Establishing the Validity of Test Accommodations and Score Interpretations for Special Education Students: A Collaboration of State-based Research
NCLB requires that states offer accommodations on the grade level assessment so that the test is accessible to as many special education students as possible. However, little research has been conducted on the validity of accommodated score interpretations or the effectiveness of test accommodations. With Connecticut serving as the lead state and with the support of the CCSSO, a large consortium of states from two SCASS groups will participate in this special research project. The purpose of the research is to establish the validity of inferences from accommodated tests based on specific student accessibility needs. Studies will be conducted in up to ten states (AK, AZ, CT, KY, MI, NV, RI, UT) on a variety of commonly-used accommodations so that the results can be used to build a shared body of evidence for the validity of the interpretation of accommodated scores.
This project will be coordinated across states, using a rigorous empirical research design. Every state will use the same 2 x 2 counterbalanced research design that has students (regular education or special education) crossed with accommodations (with or without), on one form of a pilot test to be conducted in their state. Approximately 200 – 500 students will be tested in each state’s sample, depending on the type of accommodation studied. Analyses to be conducted include statistical analysis of items and test scores, content analyses for validity evidence, and factor analysis to examine any structural changes in constructs due to accommodations. This project will result in a guidebook and associated database that provide designs, procedures, statistical data, and other information for evaluating the validity of test results from accommodated assessments. The findings and products will be disseminated to all states for their use in providing evidence on the appropriateness of their accommodations.
Obtaining Necessary Parity through Academic Rigor (ONPAR): Research, Development, and Dissemination of a Parallel Mathematics Assessment for ELLs
The Illinois State Board of Education (ISBE), on behalf of the 15-state World-class Instructional Design and Assessment (WIDA) Consortium, proposes to develop and implement a feasible, accessible, and valid assessment in mathematics for English language learners (ELLs) in the beginning stages of English language acquisition that can be used for state accountability purposes to meet the requirements of federal law. To achieve this aim, this project encompasses the following three broad goals: a) to ensure that the mathematics achievement of ELLs is assessed validly and reliably; b) to increase knowledge about English language acquisition and mathematics achievement; and c) to disseminate information about this standards-based, academic assessment and the results of related research.
The WIDA Consortium, originally established with funding from a U.S. Department of Education Enhanced Assessment Grant, currently includes Illinois and 14 additional states. Combined, the 15 WIDA partner states enroll approximately 420,000 K-12 ELLs. Since 2003, WIDA has created and adopted comprehensive English language proficiency standards (2004, 2007) that represent the second language acquisition process and the language of the content areas of language arts, mathematics, science, and social studies. Based on these standards, WIDA developed a K-12 ELP test battery–ACCESS for ELLs® –that approximately 420,000 students took in spring 2007. WIDA also provides professional development activities and maintains a Web site (www.wida.us). In 2006, Rhode Island, on behalf of WIDA, received an Enhanced Assessment Grant to develop ONPAR-Science, a parallel science assessment for grades 4, 8, and 11. The work begun with ONPAR-Science, along with other WIDA activities, will support this project to develop a Web-based, parallel assessment of mathematics–ONPAR-Math.
This project will result in a defensible, psychometrically sound, Web-based assessment that forms a common core test in mathematics for grades 3, 7, and 11. Initial piloting, field testing, and comparability studies will occur in Illinois and the New England Compact states of New Hampshire, Rhode Island, and Vermont, with the intent that the core test will be applicable to all WIDA states. The knowledge and skills developed with the core ONPAR-Math will be used to expand the assessment to include grades 3 through 8, 10, and 11. After the grant period, each state may choose to augment the core test with state-specific items to capture its full range of mathematics standards for specific grade levels.
ISBE will leverage the already strong working relationship between the Wisconsin Center for Education Research, the administrative and research home of WIDA, and the Center for Applied Linguistics, as well as other nationally recognized leaders in mathematics, assessment, measurement, and the education of ELLs. ONPAR-Math will be built on a strong foundation of research and collaboration.
Improving Methods of Analyzing Alignment of Instruction to Assessments and Standards for English Language Learners and Analyzing the Relationship of Alignment to Student Achievement
Overview: With Iowa serving as the lead state, the project will consist of a consortium of 10 states – Florida, Idaho, Iowa, Maine, Minnesota, Ohio, Utah, Virginia, Wisconsin, and one state to be named – partnering with the Council of Chief State School Officers (CCSSO) as lead contractor and three other research and technical assistance organizations: the University of Wisconsin’s Wisconsin Center for Education Research, EdCount, LLC, and WestEd, Inc. The consortium proposes a unique project integrating research, development, collaboration, and technical assistance to improve the quality and validity of state assessments designed to assess the knowledge and skills of English language learners (ELL). Each of the participating states will analyze and improve their state assessments through research and analysis of the alignment and validity of their state assessments in academic subjects for math and language arts especially in relation to English language learners, and by improving the alignment of state tests of English language proficiency to state standards. </strong
Goals and Scope: The proposal stemmed from states’ recognition of the need for additional resources and information for improving assessments and standards that apply to English language learners as identified in the CCSSO state collaborative project on ELL issues. The project proposal is also based on the states’ recognition of the potential applications of the Surveys of Enacted Curriculum (SEC) model and tools – which have been applied in more than 20 states for alignment studies and analyses of instructional alignment in several academic subjects – for analyzing alignment issues for ELL instruction, assessments, and standards. The project has five specific goals:
- Measure and report the degree of alignment between instruction, state standards, and student assessments for English language learners in a sample of classrooms and schools in the Consortium states.
- Advance and improve the alignment procedures and definitions for the Surveys of Enacted Curriculum alignment method to increase capacity for applying the method to English language development practices and materials.
- Provide expert technical assistance for consortium states to apply and use alignment analysis findings with state leaders to improve the quality and validity of assessment instruments for ELL students.
- Work with selected states to analyze growth in student achievement in relation to the degree of alignment of classroom instruction to standards, i.e., differences in “opportunity to learn,” and report the findings of the analysis to states.
- Prepare and widely disseminate products from the study for use by educators, researchers, and policymakers, including a guide for use of the alignment methods demonstrated in the study for improving assessment.
Adapting Reading Test Items to Increase Validity of Alternative Assessments Based on Modified Achievement Standards
The proposed project, Adapting Reading Test Items to Increase Validity of Alternate Assessments Based on Modified Academic Achievement Standards (ARTIIV) reflects the desire of five states to respond thoughtfully and significantly to the reading assessment needs of students eligible for the 2% option. ARTIIV will advance the states’ previous work, which resulted in a better understanding of the eligible students and promising approaches to item construction. The consortium seeks to build on those outcomes to increase the validity and accessibility of their current statewide assessments.
This project will investigate strategies that states can use to adapt their assessments based on grade-level academic achievement standards, focusing on the critical area of secondary level reading comprehension. ARTIIV will explore the systematic reengineering of assessment items based on cognitive modeling of the comprehension skill set. Cognitive modeling in the service of test development is increasingly recommended for improving the validity of assessment results interpretation (Gorin, 2006, National Research Council, 2001).
This project will utilize researched cognitive models, rework reading comprehension items to fit the model, experimentally manipulate the items to present reduced cognitive loads, and test to see whether the approach is more sensitive to the target population’s competencies without compromising psychometric properties or the intended grade level constructs. ARTIIV further proposes to study experimentally the impact of read-aloud accommodations on the performance of the reengineered items based on the work of the National Accessible Reading Assessments Projects (Cahalan Laitusis, Cook, L.L, Cline, F., King, T., and Sabatini, J., 2007) and to explore a methodology for creating modified academic achievement level descriptors.
State Academic Learning Links with Self-Evaluation for Alternate Assessment
The Pennsylvania Department of Education is the applicant for the proposed project, State Academic Learning Links with Self-Evaluation for Alternate Assessment (SALLSA). Four states (Georgia, Pennsylvania, Washington, and Wyoming) collaborate (a) to replicate and extend the Links to Academic Learning (LAL) protocol (Flowers, Wakeman, Browder, & Karvonen, 2007) and (b) to understand and improve the linkage between their state academic standards and their own alternate assessments (AA) measured against alternate achievement standards (AAs). The LAL model was designed specifically for AAs and goes beyond content match between standards and assessment and looks at the match of instruction, curriculum, and accessibility of content standards for students with significant cognitive disabilities (SWSD).
SALLSA has four goals: (1) Replicate and extend the research on the LAL alignment protocol using four states; (2) Develop and implement a process for helping states use alignment study results for planning improvements in their AA systems; (3) Conduct multiple case study to investigate implementation processes and impact of Goals 1 and 2; and (4) Disseminate project activities and findings. Perhaps the most significant need addressed is helping states interpret alignment findings and plan for improvement. Until now, the emphasis has been on conducting a study and submitting evidence for peer review. States need to take action on alignment study outcomes and SALLSA focuses on the analysis of study results for ongoing AA system improvement. As a form of validation, SALLSA replicates the LAL model in four states with unique assessment formats and investigates how states interpret alignment findings. A multiple case study will be conducted to investigate (a) how the LAL protocol is implemented within the state’s assessment system, and (b) what impact SALLSA implementation has on changes in AA systems with an emphasis on lessons learned that might guide other states in their planning for and use of alignment studies. The project will make processes and findings available through broad dissemination and the case study will maximize the quality of information made available to other states. Pennsylvania will engage Measured Progress to coordinate grant activities and conduct the alignment ratings. The authors of the LAL at the University of North Carolina at Charlotte have endorsed the replication plan, offered ongoing consultation, and agreed to be available to discuss refinements and interpretation capabilities. Western Carolina University will hold the subcontract for the project evaluation and multiple-case study. Finally, the Southeast Regional Resource Center will assist with dissemination of outcomes through the Federal Regional Resource network.
Operationlizing Alternate Assessment for Science Inquiry Skills
The South Carolina Department of Education requests $1,168,706 over an 18-month period to implement the Operationalizing Alternate Assessment for Science Inquiry Skills (OAASIS) project. Developing assessment strategies based on specific student characteristics and instructional needs is crucial if accurate inferences are to be made about what students know and can do. The OAASIS project will change and improve state assessment systems because it will inform and provide guidance to a current assessment priority—defining the target student population and developing an alternate assessment based on modified achievement standards (AA-MAS) that provides accurate results for students—in order to use the 2% flexibility option as outlined in the No Child Left Behind non-regulatory guidance document (April 2007). South Carolina’s meets the four absolute priorities and the three competitive preference priorities established for the Enhanced Assessment Grant program.
The goal of OAASIS is to investigate the process and development of an alternate assessment based on modified achievement standards (AA-MAS) by defining the target population, administering assessment strategies using multiple formats, and evaluating their accuracy in measuring student achievement. Objectives for OAASIS are 1) Establish learner characteristics and instructional needs as a basis for the design of multiple assessment strategies that increase access for students requiring AA-MAS; 2) Establish a common core of high school science inquiry standards among all participating states; 3) Design and implement three assessment strategy formats based on the common essential constructs of high school science inquiry content standards; 4) Disseminate results through diverse methods to reach widest audience possible
Partners involved in OAASIS include two universities, Vanderbilt University and the University of South Carolina; two state partners (South Dakota and Wyoming), 31 state members of the State Collaborative on Assessment and Student Standards: Assessing Special Education Students (ASES-SCASS), and the Discovery/ThinkLink Learning Corporation. Through this collaborative effort, OAASIS promises to yield significant information for all states on the processes and procedures not only to create an AA-MAS, but to create an effective, valid AA-MAS.