ACCESS for ELLs® Overview: Origins of WIDA & Development of the Field Test
In 2002, the U.S. Department Education awarded the Wisconsin Department of Public Instruction an enhancement grant to develop a K-12 English proficiency test. Under the leadership of Wisconsin DPI, three states originally formed the WIDA consortium: Wisconsin, Delaware, and Arkansas. Other early additions were the District of Columbia, Rhode Island, Maine, New Hampshire, and Vermont.
Several organizations were original partners in the WIDA project. Dr. Tim Boals and Dr. Elizabeth Cranley at the Wisconsin Department of Public Instruction led the consortium states. WIDA has since become a project within the WI Center for Education Research (WCER). The Lead Developer of the WIDA standards was Dr. Margo Gottlieb of the Illinois Resource Center. Item specification development was led by Dr. Fred Davidson of the University of Illinois at Urbana-Champaign. The Language Testing Division at Center for Applied Linguistics was responsible for test development and coordinator/administrator trainings. CAL staff at the time included Dr. Dorry Kenyon, the Project Director; Dr. Jim Bauman, the Project Manager; and research assistants Jessica Nelson and Ellen Parkhurst. The University of Wisconsin at Oshkosh coordinated the technological aspects of the project including website and database development and management and assisted with student data collection. The trainings developed by CAL reside on UWO's Desire2Learn online facility. The WIDA website is now hosted by WCER.
Development of the ACCESS for ELLs® test began in the fall of 2003, following the completion of the standards. The development followed the course of first setting test item specifications and then developing preliminary items to fit these specifications. The initial item development was largely an undertaking of teachers recruited from most of the consortium states. Items were then edited and assembled into a battery of pilot tests, which were administered in Wisconsin, Illinois, and the District of Columbia.
The purpose of piloting is to establish whether and how well the test procedures work. It also serves to give the test developers an idea of how different types of items perform and whether the test, as designed, shows a good likelihood of successfully discriminating language proficiency.
Following the pilot, the development team undertook a second round of item development. Each item was reviewed for content accuracy and appropriateness, and for cultural and linguistic bias. Following this review, the final cut of items was made and test booklets constructed for field testing.
Field testing for the ACCESS for ELLs® had four primary purposes:
- To ensure that all test items function as intended when administered to a cross-section of students representing those to whom the assessment will later be given operationally.
- To calibrate the test items (i.e., determine their difficulty levels) and create a single reporting scale both across the three tiers within a grade level cluster and across the grade level clusters. (To do so, there are some items in common both across tiers and grade levels.)
- To collect performance data to relate scores on the assessment to the WIDA English language proficiency levels and conduct preliminary validation studies.
- To have a "dry run" of all logistical procedures (e.g., test coordinator training, test administrator training) with a smaller number of individuals before the large-scale roll-out of the test.
Since the entire test battery of 24 test booklets had to be field tested, it was necessary to select students appropriate for each of these instruments and to select them in numbers high enough to provide meaningful and reliable data on the ability of each item to discriminate reliably at its designated English language proficiency level. This required that each item, and by extension each form, be administered to at least 400 students across the consortium. Since there are 24 test booklets altogether (4 grades x 3 tiers x 2 test batteries), this amounted to 9800 students.
To allocate and distribute the testing burden, each state was asked to identify a number of students proportional to its population of ELL students within the full consortium.
These students were selected and assigned to a particular test form based on the guidelines discussed earlier; that is, teacher judgment and/or the results of a screening procedure. Since the screener for the ACCESS for ELLs® test was not in place yet (in fact, the items that make up the screener were included in the field test forms), other language assessment tools currently in use in the states were used instead. To facilitate matching students to the WIDA defined proficiency scale, "crosswalks" between the scale used on those tests and the WIDA scale were provided.
Background