Reading Psychology, 21:353–372, 2000 Copyright © 2000 Taylor & Francis 0270–2711/00 $12.00 + .00 THE DILEMMA OF ERROR AND ACCURACY: AN EXPLORATION MARY ALICE BARKSDALE-LADD and JAMES R. KING University of South Florida, Tampa, Florida, USA This paper reports on the research on error and accuracy within the realms of constructivism, reading, and writing, and describes an exploratory study of teacher perceptions of approaches to addressing error and accuracy in literacy instruction. Findings demonstrate that teachers who self-reported that they were developing constructivist approaches to classroom instructional practices had contradictory beliefs about dealing with reading and writing errors. There were also contradictory beliefs about addressing reading and writing errors with moreand less-able readers and writers. The teachers were solid in their conviction that students’ inaccurate constructions of knowledge should be corrected through inquiry, but were unable to provide examples of instances in which they had done so. During this century, literacy researchers have studied errors in reading and writing extensively. Teacher treatment of errors in reading and writing has also been significantly addressed (Johnston & Allington, 1991). As instructional approaches have turned toward constructivism, there have been many changes in perspectives on the meanings of teaching and learning (Gergen, 1985; Spivey, 1997). However, little has been written about how teaching approaches founded in constructivism might lead to necessary alterations in the meanings of error and accuracy and what they can represent within the learning process. The constructivist perspective holds that people are constructive agents who learn and know through the creation of constructed products (Bruner & Feldman, 1990). Constructivists see learning as a process in which individual ways of knowing, understanding, perceiving, and valuing are at the essence of the construction of meaning. Constructivist teachers create learning experiences with Address correspondence to Mary Alice Barksdale-Ladd, 3920 Florida Ranch Blvd., Zephyrhills, FL 33541. E-mail: mabl@typhoon.coedu.usf.edu 353 354 M. A. Barksdale-Ladd and J. R. King the expectation that students may learn differently from each other and differently from the teacher’s expectations. Constructivist teachers utilize collaborative groups and different types of social interactions that allow students to learn from one another and share their constructed learning publicly. Using individual performances as products, a group sense of what counts as knowledge emerges (Spivey, 1997). If we truly invest ourselves in constructivist pedagogy, an important outgrowth becomes the question, What is error? And, what is accuracy? If the meanings learners construct through their experiences are viewed as the closest approximation to accuracy of which they are developmentally capable at a given time, can these constructions be errors? Should constructivist teachers simply ignore what they see as an error in the hope that more accurate constructions will appear later? No, we are not suggesting that teachers encourage students to construct inaccurate understandings and accept them as truth. Instead, our interest is in the study of how teachers construct understandings of errors and how they deploy these understandings during literacy instruction with children. To this end, we examine the literature on errors in reading and writing. In developing understandings of how children acquire reading abilities, why reading problems occur, and how they can be corrected, the investigation of students’ errors has been a primary route taken in the construction of knowledge by researchers (Allington, 1980, 1983; K. Goodman, 1965; Y. Goodman & Burke, 1972; Johnston, 1993). Teachers also have a history of attending to student errors. As children learn to read, errors are pointed out and children are assisted in learning how to be error-free in their reading (Allington, 1983). From early childhood, our society teaches us that our reading is at its best when there are no errors. During their preservice programs in elementary education, students are often given intensive instruction in how to assess error in reading. When we look at our societal and teacher education focus on the analysis of errors in reading, it is no surprise that teachers of reading are fixated on errors. Examination of students’ errors or miscues during their literacy work is not a new practice. For example, Powell (1970, 1980) devoted considerable attention to readers’ oral reading errors in order to determine the validity of IRI reading levels. In another example, the highly influential miscue analysis (K. Goodman, 1969; Error and Accuracy 355 Y. Goodman & Burke, 1972) as well as the currently ubiquitous Running Record (Clay, 1993) both closely examine the quality of readers’ miscues while the readers were engaged in the act of reading. In fact, the current study owes much to previous examinations that count and categorize miscues. The current study, however, differs from previous studies in that the judgment of reading and writing “error” as well as the description enabled by the “miscue” are being questioned. This study intends to question, or problematize, the very notion of error in literacy. It is fascinating that there has been so little research on teacher responses to errors in reading. Allington (1983) showed that teachers tend to interrupt weaker readers more frequently than better readers when they made oral reading errors. These teachers felt that poorer readers needed correct word identification if they were to comprehend properly, whereas stronger students were more likely to comprehend despite errors, or that they would correct their own errors. The literature does not provide evidence that correcting students’ reading errors improves their ability to read. Allington (1980, 1983) concludes that the correction of oral reading errors is disruptive to comprehension. Further, there is no direct evidence that when a teacher corrects a reading error, the likelihood of repetitions of the error are decreased. We do know that errors in reading are never random and we know a great deal about what reading errors may be able to tell teachers (Johnston, 1993). When a reader self-corrects a reading miscue, this reader is on the verge of “knowing” the word and understanding how to identify it. When a reader knows an error has occurred, but is unable to self-correct, the child is ready for instruction which will make it possible to avoid this miscue (and possibly others) in the future. When a reader makes a miscue and does not realize that it has occurred, there are two likely conclusions. One reasonable conclusion is that the reading material is simply too difficult for the child. A second likelihood is that the child is not yet developmentally ready for instruction that would make it possible to correct the miscue in the future (Johnston, 1993). Thus, careful attention to children’s responses to their own oral reading miscues could form a foundation for teacher decisions about appropriate reading instruction. However, we find no evidence that teachers engage in this type of thinking about students and their errors in reading. 356 M. A. Barksdale-Ladd and J. R. King So, how do teachers think about errors in reading? When do they feel that errors should be addressed, and how do teachers report that they address reading errors? These research questions are addressed in our study, but first, let’s move on to error in writing. Errors in writing have been studied by researchers with the same enthusiasm as errors in reading, and we know that, historically, teachers have been fastidious about error identification in writing (Haswell, 1988; Williams, 1981). Because there are specific rules for writing, it would seem reasonable to assume that the identification of errors in writing is a fairly straightforward process. It is not. Conners and Lunsford (1988) found that teachers marked 43% of the most serious errors (top 20), ignoring a whopping 57%. There was much variability in teacher perceptions of serious, markable errors, and the reasoning behind marking or not marking each error was amazingly complex. Reasons for differential marking of writing errors included (a) the perceived need of the student, (b) the stage of the composing process, (c) how serious or irritating the error was where it occurred, and (d) how much difficulty was involved in marking the error and explaining the mistake. The identification of errors in writing turns out to be a highly personal process steeped in personal impressions about students, beliefs about writing and the writing process, and instructor biases. Another aspect of the complexity of identifying writing errors involves the focus of the instructor. One approach to identifying writing errors is to focus on use of words, sentence construction, punctuation, spelling, and grammar. The problem with this approach is that it causes the instructor to lose sight of the content. When content becomes the focus, it is not possible to simultaneously attend to the words, sentences, punctuation, spelling, and grammar (Williams, 1981). It appears to be impossible to maintain a simultaneous focus on both technical aspects of writing and the content of the paper. Thus, instructor focus tends to vary during the reading of student texts. Because instructor focus varies between the content and the mechanics, and is a function of choice from moment to moment, the identification of error in students’ writing is as much about the instructor as it is about the student and the writing. Further, writing errors have two quite different foundations—the grammarian’s handbook and the instructor’s mind. Error and Accuracy 357 If errors in writing are to be identified by teachers, what is the purpose of their identification? As was the case in reading, we found no evidence that marking errors in writing leads to improved writing on the part of students. To the contrary, Haswell (1988) concluded that errors in writing are a necessary part of writing improvement, and identifying and correcting writing errors may actually be detrimental to writing development. For instance, as writers mature, they attempt to create more complex sentences. During these attempts at greater complexity, numbers of certain types of writing errors may increase (Freedman & Pringle, 1980; Steward & Grobe, 1979). As was also the case in reading, writing errors are never random, and they can be viewed as “evidence of intention” (Bartholomae, 1980, p. 255). When asked to read their written products orally, the students in Bartholomae’s study corrected their mistakes, without recognizing the mistakes or realizing that they had corrected them (this finding was also reported by Perl, 1979). When these students read their writings, they demonstrated grammatical competence. Their writing errors were reflective of a problem in the performance of writing, not in the ability to use language correctly (Bartholomae, 1980). This research may have numerous implications. In school, developing writers are generally taught to use a process approach to writing. Students select topics, write drafts, revise, edit, and share their written products with audiences. If college-level basic writers are unable to see and correct the errors in their own writing because their meanings are set and well understood, it would appear to be quite unreasonable that we would expect emergent writers to experience success in editing their own work. The studies of errors in writing that have been examined here all involved adult writing and college/university instructors. The degree to which findings regarding error in adult writing relate to error in children’s writing is unclear, and there are many possible lines of inquiry here that are yet unexplored. A first step is to examine teacher perceptions about children’s writing errors. Research Questions This study is an initial step in exploring how teachers think about responses to student errors in reading and writing. We asked the 358 M. A. Barksdale-Ladd and J. R. King following research questions: (a) how do teachers think about errors in reading? (b) when do teachers feel that reading errors should be addressed, and how do teachers self-report that they address reading errors? (c) how do teachers think about errors in writing? (d) when do teachers feel that writing errors should be addressed, and how do teachers self-report that they address writing errors? (e) how do teachers use knowledge gained from attention to student reading and writing errors in their teaching? and (f) how do teachers address inaccurate student constructions of knowledge? Method Because it was an exploratory study, a multi-method, multi-site qualitative interview format was utilized. A scenario was constructed for opening an engaging discussion on accuracy and error (Appendix A). The example was framed in language arts and related to science. Anticipating that some of the participants would struggle with the concepts in the scenario, we thought it would create a disequilibrium that might disrupt normal ways of thinking about accuracy and error and lead to an open discussion of the points addressed in the research questions. Each participant was presented with the scenario immediately prior to the interview, and after participants had completed their reading of the scenario, we began the interviews. Participants Nineteen female teachers participated in the study—a group of six, a group of three, and a group of ten. All of the teachers selected for the study self-reported that they were engaged in the development of constructivist approaches to their classroom instruction and all taught reading and writing in their classrooms. Teachers in the populations from which we drew our participants who did not feel that they met these criteria did not participate in the study. Six of the teachers were taking a course from Dr. BarksdaleLadd, author 1. They were all students in a Master’s degree program in reading and taught in schools in a southern state. This course occurs at the end of the program; thus, these teachers could Error and Accuracy 359 be considered knowledgeable about literacy and literacy instruction. These teachers participated in an audiotaped focus-group interview based upon the scenario designed for the study. This interview took two hours and fifteen minutes. A second focus-group interview was conducted with three elementary teachers who each held a master’s degree in reading. These teachers were selected to participate in the study because of their instructional focus upon writing across the curriculum and because of their involvement in a school-wide “Students as Authors” project (Barksdale-Ladd & King, 1997). These teachers also participated in an audiotaped focus group interview based upon the scenario. The interview lasted 55 minutes. The third group included ten doctoral students in a language arts theory course taught by Dr. King, author 2. The ten were all classroom teachers who taught reading and writing. These participants interviewed one another in pairs (using the scenario) and took detailed notes during the interviews. One week later, each participant submitted a detailed accounting of the results of their interview, reporting the other’s treatment of and beliefs about error and accuracy in reading and writing. The paired interviews were of an hour to and hour an fifteen minutes in length, including both interviews. Among the nineteen participants, seventeen were elementary (K–5) teachers, one was a middle school reading teacher, and one was a high school English teacher. This group of teachers can be considered exemplary in that they all held or were seeking Master’s or doctoral degrees in reading. The teachers had between 3 and 24 years experience, with a mean of 8 years of experience across the group. Analysis The tape-recorded focus-group interviews were transcribed verbatim. The transcripts were read by both researchers for categories. Recommendations for the analysis of interview data as described by Hycner (1985) were utilized. We met to compare categories and reached agreement on a limited set of themes. Then we reread the transcripts for instances of the themes. Examples of themes were checked with each other. Next, the reports written on the peer interviews were used as texts and were read by both research- 360 M. A. Barksdale-Ladd and J. R. King ers. We isolated themes in the reports and again elicited examples of the themes from the reports. Comparisons between the focus group interview themes and examples and the peer interview themes and examples were made, demonstrating that there were no qualitative differences between the two data sets. Pseudonyms have been used in reporting the results to assure participant confidentiality. Results Primary themes included teacher perceptions about: (a) addressing writing errors, (b) addressing reading errors, (c) addressing reading and writing errors with more and less skills readers and writers, (d) using knowledge gained from student errors to inform instruction, and (e) student construction of inaccurate knowledge. Writing Errors The nineteen teachers who participated in this study overwhelmingly agreed that children should be taught to use the writing process, and thus, that writing errors should be allowed and even encouraged during prewriting, drafting, and revising. What Anna said was representative: “We want them to attempt and risk. Miscues and invented spellings are encouraged without the threat of needing accuracy.” Sue furthered this perspective: “You put accuracy on the back burner because otherwise, you just squelch forever that creative flame.” The teachers provided a commonly understood rationale for allowing and encouraging errors during the early stages of the writing process. Alison explained, “In writing, the children know what their ideas are, and if the words aren’t exactly correct, they can still express their thoughts.” Because writers own the ideas about which they are writing, the teachers agreed that there was no need to be concerned about writing errors; after all, those errors did not reflect a lack of understanding about the topic. Writing errors were accepted as a natural and necessary part of the writing process. The teachers also agreed that at the later stages of the writing process—editing and publishing—writing errors should be corrected. As Mattie explained, “As the process develops, we go on Error and Accuracy 361 down to revising and editing. In editing, with the teacher or a peer editor, that’s when you start to get a little more particular about what is written as far as spelling and punctuation and grammar.” The teachers had somewhat different strategies for assisting students in correcting their errors during the editing process; however, all of the seventeen elementary teachers placed themselves in the position of final editor. For instance, Mattie said that she first used peer editing and encouraged the children to work together to correct all of the errors they could find. Then, she came in as “top editor,” because “there are some things that other kids miss, especially when it comes to making things sound correct grammatically. I have to do that.” Our understanding of this approach is that errors are objects that exist independently of the teacher. They can and should be found by the students. And if not, they should be pointed out by the teacher, who invariably served as final editor. Yet, if errors do not exist until they are identified, then students’ identification of their own errors can be seen as a writing self-destruct mechanism. Teachers are asking students to internalize an irrational selfcritique. Cora, teaching at the tenth-grade level, treated editing differently: “The idea for my kids is that they are the top editors for themselves. I don’t really have a lot to do with it. That’s a hard thing to teach, though. Sometimes I struggle with showing them how to look at someone else’s paper and find and correct all the mistakes.” Celeste felt that, for her, it was very important to have children read their writing to her before she looked at it. “That gives me a chance to listen, so that I don’t just zero in on the mistakes the minute I look at it; it give me a chance to really enjoy what they’re saying and respond to what they’ve said first.” For the teachers from the central Atlantic state, Anna, Sophie, and Sue, there was a built-in motivation for children to edit their work at the end of the writing process. Children writing books and stories in the “Students as Authors” project knew that at the end of the process, their work would be photocopied, laminated, and put in a ring binder to be sold as a book. “Students as Authors” project books are sold to parents, members of the community, and other children, and authors are paid $3.50 when a solely authored and illustrated book is sold for the first time (future pro- 362 M. A. Barksdale-Ladd and J. R. King ceeds go to the project to support copying and production costs; Barksdale-Ladd, & King, 1997). As Sue stated it, “Accuracy becomes an integral part of the whole process; it is a natural flow. They do it and they don’t realize they’re doing it . . . because when they get to the end, they want it to be the best that it can be.” Sophie continued: “When we’re writing in the book project, it’s not just for school . . . the need for accuracy comes from themselves. It just comes naturally because the students work with each other and they want it to be a perfect finished product.” Reading Errors When asked about responses to student errors in reading, all of the teachers shared some basic common beliefs. If a teacher perceived that a student’s error would interfere with the understanding of the text, the teacher felt that the error should be corrected. They indicated that it would be good to wait until the end of a sentence before correcting the error, but if the student did not self-correct by that point, the teacher should step in and either provide the correct word or help the student in identifying the correct pronunciation of the word. Winnie had a representative response: Well, if, for example, they say “cot” instead of “cat,” the whole meaning of the sentence is off, and then, if for the rest of the story they’re saying “cot,” and it’s supposed to be “cat,” they’re not going to understand what the story is about. I want them to get the correct word. I sort of talk them through it and give them strategies. ‘Look at the picture. Think about what the word might be. Look at the letters. What might it sound like?’ Then, once they get the word, I have them start the sentence again. As Winnie alluded, the teachers had a rationale for dealing with reading errors with immediate corrective actions. Alison explained, When children are reading, if they aren’t reading the correct words that are there, they’re not going to get the correct idea of what is going on. These are not ideas from their heads. This is reading. If they’re doing reading, they’re probably reading something that was written by someone else, and they don’t know what that person’s thought was, so if they can’t read those words, they’re not going to have the idea. . . . So, in reading, I’m more critical about their correctness. Error and Accuracy 363 Because word-for-word accuracy in reading was viewed as an essential aspect of comprehension, these teachers felt it was important that they encourage this type of reading accuracy. The teachers described two methods for word identification; either they told the student the word to avoid embarrassment or loss of train of thought, or they asked the students to sound out the words. Anna and Sophie were the only two teachers who had differing opinions on this point. They were less concerned with word-toword accuracy than some of the other teachers. They felt that when children encountered words they didn’t know, rather than telling the children the words or asking the students to sound out the words, the children should try to identify the words from the context. Sounding very frustrated, Anna said, No matter how you drill it, no matter how you model it, they still think that the only way they can figure out a word is to decode it. I mean, it takes me forever to teach them that you could skip that word and finish the sentence and even read the paragraph, maybe, and from context, you may find that you don’t need that word, or you may figure it out. But they are still so bent on figuring out the word! Differences for More- and Less-Able Readers and Writers We asked no interview questions about differential approaches to reading and writing errors for more- and less-able readers and writers; however this topic was brought up by all of the teachers. Interestingly, approaches to dealing with reading and writing errors differed for more- and less-able readers and writers. Better readers were more likely to be allowed to “get by” without having oral reading errors corrected, whereas reading errors for poor readers would be corrected. Bess was representative in saying, I tend to let the better readers go. If the words don’t change the meaning or something, I ignore it. I let it go, especially if they’re more fluent readers because I know that their eyes are moving faster than their mouth is . . . But with the ones that are struggling with reading, I give them help on figuring out words. I think it’s more important for them than the better readers. To the contrary, accuracy was more likely to be expected of moreable writers than of less-able writers. Errors would not be tolerated 364 M. A. Barksdale-Ladd and J. R. King for good writers. For poor writers, any effort was seen as positive, therefore errors were more likely to be accepted and not corrected. As a follow-up question to a scenario described by a participant, one interviewer asked, “What if that student showed you her writing and when reading it to you, read a word spelled ‘G I P’ as ‘pig’?” Winnie answered, Well, it depends on the student. If I know that a student knows pig, I would talk to them until they corrected it themselves. But if one of my low children, if one of them wrote pig like that, I probably would let it go. I would probably be impressed if a low child knew that those letters were in pig . . . If I expected them to do the same as one of my higher children, I would be afraid they would get totally frustrated and be turned off by the whole process. So, that’s why I try to individualize it to their developmental level. Using Knowledge Gained from Student Errors to Inform Instruction The teachers in this study all took the position that it was important to pay attention to student errors and use information gained from student errors to inform instruction. Sixteen of these teachers were not able to think of examples of ways in which they used knowledge of errors to inform literacy instruction. Three of the nineteen were able to discuss their use of errors to inform instruction. Referring specifically to writing, Celeste said, I think we can learn a whole lot from their mistakes. We’ve finally started to learn that if we’ll just listen to the kids, they’ll tell us stuff that will help us become able to teach them. . . . You have to know every single child, you have to do diagnostic work on that child, you have to have observation notes, and it takes a lot more work than just giving an assignment from the textbook. But the pay-offs are there if we’ll just stick with it. A second teacher, Cora, had developed a systematic method of making use of her students’ errors in writing: I do, kind of like a daily oral language. About twice a month I do a diagnostic writing assessment across all of the students in each class to see where they are, and I pull all of their miscues out, and then I put those in a daily language story. We go over how to correct all of those particular kinds of miscues, and I find that a lot of times, when I look at the next diagnostic, those errors are gone. So, there is some improvement by taking Error and Accuracy 365 their miscues and doing something with them. Another good thing is that if you use their mistakes, you don’t have to worry about all of that other stuff in the curriculum. If all they have problems with are commas, then you can just focus on commas. In reading, only one teacher discussed using student errors to inform instruction. Winnie worked with each of her first-graders individually once a week in reading, following steps in an early intervention instruction model. As a part of this work, she took running records with her children and used what she learned from their miscues to direct her decisions about materials and instruction in future individual sessions. The Construction of Inaccurate Knowledge We asked the teachers to talk about their responses to instances in which children constructed knowledge which was inaccurate. Clearly, all of the teachers felt that if children constructed inaccurate knowledge, it was essential that teachers take it upon themselves to assure that these inaccuracies were corrected. But the preferred method of correcting factual inaccuracies was definitely not to say, “No. You’re wrong,” and provide the corrected information. Instead, these teachers believed they should first attempt to determine why or how the students constructed their inaccurate responses. Then, it would be most appropriate to direct the children in conducting further research on the topic being considered, with the hope being that the children would discover their own inaccuracy and correct it themselves. This approach to looking at inaccurately constructed knowledge was easily verbalized, and nearly a refrain across the interviews. None of the nineteen teachers were able to provide an example of an instance in their classrooms in which students inaccurately constructed knowledge and an inquirybased model was used in dealing with the inaccuracy. Discussion This study was designed as an exploration of similarities and differences in perceptions of teachers who self-reportedly were making attempts toward the establishment of constructivist classroom environments. Our participants were students, friends, and 366 M. A. Barksdale-Ladd and J. R. King people we considered to be reflective teachers. Thus, in examining the results, we engaged in a struggle between wanting to present these thoughtful, hard working teachers in a favorable light and our own critical views of the information gleaned from the interviews. That is, our awareness of power issues and the notion of accuracy as beliefs created in a search for “truth” have been at the forefront of our thinking in examining the data. We realize that as professors and researchers, our perceptions of this data are representative of a level of power over the participants, and our critical discussions of the findings and their implications have a likelihood of being viewed as “accurate” in the eyes of these participants and others with whom our work might be shared. It is important as we begin our discussion that we point out that our perceptions are, for us, the beginning of what we believe will be a fruitful line of inquiry into error and accuracy in the classroom. We plan for more in-depth examinations of these issues, and we expect that our perceptions on error and accuracy will change considerably over time. The findings demonstrated that these teachers held accuracy as an area over which they had classroom ownership, particularly with regard to reading errors and inaccurate constructions of knowledge. These teachers found it irritating that, when students made oral reading errors, other children would interfere with the teacher’s ownership of accuracy by correcting the errors before the teacher had determined what was to be done about the errors. When students had constructed knowledge that the teachers perceived to be inaccurate, they were solid in their conviction that they should guide the students toward correct constructions. With regard to teacher perceptions of student errors in writing, it was apparent that all of these teachers had fully “bought into” the writing process model and were confident that student errors were to be accepted during the early stages of the writing process and eventually corrected through peer and teacher editing. Still, teacher ownership of accuracy was an important classroom responsibility. Teachers reported awareness of the errors in texts produced by emerging writers, but deferred action on them to make space for the students’ developing abilities with writing. It is also apparent that there was some ambiguity in what counted as writing error, and when to attend to these artifacts of the writing process. In reading, these teachers were generally intolerant of errors Error and Accuracy 367 and felt that they should be corrected with some immediacy. When comprehension errors occurred, the teachers believed they should lead students toward more correct understandings. The teachers stated that they might wait until the end of sentence to correct an oral reading error, but not much longer. There was a clear, unwavering, shared teacher belief that comprehension could not be assured unless students were able to identify individual words accurately. Reading theorists have suggested that the correction of oral reading errors by teachers may be more disruptive to comprehension than the inability on the part of the student to identify individual words (Allington, 1980, 1983; Johnston & Allington, 1991; Smith, 1994). The strength and solidness of this shared teacher belief was fascinating. Most of these teachers (13) held a Master’s degree in reading or literacy, and the remaining 6 were taking their final course in a Master’s in Reading program. Surely, at some point, these teachers had been exposed to the notion that the correction of reading errors could be disruptive to comprehension. After completing our analyses, we remained curious about why these teachers were so solid in their conviction that reading errors must be corrected, and we recommend further research on this phenomenon. One possible way of viewing writing and reading processes is to see both as having flexibly definable stages in which it might be more or less acceptable to correct errors. The writing process can be treated as a flexible process model. The teachers in our study agreed that student errors should be accepted during early stages of writing such as prewriting, drafting, and revising. Only during editing was it considered acceptable for teachers to correct student errors. The analogy to reading could posit that early encounters with the reading of a given text would be viewed in the same light as the beginning stages of the writing process. Would it be reasonable to assume that after a child encountered a text several times, he or she would have reached a more advanced process stage, and thus, it would now be acceptable for a teacher to correct reading errors? If a child had encountered a text several times, is the likelihood high that errors in the earlier readings would have been identified and self-corrected, thus any remaining errors might represent areas of interest for the teacher to examine with regard to the fruitfulness of possible instruction? Perhaps so, perhaps not—as we said, this is an exploratory study. We wonder 368 M. A. Barksdale-Ladd and J. R. King how many times a child can be reasonably expected to explore the same reading material without becoming bored with it. Continued investigations into this line of inquiry may lead to the design of a model that will support higher levels of student ownership of reading errors. The findings related to teacher approaches to errors for better and poorer readers and writers were a surprise. For better readers, oral reading errors were likely to be tolerated, and for poorer readers, oral reading errors were more likely to be corrected (as reported by Allington, 1980, 1983). On the other hand, for better writers, errors were more likely to be corrected and similar errors by poor writers were likely to be tolerated. These teachers held similar perceptions on methods of dealing with reading and writing errors that were fully contradictory across the two processes, but that they viewed as entirely appropriate. This discrepancy could be explained by teachers’ perceptions of the needs of the learners. There were indications that teachers felt that poor readers tended to make more errors and greater numbers of errors were more likely to disrupt comprehension; therefore, there was a greater need for these errors to be addressed. With regard to writing, the teachers may have felt that the better writers had sufficient mastery of written language expression to benefit from correction, whereas the poor writers would benefit from experimenting with language. A further explanation for this contradiction is that these teachers were struggling in the process of moving from more traditional to more constructivist models. While they recognized errors as reflective of inaccurate productions, they were inconsistent in their responses to these productions across the reading and writing processes. This struggle to understand was also reflected in teachers’ perceptions on making use of student errors to inform instruction. The teachers seemed strongly in agreement with the notion that they should be paying attention to errors and collecting information on student accuracy and error. However, of the nineteen teachers, only one had a systematic method of looking at student errors in writing, and only one had a similar method for examining student reading errors. These teachers knew that there was something important to be learned from student errors, but they hadn’t developed methods of utilizing information gained from the study of student errors in literacy. Error and Accuracy 369 This also held true with teachers’ perceptions about student constructions of inaccurate information. The teachers were sure that when students engaged in constructivist learning activities that resulted in inaccuracies, it was essential that they step in and assure that these inaccuracies be corrected. Yet, teacher methods for approaching the problem were described in sketchy terms, and teachers were not able to describe classroom instances in which these approaches had been used. The teachers felt that it would be best that they not merely announce that students had constructed incorrect representations of knowledge. Rather, it would be advisable to lead the students into further inquiry into the topic. If teachers redirect students into further inquiry, isn’t it obvious that a problem has been identified which necessitated further study? We wondered how students could be redirected without an implicit announcement that their constructions of knowledge were incorrect. The problem of calling for more student research on constructed knowledge without identifying an error is difficult to address. Perhaps it is because of this difficulty that these teachers were unable to provide examples of ways in which they had addressed students’ incorrect constructions of knowledge in the past. This study involved teacher self-report in many ways. The teachers self-reported that they saw themselves as being engaged in a process of becoming constructivist in their approaches to teaching. They self-reported their beliefs about errors in reading and writing, and they self-reported their practices in addressing student errors. A next step, in our view, is to move beyond self-reports, and study actual teacher practices in addressing student errors. To what degree do their self-reports match their instructional behaviors? There are other unanswered questions that yet need to be addressed. For instance, do teachers of various age groups differ in their professed beliefs in regard to correcting reading and writing errors? Does the teaching approach used in the classroom affect the respondents’ answers? Does the grade or content taught affect the teachers’ perceptions of student errors? Do current pressures to assure high test scores affect teachers’ perceptions of student errors? Were the errors of poor readers, which were more likely to be reported as receiving corrective feedback, more debilitating to textual reading than errors of better readers? How do the perceptions of the quantity of errors affect correction? Do children who 370 M. A. Barksdale-Ladd and J. R. King make a greater number of errors receive proportionally more or less correction, or the same amount of correction as those who make fewer errors? Errors are at the very heart of constructivism. It is at the point of discovering error that it becomes possible to discount specific approximations and embrace others. Without a full understanding of the roles of student error and accuracy within reading and writing processes, and the stance of teachers with regard to error and accuracy, attempts at developing constructivist-based approaches to teaching and learning may be at best misguided. We suggested at the onset that perceptions about student errors say more about the teachers who identify those errors than about their students, and these results demonstrate the value in this statement. Teachers are clearly the owners of accuracy in literacy instruction and their perceptions of approaches to addressing student error loom over student development in literacy, although these perceptions are contradictory and, in some cases, lack utility during classroom interactions. Within the constructivist domain, we have made little progress toward developing understandings of error and accuracy that can serve a solid and meaningful function within classrooms. References Allington, R. (1980). Teacher interruption behaviors during primary grade oral reading. Journal of Educational Psychology, 72, 371–379. Allington, R. (1983). The reading instruction provided readers of differing ability. Elementary School Journal, 83, 255–265. Barksdale-Ladd, M. A., & King, J. R. (1997). The worlds of a reader’s mind. Students as authors. The Reading Teacher, 50, 564–575. Bartholomae, D. (1980). The study of error. College Communication and Composition, 31, 253–269. Bruner, J., & Feldman, C. (1990). Metaphors of consciousness and cognition in the history of psychology. In D. Leary (Ed.), Metaphors in the history of psychology (pp. 230–238). New York: Cambridge University Press. Clay, M. (1993). An observation survey on early literacy achievement. Portsmouth, NH: Heinemann. Conners, R., & Lunsford, A. (1988). Frequency of formal errors in current college writing, or Ma and Pa Kettle do research. College Communication and Composition, 39, 395–409. Freedman, A., & Pringle, I. (1980). The writing abilities of a representative sample of grade 5, 8, and 12 students. (ERIC Document, ED 217 413). Gergen, K. (1985). The social constructionist movement in modern psychology. American Psychologist, 24, 266–75. Error and Accuracy 371 Goodman, K. (1965). A linguistic study of cues and miscues in reading. Elementary English, 14, 121–129. Goodman, K. (1969). Analysis of reading miscues: Applied psycholinguistics. Reading Research Quarterly, 5, 9-30. Goodman, Y., & Burke, C. (1972). Reading miscue inventory manual. Procedures for diagnosis and evaluation. New York: Macmillan. Haswell, R. (1988). Error and change in college student writing. Written Communication, 5, 479–499. Hycner, R.H. (1985). Some guidelines for the phenomenological analysis of interview data. Human Studies, 8, 279–303. Johnston, P. (1993). Constructive evaluation of literate activity. New York: Longman. Johnston, P., & Allington, R. (1991). Remediation. In R. Barr, M. Kamil, P. Mosenthal, & P. D. Pearson (Eds.), Handbook of reading research, Vol. 2 (pp. 984–1012). New York: Longman. Perl, S. (1979). The composing processes of unskilled college writers. Research in the Teaching of English, 13, 317–336. Powell, W. (1970). Reappraising the criteria for interpreting informal reading inventories. In D. L. DeBoer (Ed.), Reading diagnosis and evaluation (pp. 100– 109). Newark, DE: IRA. Powell, W., & Dunkeld, C. (1971). Validity of the IRI reading levels. Elementary English, 48, 637_642. Smith, F. (1994). Understanding reading (5th ed.). Hillsdale, NJ: Lawrence Earlbaum. Spivey, N. (1997). The constructivist metaphor: Reading, writing, and the making of meaning. New York: Academic. Steward, M. & Grobe, C. (1979). Syntactic maturity, mechanics of writing, and teachers’ quality ratings. Research in the Teaching of English, 13, 207–215. Williams, J. (1981). The phenomenology of error. College Communication and Composition, 32, 152–168. 372 M. A. Barksdale-Ladd and J. R. King Appendix Scenario Read by Participants Prior to Interviews One of the underlying assumptions of whole language is that children’s experiments with learning to read and write result in the creation of inventions that do not correspond to standard conventions. Because of their usefulness as data about children’s learning processes and value as learning occasions for the students, there has been considerable discussion about what to do with these nonstandard performances. When and how, for example, is it permissible for a teacher to “correct” children’s work? In fact, to reverse the question, when and how is it permissible for teachers to not “correct” children’s work? Is it different for reading than for writing? A related trend in elementary curricula is the integrated day in which reading and writing both occur in other subject areas. This means that students may be writing a poem about the First Thanksgiving, or acting out an invented script on photosynthesis. But what happens when students construct knowledge that teachers “know” to be inaccurate? For example, what if fifth-graders maintain that the mass of a tree comes from nutrients absorbed from the soil (rather than carbon fixing from photosynthesis)? To combine the previous two scenarios, tell about the thinking processes in which you engage to sort out “process miscues” in writing such as (a) anticipated invention with language, and (b) knowledge-based misinformation? Are there differences in the ways in which you think about and act upon these two kinds of writing process miscues? Are there similarities in the ways in which you deal with these writing process miscues? Are there differences in the ways in which you think about and act upon these two kinds of miscues in reading? Are there similarities in the ways in which you deal with these reading miscues? How do you understand this stuff when you teach; that is, what does it tell you about your students and your teaching?” How important is it, and what can you do with it as a teacher? Please tell about examples to illustrate your points when you can.