PERFORMANCE AND PORTFOLIO ASSESSMENT FOR LANGUAGE MINORITY STUDENTS

INTRODUCTION

Well-designed assessment procedures are essential to meeting the needs of language minority students acquiring English as their second language. Assessment is involved at many steps in a continuum of services for these students: in initial identification, in the placement of students into appropriate instructional programs, in monitoring the progress students make within these programs, in reassigning students to different levels within a program depending on their growth in English language skills, in moving students out of special programs and into mainstream classes, and in following the progress of these students in the mainstream. This continuum is wholly dependent at each step on the appropriate selection, use, and interpretation of relatively complex assessment procedures.

BACKGROUND

Recently, there has been a growing interest among mainstream educators in performance assessment due to concerns that multiple-choice tests, usually the only option available from test publishers, fail to assess higher order skills and other skills essential for functioning in school or work settings (Haney & Madaus, 1989; Neill & Medina, 1989; O'Neil, 1992: Wiggins, 1989). Multiple-choice tests are not authentic because they do not represent activities students typically perform in classrooms. In addition, multiple-choice tests do not reflect current theories of learning and cognition and are not based on abilities students actually need for future success (Herman, 1992). Another concern is that standardized tests cannot be used to closely monitor student progress in the school curriculum throughout the year since they are only administered once or twice annually. These concerns are no less valid for educators of language minority students.

PURPOSE AND DEFINITIONS

This publication describes performance assessment procedures and a portfolio assessment framework for monitoring the language development of language minority students in the upper elementary and middle grades. Performance and portfolio assessment may also be used to meet other purposes, such as reassignment or reclassification, as determined by teachers or other school staff. Although assessment of student progress in the content areas merits description and critical analysis, examples of performance assessment procedures provided here are limited to the monitoring of English and/or native language development. We begin by defining basic terms to be used throughout this publication: alternative assessment, performance assessment, and portfolio assessment (Baron, 1992a; Stiggins, 1987; Tierney, Carter, & Desai, 1991).

Alternative Assessment:

• is any method of finding out what a student knows or can do that is intended to show growth and inform instruction and is not a standardized or traditional test;
• is by definition criterion-referenced;
• is authentic because it is based on activities that represent actual progress toward instructional goals and reflect tasks typical of classrooms and real-life settings;
• requires integration of language skills; and
• may include teacher observation, performance assessment, and student self-assessment.

Performance assessment:

• is a type of alternative assessment;
• is an exercise in which a student demonstrates specific skills and competencies in relation to a continuum of agreed upon standards of proficiency or excellence; and
• reflects student performance on instructional tasks and relies on professional rater judgment in its design and interpretation.

Portfolio assessment:

• is the use of records of a student's work over time and in a variety of modes to show the depth, breadth, and development of the student's abilities;
• is the purposeful and systematic collection of student work that reflects accomplishment relative to specific instructional goals or objectives;
• can be used as an approach for combining the information from both alternative and standardized assessments; and
• has as key elements student reflection and self-monitoring.

Performance assessment and portfolios are complementary approaches for reviewing student language development and academic progress. Together they represent authentic assessment, continuous assessment of student progress, possibilities for integrating assessment with instruction, assessment of learning processes and higher-order thinking skills, and a collaborative approach to assessment that enables teachers and students to interact in the teaching/learning process. One of the advantages of using performance and portfolio assessment with language minority students is that the assessment can be conducted in the students' native language (O'Malley & Pierce, 1991).

Portfolios have been most widely used in the teaching of reading and writing, with a strong focus on classroom instruction, student ownership and self-evaluation, and teacher autonomy (Graves, 1983; Tierney, Carter, & Desai, 1991). More recently, portfolios have been proposed and adopted as statewide student assessment management tools in addition to or instead of standardized achievement test data (Baron, 1992b; Brewer, 1990; O'Neil, 1992; Rothman, 1991; Vermont State Department of Education, 1990). Provided that the purpose of portfolio contents is to expand understanding of a student's growth based on multiple measures, different kinds of test and non-test data can be included in a portfolio. Portfolios might also contain required information for state- or district-wide systems, but these data need not dominate or divert portfolio assessment from being used to inform classroom instruction.

Performance assessment and portfolios are typically seen as sources of teacher and student empowerment because control over assessment shifts from the administrators to those linked most closely to instruction (Graves, 1983; Tierney, Carter, & Desai, 1991). While we believe this shift of control over assessment is a positive one, we also believe that the most useful kind of assessment is that which can be shared with other teachers who interact or will interact with students. For this reason, we emphasize systematic approaches to assessment which can ensure the reliability and validity of the results.

PERFORMANCE ASSESSMENT

Performance assessment falls into two categories: achievement-related behaviors exhibited by the student (e.g., communication or reading skills) and achievement-related products that students develop (e.g., written reports or projects). Performance assessments require that the assessor: (1) observe the behavior or examine the product that is reflective of achievement, and (2) apply clearly articulated performance criteria so as to make a sound professional judgment regarding the level of proficiency demonstrated. Intuitions, impressions, and "feelings" about student performance are not a part of sound performance assessments. Like paper-and-pencil tests, performance assessments must adhere to certain rules of evidence (Stiggins, 1990). In this section, procedures are provided for the design of performance assessment tasks and instruments that can be included in portfolios of language minority students and children learning English as a non-native language. Procedures for assessing specific language skill areas (oral language, reading, and writing) and their integration are described. Suggestions are also provided for designing student self-assessment measures of language proficiency and language learning strategies. Each category below includes brief descriptions of types of assessment procedures and ways to design and administer performance tasks, with particular attention to the development of appropriate scoring procedures. Each assessment technique should be adapted as needed to match the developmental and language proficiency levels of the students. This can include assessment in the students' native language.

ORAL LANGUAGE ASSESSMENT

Purpose/Types

To determine oral language comprehension and production, teachers can administer performance assessments which reflect tasks typical of the classroom or real-life settings. In this way, assessment is authentic and aligned with both the curriculum and students' prior experience. Oral performance assessments are not limited to a single type and can take various forms depending on their authenticity in relation to classroom activities. These can include: oral interviews, story retelling, simulations/situations, directed dialogues, incomplete story/topic prompts which students are asked to complete, picture cues, teacher observation checklists, and student self-evaluations (Bachman & Palmer, 1989; Gonzalez Pino, 1988; Omaggio, 1986; Oscarson, 1989).

Design/Administration

A major obstacle to conducting oral language assessment in the classroom is the time involved in administering the assessment to students one at a time. Alternatives to individual student assessment in large classes include the use of teacher observation checklists and the assessment of pairs or small groups of students at one time (Fox & Allen, 1983; Genishi, 1985; Gonzalez Pino, 1988; Pinnell, 1985).

To conduct assessment in pairs, teachers can use cue cards and ask students to interview each other or to interact following the directions on the cue cards (Gonzalez Pino, 1988). Cue cards should be written at the reading level of the students to be assessed. In order to have enough oral language production to assign a rating, at least five or six sentences should be elicited from each student. Cue cards can easily be constructed by providing written directions, called prompts, on index cards to elicit the performance teachers wish to assess. For example, if a lesson has just been conducted on personal greetings and leave-takings, one student's cue card might read, "Greet your classmate and find out where he or she has been for the past few days." The other cue card could read, "Respond to your classmate's questions by telling him or her that you have been at home sick. Find out from your classmate what you have missed in class."

Depending upon the age and language proficiency levels of students in the class, teachers can use picture cues or topic monologues to conduct individual oral assessments. To use picture cues, the teacher presents pictures or photographs appropriate for the age and interest level of the students being assessed. >From among several pictures the teacher presents, students can be asked to choose one or two pictures that they feel they can talk about. The teacher can lead the student into talking by asking questions such as "What story does this picture tell? Has this ever happened to you? What do you think will happen next?"

In using topic monologues, the teacher provides a topic for the student to address at length. This is an authentic performance task for upper elementary and middle school students who are often asked to present oral reports and syntheses of what they have read. If a lesson has just been presented on foods to students at an intermediate level of English proficiency, for example, a topic monologue such as the following might be appropriate: "Tell me about your favorite food. Describe it and tell me why it is your favorite." For more advanced students, the topic might be: "Give me the recipe for your favorite food."

Story retelling is an effective way to integrate oral and written language skills. Students who have just read or listened to a story might be asked to retell the story or to summarize the main idea. When pictures, topic monologues, cue cards, or story retelling techniques are used, instead of the typical interview pattern of teacher-question student-response /teacher-question, teacher talk is reduced and more time is allowed for student language production.

Teachers may also want to assess students' ability to use academic language. Academic language is the cognitively demanding and contextually-reduced language of content area instruction and is critical for success in mainstream classrooms (Cummins, 1982; 1983). Academic language functions are the essential communication tasks that students must be able to perform in different content areas; they determine whether the learning task will be simple or complex (Dix, 1992; O'Malley & Chamot, 1990; Pinnell, 1985). For that purpose, teachers can identify specific language functions that are germane to an instructional goal or activity (O'Malley, 1991), such as:

• Seeking information Using language to explore the environment or acquire information (e.g.,using who, what, when, where, and how to collect information);
• Informing Reporting, explaining, or describing information or procedures (e.g., retelling a story, telling the main idea of a reading passage, summarizing);
• Analyzing Separating a whole into its parts (e.g., telling parts or features of a concept, a procedure, an object); and
• Evaluating Assessing the worth of an object, opinion, or decision (e.g., selecting or naming criteria to evaluate, prioritizing a list and explaining criteria, indicating reasons for agreeing or disagreeing).

Scoring

Teachers and/or raters should establish scoring criteria for a range of grade levels (e.g., Grades 4-6, 7-9) and identify at least three categories of proficiency: non-English proficient (NEP), limited English proficient (LEP), and fluent English proficient (FEP). Because these categories were originally intended for identification and placement purposes and consequently tend to be broadly defined, for the purposes of monitoring student language development it may be useful to differentiate bands of proficiency within each level, such as low, intermediate, and high. This will enable the teacher to design instruction more appropriate to students' needs and to monitor growth from one band to the next within levels as well as from level to level.

Scoring criteria should be holistic, with a focus on the student's ability to receive and convey meaning. Holistic scoring procedures evaluate performance as a whole rather than by its separate linguistic or grammatical features. Depending on the goals of instruction, grammatical accuracy and pronunciation may not be as important as skills critical to comprehending and conveying meaning. Students are evaluated by using detailed criteria or definitions of performance matched to a rating scale (Hamayan, et al., 1985; Navarrete, et al., 1990).

Teachers wanting to score grammar and pronunciation may wish to treat these as subscores and assign them weights which are less than the weight assigned to a subscore for overall communicative ability (Hamayan, 1985; Gonzalez Pino, 1988). Well in advance of the oral assessment, students should be provided with an explanation of how they will be rated. If teachers plan to assign a grade to the oral assessment, they will have to determine the scoring range which corresponds to specific letter grades at each grade level.

A holistic oral proficiency rating scale that teachers may want to use is the Student Oral Proficiency Rating (SOPR). The SOPR is a matrix (adapted from the Student Oral Language Observation Matrix/SOLOM developed by the San Jose Unified School District, California) which allows for rating students in five categories of oral language proficiency: comprehension, fluency, vocabulary, pronunciation, and grammar (see Figure 1). Within each category, students can be rated at five different proficiency levels. The ratings for each category are considered separate sub-scores, each ranging from 1 to 5, with 5 indicating the approximate level of proficiency for a native speaker. A total score results when sub-scores for the five categories are combined (Development Associates, 1988).

Summary

Performance assessments are often used because they provide an opportunity to measure student performance across a range of skills and knowledge over time. Many performance measures are developed directly out of local curriculum objectives and student performance standards and, therefore, may be developed by teachers or other local staff. The performance measures suggested above can be used individually, in combination, or adapted to meet developmental and language proficiency needs of students. However, the use of multiple measures can lead to problems in record keeping and interpretation if teachers and other school staff lack ways of combining the information generated by them. In the following section, a framework for addressing these issues is presented.