The comparison that is used to assign grades to pupils can influence the effort and attitude of the pupils. Norm-referenced standards, for example, tend to undermine the learning and effort of pupils who continually score near the bottom of the class, since they continually receive poor grades. Norm-referenced grading poses much less threat to the top pupils in the class (Crooks, 1988; Deutsch, 1979; Ebel and Frisbie, 1986). Competitive grading approaches such as norm-referenced grading, which make a pupil's
GRADING PUPIL PERFORMANCE 321
success or failure dependent largely on the performance of classmates, can also reduce cooperation and interdependence in study, because success for one pupil reduces the chance of success for another pupil (Crooks, 1988).
Comparison to Predefined Standards of Performance
Instead of grading by comparing a pupil's performance to that of other pupils, the teacher can compare the pupil's performance to pre-established performance standards. Performance standards define the level of mastery or performance a pupil must attain to receive a particular grade. A simple example of the use of performance standards is the driver's license test. In many states, the test for a driver's license contains two parts, a written section covering knowledge of the rules of the road and a performance section in which the applicant must drive an automobile around local roads. (Notice how a paper-and-pencil test and a performance assessment are combined in the driver's test to make certain all important behaviors related to safe driving are assessed.) The written portion of the driver's test must be passed before the performance portion is attempted, and the written test usually has ten or twenty multiple choice items that applicants answer. The test is administered to groups of applicants in much the same way paper-and-pencil tests are administered in schools. In order to pass the test in most states, an applicant must get 60 percent of the test items correct. In this case, 60 percent is the performance standard. Whether any one applicant passes or fails has nothing whatsoever to do with how the other applicants did on the test. Instead, each applicant is compared to the predefined standard of adequacy, a score of 60 percent, that defines what a passing performance is.
Grading systems that compare a pupil's performance to a predefined performance standard are called criterion-referenced, or absolute, grading systems. They are "absolute" in the sense that each pupil's performance is compared to the same performance standard in order to determine a grade. Each pupil is graded on the basis of his or her own work, independent of the work of other pupils. Since pupils are not compared to one another, it is possible for all applicants to get high grades on the test or, conversely, for no pupils to get high grades if all do poorly in comparison to the performance standard. Criterion-referenced grading is the most commonly used grading system (Ebel and Frisbie, 1986; Hills, 1981; Nitko, 1983).
There are two types of performance standards that are used in criterion-referenced grading. One type spells out in detail the specific behaviors the pupil must perform in order to receive a particular grade. This type of grading standard is used primarily for the kind of performance assessments discussed in Chapter Seven. For example, the following performance standard could be set up to grade each pupil who must give an oral speech. The teacher would observe the speech, concentrating on the specific behaviors listed in the performance standards. On the basis of the observation, the teacher would assign a grade to each pupil. Again, notice that each pupil's grade depends on how he or she performs in comparison to the standard, not how he or she performs in comparison to other pupils.
A Pupil consistently faces audience, stands straight and maintains eye contact; projects voice well and clearly; pacing and tone variation appropriate; well-organized points logically and completely presented; brief summary at end.
B Pupil usually faces audience, stands straight and makes eye
contact; voice projection good, but pace and clarity vary during talk; well-organized but repetitive; occasional poor choice of words and incomplete summary.
C Pupil fidgety; some eye contact and facial expression
change; uneven voice projection, not heard by all in room, some words slurred; loosely organized, repetitive, contains many incomplete thoughts; poor summary.
D Pupil body movements distracting, little eye contact or voice
change; words slurred, speaks in monotone, does not project voice beyond first few rows, no consistent or logical pacing; rambling presentation, little organization with no differentiation between major and minor points; no summary.
The second, more common type of standard is used mainly with paper-and-pencil achievement tests. In this standard, cutoff scores based on the percent of items answered correctly are used to award grades to pupils. Perhaps the most widely used standard of this type is one which has the following cutoff percent’s.
A 90 to 100 percent of items correct
B 80 to 89 percent of items correct
C 70 to 79 percent of items correct
D 60 to 69 percent of items correct
F Fewer than 60 percent of items correct
Many teachers have different cutoff scores than these; some teachers use 85 percent and higher as the cutoff for an A and readjust the cutoffs for the remaining grades accordingly. Other teachers
will not flunk pupils unless they get less than half (50 percent) of the items correct. Like the curve in norm-referenced grading, the performance standards that are used in criterion-referenced grading are based on a teacher's judgment about what is suitable and fair for his or her class. They should be reasonable given the ability of the class and the nature of the subject matter, and they should be academically honest and challenge the pupils (Hills, 1981).
A criterion-referenced grading system is intended to indicate how much a pupil learned of the things that were taught. An A grade in a criterion-referenced approach is intended to signify that the pupil mastered all or almost all of the things taught; B grade-conveys mastery of most of the things taught; C grade mastery of the majority of things taught; D grade difficulty with most things taught; and F grade little or no mastery of the things taught.
Because criterion-referenced grades are interpreted by most people in terms of the amount of subject matter a pupil has mastered, the use of criterion-referenced grading systems puts a premium on the construction and use of assessment instruments which provide valid and reliable information about learning in the classroom. Poor instruments that do not assess what pupils were taught, or that assess only selected aspects of what was taught, will convey an incorrect message about pupil learning. Instruments that fully assess what pupils have been taught should always be used, but because criterion-referenced grades are interpreted in terms of content mastery, it is especially important that assessment instruments adequately reflect the full range of behaviors and skills taught. Only then will the interpretation of the grades based on them be appropriate.
Regardless of whether one employs a norm- or a criterion-referenced grading approach, it is recommended that the grading curve or performance standards be defined before assessment is carried out and grades are assigned. That way teachers will have thought about expected performance and can inform pupils of what will be needed to get high grades. It must also be recognized that the curve or grading standards rest on teacher judgments about the pupils, the nature of the subject matter, and a sense of what is fair. Judgments, as we have seen, are based on less than perfect information, and therefore, are sometimes incorrect. Thus, performance standards and curves, which are based on judgments, need not be set in stone once they have been established. If a standard or curve turns out to be inappropriate or unfair for some reason, it can and should be changed before grades are assigned. If, for example, you found that many of the items on a test were not taught or were worded in a confusing manner, you might decide to discount these items when grades were assigned. In this case, the criterion-referenced standard should be changed to take into ac-
324 CLASSROOM ASSESSMENT
count the items that were not taught or poorly worded. While changes in performance standards or the curve should not be made frivolously, it is better to make occasional changes than to award grades that are unfair and incorrect. Usually, after increased experience with a class, a teacher arrives at a set of standards or a grading curve that is appropriate and that remains quite stable for the remainder of the year.
Having made this point, it must also be noted that fairness to pupils does not mean arranging standards or the curve so that all get high grades. Lowering standards to ensure high grades discourages effort and seriousness in pupils' approach to the subject matter (Rosswork, 1977). Fairness means teaching pupils the things on which they are assessed, using assessment procedures that are clear and suited to the pupils' level and classroom experiences, and establishing performance standards or a grading curve that is realistic in terms of what pupils can attain if they work hard.
Comparison to a Pupil's Ability
Teachers frequently remark that "Ralph is not working up to his ability," "Maurice is not doing as well as he can," or "Rose continues to achieve much higher than I expected she would." When teachers make statements such as these, they are comparing a pupil's actual performance to the performance they expect based on their judgment of the pupil's ability. The terms "overachiever" and "underachiever" are used to describe pupils who do better or worse than their ability suggests they should. A few teachers assign grades by comparing a pupil's actual performance to their perceived level of ability.
Under this grading scheme, pupils with high ability who do excellent work would receive high grades, as would pupils with low ability whom the teacher believed were achieving "up to their potential." Even though the actual level of performance of the low-ability pupils was well below that of the high-ability, high-achieving pupils, each group would receive the same grade because each was perceived to be achieving up to their ability. Conversely, pupils with high ability who were perceived by their teacher to be underachieving would receive lower grades than low-ability pupils who were achieving up to expectations. One of the main justifications that is advanced for using this comparison scheme in grading is that it motivates pupils to do their best and to get the most from their ability. Further, it punishes the lazy who do not work up to their perceived ability.
This approach to grading is not recommended for a number of reasons (Gronlund, 1985; Hills, 1981; Kubiszyn and Borich, 1989). First, it is heavily dependent on the teacher having an accurate indication of a pupil's ability, since this information is critical for determining what a pupil could do or is capable of achieving. Rarely do teachers have dependable enough information about pupils' abilities to permit them to predict what a pupil is capable of doing with any validity. Teachers do have a sense of pupils' abilities from their sizing up assessments and the pupils' classroom performance, but this information is hardly precise enough to determine what a pupil could do or is capable of doing for grading purposes. Also, teachers often have a difficult time separating their perception of a pupil's ability from perceptions of other pupil characteristics. Moreover, tests designed to measure ability are not given very often in schools and, like all assessments, contain enough imprecision to make them less than optimum in predicting performance levels for grading purposes.
Furthermore, even if one has valid and reliable information about a pupil's general ability level, how is that information translated into precise achievement expectations in a particular subject area? It is all but impossible to make valid predictions about what a pupil of a certain general ability level is capable of achieving in any specific subject area.
Finally, grades that are determined using this method are difficult to interpret. For example, a high-ability pupil who attained 80 percent mastery of the instruction might receive a C grade if perceived to be underachieving, while a low-ability pupil who attained 60 percent mastery might receive an A grade for exceeding expectations. An outsider viewing these two grades would probably think that the low-ability pupil mastered more in the course, because that pupil got the higher grade. Similarly, two pupils could receive the same grade even though their mastery of the subject matter was quite different.
All of these reasons argue strongly against the use of a grading system that compares actual to predicted achievement. Some report card systems allow the teacher to make separate judgments about a pupil's achievement and how well the pupil is working up to his or her capabilities. The teacher can record a grade for the subject matter based mainly on the pupil's actual achievement, and then, in a separate place on the report card, can indicate if the pupil is working up to expectations. Usually, the teacher indicates whether the pupil "needs improvement," "is improving," or "is doing best" in order to show how the pupil is performing relative to his or her ability. Even using this approach, teachers must be cautious about putting too much faith in the precision of their estimates of pupils' abilities.
Comparison to Pupil Improvement
Arriving at a grade by examining the improvement a pupil has shown over time is a grading system that is similar to comparing actual to predicted achievement. It also has many of the same difficulties (Ebel and Frisbie, 1986; Hills, 1981; Kubiszyn and Borich,
326 CLASSROOM ASSESSMENT
1989). In this system, a pupil's performance early in the term is compared to the pupil's performance later in the term. Pupils who show the most progress or growth get the highest grades, and those who show little progress or growth get the low grades. An obvious difficulty with this approach is that pupils who do well early in the term have little opportunity to improve, and thus have little chance to get good grades. Low scorers at the start of the term have the best chance for improvement and, therefore, high grades. It is not surprising that students quickly realize that it is in their best interest to do poorly on the early tests; there is an incentive to play dumb so early performance will be low and improvement can be shown easily.
Also, like comparing actual to predicted performance, grading on the basis of improvement makes grades difficult to interpret. A pupil who improves from very low achievement to moderate achievement may get an A, while a pupil who had high achievement at the start and therefore improved little may get a B or a C in spite of having mastered considerably more of the subject matter than the one who got the A grade.
Some teachers recognize this difficulty and propose the following solution: "Give the pupils who are high achievers throughout the term an A grade for their high performance, but also give A grades to those pupils who improve their performance a great deal." Certainly this suggestion overcomes the problem noted above. However, it creates a new problem. In essence, what these teachers an proposing to do is to use two different grading systems, one for high-achieving pupils and another for low-achieving pupils who show improvement. This approach provides rewards for both groups of pupils, but confuses the interpretation of the grades Logically, if we saw two pupils in the same class who received A grades, we would think that their level of subject matter master was about the same; after all, they each got an A grade. But this would not be a correct assumption in a classroom in which different grading systems were used for different subgroups. Communication, which is an essential purpose of grading, would be hand capped, because the message the grade conveyed would differ a(cording to the system of grading used for a particular pupil Whatever system of grading a teacher decides to use ought to be the same for all pupils in the class. That is the only way that grades can have a consistent, understandable message. Thus, neither a grading system based on improvement nor a grading system based on different approaches for different groups of pupils is recommended.
To summarize this section: The first judgment a teacher mu make in the grading process is what standard of comparison will I used to award grades. Norm-referenced grading and criterion-referenced grading are the two most commonly used approaches.
Basing grades on comparisons of a pupil's performance to the pupil's ability or the pupil's improvement over time is not recommended. Remember that there is no one best way to assign grades and that whatever system you select will have drawbacks, as illustrated above. Finally, the grading system that is used in a classroom ought to be the same for all pupils so that the meaning of the grades is clear and consistent.
SELECTING PUPIL PERFORMANCES
Once the comparative basis for assigning grades is decided on, it is necessary to select the particular aspects of pupil performance that will be considered in awarding the grades. If one is grading a test or a project, there is obviously only one performance to be considered when assigning a grade. If one is assigning report card grades, there will usually be many pieces of assessment information that have been gathered over the term or marking period that could be accounted for in the grade. Among the information most teachers collect that could be used for report card grading are pupils' homework, class participation, test scores, quizzes, performance assessments, and attendance, as well as the teacher's formal and informal perceptions of effort, interest, motivation, helpfulness, and classroom behavior.
The quantity and the nature of the information available to a teacher varies depending on grade level and subject area. Moreover, the range of evidence teachers collect varies depending on subject matter. For example, assigning a term grade in spelling in the elementary grades essentially involves combining the results of each pupil's performance on the Friday spelling tests, since this is usually the only kind of information most teachers have available about pupils' spelling skills. In social studies, however, a teacher may have information from quizzes, tests, homework, projects, reports, and worksheets. High school math teachers have homework papers, quizzes, and test results to consider in assigning grades, while English teachers have tests, reports, homework, quizzes, projects, and class discussion to consider. All teachers also have perceptions of effort, motivation, and other affective characteristics which they have acquired by less formal assessment means.
Each teacher must decide which of the available indicators will be included when determining report card grades. This decision is critical, because the performances that are included in the grade define what the grade means and stands for. In deciding what factors to include in a grade, the teacher is making a judgment about what the grade represents.
328 CLASSROOM ASSESSMENT
A distinction should be made between the total set of assessments a teacher has collected about pupils and the subset of those assessments that is most appropriate to use in assigning grades to pupils in a subject. Subject area grades are interpreted widely as providing information about the pupil's achievement of specific course objectives that have been taught by the teacher. Such grades are not commonly perceived to indicate effort, participation in class, or interest in school. Since the grade will be interpreted as an indication of achievement, it should be based mainly on assessment information that reflects a pupil's achievement in the subject: class grades should reflect pupils' mastery of the course objectives. Thus, the subset of assessment information that is included in pupils' grades should be composed only of information about pupil achievement of course objectives.
Effort, deportment, interest, motivation, and the like should not be part of a pupil's subject matter grade. Report card grades should not be rewards for having a pleasant personality or a good attitude, nor should they be tools of punishment for poor classroom behavior or failure to show enthusiasm during instruction. As shown in Figures 8.1 and 8.2, most schools provide a separate area on report cards for teachers' anecdotal comments about pupils or for rating pupils' affective and social behaviors. To give an A grade to a pupil who is academically marginal but very industrious and congenial would be misleading to the pupil, the parents, and others who might interpret the grade as indicating high academic performance. Such grading might affect the motivation of the academically superior pupils in the class. Pupils who work hard, are cooperative, and show greater motivation and interest than their classmates ought to be rewarded, but subject matter grades are not the proper arena for such rewards. Finally, constitutional and case law related to pupil rights suggest that grades should not be heavily dependent on deportment, interest, and attendance, and should not be used to punish pupils for behavior problems (Bartlett, 1987).
It would be easy to base grades solely on formal pupil assessments if grades were assigned by machines. But that is not the case. Teachers assign grades and teachers are not machines. Teachers know a great deal more about their pupils than any preprogrammed grading machine ever could; they know them as whole persons, not just as academic entities. Teachers understand the different effects grades will have on particular pupils in their class. They know who they can count on to be cooperative, answer questions in class, and work hard. They know about the home situations and backgrounds of many of their pupils. Because of this, only the very rare teacher can be a completely objective, dispassionate dispenser of report card grades. When assigning grades, teachers
are inevitably confronted by dilemmas that encourage them to take into account more than a pupil's performance on formal assessments.
Jerome works harder than any student in my class, but he cannot seem to overcome his lack of ability. No one tries harder, and still his tests and projects are all failures. But I just can't in good conscience give Jerome a failing grade because he tries so hard and an F would destroy him.
Melissa had a terrible term. Her test scores dropped off, her attention during instruction was poor, and she failed to complete many homework assignments. The reason for these behaviors is in her home situation. Her father left the home, her mother had to find a job, and Melissa had to assume most of the household and babysitting responsibilities because she is the oldest child. How can I not take this into account when I grade her this term?
Joe is the ultimate itch; constant motion, inattention, socializing around the classroom at inappropriate times. He drives me crazy and takes too much time away from the other kids. However, his classwork is well done and on time. When I sit down to grade him, I have to restrain myself from saying, "Okay, Joe, now I'm going to get you for being such a distraction." I have a hard time separating his academic performance from his behavior in the classroom. I know I shouldn't think this way, but you don't know Joe.
One instance in which pupil characteristics such as motivation, interest, effort, or deportment often enter into grading is when such behaviors are used to give borderline pupils the benefit of the doubt. When a teacher awards a B + to a pupil whose academic performance places her between a B and a B + grade, but who is motivated, participates in class, and works diligently, the teacher is taking into account more than just formal performance assessments. Similarly, a teacher will often give a conscientious, participating pupil a B — instead of the C + he deserves strictly on the basis of formal assessments, because the teacher feels the B — grade will motivate the pupil more than the C + grade. Strictly speaking, such adjustments distort the meaning of a grade, but all teachers make such adjustments based on their knowledge of particular pupil characteristics and needs. Grading is a judgmental process, and it is virtually inevitable that such adjustments will occur.
Minor, borderline adjustments such as those described above usually operate to the benefit of the pupil and the psychic comfort
330 CLASSROOM ASSESSMENT
of the teacher. The adjustments are usually so minor that it is hard to see the direct influence of nonacademic factors on the grades. The danger that must be guarded against is allowing effort, motivation, interest, deportment, or personality to become a dominant influence when assigning grades. If this happens, the grade is distorted; it provides little useful information about the pupil's academic achievement. When a pupil's grade is raised from a C to a B because the pupil works hard and seems interested in the subject, then nonacademic factors are being weighed too heavily. The B grade is not the accurate reflection of the pupil's academic performance it ought to be. If a pupil's grade is lowered to a D from the C that formal assessments indicate because the pupil is a behavior problem, not only is the grade not representative of the pupil's academic performance, it may be subject to legal challenge. Although few teachers can ignore nonacademic evidence when they grade, most correctly use this evidence only to make minor, borderline adjustments in pupils' grades.
Formal, academically oriented criteria such as teacher-made and textbook tests, papers, quizzes, homework, projects, worksheets, and the like are the best types of evidence to use in assigning report card grades. They are suitable in two respects. First, they provide information about pupils' academic performance, which is what grades are intended to describe. Second, they are formal, tangible products of pupils' work, which means that they can be used to defend or explain a grade if the need arises. It is defensible to say to a pupil or parent, "I gave a C grade because when I compared your test scores, projects, and homework assignments in this marking period to my grading standards, you performed at a C level." It is not defensible to say, "I gave a C grade because I had a strong sense that you were not working as hard as you could and because I have a negative perception of your daily class performance." Test scores, projects, and homework are tangible products that can be examined and averaged in determining a grade. Such evidence tends to be more objective and defensible than other, informal kinds of evidence such as a strong "sense" that someone is being lazy and a "perception" that class performance wasn't very good.
Since formal assessments of pupil achievement ought to be accorded major weight in assigning grades, it is important to stress that the grades awarded will be only as good as the formal assessment information on which they are based. The success and mean-ingfulness of grades is dependent on the success and meaningful-ness of the assessments which support them (Hills, 1981). Irrelevant, invalid evidence about pupil achievement will produce irrelevant, invalid grades. Grading as a process cannot be separated or isolated from the assessment activities teachers carry out prior to
GRADING PUPIL PERFORMANCE 331
grading. Just as we saw that good instruction could be undermined by invalid assessment, so too can a good grading system be undermined by invalid achievement indicators. Using carelessly constructed tests, gathering no formal assessment information about achievement, and using a single test or project as the sole basis for a report card grade are all practices that diminish the validity and reliability of report card grades (Hills, 1981).
Grading is the culminating step in the process of assessing pupils' achievement. As such, it ought to be based on sound evidence about pupils' achievement of the objectives they were taught. Moreover, a general rule of grading is that as long as the evidence being considered pertains to subject matter mastery, several different types of information are better than a single type because they give the pupil more of an opportunity to show what he or she can do. The fact that multiple types of evidence need to be summarized in the grading process leads to a third judgment a teacher must make during grading—namely, should different types of evidence be weighted differently in determining pupils' grades? We shall explore this question in the next section.
Дата добавления: 2015-09-28; просмотров: 21 | Нарушение авторских прав
| <== предыдущая лекция | | | следующая лекция ==> |