and retold it. They were allowed to refer to the text while retelling the story.

The final group did not receive reading instruction and did not do the reading and retelling activity.

At the end of the study, the participants completed post-tests consisting of individual interviews and two multiple choice tests that targeted “pain” related words from the reading.

Findings

Both the experimental and comparison groups performed significantly better than the control group and did not differ significantly from each other. This indicates that participation in read and retell tasks stimulated learners to process written information in a way that facilitates vocabulary acquisition.

Students whose pre-test measures indicated that they had high background knowledge on the topic gained substantially more vocabulary than those who had lower levels of background knowledge.

Comments on the Study

A possible reason why there are no significant effects for the explicit strategy instruction is that the design of the study is confounded. The treatment for the two groups should have been the same, except for the variable of strategy instruction. By adding the extra variable of allowing the second group access to the story during the retelling and not allowing that to the first group, the researcher obscured the issue of the effect of strategy instruction. However, the study does show that engaging learners in reading and retelling tasks does help them increase their vocabulary. Also, as the Carrell,1991 study in this review suggested, learners’ abilities to profit from certain kinds of instruction may be dependent on their proficiency levels. By mixing high and low English proficient students, these possible effects were lost.

Implications for Practice

Literacy tasks that promote learner engagement, such as reading and retelling tasks, can promote incidental vocabulary learning. Additionally, the learner’s prior knowledge about a topic is a predicting factor in learner vocabulary acquisition. If incidental vocabulary acquisition is the goal, pedagogic tasks should be structured around readings for which the learners have a high degree of background knowledge.

Key Words

Schema
Vocabulary

Areas for Further Research

The study could be replicated separating the effects of explicit language instruction and access to the text during the recall on incidental vocabulary acquisition.

back to table of contents

Johnson, K.E. (1992). The relationship between teachers’ beliefs and practices during literacy instruction for non-native speakers of English. Journal of Reading Behavior, 14, 83-108.

Experimental study of

ESL literacy teachers’ beliefs about teaching literacy to adult ESL learners and how these beliefs shaped their instructional styles. The researcher argues that teacher beliefs may strongly influence their decision-making processes. This article seeks to clarify the relationship between literacy beliefs and literacy teaching and thereby emphasize the importance of teacher development.

Learner Participants

The participants in this study were 30 ESL teachers from urban and suburban areas of the northeastern United States. Some were associated with K-12 education in the school districts; 8 taught in continuing education programs for adults or in programs for refugees.

Study Design

The researcher first defined three separate theories of language learning common in the second language teaching literature. These were the empiricist explanation (language learning through habit formation, in this study skills-based), the rationalist explanation (language learning through rules), and the communicative explanation (language learning through meaningful communication, in this study function-based).

In order to understand the teachers’ beliefs about ESL teaching, the teachers first recorded an explanation of what they considered the ideal ESL instructional setting. Their characterizations of an ideal teaching situation were coded as skill-based (dialogs and drills), rule-based (explicit language rules) or function-based (realistic materials and interaction). They were then given three separate lesson plans—one skills-based, one rule-based, and one function-based—designed to teach the same grammatical principle. They were asked to select the lesson plan that best reflected their own instructional beliefs or style. Finally, they were given a beliefs inventory, a set of fifteen statements about language teaching. The inventory contained five statements reflecting each of the three theories of language learning and teaching. The teachers were asked to select the five statements that best reflected their own beliefs.

Three teachers (one from each of the theoretical orientations toward teaching) were selected to be observed teaching in a literacy setting. The transcripts of these observations were divided into instructional episodes that were coded as skills-based, rule-based, or function-based.

Findings

Teachers fell into one of three categories in terms of their beliefs: those with a dominant theoretical view, those with a dual theoretical view, and those with mixed theoretical views. Almost sixty percent of the participants had a dominant theoretical view; in all, almost fifty percent had a dominant belief in function-based instruction. Twenty percent of the participants had a dual theoretical view that incorporated a function-based belief system with another of the methodologies.

The teachers that were observed overwhelmingly followed their belief systems in their instructional style. Almost seventy percent of the instructional episodes in the skills-based teachers’ class were skills based. For the teacher that professed a belief in rule-based instruction, over fifty percent of the instructional episodes in his class were rule-based. The teacher with a function-based theoretical orientation used function-based instruction in over ninety percent of the instructional episodes identified.

Comments on the Study

The study impressively establishes a link between teacher beliefs and teacher practices in the ESL literacy classroom. As the study focused on instruction in ESL literacy classrooms, the relationship between belief and practice might have been more clearly defined had the belief measures also focused on ESL literacy instruction beliefs, rather than on general ESL instruction beliefs.

Implications for Practice

Teachers’ beliefs seem to strongly impact their practices. Teachers should examine their own underlying beliefs in order to better understand their own literacy teaching practices.

Key Words

Instructional contexts
Teacher education

Areas for Further Research

A more in-depth qualitative analysis of both teacher beliefs about literacy and instructional practices in literacy classrooms could be done. This might help illuminate the ways in which beliefs influence decision-making. Teachers could also be asked to reflect on the sources of their teaching beliefs, so that the process of forming beliefs could be better understood.

back to table of contents

Jones, M.L. (1996). Phonics in ESL literacy instruction: Functional or not? Philadelphia, PA: World Conference on Literacy. (ERIC Document Reproduction Service No. ED 436 104)

Practitioner research study of

Learner Participants

The learners in the study were adult ESL students who participated in an open-entry community education class that met twice weekly over 18 months. Because of the nature of the class, many learners started and stopped attending the class during this time. Qualitative data was collected as the study progressed from the students participating at any given time. Most of the students were Spanish speakers from Mexico. Many had limited educational backgrounds and were semi-literate in Spanish.

Teacher Participants

The teacher of the course was a trained ESL teacher with a background in phonics instruction. The teacher believed that introducing adult ESL students to phonics would more effectively help them to increase their English reading and writing skills. This teacher also created the materials used to present phonics to the students.

Study Design

The researcher began the study by presenting the students with some basic phonics rules, similar to those presented to young children learning to read English. The teacher referred to these rules throughout the course, drawing student attention to the spelling of the words used in content activities. The class also participated in activities designed to draw learner attention to sound-symbol correspondences, such as learning various forms of verbs or reading and writing short paragraphs. Samples of student essays were collected, along with the teachers’ impressions of the class.

Findings

Based on an examination of the essays, the researcher concluded that the learners gained a better command of English spelling and were more willing to write in English when instructed in phonics. She observed language development in their writing as well.

Comments on the Study

Many researchers stress the importance of interactive models of reading, in which top-down predicting skills are integrated with bottom-up decoding skills. However, there is little discussion of how this is to be achieved, especially with non-literate or semi-literate students who cannot transfer their L1 decoding skills to English reading. Research into decoding instruction is valuable for the insights it gives teachers about pedagogical techniques for helping students develop literacy.

While the researcher suggests that phonics instruction improves reading ability, no direct evidence was collected on students’ reading proficiency. The effects of phonics instruction on reading have to be inferred from students’ writing ability. Also, since there was no comparison control group, it is not possible to say that the improvements in student spelling were prompted by the phonics instruction, and not by some other instructional technique.

Implications for Practice

Students with limited literacy exposure in their native language or in the Roman alphabet could most likely benefit from some phonics exposure. It is important to remember that, while bottom-up decoding skills are not the entirety of reading, they are still essential and should not be neglected in instruction.

Key Words

Decoding
Bottom-up processing

Areas for Further Research

Researchers could investigate the effectiveness of different techniques for teaching students decoding to determine whether the phonics approach or another is most useful for adult ESL literacy instruction. Also, pre- and post-testing for reading speed and comprehension could be used to quantify the effects of phonics instruction.

back to table of contents