Allaei and Connor suggest that performative assessments of ESL student writing can offer a valuable alternative to large-scale, holistically or analytically scored assessments of L2 writing proficiency. Performative assessments involve examinees in writing in response to a particular task that necessitates a specific type of writing (e.g., persuasion). This writing is then evaluated subjectively according to detailed scoring rubrics. Such tests are therefore criterion-referenced, and they seek to differentiate a variety of skills that are involved in a given writing task performance. Allaei and Connor maintain that these tests have high content validity if the rubrics used for evaluation are based on identification of skills that students actually will be using with respect to given tasks and if the test prompts elicit the same skills. They also claim construct validity for performative writing tests on the basis of this detailed identification of task dimensions along which actual student writing performances vary.
Development of an L2 performative assessment involves: (1) identification of the knowledge and skills that a target audience is expected to command with respect to specific writing tasks, (2) creation of test tasks in which content is supplied and which elicit the desired skills and knowledge, and (3) development of scoring rubrics in terms of the sub-skills required and the constitution of levels of proficiency within each of these sub-skills. (Allaei and Connor give examples of differing scoring rubrics used for several specific writing tasks.)
This kind of performative assessment serves as an effective diagnostic tool, especially in program situations where certain types of writing are emphasized and identifiable. Performative assessments should generally focus on the written representation of content and meaning, but analytic scales can be used where specific points (of grammatical accuracy, for example) need to be assessed. The strength of performative assessments lies in their descriptive depth and potential for feedback to learners and teachers. Their success hinges on development of authentic prompts that "parallel what students are expected to do in academic writing tasks" (p. 239). These should be kept specific in order to facilitate comprehensive scoring rubrics.
Arter, J. A., & Spandel, V. (1992). Using portfolios of student work in instruction and assessment. Educational Measurement: Issues and Practice, 11(1), 36-44.
Portfolios are being implemented in order to: tap student knowledge and capabilities to a greater degree, investigate student learning and production processes, align instructional and testing emphases, examine students functioning in real-life situations, provide continuous developmental feedback, encourage student engagement in and responsibility for learning. Portfolios should be purposeful (with goals and organization -- not just a folder of collected work), involve students in self-reflection (metacognitive analysis of work), employ a clear evaluation criteria schema from the inception, include guidelines for work selection, and encourage student participation in this selection process. Portfolios can be used as assessment devices, as an integrative element of instruction and assessment, and as a story-telling device reflecting student work and growth.
Aschbacher, P. A. (1991). Performance assessment: State activity, interest, and concerns. Applied Measurement in Education, 4(4), 275-288.
Aschbacher reports on the use of performance assessment at the state level across the United States. She cites the inability of traditional standardized assessment to provide the necessary information for making informed decisions about system effectiveness, and she refers to the undesirable washback effects that standardized testing has had on tertiary education. Although there exists no single method of alternative assessment, alternative or performances assessments share several common characteristics: (1) they require higher level thinking and problem-solving; (2) the tasks are also worthwhile instructional activities; (3) they involve real-world contexts or simulations; (4) they assess process as well as product; (5) performances standards and criteria are made public prior to assessment. Aschbacher provides a list of states with involvement or interest in performance assessment, and she then reflects on the major state-level concerns with implementing performance assessments.
General questions that must be answered before implementing portfolios as assessment devices include:
(1) To what extent is the portfolio representative of student work, growth, and ability?
(2) How and by whom are judgment criteria determined?
(3) What constitutes authentic work and performance, and to what extent is the portfolio reflective of the day-to-day authentic activities that should be occurring in the classroom?
(4) Who interprets the information provided by portfolios, and how are these interpreters trained?
In designing a portfolio system, the following questions should be considered:
(1) Who is responsible for the generation and design (stakeholders, top-down, students)?
(2) What is the purpose of the portfolio? Will it be used for assessment as well as instruction? Will it be used to provide information at the classroom level only, or for large-scale accountability decisions as well?
(3) What are the links to curriculum and instruction? Are criteria addressed in instruction? Are students responsible for and active in the learning process? Is self-reflection implemented in class as well as in the portfolios?
(4) What is the content involved in the portfolio? How, when, and by whom is it chosen?
(5) How is assessment incorporated? What is the goal of the assessment? What are the criteria? Is assessment standardized? Who does the assessing?
(6) Who is responsible for the following: selecting portfolio material, storing the portfolio, accessing the portfolio? Who "owns" the portfolio, and how is this ownership operationalized?
(7) How are teachers trained to implement portfolios, both in instruction and as an assessment tool?
In concluding, Arter and Spandel present a checklist for portfolio design.
States are most concerned with the costs of performance assessment. These costs include test development, training of teachers and raters, extensive rating sessions, transportation, public education about changing assessments, and score reporting, among others. Possible solutions might include remote scoring and computer networking to alleviate some of the costs. The logistics of implementing performance assessments also seems to be a big concern. Performance assessments involve large amounts of collected material, special equipment, storage and transportation of tests, and increased time commitment. Technical reliability and validity is also cited as an important issue, especially at the system level where decision-making must be defended to multiple interested parties. Finally, support for implementation is lacking in most states. This is due to the difficulty in comparing traditional inexpensive, easy-to-administer, norm-referenced tests with new alternatives. The public, which demands educational reform, is generally unaware of issues involved in test development. If performance assessment is to be applied beyond the classroom level, then extensive training and public promotion will be necessary. Further changes will also be necessitated, given the continued dominance of standardized forms of assessment with critical gatekeeping functions (e.g., the SAT, GRE, and ACT).
