|Title:||SimScientists Assessment System|
|Principal Investigator:||Quellmalz, Edys||Grantee:||WestEd|
|Program:||Mathematics and Science Education [Program Details]|
|Award Period:||4 years (7/1/2012-6/30/2016)||Award Amount:||$1,599,764|
Co-Principal Investigators: Barbara Buckley, Mark Loveland, Matt Silberglitt, and Daniel Brenner
Purpose: To date, technology has been used mainly to support the logistics of assessment administration and scoring. However, new technology has also shown promise as a tool for the development of measures of complex learning that can be useful for instruction. In particular, computer simulations are becoming a component of large-scale science assessments such as the Programme for International Student Assessment, the National Assessment of Educational Progress, and the assessments aligned with the Next Generation Science Standards are likely to include simulations. There is a need, then, to develop and validate simulation-based science assessments designed for classroom, district, and state level use.
Project: Following a three phase process of development, testing, and refinement, the researchers will develop and validate a set of simulation-based assessments for the Life Science strand of the SimScientists Assessment System. This strand will encompass three units taught in middle school: Cells, Human Body Systems, and Ecosystems.
Products: The outcomes of the project include a computerized, fully developed and validated simulation-based assessment system focusing on Life Science at middle school. Peer-reviewed publications will also be produced.
Setting: The study will be conducted in urban, suburban, and rural middle schools in California, Massachusetts, Nevada, North Carolina, and Utah.
Sample: The study sample will consist of approximately 40 middle school teachers and 1,000 students.
Assessment: The researchers will develop and validate a vertically aligned set of simulation-based assessments for the Life Science strand encompassing three units taught in middle school: Cells, Human Body Systems, and Ecosystems. Both curriculum-embedded formative assessments intended to monitor progress and summative end-of-unit benchmark assessments will be developed. These simulation-based assessments will be 45-minute problem-based, multi-task inquiries. In addition, for each of three Life Science units, the researchers will develop signature tasks (i.e., interactive computer simulations that represent significant, recurring problems addressed by scientists and to be learned by students). The signature tasks will be shorter, requiring about 5–10 minutes to complete a set of items. The six signature tasks, two for each of the three Life Science units, will be combined to create the Life Science year-end test, which can be used as a proxy for an interactive computer task component of a district or state test.
Research Design and Methods: The assessments will undergo three phases of development, testing, and refinement. In Phase 1, the researchers will work with middle school teachers to co-develop simulation assessment components (unit embedded and benchmark simulation assessments) for the Cells and Human Body Systems units. In addition, the researchers will develop the pre-test, and post-test forms for Cells and Human Body Systems that will serve as external measures. Six signature tasks, two from each of three units, will be developed. The signature tasks will be short, 5–10 minute simulation-based tasks designed to test key ideas and inquiry practices. A year-end Life Science test will be developed and will be composed of signature tasks from each of the three Life Science units and a posttest of conventional items covering the content and inquiry targets in the three units. Once the newly programmed assessments have passed through the quality assurance screening, cognitive laboratory studies with two teachers and five middle school students will be conducted. The teachers and students will work through the assessments, thinking aloud as they respond to the tasks and questions in the benchmark assessments. Following the first round of cognitive labs, the teachers will try out all of the new assessment components in their classes to examine implementation logistics and technology usability. The two teachers will have four science classes of approximately 25 students per class. Four students, one from each of the four classes, will participate in a second round of cognitive labs of the embedded and benchmark assessments to more closely examine feasibility, usability, and construct validity of the tasks and items.
In Phase 2, the researchers will pilot test and collect evidence in multiple states (Massachusetts, North Carolina, Nevada, and Utah) for initial validation of the unit assessments and the year-end Life Science test to establish their technical quality and feasibility. Pilot tests will take place in the classrooms of 10 middle school Life Science teachers. For Sample 1, it is estimated that each of the 5 teachers will have four classes of roughly 25 students, for a sample of approximately 500 students piloting the full suite of assessments and the year-end Life Science test. For Sample 2, a different set of 5 teachers with approximately 500 students will use only the year-end Life Science test of signature tasks and the posttest.
In Phase 3, the researchers will cross-validate the simulation assessments and study the predictive validity of the classroom-embedded and benchmark assessments on the Life Science year-end test. For each validity study, the researchers will randomly select one class of approximately 25 students from each of 40 teachers, for a total of 1,000 students. The classes will be randomly assigned to use the full SimScientists suite (Sample 3) or to use only the year-end Life Science test (Sample 4).
Control Condition: For cross-validation purposes, classrooms using the full SimScientists assessment suite will be compared to classrooms using only the year-end Life Science test.
Key Measures: The key measures for the study include students' responses on the developed assessments; student and teacher interviews; observations; and responses to questionnaires.
Data Analytic Strategy: Analyses of reliability and validity will be carried out using confirmatory factor analysis, stratified coefficient alpha tests, uni- and multi-dimensional item response theory, and difficulty estimates. In addition, differential item functioning analyses will be conducted examining the effect of gender, ethnicity and EL status, and subgroup analyses to examine results for low performing English learners and students with disabilities.