Formative Assessment & Standards-Based Grading. Robert J. Marzano
quizzes, class participation, and punctuality of assignments in his or her grading policy while another teacher teaching exactly the same course might include major exams, reports, effort, and attendance. Consequently, the grading schemes for the same course taught by two different teachers might be so different that grades are not comparable from teacher to teacher. In effect, individual teachers’ grades are interpretable only in the context of the grading scheme constructed by that specific teacher.
Norm-Referenced Grading
One approach to grading that has been used over the years is to report how students are performing in relation to one another. This might be called norm-referenced grading. In his doctoral dissertation, Kenneth Haponstall (2009) pointed out that this might have been the impetus for what was referred to as the “grading system” in the mid-nineteenth century, whereby students were grouped by level of knowledge and skill as well as by age so that teachers might provide more focused instruction to these homogeneous groups. He explained that James Baldwin (1884, in Haponstall, 2009) saw problems with the system even then, pointing out that no standard criteria about how students were “graded” or by whom had been established. The decision was subjective and left to anyone from the superintendent to the school secretary or a member of the board of education.
While few, if any, grading schemes currently in place use a strict norm-referenced approach, vestiges of it can be found in the practices of class rankings and grading on a curve.
Class Rankings
Class rankings are related to the concept of norm-referenced grading. Haponstall pointed out that “with districts using differing measures, including grade weighting for advanced placement classes, grade improvement for special education classes, [and] credit recovery for failed courses, there seems to be no standard method for schools to demonstrate those students who are showing academic excellence” (p. 22).
Lawrence Cross and Robert Frary (1999) noted that even though “grading is a hodgepodge of attitude, effort, and achievement at the middle and high school levels, colleges not only accept the grade point and class ranking in determining enrollment, but many are starting to use these measures exclusively” (as cited in Haponstall, 2009, p. 22). The admissions policies of many colleges exacerbate the practice of class ranking. To illustrate, David Lang (2007) pointed out that states such as California, Florida, and Texas guarantee a certain top percentage of each graduating class admission to a state school. This renders class ranks a high-stakes endeavor, particularly for those ranked too low for a guarantee of admission to a state school (as cited in Haponstall, 2009).
Grading on the Curve
When a teacher grades on the curve, he or she gives the highest grade to the student who performed best on an assessment and then gives every other student a grade by ranking his or her performance accordingly. This system essentially grades students in relation to one another. Thus, it has a basis in norm-referencing. Proponents for grading on the curve maintain that it is fair and equitable because most classes will have a normal distribution of achievement scores in any given subject area (for a discussion, see Brookhart, 2004).
Thomas Guskey (2009), however, maintained that “grading ‘on the curve’ communicates nothing about what students have learned or are able to do” (p. 11). Instead of telling teachers what a student has learned, it simply reports how much or how little he or she learned in relation to his or her fellow students. He also pointed to research by Benjamin Bloom (Bloom, 1976; Bloom, Madaus, & Hastings, 1981) indicating that student achievement does not necessarily follow a normal distribution when teachers exhibit a high level of instructional acumen. Grading students only in relation to one another, therefore, may provide information about a student’s rank in class, but it does not speak to the student’s academic achievement.
Self-Referenced Grading
Self-referenced grading occurs in relation to one’s own past performances. Proponents say that it may reduce competition in classrooms and serve to motivate students (for a discussion, see Brookhart & Nitko, 2007). On first glance, this kind of grading seems to make intuitive sense: the reference point for each student is his or her personal growth and the extent of active engagement in his or her own learning. But Brookhart and Nitko pointed out that this form of grading tends to be used primarily with low-ability students, and while heavily weighting factors such as effort, behavior, attitude, and participation might seem positive, this emphasis is one of the major criticisms of this form of grading. Mixing nonacademic competencies with academic competencies contaminates the meaning of a grade.
Standards-Based Grading
Grading that references student achievement to specific topics within each subject area is growing in popularity. This is called standards-based grading, and many consider this method to be the most appropriate method of grading (for a discussion, see Brookhart & Nitko, 2007, p. 219). Where there is interest in this system, however, there is also quite a bit of poor practice on top of considerable confusion about its defining characteristics.
As described in Marzano (2006), the origins of standards-based reporting can be traced to the concept of a performance standard. The term was popularized in a 1993 report commonly referred to as the Malcom Report in deference to Shirley M. Malcom, chair of the planning group. The report defined a “performance standard” as “how good is good enough” (National Education Goals Panel, 1993, pp. ii–iii). Since then, a popular practice has been to define student performance in terms of four categories: advanced, proficient, basic, and below basic. The scheme has its roots in the work of the National Assessment of Educational Progress. As Popham (2003) noted:
Increasingly, U.S. educators are building performance standards along the lines of the descriptive categories used in the National Assessment of Educational Progress (NAEP), a test administered periodically under the auspices of the federal government. NAEP results permit students’ performances in participating states to be compared…. Since 1990, NAEP results have been described in four performance categories: advanced, proficient, basic, and below basic. Most of the 50 states now use those four categories or labels quite similar to them. (p. 39)
The actual practice of standards-based reporting requires the identification of what we have referred to as reporting topics or measurement topics (Marzano, 2006; Marzano & Haystead, 2008). For example, consider the following common measurement topics for language arts at the fourth grade:
Reading
Word recognition and vocabulary
Reading comprehension
Literary analysis
Writing
Spelling
Language mechanics and conventions
Research and technology
Evaluation and revision
Listening and Speaking
Listening comprehension
Analysis and evaluation of oral media
Speaking applications
Here, ten measurement topics are organized under three categories (or strands, as some districts call them): reading, writing, and listening and speaking. For reporting purposes, each student would receive a score of advanced, proficient, basic, or below basic on each of the ten measurement topics. Typically, some type of rubric or scale that describes these levels is constructed for each measurement topic (we discuss this in depth in chapters 3, 5, and 6).
While this system seems like good practice, without giving teachers guidance and support on how to collect and interpret the assessment data with which scores like advanced, proficient, basic, and below basic are assigned, standards-based reporting can be highly inaccurate. Indeed, at the writing of this book, no major study (that we are aware of) has demonstrated that simply grading in a standards-based manner enhances student achievement. However, as the previous discussion illustrates, a fairly strong case can be made that student achievement will be positively