|
|
|
|
|
|
|
|
VALERIAN Postovsky and Ellen Bialystok are two of this century's most important empirical researchers in the field of second language learning. This paper focuses on Postovsky's contributions to our understanding of how best to initiate second language learning and Bialystok's contributions to our knowledge of effective teaching strategies and syllabi.
Valerian Postovsky's reputation is based on the 1974 article “Effects of Delay in Oral Practice at the Beginning of Second Language Learning,” which posits the thesis that “recognition knowledge is prerequisite for the development of retrieval knowledge” (231). In examining the effects of postponing oral drill and practice in his beginning-level Russian class at the Defense Language Institute, Postovsky empirically demonstrates that a 120-class-hour delay in oral production produces students whose ability to speak is superior to that of a control group for which oral production was mandated from the outset. Yet one would never arrive at such a conclusion if one's knowledge of Postovsky were limited to what is written about him today. Either dismissed out of hand or distorted almost beyond recognition, Postovsky has suffered a posthumous fate that illustrates the notion that De mortuis nihil dicamus or, with equal appropriateness, De mortuis nihil dicamus nisi ad proprium beneficium .
One good example of how Postovsky's findings have been misconstrued is the way they are reported by Wilga Rivers in the second edition of her magisterial Teaching Foreign Language Skills . For Rivers, an “area of controversy” is whether the learning of a language should begin with “massed listening practice” (176–81). Starting off with an error of fact—oral production for Postovsky's experimental group was delayed 120 hours of instruction, not 240, as Rivers states (177)—she proceeds to lay undue emphasis on the experimental groups' superior performance in listening comprehension and reading (“Since listening and reading involve similar processes, one would expect the development of listening strategies through intensive practice to carry over to reading … .” [178]), while minimizing their equally superior performance in writing and especially speaking. Wrongly claiming that “no significant results for a higher level of speaking skill have been reported for the experimental groups” (178), Rivers seeks to explain her interpretation by stating that the linguistic knowledge required for comprehension “may be quite different from that required for expressing our meaning in speech and writing” (179). While it can be true, as she states, that in listening for general meaning in what she terms a situation context, “expectations and inference help us create a meaning from semantic clues, so that it is often unnecessary to recognize or comprehend” the morphosyntactic fine points (179), Rivers does not acknowledge that the aural comprehension which took place in Postovsky's experimental classes was not “listening for overall content” but listening for discrete points of morphosyntaxis—not just “intensive listening” but “grammar-drill listening” for the immediate purpose of responding, in writing, to the pattern drills of an overtly audiolingual curriculum. Thus the focus of Postovsky's experimental classes was in no way at variance with what Rivers implicitly allows as a curriculum that can eventually lead to oral production with grammatical accuracy.
To gainsay Rivers's claim that Postovsky's experiment failed to demonstrate superior levels of second language oral production, as well as to set the record straight on both the significance and the nature of Postovsky's work, let us briefly review his experiment.
Postovsky repeated this experiment twice, from September to November 1969 and from January to April 1970, with two different cohorts containing sixty-two and sixty matched pairs of students. The experimental halves of each pair responded in writing, not orally, to drill material for the first four weeks of the twelve-week 360-hour introductory Russian course. (Classes and labs met for 6 hours a day, 5 days a week, or 30 hours a week.) Both experimentals and controls were exposed to the same instructors and the same curriculum, the only difference being the mode of response. At the beginning of the fifth week, experimentals were merged with controls into the traditional oral-response program that the control group had been following from the onset of instruction.
During the experimentals' four-week delay in oral practice, feedback, as I have noted, was elicited through written response. Postovsky says the development of writing skills was considered important for two reasons: first, “it provided students with a meaningful mode of response”—though to the term “meaningful” I would add “and easily quantifiable”—and, second, “it was believed that there is a high degree of positive transfer from writing to speaking, since both skills are productive” (233). The second point is one that Rivers does not think through to conclusion. (A certain amount of pronunciation practice along with drilling in the Cyrillic alphabet was given, but only during the first three days of instruction.) By “written-mode response” Postovsky means dictation in place of oral imitation drill; written, not oral, response to pattern-drill stimuli; and written reproduction from memory of the dialogues otherwise recited orally. In short, this was the standard audiolingual curriculum of the 1960s, though with an important twist. Postovsky stresses that in both groups, equal emphasis was given to the development of aural comprehension.
The nature of the tests administered, the conditions of testing, the measures taken to ensure validity and reliability, the statistical packages employed for analysis, and other similar details are found on pages 234–36 of the Postovsky article Reproduced in the appendix from these pages are the mean scores obtained on the six-week test (Postovsky's table 2) and on the twelve-week test (Postovsky's table 3). (Both comprehensive tests were “similar in format to the MLA-Cooperative Foreign Language Tests” [Postovsky 234] and were designed to provide separate measures of listening, speaking, reading, and writing skills.)
From the statistics, two conclusions are clear: intergroup differences favored the experimentals on all criterion measurers, but “the statistically significant difference at p <.01 level was observed only at the six-week level of achievement” (236). Postovsky (along, in turn, with Rivers) is not surprised that the experimentals were significantly superior to the controls in reading and writing skills, since “the E s had more written practice and they simply learned what they were taught.” Yet this conclusion cannot be applied to the observed difference in speaking, since the experiment's data “actually showed negative correlation between the amount of oral practice and quality of performance on the test” (237). For this surprising outcome, Postovsky provides a quadripartite explanation: (1) experimentals heard only authentic Russian speech and therefore possibly stored more precise acoustic data in their memory; (2) a written response “requires longer storage of auditory input in short-term memory” (237), thus promoting increased learner focus on auditory surface and also giving learners more time in which to process the message; (3) subvocalization during writing practice may play a significant role in the assimilation of linguistic structure; and (4) writing practice provides visual reinforcement, something that oral practice alone does not. It is of special interest—all the more so in the light of Rivers's comments and the bitter debate that has been raging since 1982 about the best way to achieve accuracy (as opposed to fluency) in second language production (for opposing points of view, see, e.g., Higgs and Clifford; Terrell)—that Postovsky's subsequent analysis of the seven components of the speaking score on the six-week test reveals that the two variables contributing most to the experimentals' superior and statistically significant outcome were control of grammar and reading aloud. (Experimentals also scored better in the remaining five subcategories: mimicry, answering questions, free narration, control of vocabulary, and fluency.)
Critics of Postovsky's experiment are quick to note that the twelve-week test results, though showing a continued advantage for the experimentals, were “at less than statistical level of significance,” to quote Postovsky himself (236). His somewhat lame attempts at explaining the minor decline largely center on putative effects of latency and the Cyrillic alphabet. A much more plausible explanation probably escaped Postovsky because he was modest; his article is full of self-effacing disclaimers to the effect that without additional and more detailed research specific conclusions about the second language acquisition process “would be pure speculation” (237). I myself would attribute the experimentals' twelfth-week slow-down to strictly curricular causes: at week six (the first test), experimentals were only two weeks removed from the demonstrably beneficial effects of oral production-delayed methodology, thus minimally handicapped as yet by the need to produce oral responses (“output”) before the aural cues (“input”) had had time to become intake and, processed by the brain's language acquisition devices, be made ready for ultimate and automatic production when the right moment arrived for them to do so. (My terminological debt to Stephen Krashen is clearly acknowledged here.) By way of providing a theoretical basis for his experiment, Postovsky had remarked earlier in his article, in terms that are now familiar, that
decoding capability requires development of recognition knowledge, while encoding capability requires development of retrieval knowledge. Given this difference between the two events, it would appear to be logical to assume that in the natural learning process, development of recognition knowledge would precede, not follow, the development of retrieval knowledge. (230)
Since after the fourth week of his experiment the experimental group ceased to exist as a separate entity (having merged by then with the control group), by the twelfth week the benefits of oral production delay, that is, retrieval whose development followed recognition, had begun to wear off. This, and not the Cyrillic alphabet or whatever, makes more sense to me as an explanation of why the superior development of the Defense Language Institute's students had slacked off some but was not yet negated.
These of course are speculations. Further empirical work is clearly called for. May I take advantage of the present setting to challenge the current faculty of the Defense Language Institute to redo Postovsky's experiment, though with the following change: rather than stand as a separate entity through only four of twelve weeks, experimentals would remain separated until the very end of the course. Test results would then reflect the benefits of oral production delay only, not (as in Postovsky's experiment) delayed oral production quickly followed by mandatory oral production. Alongside this new experiment I would suggest that a second variation on the Postovskian theme be attempted, one that involved a twelve-week fifty-fifty mix of written and oral response to all drilling material. From the outset, all topics would be presented, examined, and practiced silently (written mode) and then orally (spoken mode).
My belief that the second of these two proposed experiments would prove highly supportive of Postovsky's principles has been prompted by the felicitous results of testing undertaken several years ago by David E. Wolfe and Gwendolyn Jones at a Catholic high school in suburban Philadelphia. Wolfe and Jones report on an experiment conducted during the second semester of the school's level-one Spanish class for a thirteen-week period. Under investigation was James Asher's total physical response (TPR) method, whose efficacy in promoting superior degrees of receptive skills (aural comprehension and reading) has been demonstrated repeatedly over the past twenty years. (Yet to be shown—though for a glimmer of hope on the horizon see Asher, who reports on a three-year research project in the Mountain View, California, Wisman School District [16])—is the extent to which TPR produces students whose productive skills, writing and particularly speaking, are significantly superior to those produced by traditional methods.) Asher and others reported on experiments in which orthodox TPR classes were taught exclusively through the medium of the famous command forms, whereas Wolfe and Jones report on a class that combined TPR with “familiar teaching methods” requiring oral response (275). In Wolfe and Jones's experimental section, the first twenty minutes of the forty-minute class period employed the TPR strategy, presenting the chapter's vocabulary and grammar exclusively through commands. “Only when the students responded physically to the command without hesitation were they allowed to repeat,” orally, the new vocabulary and grammar contained in the commands (275). The remaining twenty minutes of the class were conducted via the oral responses of the neo-audiolingual-cum-cognitive-code method(s) as enshrined in a widely used high school textbook. The control section was taught from this same textbook in the traditional fashion: oral-response dialogue practice, oral drilling, oral answers to questions, and so on.
Results from the textbook publisher's “unit tests” administered at the end of every thirteen hours of instruction would prove quite satisfying to Postovsky, Asher, Krashen, Terrell, and all the others who are convinced that a delay in speech production promotes superior progress in a second language Wolfe and Jones report:
The experimental group significantly outscores the control group on all three unit tests. On test 1 the difference in mean scores is 5.78; on test 2 the difference is 7.38; and on test 3 there is a jump to 11.73 between the two groups. As the experimental group had more exposure to the TPR strategy, their scores on the unit tests increased, whereas the control group tended to maintain the same average throughout the study. (276)
The gap, then, widened. Of equal interest is the fate of the standard deviations: the control group's grew progressively more heterogeneous than the experimental group's, which became smaller on each subsequent exam. Thus the difference in mean scores cannot be explained by the stellar performance of a few outstanding students in the experimental class. Similar salutary changes in attitude were noted for the experimentals but not the controls. I quote part of Wolfe and Jones's final paragraph for its value as a signpost pointing toward the future; it shows that the authors, though pleased with their findings, are aware—as was Postovsky—that work with oral production-delayed second language methods represents a very promising beginning, not the last word on the topic:
Much more experimentation needs to be undertaken as to how much instruction can be conducted through an implicit approach such as TPR and when to begin instruction using explicit methods. In addition, studies are needed to determine how much instructional time needs to be given to TPR in order to achieve optimal results after an initial period of total TPR instruction …. (279)
Yet the impetus to undertake this experimentation will surely be lost unless a successful attempt is made to answer arguments such as Wilga Rivers's that minimalize or largely dismiss the potential of oral production-delayed instructional techniques. How are we to reconcile Rivers's claims—that accuracy and indeed any comprehensible level of ability in oral production cannot derive from methodologies that delay oral production—with the facts that issue forth from studies such as Postovsky's and Wolfe and Jones's? The answer, I feel, lies in a close reading of Rivers's critique of what she views as the theories underlying oral production-delayed methods.
Rivers argues, we recall, that listening and speaking “operate quite differently and may even involve quite different grammars” (179, citing Bever 66). In consequence, “the knowledge of the language we require for comprehension of oral output may be quite different from that required for expressing our meaning in speech and writing” (179). Citing I. M. Schlesinger, she writes, “In production, on the other hand, the speaker or writer expresses the intention (or idea) through the operation of the syntactic system, fine details of which may be necessary to give structure to the full semantic intention of the speaker” (267).
What Rivers overlooks initially is the existence of the two types of listening noted above in my discussion of Postovsky: listening for general meaning versus listening for morphosyntactic or other specific detail. A subsequent misstep is the false dichotomy Rivers establishes between the two types of listening, along with her view that the one necessarily precludes the other. Thus for Rivers, “close attention to surface structure features in listening interferes with the process of developing meaning. It over-loads the temporary memory store, thus hindering rehearsal of essential elements of meaning which are being retained in order to relate them to material still to come” (180). I see no reason why listening for morphosyntactic detail (“close listening”) cannot complement listening for overall content (“general listening”), even of the same passage. A passage can first be presented for developing overall content comprehension; and, once comprehension has been demonstrated (whether through physical response, picture identification, administration of a written test, or any combination thereof), the passage can be broken down into more easily managed units brief enough not to overload Rivers's “temporary memory store” but heavily seeded enough with morphosyntactic or lexical detail that the units can serve as a basis for the close attention that Rivers and others are convinced is the high road to ultimate accuracy. Krashen, Terrell, and other advocates of minimal-monitor approaches to second language acquisition would naturally disagree with such an emphasis on the need for “close attention” and the fine-toothed grammatical analysis typically attendant thereon. Thus Terrell comments, in a discussion of the proficiency movement and communicative competence:
Teachers of Spanish often seem to be so wed to the idea of accuracy before fluency that their students…spend most of their time doing exercises rather than engaging in real communication. The idea that communicative experiences in which accuracy is a secondary consideration lead inevitably to “fossilization” is not yet supported by any published evidence and is refuted by the countless numbers of adults who have acquired languages naturally, all of whom passed through stages with large numbers of grammatical errors, but who have reached high levels of fluency. The arguments to the contrary in Higgs and Clifford 1982 and in Omaggio 1983 miss the essential points entirely. (195)
What can be concluded from the above discussion is quite straightforward and highly promising, though necessarily limited. First, since we know from Postovsky that long-term delay in oral production when accompanied by written responses to traditional audiolingual drills produces students whose four second language skills, especially the speaking skill, are superior to those of students trained from the outset via oral production, then students following audiolingual curricula in intensive settings should respond in writing to the drills and exercises for an as yet not finally fixed number of class hours. Second, since according to Wolfe and Jones even a twenty-minute daily delay in oral production will be of considerable benefit to students otherwise following an essentially audiolingual curriculum, this variation on the theme of speech-production delay is also highly recommended. In fact, if we are to be practical about it, the Wolfe-Jones version should prove considerably more marketable to a suspicious profession and a public long conditioned to view immediate oral production as the sine qua non of second language instruction from day one onward.
We have thus seen that a speech production-delayed approach is successful in the initial stages of the second language curriculum and should be implemented during the first year or years of instruction in our schools. Yet as we know from information obtained from the chart “Expected Levels of Absolute Speaking Proficiency in Languages Taught at the Foreign Service Institute (revised April 1973),” as quoted in Schulz (192), commonly taught languages such as Spanish, French, and Italian require 480 hours of instruction for the student of average aptitude to achieve a 2 rating on the Interagency Linguistic Roundtable proficiency-measurement rating instrument, with German and Russian requiring 720 hours to produce a 2. Thus attention must clearly be paid to stages other than the initial ones if any progress is to be made in curricular design overall.
Here I find it profitable to turn to the work of Ellen Bialystok, specifically her 1979 contribution “The Role of Conscious Strategies in Second Language Proficiency” (all page references are to the reprint in the Modern Language Journal ). Viewed parochially, Bialystok should not be used by designers of curricula at all, since the author makes it clear that she is investigating neither the “characteristics of the learner” nor the characteristics of the learning situation but, rather, “what second language learners can do to facilitate mastery of the target language irrespective of these personal characteristics and learning circumstances” (24). Nonetheless, the conclusions she reaches concerning strategies good learners use in their mastery of a second language deserve to have curricular repercussions because the strategies she investigates closely resemble or perfectly reflect some of the methodological approaches most commonly employed today.
Bialystok identified three main strategies—practicing (subdivided into formal and functional), monitoring, and inferencing—and then measured the extent of their use by Toronto high school students who had been studying French since approximately grade six. The total sample consisted of 157 students—82 in the tenth grade and 75 in the twelfth. In her research, a distinction is made between formal and functional language, the formal roughly equivalent to Saussure's langue and the functional “concerned with the use of language in communicative situations” (24). Modality, oral or written, added another dimension to the study. Formal practice is defined as “the specific exercise of the language code for the sake of mastering the rule system,” while functional practice “occurs when the language learner increases his opportunity to use the language for communication such as going to movies, reading books, or talking to native speakers” (25). Bialystok sapiently notes that although exposure to the second language was classed as formal or functional according to a given encounter's predominant characteristic, “a strict dichotomy of this kind is both theoretically and practically impossible” (25–26). The two remaining strategies—monitoring and inferencing—are defined in ways well known to the reader: the former, hypothetically, as Krashen's mechanism for modifying responses according to specific information the speaker has learned, not acquired, and the latter as “the use of available information to derive explicit linguistic hypotheses” (26). Inferencing derives information externally from similarities between a second language and a first one or internally through analogy, and so on. Bialystok hypothesized that the use of these strategies would facilitate learning and performance and that “the effects of the strategies are specific to the type of language for which they are employed” (27). Employing both a questionnaire to provide quantitative information about each learner's use of each strategy and four achievement tests to be analyzed for the extent to which scores changed as a function of strategy use, Bialystok reached the following conclusions:
A “ceiling effect” was found for formal practice, that is, while formal practice “appeared to show a small positive relationship to achievement in Grade Ten,” formal practice showed a significantly negative effect vis-à-vis achievement by grade twelve (31–32). The negative regression coefficient obtained for formal practice in grade twelve must be analyzed carefully, writes Bialystok:
We do not interpret this finding to mean that more formal practice produces lower achievement scores; rather, additional formal practice after a particular point no longer facilitates performance. Thus, low achievers continued to engage in increasingly greater amounts of formal practice…without this additional effort improving performance. …The results of the regression analysis suggest that this time and effort may have been better spent in functional practice, for which no such ceiling effect was found. (33)
Bialystok's findings simultaneously give comfort to the partisans and the opponents of Krashen's monitor theory—a possible reason why Bialystok is seldom mentioned in American writings, so many of which are overtly polemical. For while Bialystok empirically supports the hypothesis that input-prompting meaningful conversational exchanges, which are a significant part of functional practice, do more than any other activity to promote second language achievement, she also concludes that at earlier stages of learning, other strategies (including formal practice) have roles to play as well. The curricular implications are clear. A second language curriculum that omitted formal practice at some point in its early stages might well run the risk of shortchanging the public, whereas a curriculum that placed continuing emphasis on formal topics such as grammatical theory and drilling would be wasting precious time better spent in functional practice.
My need to qualify part of this last sentence with modals and adverbs (“might well run the risk”) can only indicate the not entirely unsorry current state of affairs in research on second language teaching. While progress in achieving synthetic truth has been made—and the two excellent pieces of empirical research by Postovsky and Bialystok are prime examples of such progress, as is the less massively empirical but equally synthetic research by Wolfe and Jones—far too much of what passes for linguopedagogical science today has sought and found analytical truths, that is, conclusions backed up by logic, argumentation, analogy, and anecdote. Few case studies have been produced, and few if any attempts at empirical research. In her discussion of process research in second language classrooms, Rosamond Mitchell speaks of research on teaching in general when she writes, “the relationship between instruction and the learner's classroom experiences on the one hand, and the L2 learning process itself on the other, remains largely obscure” (346). To coin a phrase, more research is clearly needed—empirical research such as Postovsky's, Bialystok's, and Wolfe and Jones's. If, for example, half the time and effort recently expended on fighting out the battle of The Error: To Extirpate or to Ignore had been directed toward following two groups, control (errors extirpated) and experimental (errors ignored), through a full 480-hour intensive course in a setting such as the Defense Language Institute or the Foreign Service Institute, the matter might well be on its way to resolution, and other issues could be tackled. In similar fashion we are overdue for a full-dress, real-time study of one implicit approach or more to second language teaching (again, a full 480-hour intensive course appears to be the most propitious and manageable of settings). Will fossilization set in, never to be repaired? Will errors end up correcting themselves if attention is not drawn to them and adequate amounts of input are not taken in? Empirical research, once performed, will tell.
My comments end as promised with a trip—albeit quick—down methodology lane, in the light of what we have learned from three seminal contributions to research on second language teaching. Does your current methodology give short shrift—as do audiolingualism, cognitive code, and functional-notionalism—to aural comprehension as a prerequisite to oral production? If so, changes are in order, as Postovsky and Wolfe and Jones have shown. Do you interpret the emerging “proficiency-based curriculum” as a short-cut course in which phrase-book responses to phrase-book situations will provide both the means and the impetus toward communicative competence? Bear in mind that doing so runs the risk of giving birth to precisely the sort of formal practice that Bialystok proves is subject to an early-ceiling effect. Have you offered several hundred hours of strictly implicit second language instruction to students whose chances of engaging in meaningful interaction with native speakers of the second language are minimal to nonexistent? Recall the risks of error fossilization that even the most fervent advocates of natural methods would concede as likely, given such circumstances.
The ideal curriculum is well orchestrated. Not all instruments play at once, or with the same intensity or significance of role. As the symphony of second language learning begins, the leitmotivs are presented clearly by those instruments best suited for melodic line, while the rest of the orchestra “listens” or contributes to the background only. Thematic development intensifies in complexity as the symphony progresses, ultimately (and here my musical analogy suffers a fatal breakdown) toward the “artistic freedom” of functional performance (outside the orchestra hall?) in conversation with real native speakers for intrinsic reasons both public and private.
The author is Associate Professor of Spanish in the Department of Languages and Linguistics at the University of Texas, El Paso. This article is based on a paper delivered at the ADFL Seminar West, 26–29 June 1986, in Monterey, California.
Asher, James J. Learning Another Language through Actions: The Complete Teacher's Guidebook . 2nd ed. Los Gatos: Sky Oaks, 1982.
Bever, T. G. “Psychologically Real Grammar Emerges Because of Its Role in Language Acquisition.” Developmental Psycholinguistics: Theory and Application . Ed. Daniel P. Dato. Washington: Georgetown UP, 1975. 63–75.
Bialystok, Ellen. “The Role of Conscious Strategies in Second Language Proficiency.” Canadian Modern Language Review 35 (1979): 372–94. Rpt. in Modern Language Journal 65 (1981): 24–35.
Higgs, Theodore, and Ray Clifford. “The Push toward Communication.” Curriculum, Competence, and the Foreign Language Teacher . Ed. Theodore Higgs. Skokie: National Textbook-ACTFL, 1982. 57–79.
Krashen, Stephen D. Principles and Practice in Second Language Acquisition . Elmsford: Pergamon, 1982.
Mitchell, Rosamond. “State of the Art Article: Process Research in Second Language Classrooms.” Language Teaching 18 (1985): 330–52.
Omaggio, Alice. “Methodology in Transition: The New Focus on Proficiency.” Modern Language Journal 67 (1983): 330–40.
Postovsky, Valerian. “Effects of Delay in Oral Practice at the Beginning of Second Language Learning.” Modern Language Journal 58 (1974): 229–39.
Rivers, Wilga. Teaching Foreign Language Skills . 2nd ed. Chicago: U of Chicago P, 1981.
Schlesinger, I. M. Production and Comprehension of Utterances . Hillsdale: Lawrence Erlbaum, 1977.
Schulz, Renate A. “From Achievement to Proficiency through Classroom Instruction: Some Caveats.” Hispania 69 (1986): 187–93.
Terrell, Tracy D. “Recent Trends in Research and Practice: Teaching Spanish.” Hispania 69 (1986): 193–202.
Wolfe, David E., and Gwendolyn Jones. “Integrating Total Physical Response Strategy in a Level I Spanish Class.” Foreign Language Annals 14 (1982): 273–80.
| Class | Condition | Listening | Speaking | Reading | Writing |
|---|---|---|---|---|---|
| September | Experimental | 28.78 | 48.64 | 35.89 | 87.03 |
| N=28 | Control | 28.50 | 44.92 | 34.32 | 81.07 |
| January | Experimental | 31.33 | 47.15 | 36.37 | 87.37 |
| N=27 | Control | 29.96 | 45.33 | 35.15 | 82.67 |
| Class | Condition | Listening | Speaking | Reading | Writing |
|---|---|---|---|---|---|
| September | Experimental | 24.64 | 48.56 | 31.92 | 81.20 |
| N=25 | Control | 23.20 | 44.84 | 31.68 | 77.44 |
| January | Experimental | 26.66 | 49.92 | 33.08 | 79.92 |
| N=24 | Control | 23.66 | 48.38 | 32.79 | 78.67 |
Reprinted with permission from Valerian A. Postovsky, “Effects of Delay in Oral Practice at the Beginning of Second Language Learning,” Modern Language Journal 58 (1974): 235.
© 1987 by the Association of Departments of Foreign Languages. All Rights Reserved.
|
|---|
|
|
|
