A Study of High School Literature Anthologies
The literature anthology remains the central text in the majority of high school English classrooms. In a recent national study of literature instruction (Applebee, 1990), we found that fully 91 percent of a representative sample of public school teachers reported using a literature anthology, and 63 percent reported that the anthology was their primary source of materials. Some 92 percent of the teachers rated anthologies as at least adequate as a source of selections, and 88 percent rated them at least adequate as a source of teaching suggestions. Although teachers may vary considerably in the particular selections they choose to teach, and in the extent to which they use the suggested activities, the anthology clearly plays a major role in the teaching of literature. As such, any examination of the nature of the high school literature anthology must take place against a background of concern with broader issues of curriculum and instruction.
Guth (1989) has argued in discussing "The Textbook Gap" that the textbook may be the "prime suspect" in students' failure to do better in school; he speaks bitingly of "jerry-built reactionary English texts sold by marketing specialists." Boynton (1989), writing in response to Guth, has countered that schools "get what they're looking for. The best-selling texts may perpetuate bad curriculum practices..., but publishers can fairly argue that they spend a lot of time and money finding out what schools really want."
What schools really want, however, is peculiarly difficult to assess in the teaching of literature at the present time. During the past few years, literature instruction has received increasing attention both within the profession and from the public at large. Part of this attention has come from a concern that traditional cultural values have been abandoned (e.g., Hirsch, 1987), part has come from attempts to reinforce the academic curriculum (e.g., Bennett, 1988), and part has come from teachers who have begun to question whether recent changes in writing instruction may have implications for the teaching of literature as well (e.g., Andrasick, 1990). Alternative visions of the curriculum in English have been offered by a coalition of the major professional associations (Lloyd-Jones & Lundsford, 1989; Elbow, 1990), major literary theorists (e.g., Scholes, 1985), and committed teachers (e.g., Atwell, 1987).
Whether the anthologies that are widely used in classrooms today are seen as a response to or a determinant of the literature curriculum, it would seem important to examine their content and approaches carefully. Unfortunately, there have been few such studies. The last detailed analysis was carried out by Lynch and Evans (1963), as part of a reaction against the later stages of progressive pedagogy. Smaller scale analyses have been carried out recently by Appleby, Johnson, and Taylor in a series of book reviews (1989, 1990a,b) that examine current anthology series in light of Guth's (1989) critique.
Though literature anthologies have received little direct attention in recent years, a variety of commentators have criticized textbooks in general. Summing up a recent National Society for the Study of Education Yearbook on textbooks in the United States, Elliott and Woodward (1990) enumerated a variety of criticisms of the instructional quality of textbook materials:
Chief among the shortcomings researchers have identified are "mentioning," or shallow coverage of a wide range of topics; "inconsiderateness," or poor writing; emphasis on lower-level memorizing of facts and generalizations to the exclusion of problem solving and other higher-order cognitive processes; the avoidance of important topics because some consider them too controversial; and failure to promote adequate understanding of the real nature of the knowledge fields, such as science and history, that are the bases of school subjects. (p. 223)
These general criticisms can be recast in terms of two major issues that have shaped the teaching of English in recent years: one concerns the choice of selections, including the relative representation of alternative literary and cultural traditions; the second concerns the kinds of knowledge and skills that are privileged in the study apparatus and accompanying activities that are part of each anthology. Are students being introduced to an appropriate body of literature? And are they being asked to examine that literature in instructionally appropriate ways?
Such questions can only be answered within the context of a coherent theory of what "knowing" literature involves and how literature can best be taught. There is no consensus at the moment about what such a theory should consist of for elementary and secondary school instruction; indeed there are strongly divergent views within the profession about such issues as what literature to teach and how best to teach it. (Compare, for example, Bennett, 1988, and Lloyd-Jones & Lundsford, 1989.) The present study documents the content and approaches in literature anthologies without attempting to evaluate them against prespecified criteria. From the portrait that emerges, however, a variety of strengths and weaknesses in current practice will be apparent, as well as characteristics that may be seen as strengths or as weaknesses, depending upon one's point of view.
Yet no analysis is ever totally neutral, and certain assumptions underlie the features that will be examined and the way the results will be discussed: 1) that the literature course should include works of substantial quality and interest, that will promote beneficial study and discussion; 2) that the course should recognize and incorporate the contributions of diverse groups to America's shared literary heritage as well as introduce students to major works in the traditional canon; 3) that the study apparatus should emphasize reasoned and disciplined thinking rather than simply recitation of details or of interpretations confirmed by the teacher or text; and 4) that the apparatus that accompanies each selection should be coherent and cumulative, leading the reader toward a more carefully thought-through understanding of a text rather than treating a text as a series of unrelated "puzzles" to be solved.
In these assumptions, I am clearly locating myself within a constructivist tradition in the teaching of literature (Applebee, 1991), one that views productive learning as a process of constructing meaning for oneself within a shared cultural world; that values independent thinking over recitation of what others have said; and that believes that power in language, including in the reading and discussion of literature, comes through engagement in well-motivated language experiences rather than through isolated exercises in language or comprehension.
The study reported here examined the content and approaches embodied in seven popular anthology series, Grades 7 through 12. The analysis of anthology materials was itself only one part of a larger series of studies examining content and approaches in the teaching of literature in American secondary schools. Other studies in the series have examined the nature of English programs with reputations for excellence, surveyed the book-length works required at different grade levels, and surveyed the goals and approaches to literature instruction in award-winning schools and in representative national samples of public and private schools (Applebee, 1989a,b, 1990).
In an earlier study of representative samples of public and private school literature programs (Applebee, 1990), department chairs were asked whether their schools used a literature anthology, and if so, to list the anthology or anthologies currently being used at each grade level, 7 through 12. The present study focused on the seven publishers' series that had been cited most frequently (Appendix 1). It examined series aimed at average and college preparatory tracks. If earlier and related editions of each series are included in the tallies, the seven series studied accounted for 89 percent of the books cited by survey respondents. Each of the seven publishers provided a complete set of anthologies, including books targeted at literature courses in Grades 7 through 10, American literature, and British literature. (Although there is some variation among schools in the placement of courses in American and British literature, taken together these volumes comprise a typical six-year high school curriculum in literature, as envisioned by the authoring and editing teams assembled by each publisher.1) Again to insure comparability across the series, when a publisher offered alternative configurations for the British or American literature course, the more popular chronological volume was analyzed.
Because school materials such as anthologies undergo frequent revisions in response to the adoption cycles in certain large states, the study focused on the 1989 editions that had been prepared for the most recent major round of state adoptions. These editions are more recent than those actually in use in most of the schools surveyed but represent publishers' views of their most up-to-date materials at the time this study began.
Thus the main sample for the analyses of anthologized authors and selections consists of 42 volumes with 1989 copyrights, stratified by grade level and by series. The seven series are listed in Appendix 1, but specific publishers will not be identified in the tables that follow. (The ordering of the series is constant from table to table, but this ordering does not correspond to that in Appendix 1.)
For detailed analyses of the instructional apparatus that accompanies the selections, a subsample of courses and of selections within volumes was drawn. The subsample focused on courses designed for Grade 8. Grade 10, and British literature, including representative samples of 5 major types of selections from each volume: long fiction (1 per volume), plays (1 per volume), poetry (6 per volume), short fiction (6 per volume), and nonfiction (3 per volume). The targeted sample of 357 selections (17 selections x 3 courses x 7 publishers) was reduced to 350 because a number of the series did not include a complete novel or long fiction selection in the British literature course, even if they included substantial excerpts; to keep the samples comparable, this category was deleted from this course for all series. (At other grade levels, the longest fictional selection was chosen for analysis.)
The Nature of the Selections
All selections in all 42 volumes were analyzed to develop a portrait of the content of literature courses as represented by the popular anthology series. The author and title of each selection were entered into a database that allowed us to examine common authors and titles across publishers and course levels. Each selection in the database was further coded to reflect its genre, the year in which it was I written, the number of pages it took up (excluding the surrounding instructional apparatus2), the nationality or literary tradition represented by the author, and the author's gender and race/ethnicity. Information on the author was in many cases included in the anthology itself; in other cases we tracked it down through standard library reference listings (e.g., encyclopedias, Contemporary Authors). For each selection we also recorded the overall emphasis in the unit in which it was included: chronology (e.g., The Romantic Era), genre (e.g., The Short Story), thematic (e.g., The Individual in Society), individual author (e.g., Shakespeare), or literary technique (e.g., Symbolism).
The Nature of the Instructional Apparatus
The subsample of 350 selections was analyzed for a variety of features of the instructional apparatus. For purposes of the analyses, the apparatus was considered in two pieces: first in terms of the kinds of supporting materials that were included anywhere in the textbook (e.g., information about literary periods whether included with the selection or elsewhere in the text); second, in terms of the specific study activities (including prereading activities, study questions following a text, writing assignments following the text, enrichment activities, and skills practice). An activity was defined as a question, suggestion, or directive that might be separately assigned by the teacher or chosen by the student. Typically, prereading directives about what to focus on, suggestions for drawing or dramatization, and separately numbered questions following a selection were each treated as separate activities, while a series of questions embedded within a larger task (e.g., questions about intended audience, genre, and diction asked as part of a writing assignment) were considered to be part of one more extensive activity. Raters used all material available in analyzing the activities, including the selection itself and any commentary or answer keys provided in the teacher's manual.
Each selection was analyzed by one of four trained raters; to estimate interrater agreement, 29 overlapping sets of independent ratings were obtained.
For all 350 selections, each individual activity was categorized on a number of different dimensions:
Authentic vs. Recitation Activity. An activity was categorized as "authentic" (Nystrand & Gamoran, in press) if it seemed to assume that a variety of differing responses were legitimate. Activities which sought a single or correct answer were classified as "recitation" activities. Across raters and selections, interrater agreement on classification of authentic versus recitation activities was .84.
Content Emphasized. Each selection was categorized to indicate whether or not it included any attention to each of a number of different kinds of content knowledge: plot, character, or setting; theme or purpose; language or style; literary terms; cultural or historical background; and vocabulary. Many activities of course referred to several different types of content, all of which were noted. Interrater agreement on whether a selection included attention to each type of content averaged .86.
Cognitive Demand. Each activity was categorized in terms of the level or type of cognitive activity involved, whether that activity focused on personal response or on features of the text itself. Four levels of cognitive activity were distinguished:
Interrater agreement in categorizing cognitive level was .90.
Connectivity and Intertextuality. Each activity was examined for its relationship to other activities that students had been asked to do previously. Activities were categorized as: discrete (to be completed in isolation from other activities), part of a set of activities that ask for similar things but that do not build upon one another, or cumulative, building on an earlier activity. Intertextual activities, making links to other selections that a student might have read, were also tallied. Interrater agreement was .90 for connectivity to other activities and .99 for intertextuality.
Location. The placement of each activity in relation to the selection was also noted. Activities were categorized as prereading, postreading, or writing activities (requests for writing-before-reading were categorized as prereading activities). To be classified as a writing activity, there had to be an explicit reference to a written response. Interrater agreement for location was 100 percent.
Analysis of Results
Scores were summed across activities to yield totals in each category for each selection. Because there was considerable variation from selection to selection in the number of activities provided, these category totals were converted to percents based on the total number of activities for each selection. This allows a more accurate examination of the relative emphasis placed on different kinds of knowledge and skills from one selection to another.
To assess the statistical significance of observed differences, Chi-square analyses were used for categorical data. Where appropriate, three-factor ANOVAs with course and publisher as between-book factors and genre as a within-book factor were used to assess differences on continuous variables. Because long fiction selections were not included in the analyses of the British literature course, they are omitted from the statistical tests though they are included in the tables (where they represent materials for the 8th and 10th grade courses). Tabled data focus on main effects, since interactions were in general not significant.
Results will be presented in two sections, the first examining the nature of the selections included in the seven series, and the second examining the nature of the instructional apparatus.
The anthologies analyzed for the present study included a total of 38,510 pages, presenting 5,203 appearances of 2,809 different selections by 1,201 authors (including 178 authors of anonymous selections) for students to read and study. The individual volumes were massive tomes, averaging some 917 pages in length and including an average of 124 selections per volume (Table 1). As would be expected, the size of the volume and the number of pages increased with courses designed for upper grades, though there was considerable variation from series to series. The bulkiest series had 1.4 times as many pages as the smallest, though it had only 1.2 times as many selections. (The rest of its bulk reflected more extensive instructional apparatus accompanying each selection.)
All of the volumes used divisions of one sort or another to organize their selections, and most of the major divisions were further subdivided. The most typical divisions and subdivisions emphasized genre (e.g., the Short Story), literary techniques associated with particular genres (Characterization), chronology (The Romantic Era), or themes (Coming of Age). The volumes as a whole tended to be divided according to genre or (to a much lesser extent) theme in Grades 7 through 10, and according to chronology in the American and British literature courses (Table 2). The emphasis on genre is also apparent in the subdivisions within which individual selections appeared (Table 3). In the chronologically organized volumes, for example, the subdivisions within which individual selections appeared were most likely to be based on genre characteristics (47 percent) or individual authors (28 percent), with a strict chronological approach (that is, chronological subdivisions within chronological major divisions) being followed with fewer than a quarter of the selections. Anthologies whose major divisions emphasized genre, in turn, were most likely to use subdivisions that focused on literary techniques associated with particular genres (47 percent) or on cross-cutting themes (30 percent). Even within major divisions based on themes, 71 percent of the individual selections were in subdivisions highlighting the characteristics of individual genres.
Changes in Size and Organization Since 1961
To place the anthologies in the current study into some perspective, we can compare these figures with those reported by Lynch and Evans (1963). Their analyses focused on 72 texts from Grades 9 through 12; the texts they analyzed had been published between 1949 and 1961, with the great majority copyrighted in the mid to late 1950s. As a set, they represent the anthologies available to teachers in 1961. Unlike the present study, Lynch and Evans included virtually every text that they found to be in use, including successive editions of some anthologies, and series designed especially for lower track classes. The majority of the texts they analyzed, however, were quite comparable to those in the present study3. If their results are compared with those from the 9th through 12th grade volumes in the present study, the approximately 30 years between the two studies has seen the anthologies increase in length by 47 percent, and in number of selections by 21 percent.4 Confronted with the size of the volumes they studied, Lynch and Evans were moved to ask, "Why should the student, who has met literature only in the chaos and clutter of the ponderous anthology, feel inclined ever to seek it again?" (pp. 23-24). Given the significantly greater bulk of contemporary anthologies, we can but echo their query.
Lynch and Evans (1963) were also concerned about the organizational frame imposed upon the selections chosen for study. Writing out of a New Critical concern with the nature and integrity of the literary text itself, they shared the New Critics' interest in literary studies focusing upon the unique characteristics of individual genres-- a concern apparent even in Brooks and Warren's early, influential, and genre-based college text, Understanding Poetry (1938). In examining the textbooks available up to 1961, Lynch and Evans found that anthologies for 9th and 10th grade courses tended to be topically organized, while those for 11th and 12th grade courses showed a variety of forms of internal organization, though chronology was most frequent. (Lynch and Evans distinguished between , such as "Conquests of Science," and themes which focus on a human trait or quality, such as "Loyalty." In the present study both types of units were treated as thematic.) Using examples from the textbooks they analyzed, Lynch and Evans argued that a focus either on topics or on chronology tended to distort the choice of selections and the apparatus surrounding them, leading to an emphasis on nonliterary content (geography, social studies, science) or on literary history instead of on study of the literary texts themselves. As a result, Lynch and Evans recommended that the volumes for all four courses be organized around genres, and that the emphasis on American and British literature in Grades 11 and 12 be abandoned, with selections "of substance" from the British and American literature courses being redistributed across all four years.
British and American literature courses have survived in spite of these recommendations, but in the 1989 anthologies organization by genre (with all of its New Critical heritage) had driven out most other alternatives in Grades 7 through 10, and was also used to subdivide the larger chronological divisions in the anthologies for British and American literature courses.
Types of Literature Represented
Given an average of 917 pages per volume, what do the current anthology series contain? As a first step, we can examine the number and types of selections included in the anthologies, and the amount of space devoted to them. The overall figures are presented in Table 4. On average, the 42 volumes each included I novel or other long piece of fiction, 3 plays, 72 poems, 26 shorter works of fiction, 16 nonfiction selections, and 7 selections representing an assortment of other forms (including short excerpts from plays, myths, tall tales, fables, legends, and excerpts from the Bible). Although only a few plays and long fictional works were included, because of their length they took up 38 percent of the pages devoted to literary texts.
Of the 917 pages in the average volume, it is noteworthy that only 450 pages were devoted to actual selections of literature; the remaining 467 were used for the surrounding study apparatus, art work, introductory material, indices, and appendices that dealt with such things as literary terms, the writing process, and difficult vocabulary.
Table 5 summarizes the types of literature included in each course. The number of selections of each type remained relatively constant for the 7th through 10th grade courses, except for a gradual increase in the number of poems (from an average of 32 for the 7th grade course to 51 for the 10th grade course). The American literature course was marked by a doubling of the number of nonfiction selections and by an even greater increase in the number of poems (123, compared with 51 in the 10th grade course). The British literature course placed even more emphasis on poetry (with 152 poems, representing 76 percent of the selections), and gave less attention to nonfiction.
Changes since 1961 in Types Represented
A few changes over time are evident when these results are compared to those reported by Lynch and Evans (1963). In the anthologies available in 1961, only one-quarter of the fiction, compared with nearly all of the 1989 anthologies intended for the 7th through 10th grade courses or for Amer ican literature. (From their study, Lynch and Evans had concluded that the novel should be dropped altogether from the anthology series.) The proportion of nonfiction selections has dropped noticeably, from 265 to 13 percent of the selections for Grades 9-12, reflecting a sharp reduction in what Lynch and Evans termed "miscellaneous nonfiction" -- a category particularly associated with the topical mode of organization that was so prevalent in the anthologies they analyzed.
Literary Periods Represented
The teaching of literature always involves finding a balance between relatively contem- porary works, which may seem more relevant and accessible to young readers, and older works that are part of major cultural traditions. Some of the fiercest debates about the teaching of English have revolved around just how this balance should be struck (Applebee, 1974).
Table 6 summarizes the distribution of selections across centuries, separately by course and by type of literature. In the courses for Grades 7 through 10, roughly three fourths of the anthologized selections were from the twentieth century, and some 30 percent were from the past 30 years. The proportion of twentieth century works dropped to 53 percent in the American literature course, and to 27 percent in the British literature course. The proportion of works from the past 30 years also dropped sharply, to 15 percent in the course and 5 percent in the British literature course.
Table 6 also makes clear that there was considerable variation in these proportions for different types of literature. Works of fiction were drawn almost exclusively from the nineteenth and twentieth cenuries, while a substantial portion of the anthologized plays were pre-17th century (largely Shakespeare and Sophocles). Fully 60 percent of selections categorized as "other" were pre-17th century, reflecting the biblical excerpts and traditional myths, tales, and legends that were included in this category.
Changes Since 1961 in Periods Represented
Lynch and Evans (1963) also briefly examined the periods represented by the selections included in the anthologies available in 1961, separating them into 20th century and pre-20th century works (p. 150). Given the dates of the two studies, the most direct comparison is between works published in the previous 60 years (1900 and later for Lynch and Evans, 1930-1989 for the present study). The relevant data are summarized in Table 7.
The results in Table 7 suggest a significant shift away from relatively contemporary works in anthologies over the past 30 years. In the Lynch and Evans study (1963), over half of the anthologized selections had been written within the previous 60 years; by 1989, only just over a third were equivalently contemporary. A variety of factors may lie behind this shift, including Lynch and Evans' own criticisms of the "ephemeral" nature of many anthologized selections, recent calls for greater emphasis on "great works" from the Western tradition (e.g., Bennett, 1988), teachers' natural inclination to prefer selections with which they are familiar, and concern about explicit language and controversial topics in some contemporary works (a problem exacerbated by a reluctance to edit or "sanitize" a text, a practice Lynch and Evans criticized harshly in the anthologies they analyzed).
Characteristics of the Authors
Historically, the high school literary canon has reflected a mainstream Anglo Saxon tradition (Applebee, 1974, 1989a), but the past several decades have seen vigorous calls for broadening the canon to include selections from alternative literary traditions. To examine the extent to which the anthologies have responded to such calls, Table 8 summarizes the characteristics of the authors of the anthologized selections.
The data in Table 8 suggest that some effort has been made to provide balance, particularly in the materials for Grades 7 through 10. In these volumes, between 26 and 30 percent of the selections were written by women, and 18 to 22 percent by members of various nonwhite minorities.
The selections for the British literature course were much narrower, with only 8 percent by women and 1 percent by members of minority groups. To some extent, this narrowness is a result of the chronological empl,.asis in the British literature volumes, with their coverage of earlier periods when women and members of nonwhite minorities had somewhat little access to traditional avenues of publication. It results, too, from the ethnic makeup of the British population before the concept of British literature was extended to the Commonwealth. Even among selections from the past 30 years, however, only 17 percent in the British literature volumes were by women and 10 percent by nonwhite authors. The American literature volumes, in contrast, managed considerably more breadth overall in their choice of selections for study (with 24 percent of the selections by women, and 16 percent by nonwhite minorities). When the British and American volumes are compared by period, the American literature selections are broader than the British in every period.
When the selections are considered in terms of the national tradition within which they were written, authors from North American countries and from the United Kingdom accounted for 93 percent of the selections, with another 4 percent from Europe and just a smattering from other regions of the world.
In the 1989 volumes, characteristics of authors also varied by type of literature and by date of original publication (Table 9). In general, the choices of long fiction and of plays were narrower than those of shorter works. Only 10 percent of the long fiction selections were by women, and none by nonwhite authors. Women were somewhat better represented among authors of plays (18 percent),6 but nonwhite authors were not (2 percent). Overall, women were represented best in short fiction (28 percent), and nonwhite minorities in nonfiction (18 percent) and "other" (24 percent, reflecting the inclusion of myths, legends, tales, fables, and religious texts from other cultures).
When selections are considered by the period in which they were written, in general the older the selection is, the narrower the traditions represented are. Of selections written during the 17th century, only 5 percent were by women, compared with 39 percent of those written during the past 30 years (1960-89). Similarly, only 3 percent of the 17th century selections were by nonwhite authors, compared with 33 percent during the past 30 years. Pre-17th century selections showed a similar lack of representation of women (4 percent), but the proportion of traditional tales, myths, legends, and religious texts from other cultures raised the percentage of works by nonwhite authors to 18 during this early period.
Changes Since 1961 in the Characteristics of the Authors
Lynch and Evans, (1963, p. 149) were not particularly concerned with the breadth of the selections included, and report few related tabulations. (They do report the most frequently anthologized authors and selections for various genres, however, allowing some comparisons that will be presented in a later section.) They were concerned about how well British and American literature (as opposed to "foreign" works) was represented, allowing some comparison with results from the present study. As in the present study, they found a preponderance of English and American literature in the 9th and 10th grade courses, though the inclusion of some world literature in the 12th grade courses they analyzed added some breadth to the geographic representation of the selections in their volumes at that level (some 20 percent of the 12th grade selections were "foreign" rather than English or American, and another 2 percent were "classical"). Compared with Lynch and Evans' (1963) results, the 1989 anthologies show a slight decrease at Grades 9 and 10 in the proportion of selections from North America (from 74 percent in 1961, to 69 percent now), and a slight increase in the proportion from the United Kingdom (from 17 percent to 20 percent).
Between-Series Variation in Characteristics of Authors and Selections
It is possible that the averages discussed so far mask important variations in the nature of the selections available from different publishers. To examine this, Table 10 summarizes selected characteristics of each of the seven series included in the study.
The results show some variation in the average number of selections included in each volume (from a low of 115 to a high of 134), but a striking degree of consistency in the characteristics of the authors and selections. Thus the percent of selections by women varied by only 5.7 percentage points across the seven series, and that of selections by nonwhite minorities by only 6.4 percentage points. The proportion of selections drawn from each of the genres was similarly nearly identical across the seven series. Rather than offering schools a choice of emphases, all seven of these 1989 series reflected a very similar distribution representing the available literary traditions.
This consistency in emphasis represents a major change from the options available to teachers in 1961. Across series, for example, the percent of short story selections in the 9th to 12th grade courses varied from 10 to 29 percent for the ten most comparable series in Lynch and Evans' (1963) study (p. 431), compared with a range of 18 to 23 percent in the 1989 volumes. Emphasis on poetry shows a similar pattern: in Lynch and Evans' study, poetry ranged from 37 to 65 percent of the selections across the ten comparable series (p. 458); in the present study, the across-series variation was only from 56 to 62 percent.
Relationships between Anthology Selections and What Teachers Teach
The characteristics of the anthologized selections can be compared with two other Literature Center studies of the content of the literature curriculum. The first of these examined department chair's reports of the book-length works that were required reading for any group of students at each of the high school grades (Applebee, 1989a); the second examined the specific selections that teachers reported students in a specified class had been required to read in class or for homework, during any of the previous five days (Applebee, 1990). To facilitate comparisons among the studies, the results of each are summarized in Table 11.
In all three studies, the selections were relatively narrow, dominated by white male authors in the Western literary tradition. Overall, 16 percent of the selections taught in the past five days, and 19 percent of the required book-length works were written by women (compared to 22 percent of the anthologized selections). Only 7 percent of the selections taught in the past five days, and 2 percent of the required works were by nonwhite authors (compared to 14 percent of the anthologized selections). The picture is more complicated, however, when results from different genres are considered. The teachers' selections of long fiction (both those taught and those required) were broader than those in the anthologies. Teachers were teaching a narrower selection of poetry, short fiction, and nonfiction than was included in the anthologies, however.
These results suggest that the anthologies may be a moderately progressive influence in the selections for those types of literature (stories, poems, and nonfiction) which are included in reasonably large number. In choosing long fiction and plays, however, where constraints of space allow few selections to be included in the materials for any particular course, the anthologies seem to emphasize traditional selections.
Consistency in the Choice of Titles and Authors
The data discussed so far indicate considerable consistency in the general nature of the selections anthologized but indicate nothing about the inclusion of specific authors and titles. Is there a body of material that all students are expected to read? Or do the various series represent somewhat different samplings of authors and titles, even while drawing them from similar traditions?
Table 12 summarizes the number of common authors and titles across the seven series, both for the series as a whole and for each course separately. In the materials for Grades 7 and 8 there were no selections and only 1 author common to all seven series, and these numbers rose only slightly in the materials for Grades 9 and 10. For the American literature course, on the other hand, the picture is quite different, with 17 selections and 49 authors common to all seven series and another 51 selections and 70 authors common to at least six of the seven. The British literature course was similar in this respect to the American, reflecting the attempt to be comprehensive in coverage of their respective traditions.
Variety in Grade Placement of Titles and Authors
Table 13 looks at these data in terms of the variety in grade placement of authors and titles, for those that appeared in six or seven of the series. Again, authors and titles that were part of the British or American literature courses showed considerable consistency in grade placement, but others varied widely. Leaving aside the British and American volumes, all but 2 of the remaining 40 titles common to all seven series appeared at more than one grade level. Similarly (and again leaving aside the British and American literature courses), all of the 83 remaining common authors appeared at more than one grade level across the seven series. This variety in grade placement parallels a variety in grade placement of book-length works in our previous study (Applebee, 1989a), and in the placement of materials in Lynch and Evans' (1963) study of anthology contents 30 years ago. Such variety seems healthy, reflecting the many different sets of relationships that exist among authors and works (allowing them to be combined in different ways), and the many different kinds of questions that can be asked about a particular text (creating instructional contexts of varying levels of difficulty; see Purves, 1990).
Unique Selections and Shared Traditions
When Lynch and Evans surveyed textbooks available in 1961, they complained about the amount of "ephemera" and "miscellany" in the selections included for study. As one (rough) index of ephemeralness, they looked at the proportion of the selections that appeared in only a single volume out of the 72 volumes that they examined. Table 14 summarizes a similar analysis for the 1989 anthologies. Across anthologies, over a third (37 percent) of the selections that appear represented titles used in a single anthology series. In theory, that means that over one third of the selections a student reads in a typical high school course would be read by other students using the same series, but not by students using any of the other popular series. Titles were most diverse in the materials for Grade 7 (where 52 percent of the selections that appeared represented titles used in a single series), and least diverse in the British literature course (where 27 percent represented titles used in a single series). When the selections are considered by type, the least consistency occurred for nonfiction (where 53 percent of what a student reads would be unique to that series); the most consistency occurred for long fiction, where only 29 percent of the selections were unique.
The data on particular titles, however, to some extent mask a greater consistency in the authors who are read. Across anthology series, 89 percent of the selections-- or roughly 9 out of every 10 selections-- were by authors included in two or more of the series. The greatest consistency occurred for the British and American literature courses, where only 5 to 6 percent of the selections were by authors unique to one series; in the anthologies designed for Grades 7 through 10, however, considerably more variety was apparent.
Changes since 1961 in the Proportion of Unique Selections
Lynch and Evans (1963) were inconsistent in the data they reported for different genres, so that it is only possible to calculate the proportion of unique poems and unique short fiction selections in 1961. (Calculations are based on the total number of titles in a given genre that appeared in only one series, divided by the total number of appearances of that genre.) Using the data for the grades common to both studies (Grades 9 to 12), the proportion of unique poetry selections fell from 40 percent in the anthologies available in 1961 to 33 percent in the 1989 editions. The proportion of unique short fiction selections also fell, from 47 percent in 1961 to 38 percent in 1989. Since Lynch and Evans were working with a base of 72 texts compared with only 28 for Grades 9 through 12 in the present study, these comparisons understate the degree of change that has taken place. (As comparable texts are added to the analysis, the chance of repeating a selection increases and the proportion of unique texts falls.)
Such results need to be interpreted with some caution. In establishing a sense of a literary tradition, some degree of consistency is clearly important. On the other hand, there are many authors and selections that can appropriately represent the various traditions that make up America's literary heritage.
Most Frequently Anthologized Authors and Titles
Table 15 lists the 122 authors who were included at least once in all seven anthology series, in order of the total number of appearances across grades and series; the number of appearances at each grade level is also listed.
Emily Dickinson leads the list with 138 appearances, and Robert Frost follows close behind with 101. In both cases their poems were used in anthologies for all grades except the British literature course. William Shakespeare is next with 98 appearances; in his case, the works include a mixture of complete plays, excerpts of famous soliloquies, and sonnets. Langston Hughes was the most frequently anthologized minority author, with 53 appearances. Like Dickinson and Frost, his poems were used at all levels except the British literature course.
If the amount of space devoted to individual authors (reflected in the entries for "total columns" of text, Table 15, where each column is equivalent to half a page) is considered rather than the number of separate selections, the shape of the list looks quite different, with Shakespeare first, Dickens second, Steinbeck third, and Shaw fourth. In this configuration, the list looks more similar to the results from the earlier study of book-length works (Applebee, 1989a).
Table 16 provides a similar summary of specific titles anthologized in all seven series, as well as the grade levels at which they appeared. In this case the list is organized alphabetically by author. (Appendix 2 provides a fuller summary by course of all selections that are included in at least a majority [4 out of 7] of the volumes for each grade.) Here, Shakespeare has the most separate titles, with 7 different selections included in all seven series, Frost is next with 5; and Keats and Poe follow with 4. As a set, these 89 titles include traditional as well as contemporary selections, from Beowulf and the Bible to I Know Why the Caged Bird Sings.
Changes since 1961 in Most Frequently Anthologized Titles
Long Fiction. The specific titles included in these lists show a number of differences from those cataloged by Lynch and Evans (1963). In the anthologies they studied, the novel was only sporadically represented; Silas Marner was the most frequent text, and of the other titles, only Great Expectations and The Voice of Bugle Ann appeared in more than a single series. In contrast to Lynch and Evans' results, in the 1989 anthologies neither Silas Marner nor The Voice of Bugle Ann appears at all, and Steinbeck's The Pearl appears in every series (usually in Grade 10). (For easy comparison with Lynch and Evans' results, Appendix 3 summarizes the selections in each genre that were included in a majority, at least 4 out of 7, of the 1989 anthologies.) Great Expectations continues to be popular (in abridged or adapted form, in 5 series at Grade 9), as is The Call of the Wild (also in 5 series). Several long fiction selections are regularly excerpted for the anthologies (though in the analyses for the present study these were treated as short fiction). Regularly excerpted texts included Malory's Le Morte d'Arthur (in all 7 series), Mary Shelley's Frankenstein (in 6), and Swift's Gulliver's Travels (in 6). Of the long fiction selections that appear in more than one series, whether in whole or in part, only one (Mary Shelley's Frankenstein) is by a woman, and none is by a minority author.
Plays. Lynch and Evans complained about the inclusion of minor drama to the exclusion of major playwrights, and about the relative lack of attention to Shakespeare (as the greatest dramatist) even though two of Shakespeare's plays, Caesar and Macbeth, were the most frequently appearing dramas in the anthologies they analyzed. Both Julius Caesar and Macbeth appear in all seven of the 1989 series, where they are also joined by Romeo and Juliet and by Our Town. Six of the seven series include The Miracle Worker, Pygmalion, and Goodrich and Hackett's version of The Diary of Anne Frank.
Short Fiction. Short fiction shows a similar shuffling of popular titles, though some of this has involved simply trading one selection by an author for another by the same author. Thus of the ten most frequently anthologized selections of short fiction in Lynch and Evans' study, six do not appear in even a majority of the 1989 anthologies ("The Devil and Daniel Webster," "The Split Cherry Tree," "Sixteen," "The Ransom of Red Chief," "That's What Happened to Me," and "The Silver Mine.") Of the ten most frequently anthologized short storyauthors in Lynch and Evans' study, on the other hand, all were in the majority of the 1989 series, and all but 3 appeared in all 7 series (the exceptions: Jesse Stuart in 4, William Saroyan in 5, and Jessemyn West in 5).
Poetry. There are also some interesting shifts in the relative emphasis on particular poets, perhaps in response to concerns with providing broader representation of alternative literary traditions. In the anthologies available in 1961, the most frequently appearing poets were, in descending order, Whitman, Frost, Tennyson, Sandburg, Shakespeare, Dickinson, Wordsworth, Burns, Longfellow, and Robert Browning. In the 1989 anthologies, all of these poets remain prominent, but Dickinson has moved to the top of the list, and Langston Hughes has appeared as number 7. (Hughes did not appear at all in the Lynch and Evans list of 86 most frequently anthologized poets.)
Nonfiction. In examining nonfiction selections, Lynch and Evans found a great "miscellany," with little attention to "literary" works and a great deal of excerpting from book-length works. They recommended that the miscellany be deleted and that only nonfiction that could be justified on artistic grounds should be included. To a large extent their recommendations have been heeded, though there remains more diversity in nonfiction selections than in other parts of the anthologies. The six nonfiction works that were excerpted or included in all of the 1989 anthology series were Angelou's I Know Why the Caged Bird Sings, Boswell's The Life of Samuel Johnson, Bradford's Of Plymouth Plantation, Franklin's Autobiography, Donne's "Meditation 17," Pepys' , and Thoreau's Walden. Like the poetry selections already discussed, the specific titles reflect concern with representing diverse literary and cultural traditions, and even the most frequent individual selections seem noticeably broader in this respect than those cataloged by Lynch and Evans. (In their study, the four most frequently anthologized nonfiction authors were Leacock, Emerson, Thoreau, and Lamb.)
Broadening the Tradition
Tables 17 and 18 list the most frequently anthologized women and nonwhite authors. Both lists provide starting points for teachers who are consciously seeking to include a broader representation of alternative traditions in their literature courses. Both lists are dominated by relatively contemporary authors, though they include some authors from earlier periods, such as Elizabeth Barrett Browning, Anne Bradstreet, and Aesop.
THE NATURE OF THE INSTRUCTIONAL APPARATUS
The apparatus surrounding the literary selections in the typical anthology fell into several distinct parts. Usually there was some kind of introductory material before the selection, relating the selection to the theme, genre, or period of which it is a part. Sometimes this included a statement or question to focus the reading (e.g., "Notice how the characters change as the story progresses."), or a short writing activity designed to remind readers of similar, relevant experiences of their own ("In your journal, write about a time when you felt you were treated unfairly, and how you reacted to it."). Following the selection there were typically a variety of "study" or "discussion" questions; these were usually somewhat ambiguously labeled, allowing the teacher to decide whether students would be asked to write out their responses, or simply to use them to guide class or small-group discussion. Following the discussion questions, selections sometimes included one or more activities that specifically asked for a written response, study skill or vocabulary activities, or enrichment activities. The arrangement of specific study questions varied somewhat from series to series, but the questions were usually divided into several sections:
This format for questioning derives in part from Bloom's taxonomy (1956), with its hierarchical assumptions about the nature of knowing; some of the anthologies even use Bloom's categories as subheadings in either the student book or teacher's guide. The format derives in part, too, from a version of New Criticism expounded in the 1960s by the Commission on English of the College Entrance Examination Board (CEEB) (1965), which promoted close, text-based analysis as the foundation of literary discussion. The CEEB argued that "the actual study of a work will determine the order in which pressing questions rise and demand answer" (p. 58). Nonetheless their list of "fundamental questions" began with 11 "Questions about the text itself" grouped into three sets (Questions of form, Questions of rhetoric, and Questions about meaning), before getting to the final section, Questions of value (consisting of Questions about personal response and Questions of excellence, with no specific questions listed for either section). Both for the CEEB and for applications of Bloom's taxonomy to reading, there is an assumption that understanding must start at a textual level, and only when the text is fully clear, move on, perhaps, to personal response and evaluation. (For an alternative to this view, see Langer, 1985; 1989.)
These overall patterns form the background for examining the kinds of knowledge and skills implicitly valued as important by the attention they received in the activities that were provided and the responses that were expected in the various anthology materials. The discussion will begin by examining some of the variations in supporting material provided with each selection, and will continue by examining the nature of the study activities.
Supporting Material Provided with Individual Selections
The first analysis looked broadly at the kinds of supporting material provided anywhere in the text, whether as part of introductory material, as separate sections, or as information embedded in study questions or other activities. The analysis looked at information about the context of a selection, at help with problems that might be encountered while reading, at explanations or activities emphasizing literary terminology, and at the inclusion of writing activities. Table 19 summarizes variation by course in the percentage of selections that had various sorts of accompanying supporting material (whether before the selection, following the selection, or keyed to a separate unit introduction or summary).
Context. The context for the selections was provided in several ways. The most prevalent was through provision of at least a brief biography of the author (97 percent of the selections). Additional social or historical context was also provided for 93 percent of the selections in the British literature volumes, though for only 22 percent of the selections in the 10th grade course and 15 percent in the 8th grade course. Literary context-- the relationship of a work to a tradition or genre-- was also provided for many selections, particularly in the materials for the British literature course (74 percent).
Help with Reading,. Most selections also sought to help students focus on important points within the selections or to circumvent reading difficulties. Some 87 percent had a prereading activity to focus students' attention as they read. The extent of these prereading activities varied considerably. Some involved single sentences to link the selection to a theme or period; others provided guidance on what to watch for; still others involved more extended writing or discussion activities designed to prepare students by introducing unfamiliar vocabulary or emphasizing relevant personal experiences. Typical of the brief version of prereading activity was the single line, "The title of this poem helps you understand its meaning," printed just above the title for Gwendolyn Brooks' "The Children of the Poor" (McDougal, Littell, Blue Level [Grade 10]; p. 506). More involved was the activity that preceded Elizabeth Bowen's "The Demon Lover" in Prentice Hall's The English Tradition:
Like the activity preceding "The Demon Lover," many of these prereading activities were curiously detached from the selections that followed, without making the purposes of the activity clear to the students, even if they were explicated more fully in the teacher's edition.
For some three quarters of the selections, the accompanying prereading or postreading apparatus anticipated particular difficulties that a selection might pose and made suggestions for an effective reading strategy. This attention to possible reading difficulties was particularly strong in the 8th and 10th grade volumes, and decreased significantly in the British literature course. Exercises or background information dealing with difficult vocabulary (in a more extended form than notes or a glossary) showed a similar pattern, being provided for 61 percent of the selections for Grade 8 but for only 41 percent in the British literature course.
Literary Terminology. Some 86 percent of the selections were also accompanied by discussions of the specialized vocabulary of literary scholarship-- familiar terms such as plot, character, and setting as well as more exotic ones such as situational irony. Treatment of these terms ranged from brief definitions to activities requesting extended application to the selection being studied. Attention to literary terminology usually took the form of a post reading activity, though sometimes it was used to introduce a genre (e.g., haiku) or to highlight a literary technique (e.g., characterization) in a selection that followed.
Writing Assignments. Writing assignments were included as post-reading activities with three fourths of the selections. (Activities requiring writing-before-reading are included in the totals for prereading activities.)
Between-Series Variation in Supporting Material
Table 20 summarizes the amount of variation in the supporting material provided by the seven publishers. Unlike the selections, whose character was remarkably consistent across all seven series, the supporting material showed more variation. The greatest variation occurred for help with explicit vocabulary study, which ranged from only 26 percent of the selections in one series to 98 percent of the selections in another. Attention to reading strategies (36 to 100 percent) showed a similar disparity, while writing activities, prereading activities, literary terminology, and social, historical, or literary context all showed large differences among series. Only the provision of the author's biography was relatively constant across all seven publishers.
By far the most extensive material accompanying each selection consists of study activities, which may range from relatively straightforward requests to explain what happened, to extensive suggestions for library research and report writing. In counting activities, questions (Why did he kill his brother?) or directives (Summarize the story.) that were likely to be assigned separately were counted as separate activities; a series of questions embedded in a larger task (e.g., questions of audience or form in a writing activity) were treated as part of the larger task; and series of parallel exercises (e.g., metaphors to identify in a series of sentences) were treated as part of a single practice activity. Table 21 summarizes the average number of study activities accompanying selections of different types.
Number of Activities
Overall, the volumes averaged 12 activities per selection. As would be expected, the number of activities varied considerably for texts of differing lengths. Thus poems averaged only 8 activities per selection, while long fiction averaged 21. (To keep the comparisons somewhat comparable, the totals for both plays and long fiction include only the activities that preceded and followed the whole selection, the pattern that occurs for the other types of literature examined.) The great majority of activities were staged as postreading activities, with an average of 1 additional post-reading activity marked as a writing activity, and an average of less than 1 prereading activity for every 2 selections. (In these analyses, activities requiring writing before-reading were included only as prereading activities.) The British literature volumes tended to have fewer activities per selection than the volumes for Grades 8 or 10, in large part because these volumes include a higher proportion of poems. The seven series ranged from an average of 10 to an average of 14 activities per selection.
Emphasis on Recitation
Of more interest than the number of activities is the emphases that they reflect, the implicit definition of what counts as "knowing" literature. Are students asked to demonstrate their knowledge of accepted meanings or interpretations of a text? Or are they asked to engage in developing and defending alternative understandings and interpretation? To examine this, each activity was categorized as being "authentic" (Nystrand & Gamoran, in press), allowing a variety of alternative responses, or as "recitation," soliciting a presentation of a presumed common answer. (In making these judgments, raters relied on their knowledge of the selections, as well as guidance provided in the teacher's manuals.7) The results are summarized in Table 22.
Across grades, the results in Table 22 indicate an overwhelming emphasis on recitation. Overall, an average of 65 percent of the study activities tapped students' knowledge of textual detail or of accepted interpretations. This occurred both with activities that were overtly focused on memory for details and with activities that asked for analysis and interpretation. Thus, for example, the post reading activities for George Orwell's "Shooting an Elephant" in the Harcourt, Brace Jovanovich volume, Adventures in English Literature, began with a four-part "Reading Check" that asked for simple recall of details (e.g., "1. Why was the elephant out of control?"). This was followed by four questions headed "Analyzing and Interpreting the Essay" that similarly assumed right answers. For example: "What is the tone of the opening and concluding paragraphs?" (The notes in the Teacher's Edition say "Self mocking.") Only the final activity (which was given the broad heading Writing About Literature) left some room for students to develop and defend a point of their own, and then only within sharply defined limits: "Select one sentence from this essay which you think expresses Orwell's basic point better than any other. Write a composition defending your choice" (p. 889).
There was little variation in this general pattern with grade level, genre, or series. The greatest variation occurred with location of the activity: an average of 71 percent of the postreading activities expected a correct answer, compared with 16 percent of the prereading activities and 15 percent of those that were explicitly flagged as requiring a written response. This variation was tied very closely to the way these questions were used. Prereading activities tended to be used to stimulate readers' thinking, postreading activities to insure that they had correctly understood the selection, and writing activities (when they occurred) to ask students to summarize and defend the understanding that had been reached or, sometimes, to move beyond the selection.
This overwhelming emphasis on recitation activities leading to a single expected response rather than on authentic activities in which responses may legitimately vary creates a consistent image of the reading of literature as a kind of puzzle to be solved, with a set of correct responses to be derived from the text and teacher. It is not, for the most part, a context for exploring ideas and defending alternative understandings. In this sense, the emphases in the anthologies are an accurate reflection of the emphases that Purves and his colleagues (Brody, DeMilo, & Purves, 1989) have found in commercially published tests of high school literature achievement -- including the tests that typically accompany the anthology series.
Purves and his colleagues also noted that the cognitive demand of the typical examination question is relatively low. In the present study, the study activities were also categorized to reflect the type of cognitive activity involved in formulating a response. Four categories were used: activities that require only recall or paraphrase of the text; those that require analysis or interpretation; those that require students to apply what they have learned, or to relate a text to other texts or other experiences from their own lives; and those that ask them to create their own work of imaginative literature.
The results in Table 23 indicate that an average of 32 percent of the questions required recall or paraphrase; 42 percent required analysis or interpretation; 21 percent applying or relating what had been read to other experiences; and 5 percent creating works of their own. This pattern reflects the general three- or four-section structure of the post-reading activities that characterized most of the volumes, with simple recall, text-based inference, analysis, and extension of meaning each getting parallel sets of questions.
In the overall analyses, the characteristics of the post-reading activities outweighed those of the less-frequent pre-reading and writing activities. These did, however, show some variation from the general patterns. Prereading activities placed less emphasis on recall or paraphrase, more emphasis on applying and relating what they read to other experiences, and more emphasis on creating a new imaginative work. Writing activities similarly placed less emphasis on recall and paraphrase, and much more on creating a new work.
A few variations are noticeable. Recall and paraphrase questions received somewhat more emphasis at Grade 8 than in the later volumes, while analysis and interpretation activities increased somewhat across the grades. Nonfiction selections were somewhat less likely to invite analysis and interpretation, and more likely to be followed with recall or paraphrase. Poetry selections tended to have somewhat less emphasis on recall or paraphrase, and more on analysis and interpretation. All of these variations were relatively small, however, and variation among series was also relatively small.
In considering these emphases, it is important to put them in the overall context of an emphasis on recitation. Even when activities seem to require skills of analysis and interpretation, or to invite students to apply or relate what they have learned, the concomitant emphasis on an expected "right" answer may short circuit the value these activities might have. Rather than encouraging students' to think carefully about text, in most cases these activities simply test whether their reasoning is "correct."
On what aspects of the selections do the activities focus? Table 24 summarizes the proportion of selections where the activities gave any attention to what happens (plot, character, or setting), the author's meaning (theme or purpose), the way a selection is written (language or style), specialized literary terminology, the historical and cultural background of the selection, and difficult vocabulary.
The activities accompanying the great majority of selections gave some attention to what was happening (94 percent of the selections), to the theme or purpose (93 percent), and to the language or style of the piece (89 percent). Some 61 percent of the selections also included some activities focusing on the specialized vocabulary of literary criticism. Considerably smaller proportions provided activities related to the cultural or historical background of the piece (31 percent), or to vocabulary (21 percent).
Within this general pattern, a few variations are of interest. Attention to vocabulary development was higher in the 8th and 10th grade courses, and fell off sharply in the British literature course. An emphasis on the application of specialized terminology also reached a peak at Grade 10, and then similarly dropped off. Predictably enough, concern with "what happened" (plot, character, setting) was less apparent with poetry selections, though emphasis on who is speaking and the situation the poet might be in led to considerable emphasis on "what happened" even for poems (84 percent). Also predictably, nonfiction selections gave somewhat less attention to language and style. Concern with cultural background was highest for long fiction, nonfiction, and plays, and lowest for poetry.
Differences among series were substantial. All seven series gave relatively consistent attention to what happens, to what it means, and to language or style. But the percentage of selections with activities focusing on critical terminology varied from 43 to 90; cultural or historical background from 18 to 52; and vocabulary from 6 to 31.
Connections between Activities
There are two extremes in the way that the instructional apparatus surrounding a selection can be conceived. At one extreme would be a mix of activities designed to test students' understanding and knowledge. In this case there would be little or no relationship between activities. At the other extreme would be a sequence structured to support students' understanding, leading them through a set of interrelated activities to a fuller comprehension of the text. In this case, each activity is likely to be related to others, and there is likely to be some sort of discernible overall sequence. A prereading suggestion to "Notice how the different characters' react" might lead, for example, to a postreading request to discuss what causes their reactions, in turn followed with a suggestion for an essay comparing two or more of the characters. To examine the degree of connectivity, each activity was classified as building on at least one previous activity, as being part of a set of similar but not connected activities (e.g., a series involving identification of different figures of speech in a poem), or as being unconnected to other activities that accompany the selection. The results are summarized in Table 25.
The results indicate that there was very little connectivity among the activities included with each selection. On average, only 6 percent of the activities built upon previous ones, and another 31 percent were clustered in sets of similar types without any relationship among them. The nine study questions (each analyzed as a separate activity) provided for Stephen Vincent Benet's "Ballad of William Sycamore (1790-1871)" in the Scott, Foresman volume, Explorations in Literature (Grade 8) are typical in their lack of connectivity or cumulative impact:
The only connectivity within these questions is the parallel application of literary terminology (setting, point of view) in questions 4 and 5 and the sequence in time of questions I to 3; any of the 9 could be removed or reordered without affecting students' ability to answer the others.
Variations in connectivity with grade and series were not significant. Variations with genre were also small, though activities for nonfiction selections showed even less connectivity than did those for other genres.
Another kind of connectivity involves intertextuality, the connections that can be made between one selection and another. It is this sense of intertextual relationships that creates a sense of literary traditions, of texts and authors who share cultural values, genre conventions, or personal experiences. To examine intertextuality, raters tallied the percent of activities that made any reference to another work of literature, whether drawn from the students' personal experience or from the anthology itself.
The results of this analysis are summarized in Table 26, which indicates both the percent of elections that contained any questions with intertextual references, and the mean percent of that contained such references. Overall, some 30 percent of the selections were accompanied by at least one activity that referred to other works of literature, though this represented only a mean of 6 percent of the activities that were included. Prereading activities were least likely to make reference to other selections (averaging only 3 percent), though such activities could be an effective way to orient a reader toward related experiences or familiar traditions. Writing activities were somewhat more likely to make such references helping students tie their reading experiences together (17 percent). The following activities are typical, the first asking for comparisons within an author's work, the second for comparisons between works by different authors:
Variations with course, genre, and series were small. Even in materials for the British literature course, with its emphasis on a chronological presentation, only 32 percent of the selections, and 8 percent of the activities, made references to other selections that students might have read.
Variations in Treatment of Contemporary Works and Works from Alternative Traditions
The analyses of the instructional apparatus accompanying individual selections also looked at differences in the treatment of contemporary works, works by women, and works by nonwhite authors. Table 27 summarizes some of these comparisons.
There were no significant differences in the types of activities that accompanied selections that varied by period or authorship. The cognitive demand, the proportion of "recitation" versus "authentic" activities, and the content emphasized were very similar for the various groups of selections. (To illustrate these similarities, the specific results for recitation activities are included in Table 27.) There were noticeable differences in the supporting material accompanying the selections, however. In particular, contemporary selections and those by women or nonwhite authors were less likely to be situated in their literary or historical context than were older selections and those by male or white authors. (Selections of all types were likely to include an author biography, however.)
Similarly, contemporary selections and those from alternative traditions were less likely to be included in subdivisions that were organized by chronology or around a single major author; instead, they were more likely to occur in sections organized by literary techniques (e.g., Creating Suspense) or (in the case of contemporary selections) by genre or theme. These differences held across the three sets of courses examined (Grade 8, Grade 10, and British Literature), in spite of the overall differences in emphasis on context across these courses.8
The differences in supporting material seem a direct result of the "alternative" nature of these selections: They are not seen as part of the main line of literary development, and the alternative traditions are not themselves developed well enough within the anthologies to provide a context comparable to that of the mainstream tradition within which to locate the selections. In the case of contemporary selections (and many of the anthologized works from alternative traditions are themselves contemporary), there may also be a lack of a well-developed body of history and criticism around the work, for editors to draw upon in providing further contextualization.
Given these detailed analyses, how can the popular anthologies analyzed be characterized? The four assumptions about effective instruction-- and effective anthologies-- presented in the introduction will be used to organize these comments.
The first assumption was that the literature course should include works of substantial quality and interest, works that will promote beneficial study and discussions. Lynch and Evans (1963) concentrated much of their criticism on the failure of the anthologies available in 1961 to measure up to this assumption, criticizing much of what was included as "ephemeral" or "miscellaneous," displacing the works of enduring value that they wished to see instead. The present study, on the other hand, suggests that anthologies have narrowed their focus, presumably to place more emphasis on works of merit, and certainly to reduce the amount of "miscellaneous nonfiction." Although a wide variety of authors and works are included in the various series, with very few works included in all seven of the series examined, the number of unique works (those included in only a single series) has declined noticeably over the past 30 years. In a similar fashion, the proportion of works published within the previous 60 years has also been reduced, from over half of the selections examined by Lynch and Evans to about a third in the 1989 anthologies examined in the present study.
The second assumption at the beginning of this study was that the anthologies should recognize and incorporate the contributions of diverse groups to America's shared literary heritage. Over the past 30 years, literature anthologies have broadened their selections to include a wider representation of works by women and Of works from alternative literary traditions. This is particularly true in the volumes intended for use in Grades 7 through 10; those intended for American or (particularly) British literature courses remain, in comparison with those intended for the earlier grades, quite narrow in their representation both of women and of nonwhite authors.
Though representation of alternative literary traditions has increased, the amount of attention given to any one of these traditions remains very small, and the selections are less likely to be placed into a social, historical, or literary context than are selections drawn from mainstream traditions. Works by women and nonwhite minorities are most likely to be included among selections drawn from the 20th century, and least likely to appear in chronologically organized courses that emphasize older works. It is hard to imagine that the handful of selections by African American, Hispanic, Asian, or Native American authors, for example, are sufficient to leave students with a unique sense of the substance and appeal of these alternative traditions, but neither are they well-integrated into a larger, common tradition.
The third assumption was that the course as a whole should emphasize reasoned and disciplined thinking rather than simply recitation of details or of interpretations presented by the teacher or text. In contrast to this assumption, the instructional apparatus that surrounds the anthology selections is overwhelmingly text and content centered, with little attention paid to the development of students' abilities to think on their own. Following in a New Critical tradition, most anthologies base their major divisions or subdivisions on genre characteristics. Study activities emphasize text-based comprehension, beginning with simple recall and paraphrase and working from there toward analysis and interpretation. An overwhelming proportion of the study activities involve recitation, where there is a presumed single right answer; only about one third of the activities leave room for students to develop and defend their own interpretations and points of view.
The final assumption with which the study began was that the study apparatus should be coherent and cumulative, leading the reader toward a more carefully thought-through understanding of a text rather than treating a text as a series of unrelated "puzzles" to be solved. The anthologies, on the other hand, seem to assume that students build understanding out of individual details, rather than from some cumulative understanding of an evolving text. As a result, there is little connection among the activities that accompany a given selection. Across selections, an average of only 6 percent of the activities build on one another; the remainder are discrete and independent rather than cumulative.
The anthologies as they are presently constructed have responded to past concerns about content, broadening somewhat the representation of alternative traditions while at the same time reducing the amount of "ephemera" and "miscellany" about which Lynch and Evans (1963) had earlier complained. Compared with the volumes that Lynch and Evans reviewed, the selections are probably more appealing and also more teachable, in the sense of having the weight and substance to promote interesting discussion and debate. (They also may be somewhat more difficult and distant from the students' immediate experiences.)
But if many of the selections seem capable of promoting worthwhile discussion, the instructional apparatus that surrounds the selections does not. The instructional apparatus reflects a particular tradition in the teaching of literature, one that emphasizes the primacy of the text rather than the transaction between reader and text. At a time of debate and change in the profession at large, the 1989 anthologies seem remarkably consistent in their emphases, caught in an earlier tradition of text comprehension and analysis rather than attempting to implement any of the recently offered alternatives. Many of the volumes have added layers of attention to reading processes and to historical and literary context, but these remain as ancillary to the overall emphasis. What all of the texts lack is an integrated, cumulative, and coherent effort to involve students in the ongoing cultural dialogue about the human condition that literature at its best demands and to which it contributes.
1 American literature in Grade 11 followed by British literature in Grade 12 was also the most typical sequence in the secondary school English programs studied previously; this sequence was reported by half of the schools surveyed, a considerably higher proportion than reported any other sequence (Applebee, 1990, p. 40).
2 Because pages are formatted differently even within the same volume, each page was treated as consisting of two columns of text. Whether it was physically set as one or two columns, a selection (of any genre) printed alone on a page was counted as two columns. Similarly, a selection that took up half a page, with the remainder devoted to study activities, was coded as one column.
3 A subset of the volumes they studied are directly comparable to the present study. This subset includes the most recent editions of 10 publishers' series, Grades 9-12, intended for average or college-bound students. This subset will be used for direct comparison with the present study, in those cases where Lynch and Evans reported their results in enough detail -to allow individual series to be separated out.
4 That is, from an average of 689 pages per volume for the 10 comparable series in Lynch & Evans (1963, pp. 474-475) to 1011 pages per volume, and from an average of 120 selections (p. 23) to 144.7 per volume, for Grades 9 through 12.
7 There is a strong context effect at work in the way questions are presented. Activities flagged as "literal" in the text or teacher's manual were scored as recitation in this analysis-even though the same activity could be quite open-ended in a context that set different expectations, or with a slight rewording.
8 Thus for Grade 8, 51 percent of the selections by white authors were given some literary context, compared with only 23 percent of the selections by nonwhite authors. In British literature, the comparable percentages were 75 for white authors, and 0 for the nonwhite authors sampled. Similarly, at Grade 8, 19 percent of the selections by men had some social or historical background provided, compared with 6 percent for women. In British literature, the comparable percentages were 95 percent for men and 86 percent for women.
Elliott, D.L., & Woodward, A. (1990). Textbooks, curriculum, and school improvement. In D.L. Elliott and A. Woodward (Eds.), Textbooks and schooling in the United States. 89th Yearbook the National Society for the Study of Education (pp. 222-232). Chicago, IL: University of Chicago Press.
Langer, J.A. (1989). The process of understanding literature. Report Series 2.1. Albany, NY: Center for the Learning and Teaching of Literature.
A Study of High School Literature Anthologies
Arthur N. Applebee
A Study of High School Literature Anthologies is based on research conducted at the Center for the Learning and Teaching of Literature. It is now distributed by the National Research Center on English Learning & Achievement (CELA).
CELA is a national research and development center located at the University at Albany, State University of New York, in collaboration with the University of Wisconsin-Madison. Additional research is conducted at the Universities of Oklahoma and Washington.
Established in 1987, the Center initially focused on the teaching and learning of literature. In March 1996, the Center expanded its focus to include the teaching and learning of English, both as a subject in its own right and as it is learned in other content areas. CELA's work is sponsored by the Office of Educational Research and Improvement (OERI), U.S. Department of Education, as part of the National Institute on Student Achievement, Curriculum, and Assessment.
The Process of Understanding Literature is based on research supported by OERI (Grant number G008720278) and the National Endowment of the Arts (NEA). Distribution is supported in part under the Educational Research and Development Centers Program (award number R305A60005). However, the contents do not necessarily represent the positions or policies of the Department of Education, OERI, the NEA, or the Institute on Student Achievement.
For information about current publications and activities, write to CELA, University at
Albany, School of Education, B-9, 1400 Washington Avenue, Albany, NY 12222; or send email