 |
|
||||||||||
|
|
|
||||||||||
|
Interplay Between Reading Tasks, Reader Variables and Unknown Word Processing
Adina Levine and
Thea Reves
Bar-Ilan University, Israel
<revest@mail.cc.biu.ac.il>
<levina@mail.cc.biu.ac.il>
The study was set up to answer the following questions:
The subjects of the study were 42 students of an EAP (English for Academic Purposes) advanced reading comprehension course. The following instruments were adopted for the study: a word-treatment experiment, an open questionnaire, self-observation reports (verbal protocols), text-summary and verbatim translation into L1.
- To what extent are the reader's word-treatment strategies task-dependent?
- To what extent are word-treatment strategies dependent on the reader's reading profile?
In the word-treatment experiment, the subjects were asked to read a text for global and close reading comprehension and to underline words unknown to them by indicating which of the three categories they belonged to:
"Reading Profile" of each subject was composed of information obtained from a Personal Background Questionnaire, which specifically addressed individual reading habits in the L1, in the FL and/or in additional languages; reader's L1; age of beginning to read in L1 and other languages; type of texts read in L1 and in other languages.
- words they did not need for the comprehension of the text;
- words the meaning of which they could guess, so that their comprehension of the text was not impeded;
- words they had to look up in the dictionary in order to be able to go on reading and comprehending the text.
The results of the study confirm that word treatment strategies are undoubtedly dependent on the type of the [-1-] reading task and to some extent also on the reader's reading profile. Readers whose reading habits and problem solving abilities are more developed and whose inferencing skills are more alert, are also more self- confident vis-a-vis the unknown words in the text; they resort to the dictionary mainly to verify their own inferences.
Interest in the relationship between vocabulary and reading comprehension has a long history in the research of L2/FL reading. Observing the performance of FL/L2 readers, confronted with unknown vocabulary, researchers have noted the important role of vocabulary as a predictor of overall reading ability (Nation, 1990; Grabe, 1991). In fact, second/foreign language readers often cite "lack of adequate vocabulary as one of the obstacles to text comprehension" (Levine & Reves, 1990). In certain contexts, a sentence or even an entire paragraph might become incomprehensible because of the occurrence of even a small number of unknown vocabulary items. (Ulijn, 1981; Alderson, 1984; Koda 1989; Barnett, 1988; Coady 1991; Huckin et al., 1993).
Since the development of psycholinguistic models of reading, researchers and teachers alike argue that the best way to handle the unfamiliar words in the text is by drawing inferences from the rest of the text. According to Hosenfeld (1984), readiness to guess from context is what distinguishes a good reader from a bad one.
However, while most researchers find that successful L2 and FL readers can correctly guess the meaning of unknown words while reading (Carrol & Drum, 1982; Carnine, Kameenui & Gayle, 1984; Liu & Nation, 1985), others question the effectiveness of contextual guessing. Thus, Kelly (1990) claims that in anything other than a highly constrained context, guessing on its own rarely helps comprehension. Likewise, according to Bensoussan & Laufer's (1984) observations, many FL readers do not effectively use context to guess word meanings.
As to dictionary use in the process of L2/FL reading, no significant correlations were found between dictionary use and reading comprehension scores; the use of a dictionary to increase comprehension, however, seems to have rational appeal (Aspatore, 1984; Bensoussan et al., 1985).
It should be noted, that in the past researchers emphasized the benefit of dictionary use for text comprehension rather than evaluate the use of dictionary as a word-treatment strategy. This can be explained by the fact that for many years, reading research [-2-] had a static orientation and focused on the product of reading rather than on its process. One of the few studies investigating dictionary use with respect to the L2 reading process is the study conducted by J. Hulstijn (1993). When Hulstijn (1993) examined the look-up behavior of FL readers, he found no significant difference in the number of words looked up by students with varying abilities to guess word-meaning from context. On the basis of the findings, Hulstijn suggested that FL students use dictionaries even when it may not be necessary.
The findings of studies investigating readers' behavior vis-a-vis unknown words are hard to compare because researchers use different texts and tasks to test the same construct. Thus, in order to test some investigators used artificial words or cloze-type blanks instead of real texts with unknown words (Quealy, 1969; Carroll & Drum, 1982). Others used artificially constructed paragraphs with different amounts of contextual support (Carnine, Kameenui & Gayle, 1984). To provide answers to issues related to readers' word-treatment strategies (guessing word-meaning from context, dictionary use, more global type of context use) there is obvious need for authentic texts and authentic tasks to be employed in the processing research of reading.
An additional issue is the effect of readers' exposure to direct training in word-treatment strategies. Strategies are deliberate actions that learners select and control to achieve desired goals and objectives (Winograd & Hare, 1988). Most researchers emphasize the importance of a systematic and principled approach to vocabulary teaching and learning. However, little (or almost no) attention has been paid to the factors related to the reader's choice of word treatment strategies. Neither was guidance offered to FL readers in how to deal with the unknown word in the text.
Thus the present study was set up to examine the factors related to the reader's choice of word treatment strategies and specifically to answer the following questions:
The subjects of the study were 42 students of an EAP (English for Academic Purposes) advanced reading comprehension course. The following instruments were adopted for the study: a word-treatment experiment, an open questionnaire, self-observation reports (verbal protocols), text-summary and verbatim translation into L1.
In the word-treatment experiment students were given two different reading tasks: global reading comprehension and close reading comprehension of academic texts. In order to keep the level of text- difficulty constant, two different parts of the same text were chosen for the two tasks.
Task one: Global text-comprehension
The subjects were asked to read a text for global reading comprehension, to underline words unknown to them and to indicate to which of the three categories they belonged:
The subjects were asked to report on the reason of their choice regarding each word in these three categories. Global text comprehension was evaluated on the basis of a summary written in the subjects' L1.
Task two: close text-comprehension
The subjects were asked to read a shorter passage from the same text, to underline words unknown to them and to indicate to which of the three categories they belonged. The subjects were asked to report on the reason of their choice regarding each word in these three categories. Close text comprehension was evaluated with the help of a verbatim translation of the text into the subjects' L1.
Preliminary practice sessions were provided on how to categorize the unknown words; their purpose was both to raise the subjects' awareness of the importance of vocabulary in the process of reading and to instruct them in the technique of classifying the unknown words.
A "Reading Profile" of each subject was composed of information obtained from a Personal Background Questionnaire. The Questionnaire which specifically addressed individual reading habits in the L1, in [-4-] the FL and/or in additional languages; reader's L1; age of beginning to read L1 and other languages; type of texts they read in L1 and in other languages, etc.
The subjects' performance on each task, i.e., the summary written in L1, and the verbatim translation, were evaluated by two independent raters. Subjects' word-treatment strategies were graded according to the number of words in each group. The subjects' verbal reports, i.e., justification of their word-treatment strategies, were analyzed and evaluated as "rational" and "irrational" i.e., as being right, wise or sensible in terms of the context.
The following statistical analysis was performed on the data collected in the process of the study:
-Pearson Correlations
-T-test analysis
-Manova with repeated measures
-Discriminant Analysis
3.1 Quantitative Analysis
Findings of the study confirm the hypothesis of the relationship between reading task and vocabulary treatment strategies. Significant correlations (Pearson-correlation analysis) were found between criterion variables of the study (i.e., the grades on Global Reading and on Close Reading), and predictor variables (i.e., the strategies applied in the two reading tasks).
The correlations indicate that the relationship of the strategy of "Words Not Needed" with the grade attained in the task of Global Reading was positive, although rather modest. It was, however, negative and high with the grade attained in the task of Close Reading. The strategy of "Words Needed" (i.e., looked up in the dictionary) was negatively related to the grade on Global Reading and positively related to the grade on Close Reading; both correlations were high. The strategy of "Guessing the Unknown Words" correlated equally, rather moderately, with the grade on both tasks. These correlations confirm the assumption that Global Reading requires different vocabulary treatment skills than Close Reading.
In the task of Global Reading, when grouping the unknown words ("Not Needed for text comprehension", "Guessed", "Needed") the subjects consistently preferred to infer the meaning of the unknown words and did not reach for the dictionary. [-5-]
Table 1
Pearson Correlations between Reading Tasks and Reader StrategiesIn the task of Close Reading, on the other hand, the subjects consistently identified more words as "Needed". They did not take the risk of skipping the meaning of the specific word in the attempt to understand the text as a whole but rather resorted to the dictionary.Strategies TASK --------------------------------------------------------- Grade on Global Grade on Close Reading Reading --------------------------------------------------------- WNN .35 -.62 p < .05 p. < .01 --------------------------------------------------------- WG .39 .34 p < .01 p. < .05 --------------------------------------------------------- Wneed -.65 .59 p < .01 p. < .01 --------------------------------------------------------- *In all the tables the following abbreviations are used: WNN: Words not needed WG: Words guessed Wneed: Words needed (dictionary) ---------------------------------------------------------
In both tasks, Global as well as Close Reading, the "word guessing" strategy was employed by the subjects to a similar extent; it was, however, only moderately related to both reading tasks.
The difference between the task of Global Reading and Close Reading is also shown in the results of a "T-test for dependent samples" performed on the two tasks, as presented in Table 2.
Discriminant Analysis was performed on the reading strategies and scores attained in Global Reading as well as in Close Reading. The results highlighted the differences between the strategies applied to Global and Close Reading. The analysis also indicated to what extent the three strategies differentiated between those students who performed well and those who performed poorly in the two reading tasks. [-6-]
Table 2
T-test for Dependent SamplesThe data indicate that there was a statistically significant difference between the two tasks in the application of the strategies of "Words Not Needed" and "Words Needed". No statistically significant difference was found however between the two tasks in the use of "word guessing" strategy.Task Strategy x SD T-value p ---------------------------------------------------------------- Global WNN 2.40 .97 4.04 < .001 Close 1.71 1.20 Global WG 3.60 1.32 -.60 > .05 Close 3.76 1.17 ---------------------------------------------------------------- Global Wneed 3.40 1.91 -10.06 < .001 Close 5.80 1.70 ----------------------------------------------------------------
Table 3/A
Discriminant Analysis between good and poor readers[-7-]---------------------------------------------------------------- Variables Global Close (structure coeff) ---------------------------------------------------------------- WNN ** -.38 ** ** .91 ** ---------------------------------------------------------------- WG ** -.43 ** ** .41 ** ---------------------------------------------------------------- Wneed ** .86 ** ** -.83 ** ---------------------------------------------------------------- X2 28.57 23.58 ---------------------------------------------------------------- sig p < .001 p < .001 ---------------------------------------------------------------- Group centroids ---------------- poor 1.14 -.21 ---------------------------------------------------------------- good -.83 .60 ----------------------------------------------------------------
Significant structure coefficients (over 0.30) are starred (**).
As can be seen in Table 3/a, in the task of Global Reading it was the strategy of "Words Needed" which differentiated most significantly between poor performers and good performers on the task. In Close Reading, on the other hand, the strategy of "Words Not Needed" was the most significant discriminator. This confirms the above presented findings of Pearson-correlations and T-test analysis.
"Word Guessing" did not discriminate very significantly between the two kinds of readers; this corroborates the findings of the Pearson- correlation analysis, which showed that the word-guessing strategy was applied to a similar extent both in Global and Close Reading. The distinction between successful and unsuccessful performers on the strategies within the tests on Global and Close Reading may be indicative of the overall differentiation between "skilled" vs. "unskilled" readers.
Table 3/B
X and SD of variables in the Discriminant AnalysisThus findings show that readers resort to different vocabulary treatment strategies in Global Reading than in Close Reading. It will be remembered that the strategy of skipping unknown words as "not needed for text comprehension" was positively related to the grade on Global Reading and negatively to that on Close Reading.---------------------------------------------------------------- Global Close ---------------------------------------------------------------- poor good poor good (N=19) (N=26) (N=15) (N=30) ---------------------------------------------------------------- WNN 2.00 2.69 .66 2.23 (.94) (.93) (.72) (1.04) ---------------------------------------------------------------- WG 3.00 4.04 3.20 4.03 (.94) (1.39) (1.01) (1.15) ---------------------------------------------------------------- Wneed 4.84 2.35 7.20 5.10 (1.60) (1.35) (1.01) (1.54) ----------------------------------------------------------------
This confirms the generally accepted view that Global Reading requires less attention to the printed word than Close Reading. Close Reading is much more dependent on bottom-up skills, such as the skill of locating the unknown word and decoding it within the text. In Global Reading, readers disregard a considerable number of unknown words. They resort to top-down skills, aiming at a more general, overall comprehension of the text and making inferences for its implications. [-8-]
Quantitative findings of the study did not significantly confirm the direct relationship between reading profile factors and word treatment strategies. Although the readers' choice of word treatment strategies is in most cases irrespective of the reader's reading profile, some profile variables, however, do help us to explain reading behavior. Thus, more unknown words are identified as "not needed" in both Global and Close Reading by readers
Readers with this reading profile may be viewed as "skilled readers". It may be suggested that they draw on a previously developed, general reading aptitude in their attempt to solve vocabulary problems.
Some readers, on the other hand, identify more unknown words as "needed". They prefer to resort to the dictionary in both Global and Close Reading, rather than to apply the strategy of guessing the meaning of the unknown words. They are readers
An interesting relationship was found between readers' L1 and unknown word treatment strategies. Thus, the reader whose L1 is Hebrew identified more words as "needed" i.e., looked up in the dictionary-(Close X=6.04, SD=1.23; Global X=3.67, SD=1.78). The former, i.e., the Hebrew L1 speaker is apparently more text- dependent, while the latter, i.e., the L1 speaker of a different language resorts more often to the strategy of prediction, describing more unknown words as "not needed" for text comprehension. This obviously is the result of the reader's knowledge of other European languages (in our study, French, Dutch and Russian), which assisted the reader's text comprehension in English.
Readers, who according to their Reading Profile usually translate into L1 when reading in another language, needed to look up in the dictionary more unknown words in both Global reading (X=3.54 SD=1.69 vs. X=3.33, SD=2.61) and Close Reading tasks (X=5.88, SD=1.76; vs. X=5.17, SD=1.40). This might be seen as a direct consequence of their habit of translating, where the emphasis on the individual word is very strong.
The subjects of the study were asked not only to make a choice among the three word-treatment strategies, but also to verbalize on the reason of their choice.
Statistical analysis of the subjects' verbal reports shows significant relationships between rationale justification offered for the choice of word treatment strategies on the one hand, and reading tasks and reader profile variables, on the other.
In Global Reading, the use of the strategies "Words Not Needed" and "Guessing" (which were both found to be positively related to Global Reading) were sensibly explained: significant correlations were found between the grade on Global Reading and the score of [-10-] "rationality" of each verbal justification (r=.46 and r=.41). On the other hand the strategy of "Word Needed" was negatively related to Global Reading.
However, no significant correlations were found between the justification of the use of this strategy and the task of Global Reading. These two findings suggest that students cannot explain the reason for resorting to the dictionary in Global Reading.
In Close Reading, the use of the three strategies was sensibly explained; this was evidenced by the rather high relationship found between the grade on Close Reading and the justification of the use of each of the strategies ("Words Not Needed" r=.72; "Words Guessed" r=.44; "Words Needed", r=.43).
Different from correlational analysis, ANOVA analysis did show noticeable differences among readers with different reading profiles in the justification of the use of word-treatment strategies. Thus, more rational explanations were provided by early readers than by late readers (for "Words Not Needed" in Global Reading, X=2.69, Sd=1.20 vs X=2.15, SD=.88), as well as by those who read in more than one language ("Word Guessing" in Close Reading, X=3.84, SD=1.05 vs. X=2.00, SD=1.16). Those who enjoy reading in L1 justified the strategy of "Words not Needed" in Close Reading far more sensibly than those who do not enjoy reading in L1 (X=1.77 SD=.57 vs. X=1.00 Sd=.89). Similarly, those who read more than one type of texts in additional languages offered more rational explanations for the use of the strategy of Words Needed in Global Reading (X=1.917, SD=.28 vs. X=1.29 SD=.46).
It will be remembered that the instructions given to the subjects of the study to divide the words unknown to them into three groups forced them to monitor their word-treatment strategies. Qualitative analysis of the subjects' reports on their word-treatment strategies produced interesting results in each of the three groups of the unknown words.
4.1. Unknown words not needed for text comprehension
In this section the following justifications for skipping the unknown word were the most frequently listed:
4.2. Unknown words, the meaning of which could be deduced
In this section the following explanations of guessing the meaning of unknown words were the most frequently listed:
4.3. Unknown words looked up in the dictionary
In this section the subjects reported that the strategy was used as a last resort, when the unknown word was needed for text comprehension but its meaning could not be guessed.
It should be noted, however, that in quite a few cases wrong guesses and misinterpretations resulted from the attempt to guess the meaning of the unknown word and to avoid using the dictionary. Misinterpretations were evidenced in different types of guessing, be it guessing with the help of the word structure or guessing with the [-12-] help of a broader context. The misinterpretation can be illustrated with the help of the following examples:
The results of the study confirm that the treatment of unknown vocabulary is affected by the type of reading task, or, in other words, that word treatment strategies are task-dependent. Vocabulary treatment strategies are different in Global Reading than in Close Reading. In Close Reading, the skills of locating the unknown word and decoding it within the text is indispensable, while in Global Reading these bottom-up skills are less needed for text comprehension. [-13-]
As to the relationship between word treatment strategies and the reader's reading profile the study showed that only some reading profile variables explain reading behavior. Thus, variables such as, the knowledge of other languages, reading at an early age in both L1 and other languages and exposure to different types of texts, may be interpreted as contributing to the development of a reading aptitude. It may be suggested that this reading aptitude contributes to efficient reading behavior and thus fruitful and successful word treatment strategies.
The study enables us to suggest the need for multiple ways of handling unfamiliar words in the text. It can however be recommended that in the course of instruction, students should be made aware of a variety of effective strategies of the treatment of unknown words in the text. They can use a dictionary, they can try and guess the meaning of the unknown word, or they can choose to disregard the unknown word if they do not feel any breakdown in text comprehension. This kind of training both raises the students' metacognitive awareness and contributes to a more efficient and successful use of strategies.
The present study has focused on the word-treatment strategies employed by university students reading academic texts in English as a Foreign Language. Further investigation might be needed to determine whether the findings of the study could be extended to readers at different levels of language proficiency who read texts of different genre and carry out different reading tasks.
Aspatore, J. V. (1984). But I don't know all the words! Foreign Language Annals 17, 297-299.
Barnett, M. (1988). More than meets the eye: Foreign language reading. Englewood Cliffs, NJ: Prentice-Hall Regents.
Bensoussan, M. & Laufer, B. (1984). Lexical guessing in context in EFL reading comprehension. Journal of Research in Reading 7, 15-32.
Bensoussan M., Sim, D. & Weiss, R. (1984). The effect of dictionary usage on EFL test performance compared with student and teacher attitudes and expectations. Reading in a Foreign Language 2, 262-276. [-14-]
Carnine, D., Kameenui, E.J. & Coyle, G. (1984). Utilization of contextual information in determining the meaning of unfamiliar words. Reading Research Quarterly 19, 188-194.
Carrol, B. & Drum, P. (1982). The effects of context clue type and variations in content on the comprehension of unknown words. New inquiries in reading research and instruction: Thirty-First Yearbook of the National Reading Conference. Rochester, NY.
Coady, J. (1991). Rapid recognition of vocabulary in reading: Bottom-up automaticity ensures top down comprehension. Paper Presented at the Twenty-fifth Annual TESOL Conference, New York.
Grabe, W. (1991). Current developments in second language reading research. TESOL Quarterly 25, 376-406.
Hosenfeld, C. (1984). Case studies of ninth grade readers. In J.C. Alderson & A.H. Urguhart (Eds). Reading in a foreign language. New York: Longman.
Huckins, T., Haynes, M. & Coady, J. (1993). Second language reading and vocabulary learning. Norwood, NJ: Ablex.
Hulstijn, J. (1993). When do foreign language readers look up the meaning of unfamiliar words? The influence of task and learner variables. Modern Language Journal 77, 139-147.
Kelly, P. (1990). Guessing: No substitute for systematic learning of lexis. System 18, 199-207.
Koda, K. (1989). The effects of transferred vocabulary knowledge on the development of L2 language proficiency. Foreign Language Annals 22, 529-540.
Levine, A. & Reves, T. (1990). Does the method of vocabulary presentation make a difference? TESOL Canada Journal 8, 37-51.
Liu, N. & Nation, I.S.P. (1985). Factors affecting guessing vocabulary in context. RELC Journal 16, 3-42.
Nation, I.S.P. (1990). Teaching and learning vocabulary. New York: Newbury House.
Quealy, R. (1969). Senior high school students' use of contextual aids in reading. Reading Research Quarterly 4, 512-533.
Ulijn, J.M. (1981). Conceptual and Syntactic Strategies in Reading a Foreign Language. In A.K. Pugh & J.M. Ulijn (Eds.), Reading for [-15-] professional purposes: Studies and practices in native and foreign languages. London: Heinemann.
Winograd, P. & Hare, V. (1988). Direct instruction of reading comprehension strategies: The nature of teachers' explanations. In C. E. Weinstein, E.T. Goetz & P.A. Alexander (Eds.), Learning and study strategies. San Diego: Academic Press.
Adina Levine holds a Ph.D. in linguistics from the Hebrew University, Jerusalem. She is Director of the EFL Program at Bar-Ilan University and author and coauthor of a number of articles on various aspects of English reading comprehension. She is co-author of two books on reading comprehension for university students published by Collier-Macmillan.
Thea Reves holds a Ph.D. from the Hebrew University, Jerusalem. She is senior Lecturer of English at Bar-Ilan University and Supervisor Emeritus of TEFL, Ministry of Education, Israel. Dr.Reves is the co-author of an oral proficiency test-battery introduced as the national school-leaving exam. She is author and co-author of numerous articles on various aspects of language education.
|
||||||||||
|
© Copyright rests with authors. Please cite TESL-EJ appropriately. Editor's Note: Dashed numbers in square brackets indicate the end of each page in the paginated ASCII version of this article, which is the definitive edition. Please use these page numbers when citing this work. |
[-14-]