Journal of Literacy Research

0 downloads 0 Views 932KB Size Report
positive effect on reading comprehension when compared to the dictionary usage and word ..... up practice exercises involving worksheets, games, and/or team competition. The ... used when the condition of compound symmetry was met.
Journal ofhttp://jlr.sagepub.com/ Literacy Research

Enhancing Children's Reading Comprehension through Instruction in Narrative Structure Jill Fitzgerald and Dixie Lee Spiegel Journal of Literacy Research 1983 15: 1 DOI: 10.1080/10862968309547480 The online version of this article can be found at: http://jlr.sagepub.com/content/15/2/1

Published by: http://www.sagepublications.com

On behalf of: Literary Research Association

Additional services and information for Journal of Literacy Research can be found at: Email Alerts: http://jlr.sagepub.com/cgi/alerts Subscriptions: http://jlr.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://jlr.sagepub.com/content/15/2/1.refs.html

>> Version of Record - Jun 1, 1983 What is This?

Downloaded from jlr.sagepub.com by guest on October 11, 2013

Journal of Reading Behavior 1983, Volume XV, No. 2

ENHANCING CHILDREN'S READING COMPREHENSION THROUGH INSTRUCTION IN NARRATIVE STRUCTURE

Jill Fitzgerald and Dixie Lee Spiegel University of North Carolina, School of Education, Chapel Hill, NC 27514

Abstract. Twenty average and below average fourth grade readers identified as lacking a keen sense of narrative structure were randomly assigned to one of two treatments: special instruction designed to develop knowledge of story structure or instruction in dictionary usage and word study. There were two phases of instruction, a short-term intensive phase and a long-term intermittent phase. The instruction in narrative structure did enhance story structure knowledge and had a strong positive effect on reading comprehension when compared to the dictionary usage and word study group. The effects were realized by the end of phase one; they were maintained during phase two, but the group differences did not increase over time.

Narrative text structure appears to be an important variable in individuals' cognitive processing of prose. Many adults and children appear to use knowledge of structural features of narratives to understand and remember stories (cf., Kintsch, 1977; Mandler & Johnson, 1977; Rumelhart, 1978; Stein, 1979; Stein, Note 1; Stein & Glenn, Note 2) and to anticipate forthcoming information (cf., Applebee, 1978; Trabasso, Stein, & Johnson, in press; Whaley, 1981). Knowledge of narrative structure, sometimes referred to as story schema, involves identification of narrative elements and their interrelationships (cf., Mandler & Johnson, 1977; Rumelhart, 1978; Stein & Glenn, 1979; Thorndyke, 1977). Individuals' knowledge of narrative structure may be represented by a story grammar. The authors wish to thank Cheryl Knight for her assistance in gathering the data and for serving as one of the instructors for the study. The authors also wish to thank the fourth grade teachers, children, and principal of Little River Elementary School, Bahama, North Carolina. 1

Journal of Reading Behavior

For example, Mandler and Johnson's story grammar (Johnson & Mandler, 1980; Mandler & Johnson, 1977) describes six major categories of narrative information: Setting, Beginning (a precipitating event), Reaction (the protagonist's Reaction and setting a Goal), Attempt (the effort to achieve the Goal), Outcome (the success or failure of the Attempt), and Ending (the long-range consequence of the action sequence or the added emphasis). The Beginning through the Ending make up an Episode. Rules in the grammar specify temporal relationships between categories and delineate how complex stories can occur through options such as embedding of episodes. Three findings from investigations of individuals' knowledge of structural facets of stories are particularly relevant to the present investigation. First, studies have indicated that many individuals do have a sense of an idealized story, which involves knowledge of structural features (Applebee, 1978; Botvin & Sutton-Smith, 1977; Mandler, 1978; Mandler & Johnson, 1977; Rumelhart, 1978; Stein, 1979; Thorndyke, 1977; Trabasso et al., in press; Glenn & Stein, Note 3; Stein, Note 1; Stein & Glenn, Note 4). Second, there are developmental trends in the acquisition of story schema, with children's story knowledge becoming richer and more elaborate with age (cf., Applebee, 1978; Botvin & Sutton-Smith, 1977; McClure, Mason, & Barnitz, 1979; Sutton-Smith, Botvin, & Mahoney, 1976; Glenn & Stein, Note 3; Stein & Glenn, Note 2). Developmental differences can be specified in at least three areas. First, when text is disorganized, younger children's accuracy of recall deteriorates substantially in comparison to older individuals (cf., Buss, Yussen, Mathews, Miller & Rembold, in press; Mandler, 1978), suggesting that they are less able to impose order on the text as a means of facilitating recall. Next, some children, particularly younger children, may lack knowledge of specific story constituents; and finally, some may also be unaware of appropriate temporal relationships between particular story constituents (cf., Stein, in press; Stein & Glenn, in press; Stein & Policastro, in press; Glenn & Stein, Note 3). For example, when asked to produce stories, younger children more frequently generate responses that lack critical information in story episodes and that lack causal links than do older subjects (-Glenn & Stein, Note 3). Also, when critical information is either deleted from text presented to children (cf., Stein & Glenn, Note 4) or when the temporal sequence is disrupted (cf., Buss et al., in press; Stein & Glenn, in press), younger children are less able than older ones to restructure the text into an appropriate temporal order. The third and final relevant finding from prior studies is that some minimal evidence indicates that as compared to poorer readers, better readers show greater sensitivity to structural features of stories (cf., Fitzgerald, in press; Smiley, Oakley, Worthen, Campione, & Brown, 1977; Vipond, 1980; Weisberg, 1978). It is possible that better readers use knowledge of story constituents and constituent order to guide and direct their comprehension, and to permit better inferencing when specific content is later forgotten.

Enhancing Comprehension

To date at least two assertions about acquisition of knowledge about stories have not been rigorously explored. First, it may be possible and even necessary to enhance children's knowledge of story components and their interrelationships through direct instruction. Second, if knowledge of structural aspects of narratives is highly related to comprehension, then enhancing children's story knowledge may also improve their comprehension of narratives when reading. Studies involving instruction in narrative structural characteristics are rare, and among those conducted, results appear to be conflicting, with some studies finding an effect for instruction and others not (Buss et al., in press; Dreher & Singer, 1980; Singer & Donlan, 1982). The present study was designed to further investigate the possibility that direct instruction in story constituents and their relationships would enhance children's knowledge of story structure as well as their reading comprehension of narrative text. Twenty average and below average fourth grade readers identified as lacking a keen sense of narrative structure were randomly assigned to one of two treatments, special instruction designed to develop knowledge of story structure or instruction in dictionary usage and word study. There were two phases of instruction, a shortterm intensive phase and a long-term intermittent phase. Two phases were set up to see (1) what changes, if any, could be effected with a relatively short period of instruction, as opposed to a longer period of instruction, and (2) if any initial effects could be maintained and/or enhanced through additional instruction.

METHOD Subjects From a pool of 100 fourth grade students in one school who returned parental consent forms, the 50 lowest ranking readers (identified by percentile on the California Achievement Test [CTB/McGraw-Hill staff, 1977] and by teacher judgment) were screened to determine the extent of their knowledge of story structure. (Screening measures are described below.) From the pool of 50 poorest readers, the final 20 students (7 males and 13 females) identified as having the least developed knowledge of story structure were randomly assigned to treatment group, with 10 in each group. One subject was withdrawn by parental request at the end of the first phase of instruction; all of the results reported were obtained with the remaining 19 subjects. Instruments and Scoring Procedures Knowledge of story structure. To assess knowledge of story structure during screening and at interim and final testing, two tasks utilized in prior studies to reveal individuals' knowledge of story structure were used: a story production task

Journal of Reading Behavior

(cf., Applebee, 1978; Botvin & Sutton-Smith, 1977, Leondar, 1977; Sutton-Smith et al., 1976; Trabasso et al., in press; Glenn & Stein, Note 3; Stein & Glenn, Note 2) and a scrambled story recall task (cf., Buss et al., in press; Kintsch, Mandel, & Kozminsky, 1977; Mandler, 1978; Stein & Glenn, in press). These two tasks were selected because results of prior research indicated that they were useful for revealing developmental differences. Therefore, it was likely that they would also be sensitive to individual differences within a given age level. For each task, six stories or settings were developed, two each for screening, interim (end of phase one), and final (end of phase two) testing. For screening, students were tested individually, and oral responses were tape recorded and later transcribed. For interim and final testing, group-administered, written measures were used. For the production task, subjects were instructed to read a story setting and finish the story, making it into the best story they could. They were also told to ask the examiner for help on any words they didn't know when reading and not to worry about spelling when writing. A standard set of prompt statements, such as "I know it's hard, but I'd like you to try your best," was used. Approximately 15 to 20 minutes were allotted for each story. All children finished within the time limit. Individuals' knowledge of story structure was revealed by the complexity of their narrative productions. Responses were scored using a modification of a system developed by Stein and Glenn (Note 2).1 The following five degrees of complexity were scored 1 to 5 points, respectively: (1) descriptive sequence (no clear Goal; causal connections between statements rare); (2) reactive sequence (no clear Goal; causal connections between several statements); (3) abbreviated Episode (Goal stated explicitly or easily inferred; certain categories, but not all categories, stated [e.g., a Beginning, implied Goal, Outcome sequence would be an abbreviated Episode]); (4) complete Episode (Goal stated explicitly or easily inferred; all categories other than Goal and Ending explicitly stated); and (5) complex story (Goal stated explicitly or easily inferred; multiple Episodes present). Interrater reliability between the two investigators for scoring the production tasks ranged from .73 to .93 for the six stories. A supplementary category adjacency analysis was also carried out for the production task. Individuals' knowledge of temporal relationships between two specific categories was revealed by a conditional analysis in which each child's protocols were examined to see if the second category predicted by the grammar was immediately produced, given that the first member of the pair was present (e.g., given that an Attempt was present, did an Outcome follow?). Two judges separately determined the presence/absence of category adjacencies for each protocol. Agreement on this task proved difficult, and disagreements were resolved through discussion (cf., Mandler & Johnson, 1977, for similar procedures). 1

Copies of the scoring system are available from the authors upon request.

Enhancing Comprehension

For the second knowledge of story structure task, subjects were instructed to read and recall scrambled stories written at the second grade level according to the Fry (1977) readability formula. Instructions stressed that the stories the children read were all mixed up and that their task was to try to remember everything they could and put the pieces back in order. The children were also told to ask the examiner for help on any unknown words while reading and to read the story straight through once. Again, a standard set of prompt statements was used. After retelling the story 24 hours later, the children were asked if they had talked to anyone about the story. None indicated that they had done so. Three of the stories were modified versions of ones used in previous research ("The Unlucky Twins" [Mandler, 1978], "The Wolf and the Crane" [Rumelhart, 1978], and "The Fox and the Bear" [Stein & Glenn, 1979]). The remaining stories were devised by one of the investigators using the structures of the other three stories as guides so that parallel structures were presented at each testing time. Interrator reliability between the two investigators for parsing the original six stories ranged from .94 to 1.00. One of the stories is shown in Table 1. At each testing time, one of the scrambled stories was a "then-connected" two-Episode story which was interleaved in the following manner: Beginning from Episode one, then Beginning from Episode two; Reaction from Episode one, then two, respectively; Attempt from Episode one, then two, and so on (cf., Mandler, 1978). The second scrambled story at each testing time was a two-Episode story in which there was an embedded Outcome Episode. The embedded Outcome stories were all scrambled using the same random order. The two types of scrambling, one interleaved and one random, were used in order to represent a wider array of stimuli types. Individuals' knowledge of story structure was revealed by the degree to which subjects reordered the stories. Kendall's Tau rank order correlations were computed for each story for each subject (cf., Buss et al., in press; Stein & Glenn, in press; Stein & Nezworski, 1978). The Kendall's Tau scores indicated the concordance between the order of categories recalled and the original order of the categories in the canonical version of the story. Interrater reliability between the two investigators for scoring the scrambled story retellings ranged from .70 to 1.00 for the six stories. Individuals' knowledge of story constituents was revealed in three additional ways on the scrambled story task. Subjects with greater knowledge of story parts should be able to use that knowledge to guide recall, resulting in greater accuracy of recall. When forgetting does occur, individuals with greater knowledge of story parts should use that knowledge to attempt to make up forgotten material. Such attempts would be reflected in importation of new material and distortion of old material. Therefore, protocols were also scored for: (1) presence or absence in recall of statements from the stories read, (2) instances of importation of new mate-

Journal of Reading Behavior

TABLE 1 A Story Used in the Scrambled Story Recall Task3 Original Version The Wolf and the Bird With the Long Neck Setting Beginning Reaction Goal Attempt

S

Beginning Attempt

8

Outcome Ending

1

I

A Wolf was eating an animal he had killed Suddenly, a small bone in the meat stuck in his throat. He could not swallow it. He soon felt terrible pain in his throat. He wanted to stop the pain. He tried to get everyone he met to take the bone out for him. "I would give anything," he said, "If you would take it out." At last a Bird with a long neck said he would try. The Bird told the Wolf to lie down on his side. He had him open his mouth as wide as he could. Then the Bird put its long neck down the Wolfs throat. With its beak the Bird pulled on the bone. At last the Bird got the bone out. "Will you please give me the reward you said I could have?" said the Bird with the long neck: The Wolf grinned. He showed his teeth and said: "Be happy. You have put your head inside a Wolfs mouth and taken it out again safely. That is all the reward you will get."

Scrambled Version The Wolf and the Bird with the Long Neck He soon felt terrible pain in his throat. He wanted to stop the pain. Suddenly, a small bone in the meat stuck in his throat. He could not swallow it. The Bird told the Wolf to lie down on his side. He had him open his mouth as w de as he could. Then the Bird put its long neck down the Wolfs throat. With his beak, the Bird pulled on the bone. At last the Bird got the bone out. At last a Bird with a long neck said he would try. "Will you please give me the reward you said I could have?" said the Bird with a long neck. The Wolf grinned. He showed his teeth and said: "Be happy. You have put your head inside a Wolfs mouth and taken it out again safely. That is all the reward you will get." A Wolf was eating an animal he had killed. He tried to get everyone he met to take the bone out for him. "I would give anything," he said, "if you would take it out." "This story is a modified version of a story used by Rumelhart (1978).

rial, and (3) instances of distortion of the original material. Decisions were made by comparing statements in recall protocols to statements in the original stories (cf., Mandler & Johnson, 1977). For each of the three supplementary analyses, two judges separately scored the protocols. Agreement on this task proved difficult, and disagreements were resolved through discussion (cf., Mandler & Johnson, 1977, for similar procedures).

Enhancing Comprehension

Comprehension. To assess story comprehension at interim and final testing, students read stories that were approximately 400 words long and answered nine literal and eight inferential questions for each story. At each testing session, the students read two stories. They were told to read each story once silently and to ask for help on any unknown words. They were also told that after they finished reading, they would be asked to write answers to some written questions without looking back at the story. Four stories were used. The stories, all taken from a basal series, Scott, Foresman Basics in Reading (Aaron, Jackson, Riggs, Smith, & Tierney, 1978), never before used in the state, were rewritten at the mid-second grade level, according to the Fry (1977) readability formula. The stories were parsed to see if they were well-formed according to the Mandler and Johnson grammar (Johnson & Mandler, 1980; Mandler & Johnson, 1977). Interrater reliability between the two investigators for parsing the stories ranged from .78 to .96. The questions were devised using procedures developed by Bormuth (1968) and by Hansen (1981). Responses to the questions were scored allowing partial credit (scores 0 to 3).2 Alpha coefficients for estimates of reliability for the scores for the four stories ranged from .71 to .89. Interrater reliability between the two investigators ranged from .84 to .92. Design The study employed a two-group repeated measures design. One group received instruction in schematic features of narratives while the second group, a control group, received instruction primarily in dictionary usage and, to a limited extent, in word study. So that potentially intervening variables could be controlled, the following factors were equated for the two groups: (1) instructional time, (2) amount of individual versus group work, (3) degree to which the lessons were structured, (4) broad content area of instruction (language arts), and (5) amount of practice reading and writing stories. In addition, the same stories were presented to both groups. Subjects received two phases of instruction, differentiated primarily by intensity and duration. The first phase was short-term intensive instruction, consisting of six 30- to 45-minute sessions conducted over a two-week period. The second phase was long-term intermittent instruction with distributed practice, consisting of ten 30- to 45-minute sessions over a five-week period. Two tasks designed to measure story structure knowledge (production and scrambled story recall) were administered as screening tests, at the end of the first phase of instruction (interim test), and again at the conclusion of the study (final tests); and the comprehension measures were administered at interim and final test2

Copies of the scoring system are available from the authors upon request.

Journal of Reading Behavior

ing times. There were two stories for each task. Thus, each subject had four screening scores, six interim scores, and six final scores. At each testing time, measures were counterbalanced within and across tasks. Test time was treated as a repeated measure despite the fact that different stories were used at different test times. Instruction The two investigators and a trained doctoral student taught both treatment groups. Each group met with each of the three instructors an approximately equal number of times. The instruction for the story structure group centered around two purposes. The primary purpose was to help the children to form a sort of "cognitive blueprint" or structure for stories and to elaborate the "blueprint" (cf., Stein & Trabasso, in press (b); Rubin, Note 5). The secondary purpose was to increase the children's awareness of the ways in which knowing the structure of stories can help them understand stories in general (cf., Collins & Smith, in press; Brown, Campione, & Day, Note 6). As noted, there were two phases to the instruction. During phase one of instruction, each lesson focused on one story constituent and its temporal relationship to other story parts. In a typical phase one lesson, the instructor first gave a review of the story parts learned in previous lessons and an overview of the new lesson. Next, the instructor told about the story element (e.g., Attempt) by describing it, pointing out the element in a story on a wall chart, and giving two or three other examples of the element that would be appropriate for the story on the chart. Then, the instructor elicited two or three oral examples of the story element from the children. Next, the instructor gave non-examples and asked why these were not good examples of the element being studied. The non-examples might be different story parts (e.g., an Outcome for an Attempt), or they might be the right story part but misplaced within the story. Last, the students participated in one or more group or individual activities designed to reinforce understanding of the element being taught that day. Examples of reinforcement activities are prediction and macrocloze tasks. Far prediction tasks, the students read part of a story (e.g., Setting and Beginning), told what part should come next, and provided an appropriate example of that part. In the macro-cloze activities, the children filled in missing chunks of stories after reading or listening to an entire story. For example, a story might contain a Setting, Beginning, Reaction, Outcome, and Ending. After reading the story, the student would identify what part was missing and supply an appropriate Attempt. Phase two instruction consisted of individual and group activities designed to provide continued reinforcement of knowledge of story constituents and to make the children aware of the relationship between comprehension of a story and knowledge of specific story parts and their temporal relationships. Examples of

Enhancing Comprehension

additional activities used to reinforce knowledge of story constituents during phase two are scrambled story tasks, sorting tasks, and retelling activities. The scrambled story tasks required students to reorder stories in which the story had been jumbled. For the sorting tasks, the students sorted sentences and phrases from the story into piles to show which pieces "went together." They then ordered the piles to make a well-sequenced story. In the retelling activities, the children recalled stories they had just read or heard and discussed ways in which the stories deviated from well-formed stories, such as through omission of necessary elements or additions of extra elements. In each of the reinforcement activities, there were at least two or three variations for individual and small group work. To emphasize the relationship between thinking about story parts and understanding a story, students read stories, identified the story parts, and then asked and answered questions designed to tap information from each story part. When answering questions, they would identify which story part contained that answer. The instructor stressed that knowing about the story parts could lead to better understanding of and memory for the story. Instruction for the control group centered on three main dictionary skills: using phonemic respellings to determine the pronunciation of a word, using a dictionary to identify the correct spelling of an inflected word, and choosing a meaning for a word from the choice listed in the dictionary. A typical lesson plan included the following elements: (1) an overview and review were provided; (2) the instructor introduced the new skill for the day; (3) the instructor led the children through the use of the skill in a step by step manner; and (4) the children participated in followup practice exercises involving worksheets, games, and/or team competition. The stories used for instruction with the story structure group were read by the children in the control group in many of the follow-up exercises. Similarly, practice in writing stories was included in many of the follow-up exercises through activities such as writing stories incorporating the words introduced earlier in the lesson. Analyses The major analyses were multivariate and univariate repeated measures analyses of variance. Procedures outlined in Bock (1975) and Finn and Mattsson (1978) were followed. For all repeated measures analyses, preliminary tests of compound symmetry were done to determine whether the repeated measures analyses were performed as mixed models or not. Depending on the number of dependent variables, a univariate or multivariate mixed model repeated measures analysis was used when the condition of compound symmetry was met. A multivariate repeated measures model was used when the condition of compound symmetry was not met. Pre-treatment group differences on all knowledge of story structure measures and on California Achievement Test scores were assessed prior to testing posttreatment effects. There were no pre-treatment group differences on any of the measures. Group means and standard deviations are given in Table 2.

10

Journal of Reading Behavior

TABLE 2 Group Means (and Standard Deviations) on All Screening Measures

Group Story structure Dictionaryword Study

Production8 Story Total Across 1 2 Stories

Scrambled" Story Total Across 1 2 Stories

California Achievement Test reading percentiles

1.60 1.60 ( .84) ( .97)

3.20

.27 .41 ( .44) ( .47)

.68

45.80 (11.40)

1.44 1.44 ( .73) ( .73)

2.88

.39 .11 ( .42) ( .33)

.50

42.20 (12.18)

"Scores could range from 1 to 5. "Kendall's Tau Rank Order Correlations: Scores could range from 1.00 to -1.00.

Five sets of analyses were done to determine instructional effects on the children's knowledge of story structure. First, overall knowledge of story structure was examined using the production scores and the scrambled story rank order correlation scores as a multivariate set. Second, awareness of temporal relationships between story category pairs in the production task protocols was examined. This was accomplished by calculating the proportion of occasions that the second category predicted by the grammar was immediately produced, given that the first member of the pair was provided. Third and fourth, based on the hypothesis that individuals with greater knowledge of story parts should use that knowledge to attempt to reconstruct forgotten material, instances of minor distortions and then of additions of information in the scrambled story task were examined. Fifth, postulating that subjects with greater knowledge of story parts should use that knowledge to recall information more accurately, accuracy of recall in the scrambled stories was assessed. Finally, scores for the 17-item comprehension tests were analyzed to evaluate the instructional effects on story comprehension.

RESULTS Instructional Effects on Knowledge of Story Structure Overall knowledge of story structure. To assess the instructional effects on overall knowledge of story structure, a multivariate mixed model repeated measures analysis was done in which the between factor was treatment (experimental and dictionary-word study). The within factors were time (interim and final) and story

Enhancing Comprehension

11

within time (two for each dependent variable at each time). The dependent variables were production scores and scrambled story Kendall's Tau scores. There was an overall significant difference between treatment groups (multivariate F [2, 16] = 5.31, p < .02) favoring the story structure group. Means for each interim and final measure are shown in Table 3. Follow-up univariate tests indicated the difference between groups was primarily attributable to a difference on the production tasks. (Univariate F's for the production and scrambled story tasks, respectively, were: F [1, 17] = 11.25,p < .01;andF[l, 17] = 2.88,p = .11.) There were no other significant effects. TABLE 3 Group Means (and Standard Deviations) on all Interim and Final Measures Interim Group Story structure Dictionaryword Study

Group Story structure Dictionaryword Study

Production" Story Total Across 1 2 Both Stories 3.60 2.80 ( -97) (1.32)

6.40

2.11 1.67 (1.45) ( -71)

3.78

Comprehension0 Story Total Across 1 2 Both Stories

Scrambled" Story Total Across 2 Both Stories 1 .71

.37

1.08

31.10 33.00 (5.43) (6.94)

64.10

.64

17.00 27.56 (6.36)(11.62)

44.56

(.40) (.49) .31

.33

(-47) (.49)

Production3 Story Total Across 1 2 Both Stories 3.60 3.40 (1.17) (1.26)

7.00

2.33 2.11 (1.41) (1.45)

4.44

Final

Comprehension0 Story Total Across 1 2 Both Stories

Scrambled» Total Across Story 1 2 Both Stories .37

.80

1.25

34.50 36.00 (4.06) (9.04)

70.50

.70

26.78 23.44 (10.69X11.83)

50.22

(•48) (.42) .48

.22

(.50) (.44)

•Scores could range from 1 to 5. "•Kendall's Tau Rank Order Correlations: Scores could range from 1.00 to +1.00. "Scores could range from 0 to 51.

Adjacency analysis for production task. To determine whether there were treatment group differences in awareness of temporal relationships between any two adjacent story categories, adjacency figures were calculated for the proportion of occasions that the second category predicted by the grammar was immediately produced, given that the first member of the pair was provided. The story structure group demonstrated greater awareness of temporal relationships between all cate-

12

Journal of Reading Behavior

gory pairs. The proportions for the adjacency figures for the interim and final stories taken together for the story structure group were Setting-Beginning, .65; Beginning-explicitly stated Goal, .46; explicitly stated Goal-Attempt, .87; Attempt-Outcome, 1.00; and Outcome-Ending, .86. The corresponding proportions for the dictionary-word study group were .25, .22, .50, .78, and .78, respectively. There were no apparent differences in patterns over time. Number of distortions during scrambled story recall. To assess the effects of instruction on distortions during recall, a univariate repeated measures analysis was done with treatment (experimental and dictionary-word study) as a between factor. The within factors were time (interim and final) and story within time (two stories each time). The dependent variable was number of distortions. The story structure group produced more distortions than did the other group (univariate F [1, 17] = 9.29, p < .01). The mean number of distortions for the story structure group and the dictionary-word study group were .83 and .17, respectively. There were no other significant effects. Number of additions during scrambled story recall. To determine whether story structure instruction affected addition of information during scrambled story recall, a multivariate repeated measures analysis was conducted with treatment (experimental and dictionary-word study) as a between factor, and time (interim and final) and story nested within time (two stories at each time) as within factors. On the average, the story structure group produced significantly more additions in story recall (univariate F [I, 17] = 4.74, p. < .05). The overall mean number of additions for the story structure group was .90; the corresponding mean for the other group was .31. There were no other significant effects. Accuracy of scrambled story recall. To determine if story structure instruction enhanced accuracy of recall, a multivariate repeated measures analysis was done with treatment (experimental and dictionary-word study) as a between factor, and time (interim and final) and story nested within time (two stories at each time) as within factors. On the average, the group receiving special instruction in story structure accurately recalled a significantly greater proportion of statements in the scrambled stories than did the dictionary-word study group (univariate F [1, 17] = 6.68, p < .02). The mean proportion of statements accurately recalled by the story structure group was .27; the corresponding mean proportion for the other group was .11. There was a significant overall story within time effect (multivariate F [2, 16] = 4.73, p < .03). Follow-up univariate tests showed significant differences between the two interim stories (F [1, 17] = 6.73, p < .02), with story one yielding a higher mean proportion of statements accurately recalled (.18) than story two (.12). There was no significant difference between the two final stories (F [1, 17] = 1.01, p < .33). There were no other overall significant effects.

Enhancing Comprehension

13

Instructional Effects on Story Comprehension To assess the instructional effects on story comprehension, a multivariate repeated measures analysis was done with treatment (experimental and dictionaryword study) as a between factor and time (interim and final) and story nested within time (two stories at each time) as within factors. When compared to the procedures used with the dictionary-word study group, the special instruction in narrative structure had a positive effect on story comprehension (for overall treatment effect, univariate F [1, 17] = 9.55, p < .01). (Group means for comprehension scores are shown in Table 3.) On the average, regardless of group membership, individuals tended to improve on story comprehension over time (univariate F [1, 17] = 5.59, p < .04). There were also overall significant differences between stories within time (multivariate F [2, 16] = 4.33, p < .04). Follow-up univariate tests indicated the locus of the overall story difference was between stories one and two at interim testing (for interim comprehension scores, univariate F [1, 17] = 8.61, p < .01; for final comprehension scores, univariate F [1, 17] = .21, p < .66). As Table 2 shows, the finding was probably effected by the dictionary-word study group's particularly low mean score on story one. There were no other significant effects.

DISCUSSION In the present study, when compared to a control treatment (dictionary-word study), direct instruction in schematic aspects of narratives did enhance story structure knowledge of fourth graders who were average and below average readers. Importantly, the story structure instruction also had a strong positive effect on reading comprehension. Furthermore, the effects were realized very quickly, during the first phase of teaching. Although the effects were maintained during phase two, the magnitude of differences in story structure knowledge and story comprehension between the story structure group and the dictionary-word study group did not increase over time. It is important to note that time was confounded with stories in the present study. That is, there was no demonstration that all of the stories within each task were equivalent. Stories used at the final testing time may have been easier or more difficult in some way than those used for prior testing. Therefore, results of analyses of time effects, and of all interactions involving time, need to be interpreted with caution. Further testing needs to be done with stimuli of demonstrated equivalence. Several points can be made about the findings of the present study. It is particularly important to note that the positive effects of story structure instruction on knowledge of narrative characteristics were manifested in many ways: (1) overall

14

Journal of Reading Behavior

knowledge of story structure, as measured by the production and scrambled story tasks taken together and the production category score taken alone, (2) ability to infer and integrate information when specific content was forgotten, as measured by the incidence of distortions and of additions in the scrambled story recalls, (3) awareness of specific category adjacencies as shown by analysis of adjacent categories in the production task, and (4) accuracy of recall in the scrambled story recall. The finding of no significant difference between groups in knowledge of story structure when the Kendall's Tau scores for the scrambled story task were taken alone stands in marked contrast to the rest of the findings. The single discrepant result is difficult to explain, particularly because in a prior study involving training dealing with story elements and their interrelationships, significant effects were obtained for the same scoring procedure on a scrambled story recall task (Buss et al., in press). Perhaps the recall feature of the scrambled story task made it more difficult than the production task. Or perhaps the Kendall's Tau rank order correlation used to score the scrambled stories is less sensitive than the other indices. Some prior studies of cognitive processes in story knowledge indicate that different research tasks using story grammars may vary in the extent to which they reveal individuals' knowledge of narrative structure (e.g., Omanson, Warren, & Trabasso, 1978; Stein & Glenn, 1979; Stein & Nezworski, 1978). It was encouraging to find that the special story structure instruction had an early effect on development of knowledge of narrative structure and on story comprehension. Similarly rapid results for story structure knowledge were obtained in another study in which second graders were given one training session dealing with story parts and then were tested on a scrambled story recall task (Buss et al., in press). For some learners, long-term instruction is evidently not necessary to effect initial growth. Also, the relative ease with which the results were effected suggests that the instruction is practical in that it could readily be incorporated into most classroom programs. It is possible that follow-up instruction in more complex narrative structures and/or other facets of story knowledge such as understanding different types of goal formations might further develop children's knowledge of story structure and/or further enhance their story comprehension. It is important to note that the story structure instruction had a positive effect as measured by a test of both literal and inferential comprehension. A test which was not reported here showed that the story structure instruction group consistently held an advantage over the other group on both literal and inferential comprehension. It is also worth noting that although there were some story effects, there were no story by treatment interactions. Although not all differences were tested, as Table 3 shows, story structure instruction maintained an advantage over dictionary-word study instruction on all stories.

Enhancing Comprehension

15

It is impossible to identify the reasons for the overall improvement, regardless of group membership, in story comprehension over phase two. Perhaps the stories used in the final testing were easier, or there may have been a practice effect from experience with the testing procedures. Finally, although the effects of the special story structure instruction on knowledge of narrative structure and story comprehension were striking, it is not clear from the present study whether the treatment effects could be attributed more to certain aspects of the instruction than to others. For example, it is possible that either the teacher-pupil direct instruction component or abundant use of special activities was primarily responsible for the positive findings. Perhaps increasing children's awareness of the relationship between knowledge of story structure and reading comprehension was a particularly critical factor in improving story comprehension. At present, demonstration of effects of controlled, yet practical and ecologically valid, instruction is a worthwhile endeavor. Follow-up studies might now attempt to pinpoint the critical facets of such instruction. In summary, the findings of the present study suggest that direct instruction in schematic aspects of narratives is useful for developing children's knowledge of structural characteristics of narratives. It also appears to be a powerful form of classroom instruction in reading comprehension. Further instructional possibilities could also be explored with children of other ages and in other aspects of story knowledge such as knowledge of various goal structures, knowledge of necessary preconditions for and consequences of goal attainment, and knowledge of characters' motives.

REFERENCE NOTES 1. STEIN, N. L. The effects of increasing temporal disorganization on children's recall of stories. Paper presented at the Psychonomic Society Meeting, St. Louis, 1976. 2. STEIN, N. L., & GLENN, C. G. A developmental study of children's construction of stories. Paper presented at the annual meeting of the Society for Research in Child Development, New Orleans, 1977. 3. GLENN, C. G., & STEIN, N. L. Syntactic structures and real world themes in stories generated by children (Technical Report). Champaign, IL: University of Illinois, Center for the Study of Reading, in press. 4. STEIN, N. L., & GLENN, C. G. The role of structural variation in children's recall of simple stories. Paper presented at the annual meeting of the Society for Research in Child Development, New Orleans, 1977. 5. RUBIN, A. Making stories, making sense (Reading Education Report No. 14). Champaign, IL: University of Illinois, Center for the Study of Reading, 1980. 6. BROWN, A. L., CAMPIONE, J. C., & DAY, J. Learning to learn: On training students to learn from texts (Technical Report No. 189). Champaign, IL: University of Illinois, Center for the Study of Reading, 1980.

16

Journal of Reading Behavior

REFERENCES AARON, I. E., JACKSON, D. RIGGS, C., SMITH, R. G., & TIERNEV, R. Scott, Foresman Basics in Reading. Glenview, IL: Scott, Foresman and Company, 1978. APPLEBEE, A. N. Child's concept of story: Ages 2-17. Chicago: University of Chicago Press, 1978. BOCK, R. D. Multivariate statistical methods in behavioral research. New York: McGraw-Hill, 1975. BORMUTH, J. R. An operational definition of comprehension instruction. In K. Goodman & J. Fleming (Eds.), Psycholinguistics and the teaching of reading. Newark, DE: International Reading Association, 1969. BOTVIN, G. J., & SUTTON-SMITH, B. The development of structural complexity in children's fantasy narratives. Developmental Psychology, 1977, 13, 377-388. BUSS, R. R., YUSSEN, S. R., MATHEWS, S. R. II, MILLER, G. E., & REMBOLD, K. L. Development of children's use of a story schema to retrieve information. Developmental Psychology, in press. COLLINS, A., & SMITH, E. E. Teaching the process of reading comprehension. Intelligence, in press. CTB/McGRAW-HILL STAFF, California achievement test. Monterey, CA: CTB/ McGraw-Hill, 1977. DREHER, M. J., & SINGER, H. Story grammar instruction unnecessary for intermediate grade students. The Reading Teacher, 1980, 34, 261-268. FINN, J. D., & MATTSSON, I. Multivariate analysis in educational research: Applications of the multivariate program. Chicago: National Educational Resources, 1978. FITZGERALD, J. The relationship between ability and expectations for story structures. Discourse Processes, in press. FRY, E. Fry's readability graph: Clarifications, validity, and extension to level 17. Journal of Reading, 1977, 21, 242-252. HANSEN, J. The effects of inference training and practice on young children's reading comprehension. Reading Research Quarterly, 1981,16, 391-417. JOHNSON, N. S., & MANDLER, J. M. A tale of two structures: Underlying and surface forms in stories. Poetics, 1980, 9, 51-86. KINTSCH, W. On comprehending stories. In P. Carpenter and M. Just (Eds.), Cognitive processes in comprehension. Hillsdale, NJ: Lawrence Erlbaum, 1977. KINTSCH, W., MANDEL, T. S., & KOZMINSKY, E. Summarizing scrambled stories. Memory and Cognition, 1977, 5, 547-552. LEONDAR, B. Hatching plots: Genesis of storymaking. In D. Perkins & B. Leondar (Eds.), The arts and cognition. Baltimore: Johns Hopkins, 1977. MANDLER, J. M. A code in the node: The use of a story schema in retrieval. Discourse Processes, 1978, 1, 14-35. MANDLER, J. M., & JOHNSON, N.S. Remembrance of things parsed: Story structure and recall. Cognitive Psychology. 1977, 9, 111-151. McCLURE, E., MASON, J., & BARNITZ, J. An exploratory study of story structure and age effects on children's ability to sequence stories. Discourse Processes, 1979, 2, 213249. OMANSON, R. C., WARREN, W. H., & TRABASSO, T. Goals, inferential comprehension, and recall of stories by children. Discourse Processes, 1978, 1, 337-354. RUMELHART, D. E. Understanding and summarizing brief stories. In D. LaBerge & S. J. Samuels (Eds.), Basic processes in reading: Perception and comprehension. Hillsdale, NJ: Lawrence Erlbaum, 1978.

Enhancing Comprehension

17

SINGER, H., & DONLAN, D. Active comprehension: Problem-solving schema with question generation for comprehension of complex short stories. Reading Research Quarterly, 1982, 17, 166-186. SMILEY, S. S., OAKLEY, D. D., WORTHEN, D., CAMPIONE, J. C., & BROWN, A. L. Recall of thematically relevant material by adolescent good and poor readers as a function of written versus oral presentation. Journal of Educational Psychology, 1977, 69, 381-387. STEIN, N. L. How children understand stories: A developmental analysis. In L. Katz (Ed.), Current topics in early childhood education (Vol. 2). Hillsdale, NJ: Ablex, 1979. STEIN, N. L. The concept of a story. In H. Mandl, N. L. Stein, & T. Trabasso (Eds.), Learning from text. Hillsdale, NJ: Lawrence Erlbaum, in press. STEIN, N. L., & GLENN, C. G. An analysis of story comprehension in elementary school children. In R. O. Freedle (Ed.), Advances in discourse processes: New directions in discourse processing (Vol. 2). Norwood, NJ: Ablex, 1979. STEIN, N. L., & GLENN, C. G. Children's concept of time: The development of a story schema. In W. J. Friedman (Ed.), The developmental psychology of time. New York: Academic Press, in press. STEIN, N. L., & NEZWORSKI, T. The effects of organization and instructional set on story memory. Discourse Processes, 1978, 1, 177-193. STEIN, N. L., & POLICASTRO, M. The concept of a story: A comparison between children's and teachers' viewpoints. In H. Mandl, N. L. Stein, & T. Trabasso (Eds.), Learning and comprehension of text. Hillsdale, NJ: Ablex, in press. SUTTON-SMITH, B., BOTVIN, G., & MAHONEY, D. Developmental structures in fantasy narratives. Human Development, 1976, 19, 1-13. THORNDYKE, P. Cognitive structures in comprehension and memory narrative discourse. Cognitive Psychology, 1977, 9, 97-110. TRABASSO, T., STEIN, N. L., & JOHNSON, L. Children's knowledge of events: A causal analysis of story structure. In G. Bower (Ed.), Learning and motivation (Vol. 15). New York: Academic Press, in press. VIPOND, D. Micro- and macroprocesses in text comprehension. Journal of Verbal Learning and Verbal Behavior, 1980, 19, 276-298. WEISBERG, R. A comparison of good and poor readers' ability to understand explicit and implicit information in short stories based on two modes of presentation. Unpublished doctoral dissertation, Temple University, 1978. WHALEY, J. F. Readers' expectation for story structure. Reading Research Quarterly, 1981, 17, 90-114.