1
Evaluating Literacy Instruction
An Introduction
Rachael E. Gabriel and Richard L. Allington
The purpose of this book is to place a spotlight on the intersection of research on effective literacy instruction and teacher evaluation. Since 2009, 46 states and the District of Columbia have changed or revised policies related to the evaluation of teachers and school leaders. In order for new teacher evaluation policies to be used to support and develop effective literacy instruction, teachers and school leaders will need resources that connect the best of what we know about teaching literacy with current evaluation policies and support practices. As teacher evaluation policies shift and change, this book will serve as a resource for selecting and improving evaluation schemes that bring them into alignment with the best of what we know about literacy teaching.
Any teacher evaluation system requires an explanation of what constitutes effective teaching. Historic and contemporary studies of teacher effectiveness have often centered on reading instruction in particular for several important reasons. As Lipson and Wixson (2012) point out, âsome view reading and writing as the basis for most other learning. Others care about it because they view it as a general indicator of the health of public schools. Finally, some care about it because the methods of instruction may be emblematic of philosophical or personal orientations toward larger issuesâ (p. 11). Taken together, these points may explain why reading assessments and achievement are so often the focus of public policies, and why studies of teacher effectiveness so often hinge on the effective teaching of reading.
Research on effective reading instruction is central to the work of defining effectiveness in teaching because of its emblematic status as an indicator of overall educational quality. It is also practically central because of the increased role of student outcomes on measures of reading achievement (and other measures that depend on literacy skills) in teacher evaluations in the Race To The Top era. As we will describe later in this introduction, during the years since Race To The Top, teacher evaluation schemes have hinged on identifying and supporting the effective teaching of reading more than ever before.
In this introduction we begin by providing a brief outlined summary of what we view as the research base on effective reading instruction. We then discuss how reading teacher effectiveness has been researched and defined in the last century, and end by discussing its definition in our contemporary policy context, and discussing the role this text might play in linking the research base on effective reading instruction with teacher evaluation policy.
What We Know about Teaching Reading
Classroom-level influences on reading development in U.S. public schools have consistently been a topic of study since the early 1900s (e.g., Fulton, 1914; Goodykuntz et al.,1925; Betts, 1949). Though conceptualizations and methodologies have evolved over time, the central notion that teachers themselves matter has remained constant. Likewise, several aspects of reading instruction have consistently been noted as key elements of effectively developing children as readers. We will describe these but also note that none of these characteristics seem common in classroom reading lessons offered in American classrooms today. The five factors that have evidence of their influence on reading development are outlined below.
Access to Texts That Can Be Read with Accuracy
We have good evidence that enhancing childrenâs access to a wide range of texts is related to improved reading achievement (Lindsay, 2013). Classroom libraries are often the vehicle for ensuring easy access to a range of texts, but many classrooms have inadequate classroom libraries and fail to make access easy. Our view is that all classrooms, K-12, should have classroom libraries with 500 to 1,000 titles that reflect a range of genres and topics and reflect levels of text difficulty that ensure that books that can be read accurately by every student in the class.
Besides access to physical texts, opportunities to engage with text are also related to improved reading achievement (e.g., Allington et al., 2010). Even in classrooms with a large number of available texts, and literacy blocks with 60â90 minutes of allocated time, the actual time students spend engaged in reading or writing is often minimal (Hiebert, 2002; Hiebert & Mesmer, 2013). Of the time students spend with text in front of them, much is spent reading a whole-class text at or near grade level regardless of the individual levels of students, thus this text experience is unlikely to approach optimal levels of opportunity to develop literacy (Hiebert & Martin, 2009).
Since Anderson, Wilson, and Fielding (1988) noted the strength of the relationship between the volume of voluntary out-of-school reading and student reading development, we have had evidence (e.g., Langer, 2001) that reading practice is important in reading development, even in middle and high school settings. Of course, reading volume is not only a matter of allocating time for reading, but requires both ease of access to books and an orientation toward student-selection of text.
Choice of Text
We know that when children are allowed to self-select books both engagement in reading and reading achievement are powerfully affected (Guthrie & Humenick, 2004). To us this suggests that it is time to end the one-text-fits-all model of reading instruction. That is where the lesson is offered from a single text selected by the teacher (or by corporate executives many miles away). This would not necessarily end small group reading lessons but would largely eliminate whole class reading lessons.
Literate Conversations
In todayâs classrooms, if there is a focus on comprehension, that focus is on lower-order literal recall as teachers ask questions that generally require a single-word answer or sometimes a phrase or sentence segment as the response. However, we have good evidence that when classroom reading lessons focus on higher-order understanding and engage students in literate conversations students benefit enormously (Ivey & Johnston, 2013; Nystrand, 2006; Taylor, Pearson, Peterson, & Rodriguez, 2003).
Focused, Explicit Instruction
Explicit decoding instruction is best accomplished in kindergarten and first grade classrooms (National Reading Panel, 2000) with roughly 10 minutes per day allocated for providing explicit decoding instruction. The National Reading Panel also noted that there is little evidence that decoding lessons after first grade are effective even with those students who seem not to acquire decoding proficiencies in K-1 classrooms. Though researchers have long disagreed about the place and relative importance of basic skills instruction, research over the last decade has consistently highlighted the need to address bothâbasic skills, in the context of meaning-focused instructionârather than one or the other. Indeed embedded decoding lessons seem to work at least as well or better than the typical commercial decoding material lessons (Mathes et al., 2005). Though it is easier to plan, measure, and execute basic skills instruction than meaning-focused instruction, teachers cannot allow one to supplant the other if meaning-making is to be maintained as the goal of literacy instruction.
The large-scale, exemplary teacher studies (Pressley et al., 2001; Allington & Johnston, 2002) demonstrated that a balanced approach to literacy instruction that included some skill instruction with meaning-focused instruction was not only possible across settings and policy environments, it was far more powerful than instruction that privileged one or the other approach to literacy development.
Writing to Real Audiences
Students who struggle with reading skills and engagement may find their way into literacy through writing because it gives them a medium with which to accomplish something that matters to them. Like explicit comprehension instruction, writing instruction is all too often supplanted by elements of literacy that are more easily assessed (e.g., fluency, decoding), or crowded out by worksheets and brief writing assignments that are so structured and contrived they are no better than a fill-in-the-blank task. When students have control over what they are writing, to whom they are writing, and what form that writing takes they are purposefully engaged with texts of various types and purposes, and with encoding words, syntax, and imagery they will then be able to decode as readers (Graham & Hebert, 2011; Hebert, Simpson, & Graham, 2013). Both for the reciprocal skill development and potential for engagement with text, the place of real writing (where students come up with all the words) to a real audience (to whom the piece is actually delivered) cannot be overestimated.
Teachers Matter
Orchestrating opportunities to develop literacy like the five outlined above requires knowledge and commitment to literacy development that are not always supported by school and district policy environments. Still, despite the role of curricula, programs, policies, and changing student demographics, research over the past 40 years has consistently highlighted the importance of teachers in facilitating literacy development.
In the 1970s, teacher evaluation research had a process-product orientation that favored observing teacher âbehaviors,â measuring student learning on discrete and/or standardized measures, and attempting to correlate certain behaviors with certain student outcomes. The most influential study of this era was the California Beginning Teacher Evaluation Study (BTES), a multiphase project that investigated the behaviors of second and fifth grade teachers over five years. The largest contribution of the BTES studies was the conceptualization of âAcademic Learning Timeâ (ALT; Berliner, 1981), which was described as the amount of time students are engaged in tasks of the appropriate level of difficulty. ALT was strongly and positively correlated with academic achievement in both reading and math, though the strength of the correlation varies across subjects, grades, and studentsâ beginning levels of performance (Fisher & Berliner, 1985; Fisher et al., 1980). Though ALT is possible when students work independently, it is more likely to occur when teachers are actively directing activities. These findings suggested that (1) time on task is an important factor in predicting student achievement, and (2) that teacher-directed and/or teacher-supervised activities were more likely to result in engaged and appropriate time on task (ALT) than seatwork or teacher-centered instruction that is not interactive.
The notion of ALT, and the strength of its correlation to higher achievement, supported a widespread focus on ensuring âtime on task,â essentially viewing allocated time as evidence of opportunity to learn. In some states the idea that allocating time for reading instruction would increase the likelihood of higher achievement led to state laws requiring specific amounts of time for reading instruction like, for example, the 90-minute reading block.
ALT also prompted an emphasis on interactive instruction, with later phases of the BTES study noting that substantial teacher-student interaction was likely to be associated with increased engagement during interactions as well as in the group work or independent work that followed.
This process-product orientation to teacher effectiveness research influenced teacher preparation and evaluation by constructing a version of effectiveness that was centered on the orchestration of maximum ALT. This included observable factors like room arrangement, planning, and pacing, as well as elements of classroom management such as effective rules and procedures with a focus on efficiency to maximize time on task. Though states and districts often established allocations of time specifically for reading, and focused on the logistics of maximizing time on task, the orchestration of engaging and appropriate tasks requires more than logistics. This fits a general interest in efficiency in public education, supported by the increasing involvement of the business community in state and federal education policy (Cross, 2004).
Throughout the 1990s there was a shift in focus from process-product research to an interest in variables that are not behavioral or easily observable: specifically, teacher knowledge. The National Commission on Teaching and Americaâs Future (NCTAF) explicitly linked the notion of teacher quality to proxies for teacher knowledge such as test scores earned and degrees or certifications held. The first of three core premises boasts: âWhat teachers know and can do is the most important influence on what students learnâ (NCTAF, 1997, p. 8). Though researchers have struggled to find stable and meaningful connections between observable practices and proxies for teacher knowledge like degrees and test scores, the central message of the BTES was that teaching required more than just attention to efficient classroom management. Teachers need specialized knowledge and skills to generate, orchestrate, and monitor engaging and appropriate activities. According to NCTAF, this requires deep knowledge of subject matter as well as a robust understanding of how students learn. Therefore it was measured by indicators of knowledge acquisition such as degrees and scored portfolios that demonstrated proficiency on standards for what teachers should know and be able to do.
Just as the process-product approach emphasized efficiency of observable behavior for a teaching force that mostly held two- or four-year degrees in teaching, the next phase in research on teacher effectiveness emphasized teacher knowledge and was associated with a professionalization of teaching via standards and policies aimed to raise the bar for initial and continuing certification.
One high-profile example of research from this period was initiated as a study of class size, but generated findings that suggested teacher effects account for a large portion of student achievement, and that teachersâ qualifications matter. In the Tennessee Student Teacher Achievement Ratio (STAR) Project, researchers compared the achievement of students in classes with varying student-adult ratios over four years (1990â1994). Nearly 12,000 students were randomly assigned to a small class (13â17 students), regular class (22â26 students), or a regular class with an instructional aide in addition to a teacher. In general, smaller classes had higher achievement, especially in high-poverty schools (Achilles, 1999). The addition of an instructional aide did not improve achievement, and some larger classes outscored small classes, thus indicating that the individual teacher, along with class size, was a factor influencing student achievement (Boyd-Zaharias & Pate-Bain, 1998). The differences in student achievement by teacher were significant and lasted over several years, especially if students had higher performing teachers two or more years in a row (Finn, Gerber, Achillles, & Boyd-Zaharias, 2001; Nye, Konstantopoulos, & Hedges, 2004; Nye, Hedges, & Konstantopoulos, 1999). Moreover, minority and lower-performing students seemed to benefit most from a year or more with a high-performing teacher.
The variation in effect between teachers and instructional aides bolstered arguments that knowledge specific to teaching, and advanced degrees in education, were necessary to have an effect on student achievement. However, the variation within teachers, the differences in associated student achievement for some over others, also indicated that more than just a teaching degree was at work in the classrooms of highly effective tea...