Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Attention to imagery in learning and recall of French sentences in the senior secondary school Larkin, Thomas Clelland 1975

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1975_A8 L37_5.pdf [ 5.37MB ]
Metadata
JSON: 831-1.0054624.json
JSON-LD: 831-1.0054624-ld.json
RDF/XML (Pretty): 831-1.0054624-rdf.xml
RDF/JSON: 831-1.0054624-rdf.json
Turtle: 831-1.0054624-turtle.txt
N-Triples: 831-1.0054624-rdf-ntriples.txt
Original Record: 831-1.0054624-source.json
Full Text
831-1.0054624-fulltext.txt
Citation
831-1.0054624.ris

Full Text

ATTENTION TO IMAGERY IN LEARNING AND RECALL OF FRENCH SENTENCES IN THE SENIOR SECONDARY SCHOOL by THOMAS CLELLAND LARKIN B.Ed., Sec, University of British Columbia., 1970 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTERS OF ARTS (Ed. Psych.) in the Department of Educational Psychology We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April, 1975 In presenting th i s thes is in pa r t i a l fu l f i lment of the requirements f< an advanced degree at the Univers i ty of B r i t i s h Columbia, I agree tha the L ibrary sha l l make it f ree ly ava i l ab le for reference and study. I fur ther agree that permission for extensive copying of th i s thes is for scho lar ly purposes may be granted by the Head of my Department or by his representat ives. It is understood that copying or pub l i ca t i on of th is thes i s fo r f i nanc ia l gain sha l l not be allowed without my written permission. Department of The Univers i ty of B r i t i s h Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date t i i ABSTRACT OF THESIS The goal of this study was to ascertain whether attention to imagery, by grade 11 students learning French i s an effective psychological set i n learning French sentences compared to students who*learn "word for word". Thirty French sentences to be learned were presented to the students on an overhead projector. Recall of the sentences was measured by an immediate post test, and by a delay post test twenty-four hours later. Three pilot studies were carried out to verify the f e a s i b i l i t y of adapting previous research findings largely derived from verbal learning studies which had used English language. Psychological components related to learning French as well as learning materials, r e c a l l tests and subject samples were also examined and modified where necessary. Subsequent to these pilot studies different teaching methods were included in the study as an independent variable to investigate a possible inter-action with the learning techniques. One teaching method emphasised the orthography of French language expressions, during the classroom learning process, with reinforcement of the correct student response immediately following. The other method employed an aural-oral learning approach. The subjects i n the experiment were exposed to these two learning methods for six weeks preceding the learning session of the experiment. Experimental subjects consisted of grade 11 students learning French at two B.SJ.C. schools; Semiahraoo Senior Secondary and Princess Margaret Senior Secondary. At each school approximately half of the students i i i were instructed by one method and the other half by the other method. The teacher in each school was practiced i n both methods i n order to accomplish t h i s . During the learning session students i n each of the four groups were given specific instructions on how to learn the thir t y French sentences. These instructions directed students either to learn by imaging the sentence meaning or by learning "word for word". There was also a control group i n which students learned i n their own style. In the post tests, sentence subjects were given on the test sheet and sentence predicates had to be recalled. Four re c a l l measures were applied to the resulting recall data. The most d i f f i c u l t measure, which required exact recall of a French sentence to obtain credit, was found to be an unreliable measure because of i t s extreme stringency. The other three measures supported the theme that attention to imagery of sentence meaning was significantly more effective i n re c a l l than "word for word" learning. Results indicated that this treatment effect was present largely i n the immediate post test. The experimental treatments probably lacked the impact to produce a significant effect i n the long term r e c a l l . There was a significant school by recall measure interaction. That i s , students i n one school demonstrated a f a c i l i t y for recalling certain words and students i n the other school demonstrated an a b i l i t y to recall J different choice of words. The pre-experimental instructional methodology did not significantly effect r e c a l l scores. TABLE OF CONTENTS Abstract . . . . . . v CHAPTER ONE Statement of the Problem . . . . - . 1 CHAPTER TWO Imagery and the Acquiring of Meaning 7 CHAPTER THREE Three Preliminary Pilot Studies .14 CHAPTER FOUR Conclusions Relating to Comparison of Results at Princess Margaret and Semiahmoo Schools . . . 30 CHAPTER FIVE. Hypotheses and Procedure 35 CHAPTER SIX Research Results . . . . . . . 45 CHAPTER SEVEN Conclusion 77 FOOTNOTES .92 REFERENCES ' 93 Appendix A (i) Pilot Study I • .96 Appendix A ( i i ) Pilot Study I 97 Appendix A ( i i i ) Pilot Study I . . . . . . . . . . 98 Appendix A (iv) Pilot Study I 99 Appendix A (v) Pilot Study I 100 Appendix B ( i ) Pilot Study II .101 Appendix B ( i i ) Pilot Study II 102 Appendix B ( i i i ) Pilot Study II . . . . . . . . . .103 Appendix B (iv) Pilot Study I I . .«*• .104 Appendix B (v) Pilot Study II 105 Appendix B (vi) Pilot Study II 106 Appendix B ( v i i ) Pilot Study II 107 \ Appendix C ( i ) . Pilot Study III 108 Appendix C ( i i ) . P i l o t Study III . 109 Appendix C ( i i i ) . Pilot Study III 110 Appendix C ( i v ) . P i l o t Study III I l l Appendix C (v). Pilot Study III 112 Appendix D ( i ) 113 Appendix D ( i i ) . . . 114 Appendix D ( i i i ) 115 Appendix D (iv) Il6 Appendix E (l) . 117 Appendix E i i ) 120 Appendix E ( i i i ) 121 Appendix E (iv) 122 Appendix E (v) 123 Appendix E (vi) 124 Appendix E ( v i i ) 126 Appendix E ( v i i i ) 127 v i . TABLES AND FIGURES Figure 1 Summary of prompt types required to achieve optimal performance as a function of age. Schematic patterns for high SES (upper panel) and low SES (lower panel) populations. . . . 11 Figure 2 Graph Showing the Relative Value of the Treatment Means at the Two Schools "Words Correct" Snores 26 TABLE 1 "Words Correct" Means, (x) 27 Figure 3 Graph Showing the Relative Value of the Treatment Means at the Two Schools "Content Word" Scores . 28 TABLE 2 "Content Word" Means, (X) 29 TABLE 3 Mean Verbal I.Q.s for the Two Method Groups in Both Schools 43 TABLE 4 Summary of Analysis of Variance of Verbal I.Q. Scores 44 TABLE 5 Means of Recall Scores for "Total Sentence Correct" Measure 45 TABLE 6 Summary of Analysis of Variance of "Total Sentence Correct" Recall Scores 46 TABLE 7 Means (X) of "School x Post Test" Interaction . 48 TABLE 8 Bonferroni Confidence Limits for Means Contrasts "Total Sentence Correct" Scores 48 TABLE 9 Means of Recall Scores for "Content Word" Measure 51 TABLE 10 Summary of Analysis of Variance of "Content Word" Recall Scores 52 TABLE 11 Treatment Means (X) for "Content Word" Scores . 53 TABLE 12 Bonferroni Confidence Limits for Means Contrasts "Content Word" Scores 5^ TABLE 13 Means (X) of "Treatment x Post Test" Interaction Cells 55 N v i i . TABLE 14 Bonferroni Confidence Limits for Means Contrasts "Content Word" Scores 56 TABLE 15 Means (X) of School x Treatment Interaction Cells 57 TABLE 16 Bonferroni Confidence Limits for Means Contrasts "Content Word" Scores . . . . . - . . 58 TABLE 17 Means (x) of School x Post Test Interaction Cells 60 TABLE 18 Bonferroni Confidence Limits for Means Contrasts "Content Word" Scores 6 l TABLE 19 Means of Recall Scores for "correct word" Measure 6k TABLE 20 Summary of Analysis of Variance of "Correct Word" Recall Scores 65. TABLE 21 Means (x) of Treatment Scores for "Correct Word" Measure 67 TABLE 22 Bonferroni Confidence Limits for Means Contrasts "Correct Word" Measure 68 TABLE 23 Means (x) of School x Post Test Interaction Cells 69, TABLE 2k Bonferroni Confidence Limits for Means Contrasts "Correct Word" Scores 70 TABLE 25 Means of Recall Scores for "Correct Word Plus Synonyms" Measure 72 TABLE 26 Summary of Analysis of Variance of "Correct Word Plus Synonyms" Recall Scores 73 TABLE 27 Means (x) of Treatment Scores "Correct Word Plus Synonyms" Measure Ik TABLE 28 Bonferroni Confidence Limits for Means Contrasts "Correct Word Plus Synonyms" Measure . . . 75 1. CHAPTER OWE Statement of the Problem In Br i t i s h Columbia, commencing in January 197^ , the Provincial Government started scholarship examinations in French, to be written twice per year, in which the student i s expected not only to compose French discourse correctly, but also recall French expression which i s necessary to enrich i t . The goal of this study which stems from these scholarship require-ments, was to explore and possibly establish the psychological mediating process whose functioning produces "most effective" learning of French expressions. The term "most effective" to be taken as a means of predicting a b i l i t y of the learners to subsequently re c a l l the language that has been learned. Several studies concerning variables related to rec a l l of verbal material have suggested that most effective r e c a l l has been associated with verbal learning which has emphasised the semantic aspect of the language. One study by King and Russell (1966), reported that ...when subjects are instructed to learn connected meaningful material on the basis of many ideas they tend to recall proportionately more words, letters, sentences, etc., than ideas....On the other hand, when instructed to learn on an exact wording, or a word for word basis, subjects r e c a l l proportionately fewer words, letters, sentences, etc., and more ideas. (j>. 478) There was found to be no significant difference between the two groups in re c a l l of ideas. This would indicate that concentrating on meaning i s the key to solving the problem as stated above. 2 However, the author i s aware that "meaning" i s obscured and complex, as evidenced by Paivio's (1971) definition, (p.5l); i t i s "an organismic reaction with affective or motor (including verbal) or iraaginal components or a l l of these at once mediated by the neural-dispositional meaning structures activated by the (stimulus) symbol and the situation i n which i t occurs". Despite this complexity, the essential factor in meaning which has been found to be the most effective mediator of re c a l l of verbal material has been imagery. The use of imagery mnemonics goes as far back into history as around 500 B.C. to Simonides. The mnemonic process involves a series of symbolic transformations, thus, consider the task of committing to memory a speech. During the learning of the speech the main points of i t would be memorised as a series of images. That i s , the verbal symbols are transferred to imagery symbols. The imagery symbols would be retransformed to words when the speech was being delivered. Of course, the many smaller, or form words which united the ideas were more d i f f i c u l t to r e c a l l exactly, (Yates, 1966). This problem has persisted to the present day. Neverthe-less, even i n those early years the image was recognized as an invaluable aid to verbal r e c a l l . The great weight of evidence favouring the use of imagery as an invaluable aid to r e c a l l of verbal material i s derived frorfr more sci e n t i f i c studies of recent years which i n the early stages depended to a large extent upon paired-associate learning techniques. A study by W.A. Wallace, Turner and Perkins, (1957), produced dramatic results favouring imagery for learning paired-associates. Each pair of words had to be imagined linked 3. together in the same mental image. The group that learned them in this way were more successful than subjects who learned the words in separate images. This particular study i s relevant to the learning of French sentences in which subjects and predicates must be linked together in meaning--meaning which the student, is asked to picture in his mind. Further support for this unitization of imagery objects is supplied by Gordon Bower (1969) who, using concrete noun pairs, found a highly significant difference in favour of subjects who were told to learn the pair of words by imagining some interaction between the words. More closely related to the task of recalling expressions or sentences than paired associates studies are those concerned with chunking. Miller (1956) found that immediate r e c a l l of language was constant at about seven units of the language. Thus, i f the units are syllables, seven of them w i l l be recalled. If the syllables are made into word units, seven words are recalled. Words can be made into higher order units, or chunks, such as phrases. Thus, by combining words into sentence form, phrase units can be committed to memory. In this way the chunking process permits the memorising and recall of a greater number of words than can normally be retained in the memory span. Tulving and Patkau (1962) showed that both grammaticalness and meaning aided in the chunking process. However, they did not indicate the nature of the semantic variable involved in chunking. This was done by Yuille and Paivio, (1969), who held the grammatical variable constant and varied noun imagery in the sentences to be learned. The results indicated that more words were recalled as noun imagery increased. h. Some writers have declared that although sentences contain imagery many of the words such as prepositions or definite articles which are d i f f i c u l t to relate to the imagery mnemonic make research in this direction too d i f f i c u l t . However, Bugelaski (1969) postulates that imagery in sentences can be studied provided that one realizes that certain words can only be related to the image when they are not considered alone. Huttenlocher (1968), provided evidence that spatial relations as indicated by sentences, invoked spatial images. Thus entire sentence meaning may be condensed to imagery. Werner and Kaplan (1963), demonstrated the arousal of images to account for the meaning of "time" as indicated in senteces such as "He ran" and "He w i l l run". In another study by Erie (1963), sentences which were recalled after the imagery instructions frequently had ommissions of such words as " i f " and "but", obviously because of their lack of imagery content. Despite this, i f specific instructions to create an imagery set were more effective in the re c a l l of these words in French, than implicit speech mnemonics, then comparative lack of imagery content in these words may be only a p a r t i a l hinderance to learning. Paivio and Begg (1969), made use of 50 concrete and 50 abstract sentences to find out how much time was required to involve imagery and comprehension>of the sentences. Average time for the most d i f f i c u l t sentences—the abstract ones—was 2 seconds for comprehension and seconds for imagery. Students learning French sentences in the present study were given 2 0 seconds. This extended period was necessary to give students time to write the sentences down. 5. Although the research suggested in this thesis derives much of i t s direction from the King and Russell study (1966), in which "word for word" learning is compared with idea learning, the main difference lies in the type of verbal material used. King used connected verbal prose, instead of short sentences. Short sentences are used in the study because i t i s thought that acquiring French idiom and expressions can best be done, and is usually accomplished, when they are independently chosen from French literature and usually they have no connecting theme. Verbal learning generalizations are therefore most relevant for our purpose i f they are drawn from research using sentence units and from paired-associates upon which sentence research is founded. Because the emphasis i s upon how to learn, rather than how to instruct in the French language, an important feature of the methodology used in this study is the specific instructions given to the subject suggesting to him how he is expected to learn the language in the related task. This method of using specific instructions was also used by King and Russell, mentioned above, and although the learners behaviour i s not as observable as one would wish, i f significant effect i s present in the subsequent results, then such a method must be considered a relevant characteristic in the learning process. Following the learning task, a debriefing of the subjects by the experimenter should indicate to him what the subjects did, and whether they adhered closely to the instructions. The subjects involved in the above-mentioned related research have a l l been competent users of the language, and the language used has been English. Although the psychological aspects of language learning may be 6 generalizable to a large extent from one language to another, i t is s t i l l open to question whether the findings derived from these studies are applicable to Grade 11 students who are not competent users of the French language. Because of this doubt preliminary pi l o t studies had to be carried out before a more controlled experiment could be proposed. 7. CHAPTER TWO Imagery and the Acquiring of Meaning The conceptualization of meaning has been a problem to psychologists for a long time. A succinct definition of i t has not yet been made but at least great strides have been taken towards uriderstanding i t . After studying the variety of interpretations and positions held by many psycholo-gists on meaning, Paivio has derived an analysis the description of which should serve as well as a definition. This analysis credits meaning with three cognitive levels: representational, referential, and associative. Representational meaning refers to the avail a b i l i t y of the "imaginal and verbal representations corresponding to particular nonverbal and verbal stimulus units".^ At this stage the i n i t i a l stimulus representation fades rapidly but a coded trace of some kind remains which has not yet been elaborated, or changed. The referential'level refers to the arousal of a word to a nonverbal stimulus or of an image to a verbal stimulus. Lastly, the associative level of meaning i s theoretically defined as the avail a b i l i t y of words or images which are different to those in the referential stage but which are related to them in some way. In this present study which requires r e c a l l of French sentence predicates, with the verbal representation of the subjects as a given stimulus, a l l three levels of meaning w i l l be brought into operation with special emphasis upon the referential level. The central concern focusses upon two different operations which are employed to help commit meaning into the 8. memory storage. These two operations are best understood by considering the positions held by Skinner and Paivio in relation to the acquisition of meaning. In his explanation of meaning Skinner avoids treating i t as an entity but focusses upon the operational process related to i t . "The meaning of a word is...specified by the conditions under which i t occurs and apart from such a functional account nothing further need be said about meaning." But, he adds, that discriminative operant conditioning must play an important role in the acquisition of meaning. Thus, internalized behaviour which imitates a concrete stimulus, or stimulus word, is the conditioning behaviour which represents the meaning process, but the acquisition of such an internalized response is dependent upon or shaped by, the reinforcement to such responses. Thus i f the response i s known by the individual to be a correct one then the acquisition of i t is more assured and thus more readily available for recall. In answer to this analysis of meaning acquisition Paivio points out that, The only special feature of this analysis i s the emphasis on the stimulus conditions to which the sensory and motor reactions become conditioned: the reactions define the representational, referential and associative meaning of the objects and words that e l i c i t them... (but that) the images aroused by verbal clues often appear too complex and creative for them to have been acquired on the basis of actual contiguities in experience.^ The principal difference between the acquiring of word meaning held by these two men is in the emphasis by Skinner i n the need for the reinforcement process, especially i f stimulus r e c a l l i s expected. On the other hand, Paivio holds strongly to the notion of imagery as a frequent, although not 9. always essential, feature of the organismic disposition which represents internalized meaning. This controversal issue is thought to provide an explanation for the confounding results obtained in p i l o t studies I I and III and which subsequently becomes an important cohsidemtioniin the design of the experi-ment following these studies. Another facet of imagery and verbal learning which has to be considered, before the development of experimental treatments, is the concept of elaboration which has been pointed out by Rohwer (1972). He postulates that a single underlying process is responsible for the production of correct responses on noun-pair tasks. When two words which are unrelated are coupled together by the learner, this coupling is achieved by generating an event which unites them. Rohwer has named this intergrating of words by a relationship, elaboration. Unless this process of elaboration i s triggered the pair w i l l not be learned. In a series of studies in which children of different ages and of different socio-economic groups were involved Rohwer distinguished four degrees of prompt for triggering the elaborationj they were, maximal, substantial, moderate and minimal. At the maximal level actual objects representing the words to be learned are presented to the learners and these objects are v i s i b l y connected in some way. The substantial prompt requires a picture or sketch of the objects corresponding to the words to be learned. A moderate prompt is one in which imagery is used by the learner to relate the meaning of the two words and the minimal prompt simply requires that the learner try to remember the two words which he 10. sees i n written form. Schematic patterns of the four prompts-are shown i n Fig. 1. One of the conclusions derived from the results of this study-by Rohwer was that, the age of the learners was a variable which related to their f a c i l i t y to elaborate by means of these prompts and that, rather than stemming from a difference i n underlying process, the source of the phenomenon appears to be that members of one group have acquired a propensity for autonomous activation of the * elaboration process while members of the other group have not . learned this s k i l l . 4 The results of the 15 year old and the 18 year old groups depicted i n Fig. 1 are particularly relevant to this study where the mean age of the grade 11 students i s 16 years 6 months. In the treatments of the three pil o t studies and the f i n a l experiment, elaboration prompts are the same, or similar, to those described by Rohwer. Unless these prompts are developed to an equal degree i n students of this age, then mean scores derived from these treatments might show significant differences because of a lack of control of the elaboration variables. Therefore, to avoid such a poss i b i l i t y a preliminary examination of the treatment prompts to be used was necessary. In one of the treatments, i n pil o t study 1, i n which French sentences had to be learned, attention to imagery, (A.I.) corresponds to the moderate prompt i n elaboration. A second treatment employs attention to orthbgraphic features (A.O.) of the sentences. This treatment does not correspond to any of the four prompts i n Rohwer's studies. Even the minimal prompt described by Rowher i s dependent to some extent upon.word meaning. The subjects involved i n his study were competent users of the language from which the paired-associates were chosen and therefore word meaning would Age 3 1 fc I 71 r 3 I i 12 15 18 Prompt type ^Maximal || Substantial Q Moderate j j Minimal Fig. 1 Swrnmary of prompt types required to achieve optimal per-formance as a function of age. Schematic patterns for high-SES (upper panel) and low-SES (lower panel) populations. (From Rohwer 1 9 7 9 ) 12 almost certainly mediate learning and r e c a l l . Such a f a c i l i t y for perceiving word meaning may be dimished i n the case of the A.O. group i n pilot study 1 because foreign language symbols are used. However, the 5 goal of pilot study 1 i s to investigate the effect of eidetic xmagery and this phenomenon i s usually associated with young children. Therefore, i t may be assumed that the students involved i n pilot study I have long since achieved optimal performance i n the type of prompt.associated with eidetic imagery. In p i l o t studies II and III the process of elaboration i s more obviously present. The two treatments, "word for word" learning (W.W.), and attention to imagery learning both make use of the moderate prompt. The A.I. treatment corresponds very closely to the moderate prompt as described by Rohwer. The prompt used i n the W.W. treatment i s not so easily categorised. Because each item i s i n sentence form which, per se, necessitates a meaningful relationship between words, then word meaning must be a factor i n "word for word" learning. Therefore, i t i s l i k e l y that the imagery component of meaning i s present to some extent i n this treatment as well. This imagery i s probably supplemented with covert speech. In the f i n a l controlled experiment the treatment groups involved i n learning French sentences, that i s , the A.I. and W.W. learners, (the control group may be considered similar to the l a t t e r group), relate to the groups i n Rohwer*s study, which used the substantial and moderate 13. prompts. A combination of substantial prompts and moderate prompts are employed by the A.I. group of students which sketches sentence meaning ^ to assist i n learning the French verbal material. In Rohwer's study a substantial prompt i s one which uses a picture or sketch of the objects which the verbal item represents i s not shown to the learners. Such a sketch i s drawn by the learners themselves. Therefore, i t i s probable that the students w i l l be assisted i n their memorising by their own sketches as well as by the imagery generated i n the process of drawing the sketches. The other group which learns "word for word" has already been discussed i n relation to the elaboration prompt implicit i n such a treatment. On: of the basic assumptions made i n this present study i s that the process of elaboration w i l l function i n the same way as i t has done i n the Rohwer study. This implies that the relative effectiveness of substantial and moderate prompts remains f a i r l y constant for students aged 16 years 6 months. Therefore, any effective difference that may arise between the learning groups^ results may attributed to treatment rather than the a b i l i t y to spontaneously generate elaboration. Ik. CHAPTER THREE Three Preliminary Pilot Studies Anderson (1972), described the verbal learning process as having three basic stages. "Elements of text are f i r s t encoded in terms of perceptual features. Since the relevant perceptual features of text are orthographic, this can be called orthographic encoding. The next level of processing probably involves acoustic features. At this stage, which can be called phonological encoding, strings of words are rendered into implicit (or explicit) speech. Finally, there may be semantic  encoding, that i s , the person may bring to mind meaningful representation based upon the words he sees, or hears himself saying." It has already been pointed out that according to available evidence, the semantic stage holds the major mediating component relating to long term re c a l l . Baddeley, (1966), suggests that the phonographic or acoustical stage is not so important. "In long term memory (longer than 30 seconds) semantic interference is more important than acoustical interference." (The re c a l l test in this present study is completed long after 30 seconds have elapsed). It i s noteworthy that few researchers of recent years have paid any attention to the possibility that the orthographic stage could be related to long term r e c a l l and yet the phenomenon known as eidetic imagery implies that such a relation may exist. In 1920, Jaensch brought to notice the a b i l i t y of people to recall, in great detail, a picture previously seen. Kluver (1930), reports that r e c a l l of detail can also be accomplished with a 15. page of typed prose, and that such an eidetic image could he almost photographic. It was a phenomenon found to he f a i r l y common in children but rare in adults. It may be significant that i t i s largely children who are able to obtain an eidetic image of written language. It i s possibly because, not being practised readers like adults, they are more inclined to perceive the printed word as a symbol which does not immediately convey meaning. This process of reading without meaning is quite frequently encountered by the teacher who instructs the small child to read. Subsequent to apparently successful reading, questioning of the reader by the teacher, on the content of the reading matter, reveals that l i t t l e or no meaning has been imparted by the printed language. Because the subjects in this present study are not competent users of the language there was the possibility of an inhibitory effect being present i n the process transferring the symbol into meaning, just as there may be with the small child learning to read. Consequently, something akin to the child's eidetic image as a result of excessive attention being paid to the orthographic features, might have a greater influence upon retention and re c a l l than was the case with the subjects in Baddeley's study. The goal of this f i r s t study was to ascertain whether the learning of French sentences by grade 11 students is related to the orthographic stage and to compare recall scores related to "attention to the orthographic features of the language" (A.O.) with"attention to the imagery component of the meaning of language" (A.I.). It was hypothesised that the students who adopted the "attention to imagery"';would learn the French sentences more effectively than the students 16. who learned by attention to the orthography, in the two r e c a l l measures described on page 19-Post tests consisted of immediate and delay post tests,the latter taking place twenty-four hours after the immediate test. Fifty-two of the subjects were given specific instructions to concentrate upon the actual printed words of each of 30 sentences. These were the A.O. students. The rest of the sample which consisted of forty-four subjects were to pay attention solely to the task of imagery of the sentence meaning. These were the A.I. subjects. The subject sample was drawn from four grade 11 classes, totalling 9^  students, who attended Burnaby South Senior Secondary School. Two days before the learning task these students were given a l i s t of French words and expressions which were going to be used i n the learning assign-ment, together with the English meaning. The students were told that they should know the meaning of the French language on the l i s t and that they were going to be tested on i t . The order of the words and expressions in the l i s t s were scrambled so that the word sequence practice effect would not be an aid to memorizing the sentences in the experimental task. The purpose of this l i s t was to ensure that the students involved i n the A.I. group would not be handicapped in their efforts to attend to the language meaning. Each of the 30 sentences was constructed so that the syntactic structure was constant. Each sentence contained a subject which consisted of one noun, preceded by the definite a r t i c l e . The predicate consisted of k words. For example, 17. Le prisonnier f a i t semblant' de dormir SUBJECT PREDICATE The predicate contains the expression which was considered by the author to be the kind to enrich French essay-type answers. The author based his choice of expression upon his knowledge of the grade 11 and 12 French courses and upon his knowledge of essay requirements learned when marking scholarship papers in French in Victoria, 1972, 1973-Since imaging the meaning of the 30 sentences was the essence of one of the treatments the choice of the nouns to be inserted in them was considered carefully. In a study by Paivio, Yuille, and Madigan ( 1 9 6 8 ) , a l i s t of 925 nouns were drawn up by Paivio and each noun was rated for concreteness and imagery. A correlation of 0.9^ was established between these two attributes. Because of this high correlation the nouns in the t h i r t y sentences were chosen on the same basis as the nouns in Paivio's l i s t , that i s , "any word that refers to objects, materials, or persons, 7 should receive a high concreteness rating." Paivio, when discussing paired-associate learning and r e c a l l states that the f i r s t word, or stimulus word, may redintegrate the compound ' image from which the response (word) could be retrieved. It follows that the ease of learning a stimulus--response association would depend upon the image arousing value of both members of the pair, but that of the stimulus member w i l l be relatively more important. Of course, in subject-predicate association which was the concern Thanks to Mrs. J. Osborne, Chairman of French Department, Examination Marking Committee, for editing the expressions. 18. of this p i l o t study, i t is doubtful whether the predicate imagery could truly be equivalent to the paired-associate response word, nevertheless the high concrete rating of the subject noun should f a c i l i t a t e predicate r e c a l l . Immediately preceding the presentation of the 30 sentences each individual in the f i r s t two groups which had been selected by random assignment, was given the set of specific instructions on a sheet of paper, t e l l i n g him to attend to the meaning of the sentences. The other group received-the specific instructions directing them to attend to the ortho-graphic features of the sentences. Neither group was aware of the other's specific instruction. Each of these groups was further randomly assigned into two sub-groups; one to serve as the immediate post test sub-group and the other as the delay post test sub-group. Immediately following the t h i r t y sentence presentation the immediate post test group was given the re c a l l test. This consisted of the thirty sentence subjects and the students had to re c a l l the sentence predicates, thus, Le prisonnier The delay post test group was kept occupied with a scrambled letter French word puzzle to solve. Twenty-four hours later the immediate and delay post test groups reversed assignments. The method of assessing the r e c a l l scores was derived from a study by King ( i 9 6 0 ) , who attempted to validate various scoring procedures for the re c a l l of connected verbal material. He factor analyzed various inter-correlations among the scoring procedures to identify the variables of 19. importance i n the accuracy of re c a l l . He concluded, ...there are two factors involved i n determining the accuracy of written recall...these two factors are the number of content words reproduced and the quantity of words reproduced. Accordingly, two measures of r e c a l l were used in this study. After a l l of the test papers had been collected each one was marked by crediting each French word that was recalled correctly. Thus i f a test paper showed fo r t y words correctly spelled then forty was the score attributed to i t . This score was a measure of "Words Correct." When calculating the scores in the "Content Word" count, a l l of the form words, i.e., articles, conjunctions, prepositions, etc., were eliminated from the original thirty sentences and the r e c a l l responses were checked to determine the presence of the remaining content words. To be considered as present on the r e c a l l answer sheet the word had to be exactly as given on the original, except for minor spelling variations. No penalty was given for sequence errors or for the presence of words not in the original story. It i s conceivable that the groups of students who had been told to mentally picture the meaning of the sentences when memorising them, might pay less attention to form words than the A.O. group learning the sentences, hence the need for the content word count. Analyses of variance (Appendix A, Tables h, 5) for both "words correct" and for "content words" scores indicated that there was no significant difference between the treatments in both immediate and delay post tests. For both treatments there was a predictable significant difference between these two post tests. 20. The hypothesis that attention to imagery would prove a more effective treatment than attention to orthographic features of the sentences was not supported. Such a finding was not in accord with past research, mentioned in Chapter One, relating to verbal re c a l l by means of imagery mediation. Consequently, a day after the delay post test had taken place a debriefing session was held with the students who had participated in the experiment. With a l l of the students the response was the same. The A.O. group declared that they had found i t too d i f f i c u l t to pay very much attention to the orthographic features of the sentences without the meaning interfering. Without exception, they quickly abandoned hope of following i the specific instructions directing them to attend to the orthographic element of the language and they commenced learning "the way that they usually do." A common explanation of what they meant by that was d i f f i c u l t to obtain. Apparently they use different methods of memorising, but in a l l of them a knowledge of the meaning of the sentence was an essential factor. No student claimed any degree of f a c i l i t y for securing an eidetic image. On the other hand, every individual in the A.I. group declared that he had adhered to the specific instructions. Subsequent to this f i r s t p i l o t study, a second one was carried out at another school: Semiahmoo Senior Secondary School, White Rock, B.C. There were two grade eleven classes learning French. The t o t a l number of students was 51. The purpose of this study was again to verify that attention to imagery, related to the French sentence meaning was the most effective way to memorise and subsequently r e c a l l the French sentences. As a result of 21. the weaknesses learned in the f i r s t p i l o t study, this second one was changed in two essential features. F i r s t l y , a l l of the students had to write out the sentences, as they were presented to them on the overhead projector. The active response to the situation served the purpose of assuring some ' degree of student attention. In pi l o t study 1 they learned in passive fashion. This opened up the possibility of students not attending suff i c i e n t l y to the task. Secondly, because attention to orthographic features had proven too d i f f i c u l t for the students to put into effect, they were told to learn the sentences on an exact word, or "word for word" basis, as in the King and Russell study (1966). Thus instead of referring to them as the A.O. group they were changed to the W.W. (word for word) group. The A.I. group process remained the same, that i s , they had to attend carefully to the imagery of the sentences. But, this time, they also had to write the sentences instead of just reading them. The period of presentation time for each student was lengthened to 20 seconds. This provided sufficient time for the slowest writer to complete each sentence. Details of the administration of the learning task were in every other way the same as in the f i r s t p i l o t study. It was hypothesised that the students who adopted the A.I. mental set would learn the French sentences more effectively than the students who learned by the W.W. treatment. Scores were assessed by means of the same two re c a l l measures as were used in pilot study 1. Again there were two post tests, immediate and delay. Analyses of variance (Appendix B Tables 9> 10) were estimated on 22. the resulting test scores. In the "words correct" measure significantly-greater mean scores for the W.W. learners were produced in the t o t a l scores, that i s , immediate plus delay post test scores. Because of evidence of an interaction effect between treatments and post tests a Bonferroni test (Appendix B Table 12) was calculated to locate the source of the interaction. This indicated that the W.W. treatment was significantly more effective than the A.I. treatment in the immediate post test but not in the delay post test. In the content word measure there was no significant difference between the treatment means. Once again the hypothesis that attention to imagery would prove the more effective treatment had been rejected. This result, in both measures, conflicted with that of King and Russell, (1966), which found that the group which paid attention to the meaning scored higher than the one attending to the language on a "word for word" basis. It is possible that these results could have differed from those of King and Russell because the thirty sentences were not meaningfully connected as the sentences had been in the prose used by these researchers; because the language is French, not English; because the students were not sufficiently competent in i t s use; or, because this particular sample of students was a biased one. The author could find no studies which might indicate that the f i r s t three of these poss i b i l i t i e s could present a cause for this unexpected result. Indeed, time and again, studies in verbal learning tend to support the power of meaning as the most effective mediator for r e c a l l of 23. verbal material. Consequently, i t was decided to verify or reject the last-mentioned possibility, that, that the student sample used in the study were not necessarily representative of the student population. Accordingly, a third exploration study was carried out using a different student sample. The goal of this study was to compare the findings of pi l o t study 2 with the results derived from duplicating the experiment using a different subject sample. This sample consisted of grade 11 students learning French at Princess Margaret Senior Secondary School. The catchment area for this school is contiguous to that of Semiahmoo Senior Secondary School and students are of similar socio-economic status. Sixty-five students were involved. Their progress in the learning of French was equivalent to that attained by the Semisnmoo students. At both schools the students had reached the end of the sixth week in grade 11 French when the experiment was carried out. Course work in both schools was the same, being the prescribed course for senior secondary students in the province. It was verified that the rate of progress through the course was equivalent. This is to be expected since the teachers have the same period of time in which to get through an equal amount of material. The experiment had to take place in a classroom in the Princess Margaret School building but apart from that, administration of the experiment was identical to that used in pi l o t study 2. The same teacher (the author), carried out the experiment to decrease the likelihood that no accidental changes of detail might be introduced. As in the previous pi l o t studies, i t was hypothesised that students 2k. who adopted the A.I. mental set would learn the French sentences more effectively than the students who learned W.W., for both re c a l l measures. It was also hypothesised that there would be no significant difference between the t o t a l scores in each of the schools, for both re c a l l measures. This hypothesis was introduced as a control to permit a comparison of amount of learning accomplished by the two school samples and thus determine the equality of sample in the student population. Subsequent to the French sentence learning and the two post tests, analyses of variance (index C Tables lb, 17) were carried out on the results. The most noteworthy feature of the analyses was the general reversal of trends shown at Semiahmoo School. Instead of a significant superiority of r e c a l l i n the W.W. treatment in the "words correct" measure the means of the treatment groups at Princess Margaret School were not significantly different. On the other hand, in the "content word" measure the students at Princess Margaret School gave evidence of significantly more effective r e c a l l in the A.I. treatment whereas the Semiahoo students had shown no significant treatment effect in this measure. The hypothesis that the t o t a l scores for one school would not be si f n i f i c a n t l y greater than the other was supported by the significance i readings in Tables 19 and 20 (Appendix D). Following the collection of data at both schools the KR^Q r e l i a b i l i t y coefficient was calculated for both measures. For the "words correct" measure KR.2Q=0.695, a n ^ f ° r the "content word" measure KE^Q=0.787. The author was satisfied that these r e l i a b i l i t i e s were adequate for present o o 25. research purposes although higher r e l i a b i l i t i e s shall he sought for future use of the instrument. O 26. FIG. 2 Graph Showing the Relative Value of the Treatment Means at the Two Schools "Words Correct" Scores Princess Margaret A.I. Means W.W. Means Table of Means on page 27 TABLE 1 "Words Correct" Means, (X) Immediate Post test Delay Post test A.I. W.W. A.I. W.W. Semiahmoo School 2 9 > 0 ^ . l , 2 1 . 2 l 6 . 8 Princess Margaret Sch. ^ 1 > 0, 3I1.I 2 3 . 6 18.0 FIG. 3 Graph Showing the Relative Value of the Treatment Means at the Two Schools "Content Word" Scores 29. TABLE 2 "Content Word" Means, (X) Immediate post test Delay post test A.I. W.W. A.I. W.W. Semiahmoo 21.4 33-0 l 6 . 5 12.3 School Princess 27 .0 21.5 17.5 13-0 Margaret School 30. CHAPTER FOUR Conclusions Relating to Comparison of Results  at Princess Margaret and Semiahmoo Schools The outstanding feature of the studies, completed at Semiahmoo and Princess Margaret Senior Secondary Schools, was the paradoxical nature of the results derived from the two post tests. By choosing two subject samples from what was expected to be the same student population i t was anticipated that the results of one study would verify the results of the other. Interactions of a l l three variables: (I) school; ( i i ) treatment; ( i i i ) immed-delay, which are present in one or both of the analyses of variance (Appendix D Tables 19 and 2 0 ) , are further complicated by an apparent interaction related to the type of verbal r e c a l l measure used. These results must be examined in relation to the two student samples in order to find some relationship between them which could help to account for these complex interacting effects. The experimenter had a prolonged discussion with the French teacher at Princess Margaret because i t was thought that the solution to the problem must l i e in some difference, or differences, in the classroom situation. Although the detailed differences of the teaching methods used by the teachers in the two schools had to be numerous, nevertheless, there were two contrasting teaching techniques which seemed to be especially relevant. In class the French teacher at Semiahmoo School would choose French expressions, from prescribed reading sections, which he deemed worth committing to memory. This teacher, who employed certain programmed 31. learning principles, (such as active response, immediate reinforcement, and cuing), would further the learning process on the following day by writing the English translation of the French expression alongside part of the French expression on the chalkboard. The incomplete French expression acted as a cue re c a l l of the whole expression, thus: English: I have two dollars l e f t . French: II re deux dollars. The students had to respond by writing the completed French expression in their notebooks. The teacher would then write the correct answer on the board to reinforce correct responses, thus: II me reste deux dollars. The student, striving for the parts of the French expression necessary to complete the response correctly, is drawn by this instructional method to try to re c a l l the orthographic parts which are missing. In contrast, the teacher at Princess Margaret School introduced the French expression to her students by saying i t in class and then by using i t orally in a variety of ways. The students are induced to a comprehension of the expression by discriminating between the various examples which helped them to focus upon a more correct and precise meaning. The students, working in pairs, were then encouraged to use the expression in dialogue fashion. Finally, the teacher would write the l i s t of expressions on the board for the students to copy into their notebooks. By this method, the student is induced to learn by concentrating his attention more on the meaning, orally communicated, and less upon the orthography of the expression than the Semiahmoo students had done. 32. Because of the different teaching methodology employed with the two subject samples, i t is theorised that the bond between word and meaning have been established in two different ways, thus affecting strength of r e c a l l of the French language and the choice of words recalled. Such a theory concurs with the different positions held by Skinner and Paivio in how word meaning is acquired (re: Chapter Two). The former emphasises reinforcement as the key to the process and the latter supports imagery mediation. In relation to educational instruction there i s also another important difference between the positions held by Skinner and Paivio on word meaning. Skinner's reinforcement of the correct verbal response is dependent upon an external sign which represents the correct response and provides reinforcement to the learner. It i s i n providing this sign that the teacher finds that he is needed. This process which requires an interaction between teacher and learner develops into a teaching method; the method used by the French teacher at Semiahmoo School. On the other hand, Paivio's insistence that imagery is an important attribute uniting word and meaning is not dependent upon any external sign to act as reinforcement. It is essentially a process which is a subjective way of learning. The teacher is not really required. If the learner has imaged the word meaning, effective r e c a l l of the word is supposedly strengthened by such imagery. It i s theorised that an interaction effect between teaching method and learning style is the cause of the apparently discordant r e c a l l results 33. obtained in Pilot Studies 2 and 3« This i s a post hoc interpretation but in the ensuing study the relationship between two teaching methods and two learning styles are examined.in an experiment in which both method and treatment w i l l be controlled. One method is the reinforced orthographical style described above and used by the Semiahmoo French teacher. One of the treatments employed is "attention to imagery." It is s t i l l the primary goal of this study to examine the effectiveness of this treatment in the learning of French language. To contrast with the reinforcement teaching method (R. Meth.) is the aural-oral method (0. Meth.) used by the French teacher at Princess Margaret School. This latter method is the one most commonly espoused by French teachers. The basic argument in i t s favor i s that i t is a method of instructing language which parallels closely the natural method. The natural method may be represented by the young child who learns his native tongue by means of hearing i t and by repeating what he hears. He does not have to contend with the written aspect of language u n t i l a few years of this preliminary aural-oral training have passed. By this time he has become a f a i r l y competent user of the language. There are many reasons why this method may not necessarily work so well with English speaking grade 11 students who are learning French. However, these w i l l not be enumerated here. The other treatment to compare with "attention to imagery" i s the one already used in the Pi l o t Studies 2 and 3, namely, "word for word" learning. In addition to these two learning treatments a control group is 34 included. A comparison of the rec a l l scores of this group and the other two treatment groups w i l l provide an indication of the effectiveness of the learning of the latte r groups as well as the equivalence of rec a l l of the students i n the two schools. The same two schools, Semiahmoo and Princess Margaret provided the grade 11 students to act as subjects i n the experiment which took place twelve months after the Pilot Studies 2 and 3» 35 CHAPTER FIVE Hypotheses and Procedure Hypotheses In accord with findings i n past research mentioned i n Chapters One and Two, i t i s expected that attention to imagery should prove more effective than either "word for word" learning or learning by a control group. Knowledge of which language teaching method—orthographic reinforce-ment or a u r a l - o r a l — i s more effective i s largely conjectural since there i s l i t t l e evidence available from controlled studies to provide a strong indicator. Reinforcement of student response has been a technique which educators have used i n common practice and have found i t to be positively related to effective learning. On the other hand, the relative strength of covert reinforcement which i s probably present i n aural-oral language communication i s unknown. Hypothesis II reflects this controversial issue. In pilot studies 1 and 2 i t was found that there was no significant difference between the amount of learning i n the two schools. It was assumed that this would hold true i n this experiment and the amount of learning i n each school would be compared to verify this assumption. 0 Hypothesis I Students learning French sentences by consciously imaging the sentence meaning w i l l r e c a l l the material more effectively than those learning "word for word", or those i n the control group. 3 6 . Hypothesis II Students who are habituated to learning French sentences by a reinforced written method w i l l show differential r e c a l l of the sentences compared to those who normally learn by an aural-oral method. The above hypotheses w i l l be considered in relation to four different verbal re c a l l measures. These measures w i l l be described in Chapter Six. In the s t a t i s t i c a l analyses, in Chapter Seven, rejection or tenability of hypotheses w i l l be established. There w i l l be separate analyses corresponding to the four r e c a l l measures. Any discussion attached to a particular analysis w i l l merely state obvious significant variable effects without attempting to account for causes and trends. This w i l l be done in Chapter Eight' when the results of the four analyses can be related to each other and a more comprehensive perspective can be obtained. 3 7 Procedure Eight classes of grade 11 students learning French were involved. There were approximately 2 5 to 3 0 students i n each class; four classes i n each school. In these two schools, Semiahmoo Senior Secondary and Princess Margaret Senior Secondary, the experimental plan of procedure was the same. Two classes of students were taught French expressions i n an aural-oral method i n the way which has been habitually used by the teacher in the Princess Margaret school. This teaching style has been previously explained. The other two classes were taught the same expressions by using the method which involves written chalkboard expressions, active student responses, and immediate reinforcement. This method, which has also previously been described, i s the one which was normally used by the Semiahmoo teacher. Each of the teaching techniques ran concurrently during a six week pre-experimental period. Each teacher, then, taught two classes i n a fashion which he normally used and the other two classes i n a style copied from his colleague i n the other school. Much of the lesson, i n a l l classes and i n both schools was i n essence the same. A certain amount of the French text, Parler et  Lire had to be translated; a certain section of grammar had to be taught; certain exercises from a work-book had to be completed by the students. These sections of work were equated, lesson by lesson, i n both schools. This was easily achieved because there i s usually l i t t l e difference between work accomplished i n the high schools under non-experimental 38. conditions, since each teacher has a prescribed course of material to teach in a limited time. The parts of the lessons where the teachers did have to adjust their style of teaching to accord with the experimental training was in the teaching of French expressions (details of this have been given in conclusion to comparison of results at Princess Margaret and Semiahmoo^ schools pages Jo,3l)-At the end of six weeks the students were given a l i s t of French words and expressions the English meaning of which they had to know (Appendix E i ) . This verbal material was the same as that used in the thirty sentences, (Appendix E i i ) / . which the students had to learn three days later. The words and expressions were scrambled so that comprehension of them was possible but no association between them such as that present in the 30 sentences, -was apparent. In the experiment three days later, the students were assigned to learn the 30 French sentences which were shown to them on a screen by means of an overhead projector. There were two learning sessions for each class. After randomly assigning the class into three groups, one group was sent to the library. A l l other partitioning of the to t a l class sample was done by random assignment, using the l i s t of names in the register in conjunction with a l i s t of random numbers, before the. actual experiment. One of the remaining two groups was given explicit instructions to learn "word for word" (Appendix E i i i ) and to ensure that each word was attended to by the students they were also instructed to copy each of the thirty sentences on a sheet of paper provided for this purpose. This was the W.W. group. These students were seated at one side of the classroom. 39. The other group, which was the control group, situated at the other side of.the class, (a gap between these groups was created by the exit . of students to the library), was given instructions (Appendix E iv) to learn the sentences in the fashion which they customarily employed. They were also provided with paper so that they could write or draw as they wished i f they f e l t that i t would assist them in the learning process. These two groups, the W.W. group and the control group, took part in the f i r s t of the two learning sessions. Explicit instructions for each group were typed upon paper similar in appearance to avoid suspicion or curiosity between members of the groups. The time of exposure of each sentence to be learned was 20 seconds. .9 Paivio and Foth exposed paired associates for 15 seconds to college students for a similar type of experiment but in pilo t studies run at Semiahmoo i t was found that many of the grade 11 students required up to 20 seconds to copy the sentences or draw the meaning implied by them. Immediately following the sentence presentation approximately half of the students were given the test sheets to complete, (Appendix E, v i ) . This test sheet consisted of 30 stimulus words which were the subjects of each sentence. Students were expected to rec a l l the predicates of the 30 sentences and to write them down. The other half of the class was given a French word puzzle to solve (Appendix E v i i ) . This group of students formed the delay post test group. They had to re c a l l the French predicates 2k hours later. When the test sheets were completed, a l l sheets (specific instructions, tests, drawings, and puzzle sheets), were collected. These students 40. were displaced by those who had waited in the library and whose turn i t was in the experimental situation. This group was the A.I. group which had to attend to the imagery of the sentences and who received specific instruction sheets urging them to do that (Appendix E v). They were also provided with sheets on which to draw a sketch of the "sentence meaning" for each of the 30 items (Appendix E v i i i ) . This ensured that they did strive to conjure up an image to f i t the meaning of the sentence. The quality of the picture was not considered too relevant and the students were informed of this. Before presenting 30 sentences a demonstration of what was expected of them was given by the experimenter, on the chalkboard. For example, i f the sentence was: La v i e i l l e regarde par la fenetre (The old lady looks out of the window). an appropriate drawing may simply be like this: Three examples such as this were demonstrated and pertinant questions from the students were answered. When a l l 30 items had been completed by the students those of them who had been assigned in the immediate post test groups completed the same test as the W.W. and control groups and the rest worked at the French puzzles. When the tests were completed, once again, a l l papers were collected. Questions from curious students relating to the experiment were deferred for a couple of days u n t i l following the delay post tests. On the following day the delay post test groups were told to complete the same form of test as had been used on the previous day. Data Collection There were four measures of recall. Two of these have been previously used in King and Russell's study ( 1 9 5 6 ) . One measure was "words correct" score in which the numbers of words which are correct were totalled for the thirty items. The other one consisted of a count of "content words", which included nouns, verbs and adjectives, and excluded form words such as articles, conjunctions and prepositions. A third measure, which was the most stringent one, credited only predicates which were perfectly recalled; a spelling error or an incorrectly placed word rendered the item incorrect. The fourth measure gave positive credit to a word or an expression which, although i t may have been different to the originally learned material, indicated some measure of recall. For example, synonyms and verbal substitutes which were not too remote from the correct meaning were scored positively. Included in this "synonym group" were words which, but for a minor misspelling would have been correct. This measure required more arbiting in item scoring than the other three measures but a consistency of scoring judgement was applied throughout the two school samples to permit sound comparative relations to be made. 42. Randomness in the Class Samples Although random assignment and s t a t i s t i c a l analysis based upon randomness does obviate the influence of variables due to individual differences, these French class groups had been previously selected so that individual differences in students programmes could be accommodated into the school timetable. Consequently, i t was l i k e l y that the class groups had not been randomly chosen. Therefore, a comparison of the verbal intelligence scores of the students in each of the classes had to be made to verify whether they were similar in this regard. One of the basic assumptions implicit in taking this particular measure is that the learning of French sentences is a variable which correlates with verbal intelligence. The verbal intelligence scores were obtained from the students record cards which are on f i l e in the school administration office. A l l of the verbal I.Q.'s were not available because of incomplete record cards but sufficient data was obtained to permit a valid group comparison (Tables 3 and k). TABLE 3 Mean Verbal I.Q.s for the Two Method Groups in Both Schools Teaching method Semiahmoo School Princess Margaret School Aural-oral 120.10 (N=45) 121.10 (N=45) Reinforcement 119.86 (N=42) 119.80 (N=50) An analysis of variance (Table k) was calculated to estimate equivalence of means. 44. TABLE k Summary of Analysis of Variance of Verbal I.Q. Scores Source S.S. D.F. M.S. F SIG School (S) 7.0 1 7.0 0 . 2 N.S. Method (M) 2 5 . 0 1 2 5 . 0 0 . 8 N.S. Interaction SXM 14.0 1 14.0 0 . 4 N.S. Error (within cells) 5446.0 170 32.0 TOTAL 5492.0 173 As Table 4 indicates the verbal I.Q.s were not significantly different. This was not surprising since no apparent selection system was used for placing students into different classes except for student convenience. CHAPTER SIX Research Results These results are organised i n four sections. Each section i s concerned with one of the particular verbal r e c a l l measures used. In each case mean scores of a l l the sample subgroups are presented followed by an analysis of variance of the four variables involved i n the experiment. Support or rejection of the hypotheses and any pertinent comment relating to the analysis of the results are also stated. I. Analysis of Scores for "Total Sentence Correct" Measures TABLE 5 Means of Recall Scores for "Total Sentence Correct" Measure Post Semiahmoo School Princess Margaret School method test A.I. W.W. Control A.I. W.W. Control Reinforced Immed. 5.5 4.6 3.4 . (N=9) (N=9) (N=9) 2.4 (N=ll) 1.5 (N=12) 2.2 (N=13) Delay 1.0 0.7 2.5 (N=8) (N=9) (N=8) 1.5 (N=7) 0.7 x (N=ll) 1.4 (N=9) Aural-• oral Immed. 2.8 4.2 3.5 (N=10) (N=10) (N=9) 4.0 (N-9) 1.1 (N-8) 2.6 (N=12) Delay 1.2 1.1 2.2 (N-ll) (N=8) (N=9) 1.0 (N=12) 2.7 (N=8) 1.75 (N=7) 46. TABLE 6 Summary of Analys is of Variance of " Total Sentence Correct" Recall Scores Source S.S. D.F. M.S. F Sig. Schools (S) 36.30 1 36.30 4.93 p< 0 . 0 5 Treatments (T) 6.69 2 3.36 0.46 N.S. Teaching Method (M) 0 . 5 1 , 1 0 . 5 1 0 . 0 7 N.S. Immed:Delay post Tests (P) 158.85 1 158.85 21.58 p< 0 . 0 0 1 Interaction SxT 6.91 2 3.45 0.47 N.S. SxM 12.89 1 12.89 1.75 N.S. SxP 49.75 1 49.75 . 6 . 7 6 p< 0 . 0 1 TocM 8.70 2 4.35 0 .59 N.S. TxP 23.36 2 11.68 1.59 N.S. MxP 2.54 1 2.54 0.35 N.S. SxTxM 4.50 2 2.25 0 . 3 1 N.S. " SxTxP 27.00 2 13.50 I . 8 3 N.S. SxMxP 8 . 0 0 1 8 .00 1.09 N.S. " TxMxP 38.15 2 19.08 2.59 N.S. SxTxMxP 0.72 2 0 . 3 6 0 . 0 5 N.S. Error (within cells) 1501.03 204 7.36 TOTAL I 8 8 5 . 9 0 227 47. HYPOTHESIS 1 This hypothesis i s rejected. The students who learned the sentences hy consciously attending to the imagery of the sentence meaning did not learn significantly better than the W.W. and the control groups. HYPOTHESIS 11 This hypothesis i s also rejected. The habitual method of teaching French—whether i t be reinforced or aural-oral, has no significant effect upon the re c a l l of the tot a l correct sentence. Throughout the rest of Chapter Seven the Bonferroni t test was used: (i) to determine whether the A.I. treatment i s significantly superior in r e c a l l effect to both W.W. and control treatments; ( i i ) to locate the source of variable interactions indicated in the analyses of variance. 48 The analysis of variance i n Table 6 indicates a significant interaction between school and post test. TABLE 7 Means (X) of "School X Post Test" Interaction Semiahmoo School Princess Margaret School (N = 109) (N = 119) Immediate post test X = 3.89 (S.D. = 2.9) X = 2.29 (S.D. = 1.6) Delay post test X = 1.36 (S.D. « 0.9) X = 1.44 ,(S.D."=- l . l ) The source of the interaction between schools and post test was determined by means of the Bonferroni t test. S t a t i s t i c a l hypothesis: Ho : Semiahmoo^ Princess Margaret rec a l l • recal (Bonferroni table of 2 t a i l probabilities) TABLE 8 Bonferroni Confidence Limits for Means Contrasts "Total Sentence Correct"Scores Lower Upper Comparing limit l i m i t s ig. Semiahmoo immediate n , _s C** o rt ^ Princess Marg. immediate + °' 6 ^ — + 2 , 8 p < °'° 5 Semiahmoo delay: _ 0.4 ^ V ^ + 2.0 N.S. Prxncess Marg, delay ^ h9. The n u l l hypothesis (Ho) is supported by the immediate post test results but i t is not supported by the delay post test results. DISCUSSION In this r e c a l l measure attention to imagery has not been found to be significantly superior compared to the other treatments. Different teaching methods which were used for six weeks preceding the experiment have not shown differential effect. The variable which did produce a significant difference in r e c a l l scores was the school variable. This result is confounding because of the basic assumptions in this study was that there would be no significant difference between the means of the r e c a l l scores of the two schools. Because there is evidence of an interaction effect between the schools and the post tests, the Bonferroni calculation '(Table 8), was used to ascertain i n which test, immediate or delay, the greater r e c a l l effect was located. The Bonferroni results indicate that the r e c a l l scores of the Semiahmoo students were significantly superior only in the immediate post test. The reason for this result i s not at a l l clear, and i t w i l l be considered later when i t can be related to the results of the other r e c a l l measures. This "total sentence correct" measure, which demands a perfectly recalled sentence in order to obtain credit, i s obviously a very stringent one. Out of the twelve cells, which show the means of the immediate re c a l l scores, in Table 5, the highest mean is only 5.h7. The maximum possible correct is 3 0 . The KRgQ r e l i a b i l i t y coefficient for the test 5 0 . was r, , = 0.43. It is possible that the d i f f i c u l t y of recall, as XtXi measured here, together with the poor r e l i a b i l i t y of the measure, is responsible for the enigmatic result. 51. II Analysis of Scores for "Content Word" Measure TABLE 9 Means of Recall Scores for "Content Word" Measure Teaching Post Semiahmoo School Princess Margaret School method test A.I. W.W. Control A.I. W.W. Control Reinforced Immed. 44.5 (N=9) 32.9 (N=9) 35.8 (N=9) 35.8 (N=ll) 21.3 (N=12) 25.3 (N-13) Delay 20.6 (N=8) 19.0 (N=9) l 6 . l (N=8) 23.2 (N=7) 16.5 (N=ll) 15.8 (N=9) Aural-oral Immed. 31.7 (N=10) 32.6 (N=10) 33.7 (N=9) 38.0 (N=9) 24.3 (N=8) 22.6 (N=12) Delay 21.0 (N=ll) 17.5 (W=8) 14.3 (N=9) 22.5 (N=12) 13.3 (N=8) 15.8 (H=7) TABLE 10 Summary of Analysis of Variance of "Content Word" Recall Scores Source S • S • D.F. M.S. F SIG. Schools (S) 774.00 1 774.00 9.44 p < 0.01 Treatments (T ) 2582.50 2 1291.25 15.75 p < 0.001 Teaching method (M) 109.90 1 109.90 1.34 N.S. Immed:delay post tests (P) 9368.90 1 9368.90 114.25 p< 0.001 Interaction SxT 531.50 2 265.75 3.24 P < 0.05 SxM 160.11 1 l 6 0 . l l 1.95 N.S. SxP 975.38 1 975.38 II .89 P ^ : 0.001 " TxM 319.60 2 159.80 1.90 N.S. " TxP 542.60 2 271.30 3.30 p,< 0.05 MxP 43.25 1 43.25 0.53 N.S. SxTxM 6.70 2 3.35 o .o4 N.S. SxTxP 97.59 2 48.80 0.60 N.S. SxMxP 43.00 1 43.00 0.52 N.S. " TxMxP 34.00 2 17.00 0.20 N.S. SxTxMxP 335.97 2 167.99 2.05 N.S. Error '(within cells) 16742.00 204 82.07 TOTAL 32667.00 227 53. HYPOTHESIS 1 This hypothesis i s supported. Attention to imagery is the most effective treatment, (p< 0.001). HYPOTHESIS II This hypothesis is rejected. No significant difference between the mean scores of "reinforced learners" and "aural-oral learners" i s apparent. The analysis of variance in Table 10 indicates a significant difference between the treatment means. TABLE 11 Treatment Means (X) for "Content Word" Scores  A.I. W.W. Control x = 27.78 x = 22.27 x = 22.83 54 The Bonferroni calculation was used to determine which of the treatments produced significantly higher mean scores. St a t i s t i c a l hypotheses: 1 2 3 Null Ho : A.I.^. W.W. Research Ho : A.I. > W.W. (1 t a i l test) Ho : A.I.-= Control Ho : A.I.> Control ( l t a i l test) Ho : W.W. ^ -Control Ho : W.W. = Control (2 t a i l test) TABLE 12 Bonferroni Confidence Limits for Means^Contrasts  "Content Word" Scores Comparing Lower Upper Sig. limit limit A.I. : W.W. +2.3 ^ ^ — ° ° . p^.0.05 A.I. : Control +1,7 6 ^ 6 « p < 0.05 W.W. : Control - 3.0 <a ¥ & + 4.2: N.S. The null hypotheses 1, 2, and 3 are a l l rejected. / The analysis of variance in Table 10 indicates a significant interaction between treatments and post tests. TABLE 13 Means (X) of "Treatment x Post Test" Interaction Cells Treatments Tests A.I. W.W. Control Immediate X = 37.26 X = 27 .46 X = 28.46 Delay X = 21.84 X = 16.64 X = 15.48 56 The significant interaction effect, indicated i n Table 10, between treatment and post test was examined by means of the Bonferroni t test i n order to locate the source of the interaction. S t a t i s t i c a l hypotheses: 1 Null Ho: u A. I. :•=: W.W. Research Ho: A.I. W.W. ( l t a i l test) Ho : A . I . C o n t r o l Ho : W.W. ^/Control Ho : A.I.> Control ( l t a i l test) Ho : W.W. = Control (2 t a i l test) TABLE 14 Bonferroni Confidence Limits for Means Contrasts "Content Word"Scores Comparing Lower lim i t Upper limit Sig A.I. CD +J - p CQ cd CD •H - P CD - P fl O M & W.W. A.I. W.W. + 4.7 Control +3.8 Control - 4.5 + 13.8 p ^ 0.05 p 0.05 N.S. -p A.I. : W.W. + 0.2 A.I. : Control + 0.9 Cd -4-> H CQ o B, W.W. : Control -10.2 A 4=. if/ A + 12.6 p < 0.05 p< 0.05 N.S. The null hypotheses 1, 2, and 3 are rejected. The Bonferroni contrasts between immediate and delay post test means on the same variable have been ommitted i n this Table and i n others where they are a l l clearly significantly different i n the one direction. 57. The analysis of variance in Table 1 0 indicates a significant interaction between the means of treatments and schools. TABLE 1 5 Means (X) of School x Treatment Interaction Cells School A.I. W.W. Control Semiahmoo X - 2 9 . 3 X = 2 5 . 9 X = 25.2 Princess Margaret X = 3 0 . 0 X = 1 8 . 9 X = 2 0 . 8 58 The Bonferroni calculation was used to determine the source of the interaction. S t a t i s t i c a l hypotheses: 1 Null Ho : A . I . ^ W.W. Research Ho : A.I. W.W. (1 t a i l test) Ho : A.I.. 2 Control Ho : A.I. 7 * Control (1 t a i l test) Ho : W.W. ^ Control Ho : W.W. = Control (2 t a i l test) Null Ho : Semiahmoo / Princess Margaret School ^ School Research Ho : Semiahmoo .^ Princess Margaret School School TABLE 16 Bonferroni Confidence Limits for Means Contrasts "Content Word" Scores Comparing Lower limit Upper limit Sig. CO N.S. N.S. + 6.4 N.S. p < 0.05 OO p< 0.05 + 7.5 N.S. + 6.6 N.S. + 13.9 p < 0.05 + 9.3 N.S. Semi. A.I. : Semi. A.I. : Semi. W.W. : P.Marg. A.I. P.Marg. A.I. P.Marg. W.W. Semi. A«I.« : Semi. W.W. : Semi. Control Semi. W.W. Semi Control Semi Control 1 P.Marg. W.W. : P. Marg. Control : P. Marg. Control P.Marg. A»I. P.Marg. W.W. P.Marg. Control - 2.0 - 1.3 -5.0 + 5.8 + 4.0 - 3.1 - 5.3 + 2.1 - 0.4 A If % The null hypotheses 1 and 2 are supported at Semiahmoo School but hypothesis 3 i s rejected. 59. At Princess Margaret School hypotheses 1 and 2 are not supported. Once again hypothesis 3 is rejected. When the treatment means of the two schools are compared, the null hypotheses for the A.I. and the control treatments are rejected. The n u l l hypotheses for the W.W. treatment i s supported. 6o. The analysis of variance in Table 10 indicates a significant interaction between the means of post test and school TABLE 17 Means (X) of School x Post Test Interaction Cells Post Test Semiahmoo School Princess Marg. School Immediate X" = 35 .1 X = 27.5 Delay X = 18.2 X = 18.1 6 1 T h e s o u r c e o f t h e i n t e r a c t i o n b e t w e e n s c h o o l a n d p o s t t e s t w a s i n v e s t i g a t e d b y u s i n g t h e B o n f e r r o n i t t e s t . S t a t i s t i c a l h y p o t h e s i s : N u l l H o : S e m i a h m o o ^ P r i n c e s s M a r g a r e t R e s e a r c h H o : S e m i a h m o o = P r i n c e s s M a r g a r e t ( 2 t a i l t e s t ) T A B L E 1 8 B o n f e r r o n i C o n f i d e n c e L i m i t s f o r M e a n s C o n t r a s t s " C o n t e n t W o r d " S c o r e s C o m p a r i n g L o w e r l i m i t U p p e r S i g . l i m i t S e m i a h m o o ' I m m e d i a t e : P r i n c e s s M a r g a r e t I m m e d i a t e S e m i a h m o o D e l a y : P r i n c e s s M a r g a r e t D e l a y + 4.0 y/ ^ + 11.3 p ^ - 0.05 - 3 . 8 ^ ~P + 4 . 1 N . S . T h e n u l l h y p o t h e s i s i s s u p p o r t e d i n t h e i m m e d i a t e p o s t t e s t , b u t i t i s r e j e c t e d i n t h e d e l a y p o s t t e s t . 62. DISCUSSION Recall of the content words in the French sentences i s clearly-superior when they are learned by attending to the imagery that they invoke. Not only i s this r e c a l l superior to that of the W.W. learning group but also to the control group of students who learned the sentences in their own individual fashion. It i s quite probably that the method of learning of both W.W. and control group students is similar; there being hardly any difference between their r e c a l l scores. Once again, the Semiahmoo students' r e c a l l scores were higher than the scores of the students of the other school. This superior r e c a l l l i e s in the W.W. treatments but not in the A.I. nor in the control treatments, (Table l6). The scores of the Princess Margaret A.I. group which are significantly higher than those in the W..W. and control groups of the same school supports hypothesis 1 but the same support is not evident in Semiahmoo where the three treatments are not significantly different. This inequality between the scores of the two schools and the interaction between treatments is the single most confounding feature of the study. A possible explanation for these results i s given in Chapter 8. Like the "total sentence correct" measure, the greater mean scores of the Semiahmoo students are located in the immediate post scores; the difference in the delay post test scores being not significantly different (Table 18). It may be assumed that the reason for the significant difference in the school variable in the "total sentence correct" measure i s causally related to the difference in the same variable in the "content 63. word measurerj; the items in the former measure being more global and being, to some extent, composites of several items of the latter measure. In this measure the r e l i a b i l i t y coefficient for the test i s considered by the author to be adequate, KR 2 Q = O.78. A higher r e l i a b i l i t y may be sought for future use of the test. 64. III. Analysis of Scores for "Correct Word" Measure TABLE 19 Means of Recall Scores for "Correct Word" Measure Teaching Post Semiahmoo School Princess Margaret School method test A.I. W.W. Control A.I. W.W. Control • ^ 03 Immed. 60.I 49.7 46.2 55.8 34.5 4 2 . 6 0 u 0 (N=9) (N=9) (N=9) (N=ll) (N=12) (N=13) a • H Delay 33.4 26 .0 27.5 .35.8 2 6 . 3 25 .0 CO « (H-8) (N=9) (N=8) (N=7) (N=ll) (N=9) Immed.v 5 2 . 0 50.3 46 .3 58.0 36.0 36.0 1 r-i Cd r-i (N=10) (N=10) (w=9) (N=9) (W=8) (N=12) U cd 3 U < O Delay 34.4 25.5 23 .2 35.0 21.7 25 .8 (N=ll) (K=8) (N=9) (N=12) (N=8) (N=7) 6 5 . TABLE 20 Summary of Analysis of Variance of "Correct Word" Recall Scores Source SS. D.F. M.S. F Sig Schools (S) 4 6 7 . 8 0 1 4 6 7 . 8 0 2 . 2 0 N.S. Treatments (T) 6 0 4 1 . 8 2 2 3 0 2 0 . 9 1 1 4 . 2 3 p<r 0 . 0 0 1 Teaching method (M) 3 5 . 4 9 1 3 5 . 4 9 0 . 1 7 N.S. Immed.: delay post tests (P) 2 0 3 0 4 . 8 9 1 2 0 3 0 4 . 8 9 9 9 . 5 3 p< 0 . 0 0 1 Interaction SxT 976.OO 2 4 8 8 . 0 0 2 . 3 0 N.S. SxM 5 9 . 5 1 1 5 9 . 5 1 0 . 2 7 N.S. SxP 8 2 6 . 6 4 1 8 2 6 . 6 4 3 . 8 9 P< 0 . 0 5 TxM 3 1 6 . 9 6 2 1 5 8 . 4 8 0 . 7 5 N.S. " TxP 9 2 6 . 0 9 2 4 6 3 . 0 4 2 . 1 8 N.S. MxP 5 . 5 8 1 ' 5 . 5 8 0 . 0 1 N.S. SxTxM 3 5 . 2 2 2 1 7 . 6 1 0 . 0 8 N.S. SxTxP 1 2 7 . 5 6 2 6 3 . 7 8 0 . 3 0 N.S. SxMxP 1 9 . 8 9 1 1 9 . 8 9 0 . 0 9 N.S. " TxMxP 3 2 . 0 0 2 1 6 . 0 0 0 . 0 7 N.S. SxTxMxP . 1 1 4 . 6 0 2 5 7 . 3 0 0 . 2 7 N.S. Error (within cells)4 3 3 1 1 - 7 5 2 0 4 2 1 2 . 3 1 TOTAL 7 3 6 0 1 . 8 0 227 66. HYPOTHESIS 1 This research hypothesis is supported. Attention to imagery again proves the most effective treatment. HYPOTHESIS 11 This hypothesis i s rejected. Different methods of teaching French expressions during the six months pre-experimental period produced no significant difference in rec a l l scores. 67. The analysis of variance in Table 20 indicates a significant difference between the means of treatments. TABLE 21 Means (X) of Treatment Scores for "Correct Word" Measure A.I. W.W. Control X = 45.7 X = 35-7 X = 34.0 68. The source of the significant difference between the treatment means was investigated by the Bonferroni t test. S t a t i s t i c a l hypotheses: 1 Null Ho : A . I . ^ W.W. Research Ho : A.I.> l.W. (1 t a i l test) Ho : A.I. ^ C o n t r o l Ho : A. I.> Control (1 t a i l test) Ho : W.W. ^ Control Ho : W.W. = Control ( 2 t a i l test) TABLE 2 2 Bonferroni Confidence Limits for Means Contrasts "Correct Word" Measure Comparing l Lower Upper Sig A.I. : W.W. A.I. : Control W.W. : Control + 4.9 + 6.5 - 4.0 A A A + 7.4 p < 0.05 p < 0.05 N.S. Null hypotheses 1, 2 and 3 are rejected 6 9 . The analysis of variance in Table 20 indicates a significant interaction between the means of school x post test. TABLE 23 Means (X) of School x Post Test Interaction Cells Post test Semiahmoo School Princess Margaret School Immediate X = 5 0 . 9 X = hk.h Delay X = 2 8 . 5 X = 2 8 . 5 7 0 T h e s o u r c e o f t h e s i g n i f i c a n t i n t e r a c t i o n b e t w e e n t h e s c h o o l s a n d t h e p o s t t e s t s w a s a s c e r t a i n e d b y m e a n s o f t h e B o n f e r r o n i t t e s t . S t a t i s t i c a l h y p o t h e s i s : N u l l S e m i a h m o o S c h o o l ^ P r i n c e s s M a r g a r e t S c h o o l R e s e a r c h H o : S e m i a h m o o S c h o o l = P r i n c e s s M a r g a r e t S c h o o l T A B L E 2 4 B o n f e r r o n i C o n f i d e n c e L i m i t s , f o r M e a n s C o n t r a s t s " C o r r e c t W o r d " S o u r c e C o m p a r i n g L o w e r l i m i t s U p p e r l i m i t s S e m i a h m o o : I m m e d . S e m i a h m o o : D e l a y P r i n c e s s M a r g . I m m e d . P r i n c e s s M a r g . D e l a y + 0 . 5 - 6 . 1 r — + 6 . 2 S i g ^ = + 1 2 . 6 p < 0 . 0 5 N . S . T h e n u l l h y p o t h e s i s i n t h e i m m e d i a t e p o s t t e s t i s s u p p o r t e d . T h e n u l l h y p o t h e s i s i n t h e d e l a y p o s t t e s t i s r e j e c t e d . 7 1 . DISCUSSION The attention to imagery learning once again proved more effective than the learning accomplished by the students in the "word for word"- or the control groups. There was no significant difference between the r e c a l l means of the two latter groups (Table 22). Compared to the two previous r e c a l l measures the results of this one suggest that an equalising influence is present which has reduced the disparity between the scores of the students in Semiahmoo and Princess Margaret, for this time there is no significant difference between the means of the two schools. Nevertheless, the Bonferroni test (Table 2k), reveals that in the immediate post test there i s s t i l l more effective r e c a l l occurring with the students at Semiahmoo than with the students -at Princess Margaret School. KRgQ for "correct word" measure was 0;69, which the author considers a reasonably high r e l i a b i l i t y coefficient for present research purposes. 7 2 . IV Analysis of Scores for "Correct Word Plus Synonyms" Measure. TABLE 2 5 Means of Recall Scores for "Correct Word Plus Synonyms" Measure method Post Semiahmoo School Princess Margaret School test A.I. W.W. Control A.I. W.W. Control Immed. 8 0 . 0 5 7 . 0 5 6 . 5 7 0 . 2 5 7 . 3 5 1 . 2 (N=9) (N=9) (N=9) (N=ll) (N=12) (N=13) Delay 48.2 3 2 . 3 3 4 . 5 48.8 3 5 . 5 3 1 . 4 (W=8) (N=9) (N=8) (N=7) (N=ll) (H=9) Immed. 7 7 . 5 5 6 . 9 6 0 . 3 7 6 . 5 5 6 . 9 5 0 . 8 (N=10) (N=10) (N=9) (N=9) (N*8) (N=12) Delay 5 5 . 0 3 3 - 3 2 9 . 1 5 1 . 0 2 8 . 3 3 8 . 5 (N=ll) (N=8) (N=9) (N=12) (N=8) (N=7) O U O -P <H a C <L) • H S <D K I r - l 1 ° TABLE 26 Summary of Analysis of Variance of "Correct Word Plus Synonyms " Recall Scores Source S.S. D.F. M.S. F Sig Schools (S) 253.95 1 253.95 0 . 7 6 N.S. Treatments (T) 17271.40 2 .8635.70 25.88 p < 0 Teaching method (M) 85.59 1 85.59 0.26 N.S. Immed.: delay post test (P) 29030.13 1 29030.13 86.99 p < 0 Interaction SxT 6.40 2 3.20 0 . 0 1 N.S. SxM 14.34 1 14.34 0.04 N.S. SxP 538.93 I 538.93 1.62 N.S. TxM 28.19 2 14.09 0.04 N.S. '' TxP 1827.73 2 913.87 2.74 N.S. MxP 172.98 1 172.98 0.52 N.S. SxTxM 180.33 2 90.17 0 .27 N.S. SxTxP 194.70 2 97-35 0 .29 N.S. " SxMxP 7.99 1 7.99 0 . 0 2 N.S. " TxMxP 79.40 2 39.70 0.19 N.S. SxTxMxP 527.01 2 263.51 0 .79 N.S. Error (within cells) 68077.13 204 333.70 TOTAL 118296.20 227 74. The analysis of. variance Table 26 indicates a significant difference between the treatment means. TABLE 27 Means (X) of Treatment Scores  "Correct Word Plus Synonyms" Measure A.I. W.W. Control X = 63.6 X = 45.0 X = 45.5 75. The source of the significant difference between the treatment means was determined by using the Bonferroni t test S t a t i s t i c a l hypotheses Null Ho: A.I. W.W. Ho : A.I. ^ Control Ho : W.W. -f Control Research Ho : A.I.> W.W. Ho : A.I. y Control Ho : W.W. £ Control (1 t a i l test) (1 t a i l test) (2 t a i l test) TABLE 28 Bonferroni Confidence Limits for Means Contrasts  "Correct Word Plus Synonyms" Measure Comparing , Lower Upper Sig limit limit A.I. : W.W. + 12:3 — Y ^ 0 0 P < 0.05 A.I. : Control + 11.8 ^ Y ~ ° ° P < °*° 5 W.W. : Control - 6.5 — V ~ + 7 ' 7 N , S ' The null hypotheses 1, 2 and 3 are rejected. 76. DISCUSSION Unlike the previous measures, no interaction between schools and post test i s apparent, indicating that any equalising influence between the two school r e c a l l scores i s stronger s t i l l than in the preceding r e c a l l measure. KRpQ r e l i a b i l i t y coefficient for the "correct words plus synonyms" measure was 0.73; considered an acceptable index of r e l i a b i l i t y by the a uthor. 77. CHAPTER SEVEN Conclusion Attention to Imagery Three out of four of the rec a l l measures used i n this experiment strongly support the theme that attention to imagery more than any other attribute associated with verbal learning, i s most effective i n r e c a l l . The "content word", "correct word" and "correct word plus synonym" rec a l l measures a l l produced significant differences i n mean scores indicating that the students who imaged the French sentence meaning had succeeded i n their learning task better than the students who learned "word for word'! or the students of the control group. Both the low r e l i a b i l i t y and the positively skewed distribution of scores i n the "total sentence correct" measure make the results derived from them of doubtful worth. A longer period of learning time would have been required to have produced larger scores which could have been more effectively compared. However, such a change might have adversely affected the three other measures especially the least stringent re c a l l measure— the "correct word plus synonym" measure. Verification of excessive d i f f i c u l t y i n such a measure may be of value i n determining the type of measures to be used i n future studies when recall of a foreign language, or time required i n such r e c a l l , are being considered. The overall effectiveness of "attention to imagery" i n the three more reliable measures i s confirmed i n a l l of the immediate post tests as well as i n the delay post test of the "content word" measure. Such a 78. result i s understandable when one considers that long terra r e c a l l has been found to be related positively to imagery, and as Paivio has stated, (p. concrete nouns, such as those used i n the thir t y French sentences, have high imagery value. Accordingly, attention to sentence imagery should more surely produce a superior r e c a l l effect i n the "content word" measure than in the other measures where form words, which are low in imagery, are taken into account i n assessing scores. The present finding that content words positively relate to long term re c a l l is a more practical extension to Paivio's study of this phenomenon. In his research he used paired-associates. That i s , the content words were not embedded in sentence structures as is the case in these thirty French sentences. In this study, the learning of French verbal material, by the A.I. students, was carried out by the process of "chunking" (Tulving and Patkau, I962, i n Statement of the Problem). That i s , the learning "unit" may consist of several words of the sentence. The individual word meanings are united by the t o t a l sentence imagery conjured up i n the mind of the learner. Since chunking imagery necessitates that form words are symbolically represented i n the imagery as an integral part of the sentence meaning, when the sentence r e c a l l i s required the f orm words, connecting the concrete words, are also transformed into word symbols. An examination of the less effective verbal mediation used by students in the other two treatment groups gives an idea of how students usually learn French sentences. Students i n the W.W. and control groups obtained more practice at the orthography of the verbal material. A l l of 79. them in the W.W. group wrote out each sentence as i t was presented to them and many of the control group did the same. It is interesting to note that not one of the learners in the control group drew sketches to represent sentence meaning, as> did a l l of the A.I. learners. It may be that those in the control group who were not copying the sentences were concentrating upon sentence meaning but in a more passive manner than the A.I. group. It is very l i k e l y that both W.W. and control groups also got more practice at reciting sentences since recitation is an automatic process when;..sentence copying is done. In this way they could transform the'verbal material into speech patterns. Fur those students the chunking process would function moreso in covert speech mediation than i t had done with the A.I. learners. Nevertheless, i t is probable that the W.W. and control group students also imaged the meaning to some extent, although not as much as the students who were obliged to sketch the imagery. When the learning process carried out in this research is seen in terms of the three stages--orthgraphical, phonological and semantic—the results of this study must favour emphasis of the last stage for successful learning and subsequent recall. This is especially so i f the imagery attribute of the semantic process is where the emphasis is lai d . However, in endeavouring to encourage students to attend zo imagery, both teacher and student should realize that, in the i n i t i a l steps of the process, an active response, such as sketching, may be essential to the success related to the imagery learning. If the learner i s merely l e f t to mentally picture sentence meaning, without the active response, i t is possible that 80. retransforming the imagery into verbal form may not meet with the same success. It may be necessary to spend some time i n training the learner to image sentence meaning without the support of the sketching. The control group scores, i n a l l four measures, were not significantly different to the W.W. group scores. Nor was the manifest behaviour of the students i n the control group any different to that of the W.W. students, for the majority of them also wrote out the sentences to be . learned on the sheet of paper provided for them, which they had the choice of using or not. It would seem reasonable to conclude that there was l i t t l e , i f any, difference between the W.W. and control groups and that there were only two treatments at work rather than three. The implication of this would be that "attention to imagery" was being compared with the students" habitual style of learning French expressions. The psychological process involved i n the two treatments might very broadly be interpreted as an emphasis on imagery with minimal attention being paid to the orthographical and phonological stages of learning — t h i s i s the A.I. treatment. The other two treatments may be considered a synthesis of variable amounts of emphasis applied when the students attended to each of the three learning stages. The Two Schools Among the variables studied the one which produced the most confounding results was the school variable. It was assumed that the two schools would not score significantly different to each other. This was true i n 81. two out of the four re c a l l measures but in the other two measures Semiahmoo scored significantly higher, suggesting that there has been a "school x r e c a l l measure" interaction. That i s , students in one school have a f a c i l i t y for recalling certain words in the French sentence--as indicated by the specificity of word selection of which a particular r e c a l l test was composed—and students in the other school have compensated by recalling a different choice of words. The most pronounced superiority of re c a l l was provided by the Semiahmoo students in the "content word" scores (p<O.Ol). Because "content words" have high imagery i t should follow that these same students should have performed better in the treatment which required attention to imagery. This is not borne out by the significant "school x treatment" interaction which indicates that the Semiahmoo students had scored significantly higher in the W.W. treatment (Tables 15, l6). Even the control group of Semiahmoo gave evidence of high r e c a l l relative to the Princess Margaret control group. The "correct words" measure did not support the superiority of r e c a l l by Semiahmoo. There was no significant difference between the scores of the two schools in this test. Nor was there any significant difference between the scores of the two schools in the "correct word plus synonym" reca l l measure. From the overall view of the results, in the four different measures, the nature of the tendency for re c a l l of a particular class of word by students in the two schools, becomes apparent. Semiahmoo has been more effective in recalling the "content words"; Princess Margaret students 82. have compensated in part by form words (as indicated by the "words correct" results) and in part by the synonyms. These results indicate an almost complete reversal of the trend shown in pil o t studies 2 and 3 where there had been a similar "school x word measure" interaction but in a different direction. That i s , the Semiahmoo learners had recalled more effectively in the "exact words" measure. Princess Margaret had scored higher in the "content words" measure. In order to find the reason for this reversal of trend an examination of the changes made between the pil o t studies and this one might provide some clue. There were two major changes. F i r s t l y , both teachers in this latest study had to teach by using two different methods instead of just one, and secondly the A.I. students had to sketch the imagery of the sentence meaning instead of just passively imaging i t . The possibility that the active response of sketching may have been responsible for the "school x r e c a l l measure" interaction must be rejected. Sketching was an activity relating to only one of the treatments, "attention to imagery", not to the other two. The students of Semiahmoo did not reca l l significantly better in the A.I. group. The superior effect was more apparent in the other two treatments (Tables 15, 16). Some reason other than sketching must be responsible. It i s more likely that the "teacher x method" interaction would cause the re c a l l of a different choice of words in the two schools. A glance at the r e c a l l results in the ANOVA for a l l four measures indicates that in both schools the two methods, aural-oral and reinforcement, have produced 83. no significant difference between the means in the scores resulting from them. The factor influencing the scores to produce the reverse trend must not be confined to a particular method but must be effecting both methods in both of the schools. It i s theorised that some of the ideas implicit in the new method, which the teacher has copied from his colleague at the other school, are being transferred unwittingly into the method which he has habitually used in the past. Thus, the Semiahmoo school teacher, in his efforts to teach the aural-oral method, must have carried over some of the methodological features into the reinforcement method. In like manner, the teacher at Princess Margaret, trying to teach the reinforcement method, may also have retained some of i t s principles and have applied them in the aural-oral method. If this i s true, then the research design which planned for a balanced instruction in both methods at the two schools has not been successful. An absolutely equal effect on pupil learning by both teachers and in both methods has been too d i f f i c u l t to achieve. The fact that the methods have not produced significantly different r e c a l l scores does not necessarily imply that they have equal effect unless the teachers have been able to produce the balance in teaching the two methods that had been originally planned. The failure to control the methods variable i s not entirely negative i f in fact there are signs of a "method x treatment" interaction in the research. This would seem to be the case. Further evidence is necessary to establish the relationship and also to provide information on the nature of this relationship. 84. An interaction effect of relatively less import i s shown between school and post test (Tables 10, 20). Semiahmoo's significantly greater scores i n the "content word" and in the "correct word" measures are found to occur in the immediate post test (Tables 18, 24). This is not surprising. In only one of the analyses (Table 14) so far has there been an indication of a delay post test significant effect. The nature of the learning task may have been such that i t lacked the impact to differentiate the other groups in the delay test, carried out one day later. Reinforcement Versus Aural-Oral Method In not one of the four r e c a l l measures does the pre-experimental teaching method produce any significant effect. One of the features of the research design which had to be carefully considered before commencing this last study was the methodological balance in the two schools during the six weeks preceding the learning of the thirty French sentences. It was expected that the teacher in Semiahmoo would teach the reinforcement method effectively since i t was the method that he habitually used in the past. The aural-oral method was not his customary method of teaching and had to be mastered by practice. One might have anticipated that this method had less chance of being taught so well. With the Princess Margaret teacher, in respect of teaching method, the opposite argument could hold true. Nevertheless, the fact that re c a l l scores derived from these "school x method" groups did not interact in any of the four r e c a l l measures is not necessarily a sign that both methods have equal impact. This has been discussed above, on page 85. 8 5 . Limitations of the Study A problem which is common to nearly a.ll research findings is the assessment of the extent to which the findings, are generalisable. In this study two senior secondary school student samples which may be assumed to be of similar socio-economic status, were involved. Residential areas of the students are about twenty miles from the city of Vancouver. Although access to the city for these students is f a i r l y easy i t is doubtful whether they could be thought to be of the same socio-economic status as students who reside in the urban areas. Nor, are they so removed from the town that they could be considered representative of more distant rural areas. Because of this, any educational implementation of French language learning should view the present research findings as having limited application u n t i l further confirmation of them can be obtained from a wider population. A further consideration relating to generalisability to student population is the extent of the student exposure to French language usage. If students live in an area where French is spoken there is greater probability that they have acquired both a larger number of speech patterns and a greater f a c i l i t y for imaging French verbal material. For such students the effect of attention to imagery and of "word for word" learning may differ from that produced in this study. The chunking of words into "units" during the learning i s also a process which should vary according to subject sample. Tulving and Patkau (1962) found that "adopted chunks", that i s , groups of words recalled, which were identical to those learned, varied with Thorndike-Lorge frequency of 86. the words. There i s no study comparable to that of Thorndike and Lorge to provide an indication of the frequency of French words used by English speaking students learning French. In this study, words used in the thirty sentences to be learned were chosen from the text books upon which the grade 11 French course was based. If students were very familiar with certain words the number of "adopted chunks" might increase. This could possibly result in an increase in the number of form words recalled relative to content words. The variety of methods used by teachers of French is also a variable which must be taken into account. Although the two different teaching methods studied in this experiment have not provided conclusive evidence of effect, the p i l o t studies did indicate that teaching methods probably present a covarying factor with learning technique. Until more studies of similar nature are completed, in which a variety of methods and student samples are used, French teachers should regard present research findings with some circumspection. The fixed period of time permitted to the students to learn each sentence also imposes a limitation to the geS§ralizability of the research findings. Only in one of the rec a l l measures was there any sign of effective r e c a l l over a period of twenty four hours. This was in the "content word" measure where imagery mediation was positively related to reca l l . Such was not the case with the verbal mediation used by the other two treatment groups. One implication i s that the learning procedure in the experiment may lack the impact to commit a more permanent memorisation of words other than words of high imagery. "Word for word" and the 87. control group of learners probably used covert speech mediation to a greater extent than the group which was instructed to attend to word imagery. If covert speech patterns were given more learning time than just twenty seconds to become imprinted in the memory then i t is possible that greater r e c a l l would result. The value of these speech patterns could l i e in the inclusion not only of content words but also of form words which are comparatively lacking in the semantic attribute of imgery. Speech patterns may play a major role in the verbal repertoire of young children whose a b i l i t y to image verbal symbols is not so highly developed. As Rohwer has already pointed out, the use of imagery mediation for memorising verbal material does not reach optimal performance u n t i l , on average, the age of nine years. Therefore, teaching and learning techniques which are developed with attention to imagery as a basic fundamental should only be used for students above this age. Below this age, to support covert speech habits, i t is li k e l y that more explicit prompts are needed. Suggestions for Further Research In considering this present study and research related to i t , i t must be born in mind that i t is concerned with the learning of a second language. Findings from past research in verbal learning may provide useful clues on how to f a c i l i t a t e the learning process but direct application of these findings in relation to the learning process may need modification. For example, the chunking process is dependent both on the syntactic and semantic structure. It is possible that attention to imagery in learning 88. the second language could be as potent as in the learning process of the student's native language. On the other hand, the syntactic structure of a second language may be very different and successfully learned only after years of continuous usage. Further studies are needed, using the chunking process, to establish the relationship between imagery and sentences of increasing syntactical complexity. If, as the grade 11 students in this study indicated, attention to imagery is a better way of learning the French language, i t is to be wondered why they do not use this learning method to a greater extent than they do. The reason why less effective learning techniques are used might be investigated. It may be that their learning styles have been found to be efficient in their native tongue where covert speech patterns are well established and new speech associations can be easily made. In a second language this may be much more d i f f i c u l t to achieve. Knowledge of precisely what mediators for retention of language they do use would provide a sound base for further study in this regard. One of the major problems in research, involving students in a learning session, is to devise a method which w i l l have sufficient impact to produce significant learning effect which w i l l endure for more than a few hours. The average student w i l l show only a limited enthusiasm and willingness to learn in a style in which he is not accustomed. Increasing the length of the learning period or repetition of a learning task can produce boredom and fatigue. Alternatively, a more sensitive measuring instrument might be used. In any case, i t is a consideration which is of great import in respect to learning a second language where long term 89. retention relates more r e a l i s t i c a l l y to the nature of the task in everyday l i f e . Attention to imagery may be considered a f a i r l y passive behaviour. In the present study i t was accompanied by the active response of sketching. An evaluation of these two factors in verbal re c a l l might more accurately pinpoint which of the two attributes contributes most to successful verbal learning. An aspect of learning French which is s t i l l open to much study is the relationship of the teaching method to the learning process. It has already been pointed out that great care is required to avoid interaction with other variables such as teacher competence and style of teaching. Perhaps the use of teaching machines or tape recorders, where teacher personality is removed from the teaching method, might produce more reliable results. With such equipment replacing the teacher i t might be possible to compare the effect of overt written reinforcement compared to overt aural-oral reinforcement. In the more usual aural-oral language learning lesson, in the classroom, l i t t l e is known of the nature and strength of the reinforcement in i t . Some knowledge of correctness of response must accompany the discriminant operant conditioning implicit in language learning which is dependent upon interpersonal communication. Such reinforcement is relatively covert and less obvious. The strength of this reinforcement compared to the more overt reinforcement, such as that used by the teacher, i s not known. One of the basic assumptions incorporated in this thesis is that autonomous activation of the process of elaboration with the age of the 90. learner is the same as that indicated in Rohwer's study, (cf. p. 10) Such an assumption seems reasonable since elaboration seems to relate more with the individual's mental development through time rather than with the words themselves which provide the stimulus for e l i c i t i n g the process. Thus French words may replace English words without the elaboration process being affected. However, a study, such as that carried out by Rohwer, in which the paired-associates use French words rather than English would help to establish that the above assumption is well founded, and in so doing, open up decisive research related to French instructional methodology. This last suggestion for a research project probably holds high p r i o r i t y over a l l others. As Rohwer states the case: A strategy for gaining the information prerequisite to decisive research has been illustrated in terms of the case study. The strategy could be further improved by analyzing school subject tasks to reveal their essential psychological components, creating non-school tasks that are isomorphic with these components, and conducting coherent programs of research using the new tasks. The reason for insisting on non-school tasks, of course, is that school-subject tasks cannot be used in such research u n t i l the present lock on the timing of their introduction is opened. And that lock cannot be opened u n t i l we know enough to make a key. , , N (p. 121) For many years, in B r i t i s h Columbia, French instruction has been largely confined to secondary schools, starting at grade 8 level. Wow, in certain school di s t r i c t s , the teaching of French is beginning in the primary schools. Because so many students at various ages are now learning French, a study, using French paired-associates, could be more easily accomplished. 91. The research could determine the student age when optimal performance i s reached, using prompts of different degrees of explicitness. Such a study could be a f i r s t step in providing direction for the organization of French teaching methodology in different school grades in the province. 9 2 . FOOTNOTES 1 Paivio, A., 1 9 7 1 . pp. 73^77 2 Cited by Paivio, A., 1 9 7 1 . p. 3 Paivio, A., 1 9 7 1 . p. 51. ^Rohwer, W.D. (Jun.) 1 9 7 2 . p. 116 . ^Eidetic imagery, Chap. Three. 6 P. 7Paivio, A., Yuille, J.C., and Madigan, S.A., ( 1 9 6 8 ) . p. 5-King, D.J. ( i 9 6 0 ) , p. 1 13 . 9 Paivio and Foth ( 1 9 7 0 ) . pp. 3 8 4 - 3 9 0 . 9 3 . REFERENCES Anderson, R.C. "How to Construct Achievement Tests to Assess Comprehension," Review of Educational Research, 1 9 7 2 , II.. p 1 4 6 . Baddeley, A.D. "The Influence of Acoustic and. Semantic Similarity on Long Term Memory for Word Sequences, " Quarterly Journal of Experimental Psychology, I966, 1 8 . pp 3 0 2 - 3 0 9 Baddeley, A.D. and Dale, H.C.A.. "The Effect of Semantic Similarity on Retroactive Interference on Long and Short Term Memory," Journal of Verbal Learning and Verbal Behaviour, I 9 6 5 , 5« pp 4 1 7 - 4 2 0 Bower, G.H. "Mental Imagery and Associative Learning," F i f t h Annual Symposium on Cognition, Carnegie Mel. University, Pittsburg, Pennsylvania, 1 9 6 9 . Bugelski, B.F. Learning Theory and the Reading Process, University of Pittsburg Press, 1 9 6 9 . Erie, R. (cited by Werner and Kaplan, I963, 454-466). Holland, J.G. and Skinner, B.F. Analysis of Behavior, McQraw H i l l , New York, I96I. Huttenlocher, J. "Constructing Spatial Images: a Strategy in Reasoning," Psychological Review, 1 9 6 8 , 7 5 , 5 5 0 - 5 6 0 . King, D.J. "On the Accuracy of Written Recall; A Scaling and Factor Analytic Study," The Psychological Record, i 9 6 0 , 10 , 1 1 3 - 1 2 2 . King, D.J. and Russel, W.R. "A Comparison of Rote and Meaningful Learning of Connected Meaningful Material," Journal of Verbal Learning and Verbal BehaviorS, 1 9 6 6 , 4 7 8 - 4 8 3 . 3k. Kluver, H. "An Experimental Study of the Eidetic Type," Genetic Psychology Monographs, 1 9 2 6 , 1. Montagne, W.E. Adams, J.A. and Keiss, H.D. "Forgetting and Natural Language Mediation, " Journal of Experimental Psychology, 1 9 6 2 , v o l . 7 2 , p 829 Miller, G.A. "The Magical Number Seven, plus or Minus Two," Psychological Review, 1 9 5 6 , 6 3 , 8 I - 9 7 . Paivio, A. Imagery and Verbal Processes, Holt, Rinehart and Winston, 1971 , p. 5 1 . Paivio, A. and Begg, I. "Imagery and Comprehension Latencies as a Function of Sentence Concreteness and Structure," Research Bulletin No.154, Department of Psychology, University of Western Ontario, 1 9 7 0 . Paivio, A. and Foth, D. "Imaginal and Verbal Mediators and Noun Concreteness in Paired-Associate Learning: the Elusive Interaction," Journal of Verbal Learning and Verbal Behavior, 1970, 9- p 366 Paivio, A., Yuille, J.C. and Madigan, S.A. "Concreteness, Imagery and Meaningfulness Values, " Journal of Experimental Psychology, 1 9 6 8 , 7 6 , 1, part 2 . Rohwer, W.D. Jr. "Decisive Research; A Means for Answering Fundamental Questions about Instruction," Educational Research, 1972, I No. 1, 5 - 1 1 . Wallace, W.H. Turner, S.H. and Perkins, C.C. "Preliminary Studies of Human Information Storage," Signal Corps Project No. 1 3 2 0 , Institute for Cooperative Research, University of Pennsylvania, 1 9 5 7 . Werner, H. and Kaplan, B. Symbol Formation: An Organismic-Developmental Approach to the Psychology of Language and the Expression of Thought, 95. New York, Wiley, 1963. Yates, F.A. The Art of Memory, London, Routledge and Kegan Paul, I966, 2 4 - 2 5 . Yuille, J.C. and Paivio, A. 'Abstractness and Recall of Connected Discourse,' Journal of Experimental Psychology, 1969, 62. pp 467-471 96. Appendix A ( i ) . Pilot Study I Means (X), Standard Deviations (S.D.), and Sample Sizes (N), for the Immediate Post Test A.I. Group A.O. Group "Correct Word" "Content Word" "Correct Word" "Content Word" X = 48 .2 X = 34 .3 X = 53-9 X = 35.0 S.D. = 31.0 S.D. = I 6 . 9 S.D. = 33.0 S.D. = 18.0 N = 13 W = 13 N = 15 N = 15 97. Appendix A (ii).. Pilot Study I Means (x), Standard Deviations (S.D.) and Sample Sizes (N) for the Delay Post Test A.I. Group A.O. Group "Correct Word" "Content Word" . "Correct Word" "Content Word" Scores Scores Scores Scores X = 35.6 X = 23.0 X = If 1.6 X = 26.0 S.D. = 12.5 S.D. = 7.7 S.D. = 14.7 S.D. = 8.9 N=9 N=9 N = 11 N = 11 9 8 . Appendix A ( i i i ) . P i l ot Study I Before calculating an analysis of variance on the scores Bartlett's test was applied to ascertain whether there was any significant difference in the variances. Summary of Results of Bartlett's Test Bartlett Recall measures Statistic D.F. Chi. Sq. Sig "Correct Word" scores b = 3.4 V = 3 7 . 8 N.S. "Content Word" scores b = 2 . 8 V = 3 7 . 8 N.S. 9 9 . |Appendix A (iv) Pilot Study I Summary of Analysis of Variance of "Correct Word" Scores Source S.S. D.F. M.S. F. Sig. Treatments (T) 2 3 6 . 2 1 2 3 6 . 2 O .38 N.S. Immed.:Delay (P) 2 2 3 3 . 3 1 2 2 3 3 . 3 3 . 6 N.S. Interaction TxP 42.3 1 42.3 0 . 0 7 N.S. Error (within ce l l s ) 2 8 2 6 4 . 2 46 614 TOTAL 30776.0 1 0 0 . Appendix A (v). Pilot Study I Summary of Analysis of Variance of "Content Word" Scores Source S.S. D.F. M.S. F Sig Treatments (T ) 22.5 1 2 2 . 5 0 . 1 2 5 N.S. Immed'.): Delay (P) 1175-0 1 1 1 7 5 . 0 6 . 4 < 0 . 0 5 Interaction TxP 7 . 3 •1 • 7 . 3 0.04 N.S. Error (within cells) 8 4 2 1 . 7 46 1 8 3 . 0 TOTAL 9 6 2 6 . 5 49 X 1 0 1 . Appendix B ( i ) . Pilot Study II Means (X), Standar Deviations (S.D.) and Sample Sizes (N) of the Immediate Post Test W.W. Group Correct word Content word scores scores X = 1+9.4 X = 3 3 . 0 S.D. = 2 0 . 8 S.D. = 1 0 N = 1 3 N = 1 3 A.I. Group Correct word Content word scores scores X = 2 9 . 0 S.D. = 1 9 . 2 N = 1 4 X = 2 1 . 4 S.D. = 1 0 . 4 N = 1 4 1 0 2 . Appendix B ( i i ) . Pilot Study II Means (X), Standard Deviations (S.D.) and Sample Sizes (N) of the Delay Post Test A.I. Group "Correct word" "Content Word" scores scores W.W. Group 'Correct Word" "Content Word' scores scores X = 2 1 . 2 S.D. = 12.5 N = 13 X = 16.5 S.D. = 1 0 . 7 H = 13 X = 1 6 . 8 S.D. = 14.3 N = 11 X = 1 2 . 3 S.D. = 8 . 5 N = 11 1 0 3 . Appendix B ( i i i ) . Pilot Study II As in Pilot Study I Bartlett's test for the equality of variance was estimated. Summary of Results of Bartlett's Test Recall measures Bartlett's D.F. Chi. Sq. Sig. Statistic  Correct word scores b = 1 . 0 V = 3 7 . 8 N.S. Content word scores b = 0 . 5 V = 3 7 . 8 N.S. Since the variances in the subgroups for both measures were not significantly different the total scores were then subjected to analyses of variance (Appendix B iv, v). 104. Appendix B (iv ) . Pilot Study II Summary of Analysis of Variance of "Correct Word" Scores Source S.S. D.F. MS. F Sig. Treatments (T) 1158.0 1 1158.0 4 .2 P< 0.05 Immed.: delay (P) 4905.0 1 4905 .0 17.8 P < 0.05 Interaction TxP 1837.0 1 1837.0 6.7 P C 0 . 0 1 Error (within cells) 12973.0 47 276.0 TOTAL 20873.0 50 1 0 5 . Appendix B (v). Pilot Study II Summary of Analysis of Variance of "Content Word" Scores Source S.S. D.F. M.S. F Sig Treatments (T) 2 0 5 . 3 1 2 0 5 . 3 0 . 6 N.S. Immed.:delay (?) 1 7 9 3 . 2 1 1 7 9 3 . 2 5 . 6 p <" 0 Interaction TxP 8 4 0 . 5 1 8 4 0 . 5 2 . 7 N.S. Error (within c ells) 4 8 0 9 . 1 47 3 1 8 . 6 TOTAL 7 6 4 8 . 1 50 106. Appendix B ( v i ) . Pilot Study II The analysis of variance (Appendix B (iv)) indicates a significant interaction effect between the means of treatment x post test. Means (X) of Treatment x Post Test Interaction Cells Post test Treatments A^I. VMJ. Immediate X = 29.0 X = h9.k Delay X = 21.2 X = 16.8 107 Appendix B ( v i i ) . H o t Study II The Bonferroni calculation was used to determine the source of the interaction indicated i n the Table (Appendix B ( i v ) ) . S t a t i s t i c a l hypothesis: Null Ho : A.I. ^  W.M. Research Ho :.A.I. ;> W.W. (1 t a i l test) Benferroni Confidence Limits for Means Contrasts  "Correct Word" Scores Comparing Lower Upper Sig limit l i m i t A.I. immed. : W.W. immed. " + 5.4 ^ if/ ~ 0 0 p <c 0.05 A.I. delau : W.W. delay - 11.6 ^ ^ — ® ° N.S. The n u l l hypothesis (Ho) i s rejected 1 0 8 . Appendix C ( i ) . Pilot Study III Means (X), Standard Deviations (S.D.) and Sample Size (W) of the Immediate Post Test A.I. Group W.W. Group "Correct word" "Content word" "Correct word" "Content word" scores scores scores scores  X = 1+1.9 X = 2 7 . 0 X = 3 4 . 1 X = 2 1 . 5 S.D. = 2 0 . 8 S.D. = 1 0 . 1 S.D. = 1 5 . 2 S.D. = 7-5 N = 18 N = 18 N = 19 N = 19 1 0 9 . Appendix C ( i i ) . Pilot Study III Means (X), Standard Deviations (S.D.) and Sample Size (N) of the Delay Post Test A.I. Group W.W. Group "Correct Word" "Content Word" "Correct Word" "Content Word" scores scores scores scores  X = 2 3 . 6 X = 1 7 . 5 X = 1 8 . 0 X = 1 3 . 0 S.D. = 1 7 . 0 S.D. = 8 . 9 S.D. = 1 2 . 8 S.D. = 7 . 0 N = 13 N = 13 N = 15 N = 15 110. Appendix C ( i i i ) . Pilot Study III Bartlett's test for the equality of variance was calculated for the subgroups in this study. Summary of Results of Bartlett's Test Recall measure Bartlett's D.F. Chi.Sq. Sig st a t i s t i c  Correct word scores b = 0.5 V = 3 7 . 8 N.S. Content word scores b = 2 . 8 \) = 3 7 . 8 N.S. Since the variances in the subgroups for both measures were not significantly different the total scores were then subjected to analysis of variance (Appendix C (iv)). 111 . Appendix C (iv). Pilot Study III Summary of Analysis of Variance of "Correct Word" Scores Source S.S. D.F. M.S. F Sig Treatments (T) 7 7 4 . 3 1 7 7 4 . 3 2 . 7 N.S. Immed.:delay (P) 4 5 5 9 . 3 1 4 5 5 9 . 3 1 6 . 4 p C 0 . 0 1 Interaction TxP 5 6 . 1 1 5 6 . 1 0 . 2 N.S. Error (within c ells) 1 6 9 2 6 . 7 61 2 7 7 . 5 TOTAL 2 2 3 1 6 . 4 64 112. Appendix C (v). Pilot Study III Summary of Analysis of Variance 1 of "Content Word" Scores Source S.S. D.F. M.S. F Sig Treatments (T) 433.1 1 433.1 6 .1 p < 0 .05 Immed.:delay (P) 1275.9 1 1275.9 18.0 p< 0 . 0 1 Interaction TxP 30.0 1 30.0 0 . 4 N.S. Error (-within cells) 4341.4 61 71.2 TOTAL 6080.4 64 Means (X), Standard Deviations (S.D.) and Sample Size (N) of Post Test Results in the Two Schools for "Correct Word" and "Content Word" Measure (-A.I.=Attention to Imagery. W.W. = Word for Word) Semiahmoo School Princess Margaret School Words Correct Scores Content Word Scores Words Correct Scores Content Word Scores A.I. W.W. A.I. W.W. A.I. W.W. A.I. W.W. Immediate x = 29.0 X = 49.4 X = 21.4 X = 33.0 X = 41.9 X = 34.1 x = 27.0 X = 21. 5 POst S.D.=19.2 S.D. =20.8 S.D.=10.4 S.D.=10.1 S.D.=20.8 S.D. =15.2 S.D.=10.1 S.D.-7.5 Test N = 14 N = 13 N = 14 N = 13 H = 18 N = 19 N = 18 N = 19 Delay X = 21.0 X = 16.8 X = 17.5 x = 12.3 x = 23.7 x = 18.3 x = 19.5 x = 13.0 Post S.D. =12. 5 S.D.=14.3 S.D.=10.7 S.D.=8.5 S.D.=l6.9 S.D.=12.8 S.D.=8.9 S.D.=7.0 Test N = 13 N = 11 N = 13 N = 11 N = 13 N = 15 N = 13 N = 15 114. Appendix D ( i l ) Before analysing the scores of pilot studies I and II Bartlett's test for the equality of variance was estimated. Summary of Results of Bartlett's Test Bartlett Recall Measures Statistic D.F. Chi.Sq. Sig. "Correct Word" scores J> = 2 . 2 V =7 1 8 . 5 N.S. "Content Word" scores o = 1 . 3 \) =7 14.1 N.S. 115 . Appendix D ( i i i ) Summary of Analysis of Variance for "Correct Word" Scores for Immediate and Delay Tests (both Schools) Source S.S. D.F. M.S. F Sig Schools (s) 1 6 . 0 1 1 6 . 0 0 . 1 N.S. Treatments (T) 1 7 5 . 0 1 1 7 5 . 0 0 . 6 N.S. Immed.:delay (P) 9 4 3 8 . 0 1 9 4 3 8 . 0 3 4 . 0 p<" 0 . 0 0 1 Interaction SxT 1747 .0 1 1747 .0 6 . 2 p < 0 . 0 5 M SxP 3 7 . 0 1 3 7 . 0 0 . 1 N.S. ri TxP 423 . 0 1 4 2 3 . 0 1 .5 N.S. SxTxP 1 2 7 5 . 0 1 1 2 7 5 . 0 4 . 6 p < 0 . 0 5 Error (within cells) 3 0 2 4 6 . 0 108 2 8 0 . 0 TOTAL 4 3 3 5 7 . 0 115 l i b . Appendix D (iv) Summary of Analysis of Variance for "Content Word" Scores for Immediate and Delay Tests (both Schools) Source S.S. D.F. M.S. F Sig. Schools (S) 3 9 . 3 l 3 9 - 3 0 . 1 N.S. Treatments (T) 40.7 1 40.7 0 . 1 N.S. Immed.: delay (P) 2 7 4 8 . 6 l 2 7 4 8 . 6 3 4 . 4 p«c 0 . 0 0 1 Interaction SxT 6 7 7 . 2 1 6 7 7 . 2 8 . 5 p< 0 . 0 1 SxP 3 2 0 . 2 1 3 2 0 . 2 4 . 0 p< 0 . 0 5 II TxP 5 3 6 . 5 1 5 3 6 . 5 6 . 7 p < 0 . 0 5 SxTxP 1 7 6 0 . 0 1 1 7 6 0 . 0 2 2 . 0 p< 0 . 0 0 1 Error (within cells ) 8 6 3 5 . 0 " 108 7 9 . 0 TOTAL 14757-5 115 117. Appendix E (l) INSTRUCTIONS: You are to get a test which w i l l include these French words and expressions. You must know their meaning. FRENCH ENGLISH chou cabbage chanter to sing rire to laugh bavarder to gossip entrer to enter chateau castle monter to go up proteger to protect dormir to sleep garcon boy juge judge dame lady Allemand German correspondant correspondent ciseaux scissors jouer to play coiffeur hairdresser aimer to like europeen European habiter to live cultiver to grow donnerc to give troisieme third odeur smell beaucoup much (lots) apres after programme programme contre against journee day chef chef aviateur flyer merci thank you avancer to advance manger to eat robe dress cuire to cook ocean ocean animal animal Appendix E cont'd. pupitre classe argent belle nuit pluie bon cage vert fleur rester sonner village prince fermier pays etage pendant bombe prisonnier li v r e parapluie paysanne veut chi en viande laisser t r i s t e ascenseur touriste porter enfant repondre toute beau reine plage ours tarte etranger mere coq chat tomber telephone boulanger lettre eleve desk class money beautiful night rain good cage green flower to stay to sound (to ring) village prince farmer country floor (storey) during bomb prisoner book umbrella peasant (f) wants dog meat to let sad elevator tourist to wear child to reply a l l handsome queen beach bear pie foreigner mother rooster cat to f a l l telephone baker letter pupil Appendix E (i) cont'd.. EXPRESSIONS loin de far from chez moi at my home avoir envie to want (to long for) beaucoup de lots of au s o l e i l in the sun faire semblant to pretend a* l a television on the television a l l e r chercher to go for veut dire to mean rentrer to go home aimer mieux to prefer de bonne heure early avoir l ' a i r to look faire attention to pay attention 1 2 0 . Appendix E ( i i ) List of Thirty Sentences 1. L'ours reste loin du village. 2 . La reine a l ' a i r t r i s t e . 3 . Le boulanger a beaucoup d'argent. k. Le prisonnier f a i t semblant de dormir. 5 . Le prince est un beau garcon. 6. L'e"leVe rentre apres l a classe. 7 . La fermier cultive de bons choux. 8 . Le coq. chante de bonne heure. 9 . L'ascenseur monte au troisieme etage. 1 0 . Le juge a envie de r i r e . 1 1 . La fleur donne une belle odeur. 1 2 . Le programme est et l a television. 1 3 . Le parapluie protlge contre l a pluie. Ik. La paysanne aime mieux bavarder beaucoup. 1 5 . Le telephone sonne pendant la nuit. 1 6 . Le touriste entre dans le chateau. 17 . Le l i v r e est sur le pupitre. 1 8 . Le chef f a i t cuire des tartes. 1 9 . La dame porte une robe verte. 2 0 . L'aviateur laisse tomber l a bombe. 2 1 . L'animal dort dans l a cage. 2 2 . L'allemand veut habiter chez nous. 2 3 . La France est un pays europeen. 2k. Le chat aime rester au s o l e i l . 2 5 . La-, correspondante repond a* la lettre 2 6 . L'ocean avance sur l a plage. 2 7 . L'etranger veut dire merci beaucoup. 2 8 . Le coiffeur va chercher les ciseaux. 2 9 . Le chien mange toute l a viande. 3 0 . L'enfant f a i t attention aux instructions. 121. Appendix E ( i i i ) Learning Instructions There w i l l be a test following the presentation of 30 sentences, on the overhead projector, to see i f you can re c a l l them. The test w i l l be one requiring exact wording of the sentences. Write each sentence on the sheet provided and try to remember them "word for word". 122. Appendix E (iv) Learning Instructions There -will be a test following the presentation of 30 sentences, on the overhead projector, to see i f you can r e c a l l them. The test w i l l be one requiring exact wording of the sentences. You should learn the way you usually do. 1 2 3 . Appendix E (v) Learning Instructions There w i l l he a test following the presentation of 3 0 sentences, on the overhead projector, to see i f you can rec a l l them. The test w i l l he one requiring exact wording of the sentences. Buts, as you are writing each sentence, try to picture in your mind, as clearly as possible, the image or picture that the sentence i s trying to convey to you. Draw a l i t t l e picture of the meaning of each sentence on the sheet  provided to help you to make the sentence more vivid in your mind. 12k. Appendix E (vi) Date: FRENCH. / NAME: Predicate. Below you are given the f i r s t two words of each of the sentences which you have just seen. Complete the sentence by writing in the remaining four French words. 1. L'ours 2 . Le boulanger 3 . La reine k. Le prisonnier 5 . Le prince 6 . L' e*leve 7. Le fermier 8 / Le coq 9 . L'ascenseur 1 0 . Le juge l i La fleur 1 2 . Le programme 1 3 . Le parapluie ik. La paysanne 1 5 . Le telephone 16 . Le touriste 17. Le l i v r e 1 8 . Le chef 19- La dame 2 0 . L'aviateur 2 1 . L'animal 2 2 . L'Allemand 2 3 . La France 2k. Le chat 2 5 . La correspondent 2 6 . L'ocean 2 7 . L'etranger 2 8 . Lecoiffeur 2 9 . Le chien 3 0 . L'enfant 126. Appendix E (vii) NAME: DATE: How many of these following puzzles can you get correct? You have to rearrange the letters to make a French word. Example: TNSO Answer: SONT If one word seems to be d i f f i c u l t , try another one and return to the d i f f i c u l t one later. 1. LI 2. IAVS 3. ONN '  4. URM 5. OIICV 6. ESRUO 7. EBNI 8. SUASI 9 . MIA 10. TOEPR 11. NPIA 12. CNERT 13. SONLA 14. OSINAM 15. OERCUUL 1 2 7 . Appendix E ( v i i i ) Draw a picture of the sentence meaning in the spaces below. There are more spaces over page. NAME: 128. Appendix E ( v i i i ) cont'd... Appendix E ( v i i i ) cont'd.. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0054624/manifest

Comment

Related Items